Human astroviruses (HAstVs) are common etiological agents of acute gastroenteritis in children, the elderly, and immunocompromised patients; some virus strains have also been associated with neurological disease. Despite their medical importance, the study of these pathogens has advanced at a slow pace. In this work, we produced neutralizing antibodies to the virus and mapped the epitopes they recognize on the virus capsid. These studies provide the basis for development of subunit vaccines that induce neutralizing antibodies, as well as tools to explore the development of a specific antibody therapy for astrovirus disease. Our results also establish a platform to advance our knowledge on HAstV cell binding and entry.
KEYWORDS: astrovirus antigenic structure, human astrovirus, neutralizing monoclonal antibodies
ABSTRACT
Human astroviruses (HAstVs) cause severe diarrhea and represent an important health problem in children under two years of age. Despite their medical importance, the study of these pathogens has been neglected. To better understand the astrovirus antigenic structure and the basis of protective immunity, in this work we produced a panel of neutralizing monoclonal antibodies (Nt-MAbs) to HAstV serotypes 1, 2, and 8 and identified the mutations that allow the viruses to escape neutralization. We first tested the capacity of the recombinant HAstV capsid core and spike domains to elicit Nt-Abs. Hyperimmunization of animals with the two domains showed that although both induced a potent immune response, only the spike was able to elicit antibodies with neutralizing activity. Based on this finding, we used a mixture of the recombinant spike domains belonging to the three HAstV serotypes to immunize mice. Five Nt-MAbs were isolated and characterized; all of them were serotype specific, two were directed to HAstV-1, one was directed to HAstV-2, and two were directed to HAstV-8. These antibodies were used to select single and double neutralization escape variant viruses, and determination of the amino acid changes that allow the viruses to escape neutralization permitted us to define the existence of four potentially independent neutralization epitopes on the HAstV capsid. These studies provide the basis for development of subunit vaccines that induce neutralizing antibodies and tools to explore the possibility of developing a specific antibody therapy for astrovirus disease. Our results also establish a platform to advance our knowledge on HAstV cell binding and entry.
IMPORTANCE Human astroviruses (HAstVs) are common etiological agents of acute gastroenteritis in children, the elderly, and immunocompromised patients; some virus strains have also been associated with neurological disease. Despite their medical importance, the study of these pathogens has advanced at a slow pace. In this work, we produced neutralizing antibodies to the virus and mapped the epitopes they recognize on the virus capsid. These studies provide the basis for development of subunit vaccines that induce neutralizing antibodies, as well as tools to explore the development of a specific antibody therapy for astrovirus disease. Our results also establish a platform to advance our knowledge on HAstV cell binding and entry.
INTRODUCTION
Human astroviruses (HAstVs) are common etiological agents of acute gastroenteritis in children, the elderly, and immunocompromised patients (1–3). They are estimated to be responsible for 2% to 9% of cases of acute, nonbacterial childhood diarrhea (4). Recently, the epidemiology of HAstV in community settings in 8 low- and middle-income countries with high prevalence of diarrhea and undernutrition was reported (5); the prevalence of HAstV in diarrheal stools was 5.6%, and remarkably, the disease severity exceeded that of all enteropathogens tested except rotavirus. In a related study, norovirus GII, rotavirus, and astrovirus exhibited the highest attributable burdens of diarrhea in the first 2 years of life (6). These studies show that HAstV is an overlooked cause of diarrhea among vulnerable children worldwide, and they highlight the need for future research as well as the importance of this virus as a target for vaccine development (5, 6).
HAstVs are classified into 8 classical serotypes (HAstV-1 to -8) associated with acute gastroenteritis; however, novel HAstV strains have been recently described and are also associated with neurological disease, including meningitis and encephalitis in immunocompromised patients (4). HAstVs are small, nonenveloped viruses with a single-stranded positive-sense RNA genome of about 6.8 kb comprising three open reading frames (ORFs). ORF2 encodes a capsid precursor structural polyprotein of about 90 kDa that self-assembles intracellularly and is processed by caspases to produce a virus particle formed by a 70-kDa protein (VP70) (7). This immature particle egresses the cell and is further processed by extracellular proteases; in vitro, the treatment of this particle with trypsin yields a mature infectious virus that, in HAstV-8, contains three protein species of approximately 34 (VP34), 27 (VP27), and 25 (VP25) kDa (7), although it has been recently shown that only VP34 and VP27 are relevant for HAstV-8 infectivity (8).
Cryo-electron microscopy of mature HAstV virions reveals particles of 44 nm comprised of a T = 3 icosahedral shell and 30 globular spikes (9). The crystal structures of both VP34 (10, 11) and VP25 (11–14) have been recently determined; VP34 folds into a domain that constitutes the shell of the virus capsid (core domain), while VP27/VP25 form dimeric capsid spikes (spike domain). VP34 is derived from the highly conserved N-terminal region of VP70, while VP27/VP25, which differ only at their amino termini, are derived from the hypervariable region of VP70 (15). Despite important advances in the structural characterization of the HAstV particle, the functional sites on the virus capsid, including the location of the receptor-binding site and the antigenic determinants where neutralizing antibodies bind, have been poorly characterized. Only one neutralizing antibody epitope, located on the capsid spike domain, has previously been defined by X-ray crystallography (12).
To better understand the astrovirus antigenic structure and the basis of protective immunity, in this work we produced neutralizing monoclonal antibodies (Nt-MAbs) to HAstV serotypes 1, 2, and 8 and identified the mutations that allow variant viruses to escape neutralization. This allowed us to define the existence of four potentially independent neutralization epitopes on the HAstV capsid. These studies provide the basis for the development of subunit vaccines to induce neutralizing antibodies and tools to explore the possibility of developing a specific antibody therapy for astrovirus disease. Our results also establish a platform to advance our knowledge on HAstV cell binding and entry.
RESULTS
The HAstV capsid spike but not the capsid core domain induces neutralizing antibodies.
To identify the neutralization epitopes on the HAstV capsid, we first evaluated the capacity of the core and spike domains to induce total (determined by enzyme-linked immunosorbent assay [ELISA]) and neutralizing antibodies. These two structural domains of the virus can be produced in Escherichia coli in a soluble and correctly folded form (10–14). Rabbits and mice were immunized with either the purified, recombinant core or the spike domain of HAstV-1 (core1 or spike1). As expected, both domains were immunogenic and induced a robust antibody response (Fig. 1A). We then tested the capacity of the polyclonal sera directed to either domain to neutralize the infectivity of HAstV-1. The core1 did not induce a detectable level of neutralizing antibodies to either HAstV-1, HAstV-2, or HAstV-8 (Fig. 1B), while the spike1 domain induced a high (>1/50,000) neutralizing response that was specific for HAstV-1 (Fig. 1B) Of interest, despite the fact that the anti-spike1 mouse polyclonal serum (both rabbit and mice sera behaved equally in all assays tested) was specific for HAstV-1 by neutralization, it also recognized the heterologous capsid spikes of HAstV-2 and -8 (spike2 and spike8) as determined by ELISA, although about 50 to 100 times less efficiently (Fig. 2A). The anti-spike2 antibodies (see Materials and Methods) also interacted more weakly with spike1 and spike8 (Fig. 2B). On the other hand, the anti-spike8 serum better recognized the spike1 (with about a 5- to 10-fold difference) than the spike2, which was bound about 100 times less efficiently (Fig. 2C). In turn, the anti-core1 polyclonal serum cross-reacted efficiently with the core domains of both HAstV-2 and HAstV-8 as determined by Western blotting, and it also recognized cells infected with HAstV serotype 1, 2, or 8 as determined by both immunofluorescence and immunoblotting (Fig. 2D and E).
FIG 1.
Immunogenicity and induction of neutralizing antibodies by recombinant HAstV capsid core and spike domains. (A) Recombinant capsid core and spike domains of HAstV serotype 1 (core1 and spike1) were immobilized on ELISA plates and then incubated with the indicated dilution of rabbit preimmune (PI) or hyperimmune (HI) polyclonal serum (anti-core1 and anti-spike1). The interaction was determined as described in Materials and Methods, and the optical density at 405 nm (OD405) of the developed color was detected. Experiments were performed in biological duplicates carried out in duplicate. The data were normalized to the maximal OD obtained in each experiment and represent the mean ± standard error of the mean (SEM). The maximal OD405 for the anti-spike1 serum was 2.49, and that for the anti-core1 serum was 2.39. (B) The indicated HAstV strains were preincubated with the HI rabbit serum directed to either the core1 or spike1 protein at the indicated dilutions, and the infectivities of the viruses was determined as described in Materials and Methods. All the infectivity experiments were performed in biological triplicates carried out in duplicate. The data are expressed as percentages of the value for the positive control (viruses not incubated with antibodies) and represent the mean ± SEM.
FIG 2.
Interaction of anti-spike polyclonal antibodies with recombinant spikes as determined by ELISA and of anti-core polyclonal antibodies with viral proteins as determined by immunofluorescence and Western blotting. (A to C) Recombinant spikes derived from HAstV serotype 1, 2, or 8 were immobilized in ELISA plates, and they were later incubated with the indicated dilution of mouse anti-spike1 (A), anti-spike2 (B), or anti-spike8 (C) antibodies. The interaction was determined as described in Materials and Methods, and the optical density (OD) of the color developed was detected at 405 nm. Experiments were performed in biological triplicates carried out in duplicate. The data were normalized to the maximal OD obtained in each experiment and represent the mean ± SEM. The maximal OD405s were 2.77 for the anti-spike1 serum, 2.91 for the anti-spike2 serum, and 2.87 for the anti-spike8 serum. (D) Caco-2 cells were infected with either HAstV serotype 1, 2, or 8 at an MOI of 0.08, and at 20 hpi they were fixed with formaldehyde and incubated with anti-core1 polyclonal antibodies at a 1:200 dilution. The cells were subsequently incubated with Alexa 488-labeled anti-IgG antibodies and observed for immunofluorescence. (E) Caco-2 cells were infected with either HAstV serotype 1, 2, or 8 at an MOI of 0.3, and at 20 hpi they were lysed and the viral proteins were analyzed by Western blotting using anti-core1 polyclonal antibodies at a 1:500 dilution, followed by incubation with Alexa 647-labeled anti-IgG antibodies, and detected in a Typhoon FLA 9500 instrument.
Isolation and serotype specificity of Nt-MAbs.
Given that the spike was the only domain that induced neutralizing antibodies, we used it to immunize mice to generate neutralizing monoclonal antibodies (Nt-MAbs). Since we were interested in isolating cross-reactive, heterotypic Nt-MAbs, we immunized mice with a mixture of spike domains of HAstV serotypes 1, 2, and 8, following the immunization scheme described in Materials and Methods. The primary screening for hybridomas secreting astrovirus-specific MAbs was carried out by ELISA, using as antigens the individual spike proteins from the three different HAstV serotypes. We selected 54 positives hybridomas from 480 wells initially screened: 21 for spike1, 1 for spike2, and 16 for spike8. Thirteen were cross-reactive with at least 2 serotypes, and three recognized the spike domains of the 3 serotypes. Twenty-two of these hybridomas were further screened by a neutralization assay. Twelve of these MAbs had neutralizing activity, and five of them that grew stably were cloned and expanded for further analysis. Of these, two Nt-MAbs were directed to HAstV-1 (MAbs 3B4 and 3H4), one to HAstV-2 (MAb 4B6), and two to HAstV-8 (MAbs 2D9 and 3E8). All five Nt-MAbs were serotype specific as determined by ELISA (not shown) and neutralization (Fig. 3) and also specifically recognized their antigen as determined by immunocytochemistry and immunofluorescence in infected Caco-2 cells (not shown).
FIG 3.
Specificities of neutralizing monoclonal antibodies. HAstV serotype 1, 2, or 8 was preincubated with MAb 3B4 (A), 3H4 (B), 4B6 (C), 2D9 (D), or 3E8 (E) or with mouse hyperimmune (HI) polyclonal sera directed to the indicated spike proteins (F) at the indicated dilutions. The infectivities of the viruses were determined as described in Materials and Methods. All experiments were performed in biological triplicates carried out in duplicate. The data are expressed as percentages of the value for the positive control (viruses not incubated with antibodies) and represent the mean ± SEM.
In general, all selected MAbs neutralized the cognate virus to high titers. Those specific for HAstV-1 had a neutralization titer higher than 1/10,000, which was slightly lower than the neutralization titer achieved with the hyperimmune polyclonal serum to spike1 (Table 1; Fig. 3A, B, and F). MAb 4B6, directed to HAstV-2, had a neutralization titer of more than 1/100,000, similar to that of the hyperimmune polyclonal serum to spike2 (Table 1; Fig. 3C and F); MAbs 2D9 and 3E8, directed to HAstV-8, reached a neutralization titer of more than one million, a titer about 10- to 50-fold higher than that reached with the polyclonal antibodies to spike8 (Table 1; Fig. 3D, E, and F).
TABLE 1.
Characteristics of the antibodies described in this work
| Antibody | Virus recognized | Peroxidase/IFa | Neutralization titerb |
|---|---|---|---|
| Monoclonal antibody | |||
| 3H4 | HastV-1 | + | 1/20,000 |
| 3B4 | HAstV-1 | + | 1/16,000 |
| 4B6 | HAstV-2 | + | 1/800,000 |
| 2D9 | HAstV-8 | + | 1/5,000,000 |
| 3E8 | HAstV-8 | + | 1/1,000,000 |
| Polyclonal serum | |||
| Anti-spike1 | HAstV-1 | + | 1/250,000 |
| Anti-spike2 | HAstV-2 | + | 1/250,000 |
| Anti-spike8 | HAstV-8 | + | 1/63,000 |
| Anti-core1 | HAstV-1 | + | Not neutralizing |
| HAstV-2 | + | ||
| HAstV-8 | + |
Reactivity of antibodies as determined by immunocytochemistry and immunofluorescence in cells infected with the indicated virus.
Reciprocal of the antibody dilution that reduces at least 50% of the infectious virus foci of the indicated virus in the neutralization assay described in Materials and Methods.
Isolation of virus variants resistant to neutralization by monoclonal antibodies.
To determine the immunogenic topology of the viral capsid, we selected virus variants resistant to neutralization by each of the five MAbs characterized. Each neutralization escape variant was completely resistant to the corresponding Nt-MAb even at a 10−3 dilution of the ascites fluid (Fig. 4), with the exception of the escape variant for MAb 2D9 (vHAstV-8/2D9), which was slightly neutralized by MAb 2D9 (titer of 5 × 10−3) (Fig. 4D), although more than 1,000-fold less efficiently than the wild-type (wt) HAstV-8 virus (titer of 8 × 10−6) (Fig. 4D). On the other hand, all escape variants were still efficiently neutralized by the corresponding hyperimmune anti-spike serum (Fig. 4), indicating the existence of more than one neutralization epitope on the spike domain of HAstV.
FIG 4.
Neutralization of HAstV escape variants. Escape variants vHAstV-1/3B4 (A), vHAstV-1/3H4 (B), vHAstV-2/4B6 (C), vHAstV-8/2D9 (D), and vHAstV-8/3E8 (E) were preincubated with mouse hyperimmune (HI) polyclonal serum or with monoclonal antibodies at the indicated dilutions. The infectivities of the viruses were determined as described in Materials and Methods. All experiments were performed in biological triplicates carried out in duplicate. The data are expressed as percentages of the value for the positive control (viruses not incubated with antibodies) and represent the mean ± SEM.
When cross-neutralization of escape variants of the same HAstV serotype was tested, i.e., neutralization of vHAstV-1/3B4 by MAb 3H4 and neutralization of vHAstV-1/3H4 by MAb 3B4, an efficient inhibition of infectivity was still observed, suggesting that the these two MAbs recognize two different epitopes on the virus (Fig. 4A and B). Similar results were obtained when HAstV-8 neutralization escape variants vHAstV-8/2D9 and vHAstV-8/3E8 were evaluated for inhibition by MAbs 2D9 and 3E8, since they were neutralized with titers higher than 10−6 by the “nonhomologous” MAb, suggesting again that these MAbs might interact with two different antigenic determinants on the HAstV-8 capsid (Fig. 4D and E). Of interest, it was also possible to isolate HAstV-1 variants that were highly resistant to neutralization by both MAbs 3B4 and 3H4, as well as HAstV-8 double-escape variants that were resistant to inhibition by both 2D9 and 3E8 MAbs (Fig. 5), showing the resilience of the virus to withstand mutations in more than one site in order to escape antibody neutralization. In addition, the double-escape variants were still efficiently neutralized by the hyperimmune polyclonal sera, suggesting that there are more than two neutralizing antigenic regions on the virus.
FIG 5.
Neutralization of double-escape variants. Double-escape variants vHAstV-1/3B4/3H4 (A), vHAstV-1/3H4/3B4 (B), vHAstV-8/2D9/3E8 (C), and vHAstV-8/3E8/2D9 (D) were preincubated with mouse hyperimmune (HI) polyclonal serum or with monoclonal antibodies at the indicated dilutions. The infectivities of the viruses were determined as described in Materials and Methods. All experiments were performed in biological triplicates carried out in duplicate. The data are expressed as percentages of the value for the positive control (viruses not incubated with antibodies) and represent the mean ± SEM.
Identification of the amino acid changes that confer viruses resistance to neutralization by MAbs.
To understand the localization of the neutralization epitopes, we sequenced the genome regions encoding the spikes of the virus escape variants and identified amino acids changes that confer virus resistance to the Nt-MAbs. A single amino acid change was identified in the spike sequence of each single variant, with the exception of vHAstV-2/4B6, that had two contiguous amino acids changed (Table 2; Fig. 6A). The identified neutralization escape mutations were then mapped onto the crystal structure of the spike1 domain (11); all of the mutated amino acids were located on the surface of the spike (Fig. 6B). In agreement with the results described above that suggest that the MAbs to HAstV-1 recognize different antigenic sites (Fig. 4), the amino acid residues that confer HAstV-1 resistance to neutralization by MAbs 3H4 and 3B4 were located in two very different regions of the spike (Fig. 6B; Table 2). A K504E change, located in loop 2 on the inferior/lateral region of the spike, conferred resistance to MAb 3H4, while the S560P mutation, which results in the loss of neutralization by MAb 3B4, is located on the top of the spike in the middle of the long loop 3 that extends over the top, adjacent to the monomer-monomer contact in the spike and away from the K504E amino acid change.
TABLE 2.
Positions of mutations in the escape variants
| Virus | Monoclonal antibody | Neutralization escape variant | Amino acid change(s) |
|---|---|---|---|
| HAstV-1 | 3H4 | vHAstV-1/3H4 | K504E |
| 3B4 | vHAstV-1/3B4 | S560P | |
| 3H4/3B4 | vHAstV-1/3H4/3B4 | K504E/S560P | |
| 3B4/3H4 | vHAstV-1/3B4/3H4 | K504E/S560P | |
| HAstV-2 | 4B6 | vHAstV-2/4B6 | D564E/N565D |
| HAstV-8 | 3E8 | vHAstV-8/3E8 | Y464H |
| 2D9 | vHAstV-8/2D9 | D597Y | |
| 2D9/3E8 | vHAstV-8/2D9/3E8 | Y464H/D597Y | |
| 3E8/2D9 | vHAstV-8/3E8/2D9 | Y464H/D597Y |
FIG 6.
Localization of point mutations that confer escape from neutralization by monoclonal antibodies. (A) Sequence alignment of HAstV-1-8 CP spike domains. Conserved, similar, and nonconserved amino acids are colored red, pink, and black, respectively. The amino acid residues that changed in the variants that escaped neutralization by MAbs 3H4 and 3B4 to HAstV-1, 4B6 to HAstV-2, and 3E8 and 2D9 to HAstV-8 are indicated with green, gold, and blue arrows, respectively. The accession numbers for the spike sequences shown are as follows: GenBank L23513 (serotype 1) and UniProt Q82446 (serotype 2), Q9WFZ0 (serotype 3), Q3ZN05 (serotype 4), Q4TWH7 (serotype 5), Q67815 (serotype 6), Q96818 (serotype 7), and Q9IFX1 (serotype 8). (B) spike1 structure shown from the top and one side. One half of the dimer is gray, and the other half is white. The positions of the mutations that confer escape from neutralization by the various monoclonal antibodies are colored green for HAstV-1, gold for HAstV-2, and blue for HAstV-8. The mutations are shown in only one protomer in the dimer.
The escape of neutralization of HAstV-2 by MAb 4B6 was the consequence of two contiguous amino acid changes (D564E and N565D) located in loop 3, on the top of the spike and very close to the amino acid that defines the resistance of HAstV-1 to MAb 3B4 (Fig. 6; Table 2). In the case of HAstV-8, the mutations that confer resistance to MAbs 3E8 and 2D9 were found to map on different sides of the spike, opposite to the dimeric subunit. These changes, Y464H for MAb 3E8 and D597Y for MAb 2D9 (Fig. 6; Table 2) are located, respectively, at the end of the large loop 1 and at the beginning of loop 4, and they may represent two independent epitopes.
Sequence analysis of the double-escape variants of HAstV-1 (vHAstV-1/3H4/3B4 and vHAstV-1/3B4/3H4) confirmed the mutations of the single-escape variants, bearing in both cases the two mutations identified individually in vHAstV-1/3H4 and vHAstV-1/3B4 (Table 2). A similar finding was obtained when the double-escape variants of HAstV-8 were characterized (Table 2).
DISCUSSION
To gain insight into the antigenic sites of the HAstV capsid, Nt-MAbs directed to serotypes 1, 2, and 8 were produced, and the sites of amino acid mutations that allow the viruses to escape neutralization were mapped onto the virus capsid spike. Only two reports, published over 20 years ago, have described the isolation of Nt-MAbs to HAstV. In the first study, one MAb specific for HAstV-2 (MAb PL-2) was isolated (16). In the second work, three Nt-MAbs targeting HAstV-1 were reported (17). Despite the fact that in those studies purified virus was used as the immunogen, all neutralizing MAbs isolated were directed to the spike domain, in agreement with our findings that the spike, but not the core domain, is able to induce Nt-MAbs. Also, the crystal structure of the HAstV-2 spike bound to the Fab fragment of MAb PL-2 was recently reported (12), confirming the spike domain as a target for neutralizing antibodies. In the work by Bass and Upadhyayula (17), the three MAbs to HAstV-1 were found to be topographically in close proximity as determined by a competition ELISA. All four Nt-MAbs previously reported were able to prevent the attachment of the virus particles to Caco-2 cells, suggesting that the spike domain also contains the receptor-binding site of the virion. These results indicate that a subunit vaccine should focus on the spike domain as the main immunogen to induce a potential protective antibody response.
Of interest, two of the Nt-MAbs reported by Bass and Upadhyayula (17) neutralized more than one serotype, with one of them being broadly neutralizing to all seven HAstVs tested, although this MAb was of the IgM isotype and probably was of low affinity. None of these MAbs are currently available. In this regard, we found that about 1 out of 4 of the initially positive hybridomas neutralized more than one serotype, indicating again the existence of cross-neutralizing epitopes in the spike domain. It cannot be ruled out, however, that a proportion of the cross-reacting hybridomas were the result of a mixture of two or more independent hybridoma cell clones having different serotype specificities. Regardless of this fact, the cross-neutralization epitopes do not seem to be immunodominant, at least in mice, since the anti-spike1 hyperimmune polyclonal serum was serotype specific as determined by neutralization and recognized the HAstV-2 and HAstV-8 spikes about 10-fold less efficiently as determined by ELISA.
It is important to note that the infectivity of the neutralization escape variants was blocked by the hyperimmune serum to the corresponding spike to a titer similar to that found for the wild-type virus. It would be interesting to test whether the neutralization escape variants are also efficiently neutralized by convalescent-phase sera from patients naturally infected with astrovirus, since in the case of rotavirus it has been shown that marked differences can characterize the immune responses to parenteral hyperimmunization and to intestinal infection (18).
Analysis of the structure of the recombinant HAstV-2 spike domain bound to scFv PL-2 (12) showed that this neutralizing MAb recognizes a quaternary epitope on each side of the dimeric capsid spike, and a deletion of three amino acids (amino acids 460 to 462) in loop 1 abolished the binding of PL-2. The location of the MAb3E8 escape mutation in loop 1 suggests that MAb 3E8 (anti-HAstV-8) may recognize an epitope that overlaps with the MAb PL-2 (anti-HAstV-2), suggesting in turn that different HAstV serotypes may share neutralization epitopes on the spike domain. In this regard, the escape mutations that allow HAstV-1 and HAstV-2 to escape neutralization by MAbs 3B4 and 4B6, respectively, are only 4 amino acids apart in loop 3 (Table 2 and Fig. 6), suggesting again that potentially this could be the same epitope shared by these two HAstV serotypes. On the other hand, MAbs 3H4 and 2D9, localized in loops 2 and 4, respectively, seem to define two new independent epitopes on the astrovirus capsid, which have not been previously reported. It is important to keep in mind that the amino acid changes identified in the neutralization escape variants represent only mutations that allow the virus to escape neutralization and that they are not necessarily located in the epitope recognized by the Nt-MAb, since distal mutations can, in principle, alter antibody binding. Thus, in this work we have defined the potential existence of 4 different neutralization epitopes on the astrovirus capsid, although some overlapping or indirect effects of escape variants among them cannot be ruled out.
Several patches of highly conserved amino acids among the eight HAstV serotypes, composed of amino acids that are distant in the linear sequence of the spike, have been identified (12, 13); they have been proposed to represent potential functional sites on the virus surface, including the so-far-unknown receptor-binding site. Of interest, all of the neutralizing escape mutations identified in this work, with the exception of K504E (vHAstV-1/3H4), map close to (only between 5 and 10 amino acids away from) the conserved amino acids that have been proposed to line the P and S sites, which are potentially functional sites identified in the HAstV-8 spike (13). It is likely that the footprint of the MAbs that recognized these epitopes overlaps the proposed functional P and S sites. Also, the footprint of MAb PL-2 on the HAstV2 spike was described to overlap four conserved amino acid regions and to block the binding of the recombinant spike2 to Caco-2 cells, suggesting that this MAb could also inhibit the attachment of the virus to the cell surface through interfering with the interaction of the virus with the cell receptor (12).
When screening the GenBank database to look for wild-type viruses that could have the mutations identified in our neutralization escape mutants, we found 24, 9, and 3 complete spike sequences, which were annotated as HAstV serotypes 1, 2, and 8, respectively. None of the 24 HAstV-1 strains found had a change at Ser-560. In contrast, Lys-504 was present in only one of the 24 viruses; one additional virus had a Pro in that position, and 22 showed a Gln. Regarding HAstV-2, four of the spike sequences found showed Asp-Asn-Asn at amino acid positions 564 to 566, like the wild-type HAstV-2 virus characterized in this work, while five had a deletion of one of the two Asn residues. Finally, the three additional serotype 8 HAstV strains had the same amino acids as the parental Yuc8 virus at the amino acid residue positions where escape mutations to MAbs 2D9 and 3E8 were mapped. Overall, the finding of amino acid differences at some of the positions that allow our variant viruses to escape neutralization by the selected MAbs has to be interpreted cautiously and cannot be attributed to have occurred as the consequence of selective pressure in humans, since our MAbs recognize a mouse epitope, which could or could not be immunogenic in humans. Whether these amino acid changes actually allow the viruses to escape neutralization by the MAbs described in this study needs to be determined.
Work in progress in our labs is directed to define the mechanism of neutralization of the antibodies characterized in this work and to use structural biology to define the antigenic sites of the Nt-MAbs described in this study. Altogether, our results advance our knowledge on the antigenic structure of HAstVs, the basis of HAstV neutralization, and the functional sites relevant for virus entry, providing a basis for the development of vaccines and therapeutic measures against astrovirus disease.
MATERIALS AND METHODS
Cells and viruses.
Caco-2 cells, clone C2Bbe1 (ATCC), were propagated in high-glucose Dulbecco’s modified Eagle’s medium (DMEM-HG) (Sigma) supplemented with nonessential amino acids (Gibco) and 15% fetal bovine serum (FBS) (Cansera) in a 10% CO2 atmosphere at 37°C. HAstV serotype 1 (HAstV-1-RIVMb) was obtained from Susana Guix (Department de Microbiologia, Facultat de Biologia, Universitat de Barcelona), and HAstV serotype 2 strain Oxford (HAstV-2-Oxford) was obtained from J. B. Kurtz (Department of Virology, John Radcliffe Hospital, Oxford, United Kingdom). HAstV serotype 8 strain Yuc8 was isolated in our laboratory. All viral strains were activated with trypsin and grown as described previously (19).
HAstV-1 CP core production.
cDNA corresponding to HAstV-1 capsid protein (CP) residues 80 to 429 (accession number AAC34717.1) was cloned into pET52b in frame with a C-terminal thrombin cleavage site and a 10-histidine purification tag sequence. The plasmid was transformed into Escherichia coli strain BL21(DE3)pLysS, and HAstV-1 CP core expression was induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) at 18°C for 16 h. E. coli cells were lysed by ultrasonication in 20 mM Tris (pH 7.5)–1 M NaCl–5% (vol/vol) glycerol–1 mM dithiothreitol (DTT)–20 mM imidazole containing 2 μM MgCl2, 1,250 U Benzonase (Millipore), and 1× protease inhibitor cocktail (set V, EDTA free) (Millipore). The HAstV-1 CP core was purified from soluble lysates by HisTrap metal affinity chromatography. The HAstV-1 CP core was further purified by size exclusion chromatography on a Superdex 75 column in 20 mM phosphate buffer (pH 7.6), 1 M NaCl, and 5% (vol/vol) glycerol.
HAstV spike production.
Synthetic genes codon optimized for Escherichia coli encoded HAstV-2-Oxford capsid spike amino acids 431 to 674 (GenBank accession number KY964327), HAstV-1 capsid protein residues 429 to 645 (accession number AAC34717.1), or HAstV-8 CP spike amino acids 429 to 647 (UniProtKB entry Q9IFX1). To make spike expression plasmids, genes were cloned into pET52b (Addgene) in frame with a C-terminal thrombin cleavage site and a 10-histidine purification tag. All plasmids were verified by DNA sequencing. Plasmids were transformed into E. coli strain BL21(DE3), and protein production was induced with 1 mM IPTG at 18°C for 16 h. E. coli cells were lysed by ultrasonication in 20 mM Tris-HCl (pH 8.0)–500 mM NaCl–20 mM imidazole containing 2 μM MgCl2, 1,250 U Benzonase (Millipore), and 1× protease inhibitor cocktail (set V, EDTA free) (Millipore). Proteins were purified from soluble lysates by HisTrap metal affinity chromatography. Proteins were buffer exchanged into phosphate-buffered saline (PBS) and further purified by size exclusion chromatography on a Superdex 200 column in PBS.
Polyclonal serum production.
Rabbit anti-spike1 polyclonal serum was generated by immunization of New Zealand rabbits with 250 μg of recombinant HAstV-1 capsid spike in Freund’s complete adjuvant (FCA) followed by three immunizations, every 2 weeks, with 250 μg of the protein in Freund’s incomplete adjuvant (FIA). Rabbit anti-core1 polyclonal serum and rabbit anti-HAstV-8 have been previously described (19). Mouse polyclonal serum to either spike1, spike2, or spike8 was generated by immunization of BALB/c mice with 50 μg of the corresponding recombinant protein in FCA followed by three immunizations, every 2 weeks, with 50 μg of the same protein in FIA.
MAb production.
Eight-week-old BALB/c mice were immunized with 50 μg of spike1 recombinant protein at 1:1 with FCA, and three more immunizations were performed every 2 weeks using a mix containing 50 μg each of spike1, spike2, and spike8 at 1:1 in FIA. Four days after the last immunization, the spleen was isolated and splenocytes were fused with Fox myeloma cells using 50% polyethylene glycol; the cells were suspended in hypoxanthine-aminopterin-thymidine (HAT) medium and directly plated in 96-well plates. Hybridomas secreting antibodies to either recombinant spike (1, 2, or 8) were screened by an ELISA (see below), and those positive by the ELISA were then assayed in a neutralization assay (see below). The hybridomas of interest were cloned three times by limiting dilution using thymocyte feeder layers. Selected MAbs were amplified as mouse ascites fluid.
ELISA.
Purified spike proteins at a concentration of 2 μg/ml in PBS (50 μl total) were incubated for 1 h at 37°C in 96-well ELISA microtiter plates. The plates were then washed three times with PBS containing 0.1% Tween 20 (PBST). The wells were blocked by adding 100 μl of 1% bovine serum albumin (BSA) in PBS and incubating at 37°C for 1 h, followed by three PBST washes. Conditioned medium from the hybridoma cells or the indicated dilution of mouse or rabbit polyclonal serum was added, incubated at 37°C for 1 h, and then washed three times with PBST. Plates were incubated for 1 h at room temperature with 50 μl of goat anti-mouse antibody conjugated to alkaline phosphatase (KPL) diluted 1:1,000 in PBS or with goat anti-rabbit antibody conjugated to alkaline phosphatase (KPL) diluted 1:1,000 in PBS. The plates were washed three times with PBST, and the reaction was developed by adding the phosphatase substrate (Sigma 104; 1 mg/ml) diluted in diethanolamine buffer (100 mM diethanolamine [pH 9.4], 1 mM MgCl2, 5 mM sodium azide). The absorbance was recorded at 405 nm in a FLUOstart Omega instrument (BMG Labtech).
Neutralization assays.
The indicated concentration of antibody was preincubated at a multiplicity of infection (MOI) of 0.002 of the indicated HAstV strain for 1 h at room temperature. The virus-antibody mixture was then added to confluent C2Bbe1 cell monolayers grown in 96-well plates and incubated for 1 h at 37°C. After this time, the cells were washed three times with minimum essential medium (MEM) without serum, and the infection was left to proceed for 18 h at 37°C. Infected cells were detected by an immunoperoxidase focus-forming assay, as described previously (19). The neutralization antibody titer was defined as the antibody dilution that blocks at least 50% of the input virus.
Sequencing.
To determine the nucleotide sequences of the various HAstV spikes, RNA was isolated from viral lysates using the PureLink viral RNA/DNA minikit (Invitrogen). cDNA was then synthesized with SuperScript III reverse transcriptase (Thermo Fisher Scientific) using the following primers: 5′-CGGCTTCCAGATGTGCAG-3′ (HAstV-1Lw), corresponding to nucleotides 6658 to 6675 of HAstV-1 strain Oxford (accession number L23513.1); 5′-GCGGTCTCCAGAAAGTTTG-3′ (HAstV-2Lw), corresponding to nucleotides 2369 to 2387 of HAstV-2 strain Oxford (accession number L06802.1); or 5′-GCTTCCAGATAGTTGCAG-3′ (HAstV-8Lw), corresponding to nucleotides 6639 to 6656 of HAstV-8 strain Yuc8 (accession number AF260508.1). For PCR amplification, Vent DNA polymerase (New England BioLabs) was used, with the following primers: HAstV-2Up (5′-CAGTTCACTCAAATGAACCA-3′), corresponding to nucleotides 1215 to 1234 of the HAstV-2 capsid gene (accession number L06802.1), for the HAstV-1 RIVMb and HAstV-2 Oxford strains and HAstV-8Up (5′-CAGTTTACACAGATGAATCA-3′), corresponding to nucleotides 5531 to 5551 of HAstV-8 strain Yuc8 (accession number AF260508.1) for HAstV-8 Yuc8. The PCR product was purified using the DNA Clean and Concentrator-5 kit (Zymo Research) and sequenced using Sanger chemistry in the sequencing facility of the Instituto de Biotecnología, Universidad Nacional Autónoma de México.
Isolation of neutralization escape variants.
For selection of escape variants, we incubated viral lysates (with at least 1 × 107 focus-forming units [FFU]/ml) with the appropriate MAb at a dilution of 1:1,000 to 1:100 of the ascites fluid for 1 h at room temperature. The mix was then used to infect CaCo-2 monolayers grown in 6-well plates for 1 h at 37°C, and the unbound virus was removed by washing three times. The cell monolayers were incubated at 37°C for 48 to 72 h postinfection (hpi) in the presence of trypsin (5 μg/ml), tetracycline (1 μg/ml), and the corresponding MAb diluted 1:1,000. Viral lysates were then prepared by three cycles of freeze-thawing, and the procedure was repeated at least three times before confirming the phenotype. The phenotypes of the variant viruses were evaluated by neutralization assays, and the selection was repeated until detection of the neutralization escape variant (this usually took 4 or 5 passages in the presence of the Nt-MAb). The double-escape variants were selected as described above but starting with the single variant instead of the wt virus.
Accession number(s).
The sequences determined here were deposited in GenBank with accession numbers MH763691 (HAstV-1), L06802 (HAstV-2), and MH763692 (HAstV-8).
ACKNOWLEDGMENTS
This work was partially supported by grants from DGAPA/UNAM (IG200317 and IN208317) and the National Council for Science and Technology-Mexico (CONACYT) (A1-S-9477). This work was also supported by the National Institutes of Health, National Institute of Allergy and Infectious Diseases, grant AI130073 (to R.M.D.).
We thank E. Mata and the animal house personnel for help during mouse immunization and monoclonal antibody production.
REFERENCES
- 1.Bosch A, Pinto RM, Guix S. 2014. Human astroviruses. Clin Microbiol Rev 27:1048–1074. doi: 10.1128/CMR.00013-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Cortez V, Freiden P, Gu Z, Adderson E, Hayden R, Schultz-Cherry S. 2017. Persistent infections with diverse co-circulating astroviruses in pediatric oncology patients, Memphis, Tennessee, USA. Emerg Infect Dis 23:288–290. doi: 10.3201/eid2302.161436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Méndez E, Arias CF. 2013. Astroviruses, p. 1347–1401. In Knipe DM, Howley PM (ed), Fields virology, 6th ed, vol 1 Lippincott Williams & Wilkins, Philadelphia, PA. [Google Scholar]
- 4.Vu DL, Bosch A, Pinto RM, Guix S. 2017. Epidemiology of classic and novel human astrovirus: gastroenteritis and beyond. Viruses 9:33. doi: 10.3390/v9020033. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Olortegui MP, Rouhani S, Yori PP, Salas MS, Trigoso DR, Mondal D, Bodhidatta L, Platts-Mills J, Samie A, Kabir F, Lima A, Babji S, Shrestha SK, Mason CJ, Kalam A, Bessong P, Ahmed T, Mduma E, Bhutta ZA, Lima I, Ramdass R, Moulton LH, Lang D, George A, Zaidi AKM, Kang G, Houpt ER, Kosek MN, MAL-ED Network . 2018. Astrovirus infection and diarrhea in 8 countries. Pediatrics 141:e20171326. doi: 10.1542/peds.2017-1326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Platts-Mills JA, Babji S, Bodhidatta L, Gratz J, Haque R, Havt A, McCormick BJ, McGrath M, Olortegui MP, Samie A, Shakoor S, Mondal D, Lima IF, Hariraju D, Rayamajhi BB, Qureshi S, Kabir F, Yori PP, Mufamadi B, Amour C, Carreon JD, Richard SA, Lang D, Bessong P, Mduma E, Ahmed T, Lima AA, Mason CJ, Zaidi AK, Bhutta ZA, Kosek M, Guerrant RL, Gottlieb M, Miller M, Kang G, Houpt ER, MAL-ED Network Investigators . 2015. Pathogen-specific burdens of community diarrhoea in developing countries: a multisite birth cohort study (MAL-ED). Lancet Glob Health 3:e564–e575. doi: 10.1016/S2214-109X(15)00151-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Arias CF, DuBois RM. 2017. The astrovirus capsid: a review. Viruses 9:15. doi: 10.3390/v9010015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Aguilar-Hernández N, López S, Arias CF. 2018. Minimal capsid composition of infectious human astrovirus. Virology 521:58–61. doi: 10.1016/j.virol.2018.05.021. [DOI] [PubMed] [Google Scholar]
- 9.Dryden KA, Tihova M, Nowotny N, Matsui SM, Mendez E, Yeager M. 2012. Immature and mature human astrovirus: structure, conformational changes, and similarities to hepatitis E virus. J Mol Biol 422:650–658. doi: 10.1016/j.jmb.2012.06.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Toh Y, Harper J, Dryden KA, Yeager M, Arias CF, Mendez E, Tao YJ. 2016. Crystal structure of the human astrovirus capsid protein. J Virol 90:9008–9017. doi: 10.1128/JVI.00694-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.York RL, Yousefi PA, Bogdanoff W, Haile S, Tripathi S, DuBois RM. 2015. Structural, mechanistic, and antigenic characterization of the human astrovirus capsid. J Virol 90:2254–2263. doi: 10.1128/JVI.02666-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bogdanoff WA, Campos J, Perez EI, Yin L, Alexander DL, DuBois RM. 3 January 2017. Structure of a human astrovirus capsid-antibody complex and mechanistic insights into virus neutralization. J Virol 91:e01859-16. doi: 10.1128/JVI.01859-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dong J, Dong L, Mendez E, Tao Y. 2011. Crystal structure of the human astrovirus capsid spike. Proc Natl Acad Sci U S A 108:12681–12686. doi: 10.1073/pnas.1104834108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.DuBois RM, Freiden P, Marvin S, Reddivari M, Heath RJ, White SW, Schultz-Cherry S. 2013. Crystal structure of the avian astrovirus capsid spike. J Virol 87:7853–7863. doi: 10.1128/JVI.03139-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bass DM, Qiu S. 2000. Proteolytic processing of the astrovirus capsid. J Virol 74:1810–1814. doi: 10.1128/JVI.74.4.1810-1814.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sanchez-Fauquier A, Carrascosa AL, Carrascosa JL, Otero A, Glass RI, Lopez JA, San Martin C, Melero JA. 1994. Characterization of a human astrovirus serotype 2 structural protein (VP26) that contains an epitope involved in virus neutralization. Virology 201:312–320. doi: 10.1006/viro.1994.1296. [DOI] [PubMed] [Google Scholar]
- 17.Bass DM, Upadhyayula U. 1997. Characterization of human serotype 1 astrovirus-neutralizing epitopes. J Virol 71:8666–8671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Giammarioli AM, Mackow ER, Fiore L, Greenberg HB, Ruggeri FM. 1996. Production and characterization of murine IgA monoclonal antibodies to the surface antigens of rhesus rotavirus. Virology 225:97–110. doi: 10.1006/viro.1996.0578. [DOI] [PubMed] [Google Scholar]
- 19.Bogdanoff WA, Perez EI, Lopez T, Arias CF, DuBois RM. 2018. Structural basis for escape of human astrovirus from antibody neutralization: broad implications for rational vaccine design. J Virol 92:e01546-17. doi: 10.1128/JVI.01546-17. [DOI] [PMC free article] [PubMed] [Google Scholar]