Impact of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Variant-Associated Receptor Binding Domain (RBD) Mutations on the Susceptibility to Serum Antibodies Elicited by Coronavirus Disease 2019 (COVID-19) Infection or Vaccination

Abstract

Background

Several severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) lineages with mutations at the spike protein receptor binding domain (RBD) have reduced susceptibility to antibody neutralization, and have been classified as variants of concern (VOCs) or variants of interest (VOIs). Here we systematically compared the neutralization susceptibility and RBD binding of different VOCs/VOIs, including B.1.617.1 (kappa variant) and P.3 (theta variant), which were first detected in India and the Philippines, respectively.

Methods

The neutralization susceptibility of the VOCs/VOIs (B.1.351, B.1.617.1, and P.3) and a non-VOC/VOI without RBD mutations (B.1.36.27) to convalescent sera from coronavirus disease 2019 (COVID-19) patients or BNT162b2 vaccinees was determined using a live virus microneutralization (MN) assay. Serum immunoglobulin G (IgG) binding to wild-type and mutant RBDs were determined using an enzyme immunoassay.

Results

The geometric mean neutralization titers (GMT) of B.1.351, P.3, and B.1.617.1 were significantly lower than that of B.1.36.27 for COVID-19 patients infected with non-VOCs/VOIs (3.4- to 5.7-fold lower) or individuals who have received 2 doses of BNT162b2 vaccine (4.4- to 7.3-fold lower). The GMT of B.1.351 or P.3 were lower than that of B.1.617.1. For the 4 patients infected with B.1.351 or B.1.617.1, the MN titer was highest for their respective lineage. RBD with E484K or E484Q mutation, either alone or in combination with other mutations, showed greatest reduction in serum IgG binding.

Conclusions

P.3 and B.1.617.1 escape serum neutralization induced by natural infection or vaccine. Infection with 1 variant does not confer cross-protection for heterologous lineages. Immunogenicity testing for second generation COVID-19 vaccines should include multiple variant and “nonvariant” strains.

Since the beginning of the coronavirus disease 2019 (COVID-19) pandemic, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has evolved into many lineages [1, 2]. Viruses carrying spike D614G mutation have spread globally since March 2020 [3]. A single novel lineage can cause large outbreaks in places with low incidence. For example, the B.1.1.63 lineage and B.1.36.27 lineage dominated the third and fourth wave in Hong Kong, respectively [4, 5].

The spike protein receptor binding domain (RBD) is responsible for binding to host cell receptor angiotensin-converting enzyme 2 (ACE2) and is the major target of neutralizing antibodies [6]. RBD mutations may increase infectivity and transmissibility, confer an increased risk of reinfection, and reduce vaccine efficacy [7]. In late 2020 and early 2021, variants with N501Y mutations at the RBD were reported to spread rapidly in the United Kingdom (B.1.1.7), South Africa (B.1.351), and Brazil (P.1). The B.1.1.7 variant (alpha variant; VOC 202012/01) has been associated with increased transmissibility and mortality in epidemiological studies [8, 9]. B.1.1.7 variant has better fitness than earlier D614G lineages in vitro and replicates better in the upper respiratory tract in a hamster model [10]. The B.1.351 (beta variant; VOC 202012/02) and P.1 (gamma variant; B.1.1.28.1; VOC202101/02) variants have additional mutations at spike amino acid residues 417 and 484, and have been found to be 4.5- to 8.6-fold less susceptible to neutralization by vaccine or natural infection induced antibodies [11]. The P.3 (theta-variant; B.1.1.28.3) variant, which has the same ancestor as P.1 and carries mutation N501Y and E484K, is prevalent in the Philippines [12]. At the time of writing, 80% of the P.3 lineage viruses were found in the Philippines [13]. The World Health Organization has classified B.1.1.7, B.1.351, and P.1 as variants of concern (VOCs), and P.3 as a variant of interest (VOI) [14].

Apart from variants with N501Y mutation, viruses with spike L452R mutation have also been found. The variants B.1.427/B.1.429 (epsilon variant) with L452R spread rapidly in California and have reduced neutralizing antibody titer [15]. Since March 2021, there has been a rapid increase in the incidence of COVID-19 in India [16]. The number of cases exceeded 400 000 per day. This massive outbreak was associated with a novel lineage B.1.617 with L452R and/or E484Q mutation at the spike RBD. On 11 May 2021, the World Health Organization has declared B.1.617 as the fourth VOC [14]. B.1.617.1 (kappa variant) was later reclassified as a VOI on 31 May 2021. This study systematically compared the effect of neutralization and immunoglobulin G (IgG) binding that are conferred by these variants or RBD mutations. We included serum specimens from recovered COVID-19 patients, including a patient with reinfection we reported previously [17], or BNT162b2 messenger RNA (mRNA) vaccine recipients, including 2 with prior COVID-19.

METHODS

Clinical Specimens

Serum specimens were collected from recovered COVID-19 patients of Queen Mary Hospital or Princess Margaret Hospital and from BNT162b2 vaccine recipients. COVID-19 patients were diagnosed by reverse transcription polymerase chain reaction (RT-PCR) at the Clinical Microbiology Laboratory of Queen Mary Hospital or at the Public Health Laboratory Centre of Hong Kong. BNT162b2 vaccine recipients received the vaccine in Hong Kong. For COVID-19 patients, serum specimens were randomly selected for live virus microneutralization assay (MN) or RBD assay. For vaccine recipients, all recruited patients as of 15 May 2021, were included. Written informed consent was obtained.

All SARS-CoV-2 viruses were isolated from respiratory specimens collected from patients in Hong Kong (Figure 1). The B.1.36.27 lineage virus was isolated from the nasopharyngeal swab of a 4-year old male in February 2021, who acquired the virus locally (GISAID EPI_ISL_2423555). The B.1.351 lineage virus was isolated from the posterior oropharyngeal saliva of a 43-year-old female who traveled from the Philippines to Hong Kong in March 2021 (GISAID EPI_ISL_2423556). The P.3 lineage virus was isolated from the combined nasopharyngeal/throat swab of a 37-year-old female who traveled from the Philippines in January 2021 (GISAID EPI_ISL_2423558). The B.1.617.1 lineage virus was isolated from a nasopharyngeal swab specimen from a 2-year-old male patient who returned from India in April 2021 (GISAID EPI_ISL_2423557). Written informed consent was waived by the IRB committee for the use of archived clinical specimens for viral culture. This study was approved by Institutional Review Board of the University of Hong Kong/ Hospital Authority Hong Kong West Cluster (HKU/HA HKW IRB) (UW 13–265 and UW 21–214) and the Kowloon West Cluster REC (KW/EX-20–038[144-26]).

Virus Culture and Live Virus Microneutralization Assay

Viral culture and live virus MN assay were conducted in the Biosafety Level 3 facility at the University of Hong Kong as we described previously [18]. Viral culture was conducted using TMPRSS2-overexpressing VeroE6 cells (VeroE6/TMPRSS2 cells) (JCRB Cell Bank Catalogue no. JCRB1819) to avoid mutations which may arise during culture in normal VeroE6 cells.

Live virus neutralization assay was performed as we described previously with modifications [18, 19]. Briefly, serum specimens were serially diluted in 2-folds starting from 1:10. Duplicates of each serum dilution was mixed with 100 TCID₅₀ of B.1.36.27, B.1.351, P.3 or B.1.617.1 virus isolates for 1 hour, and the serum-virus mixture was then added to VeroE6/TMPRSS2 cells. After incubation for 3 days, cytopathic effect was visually scored for each well by 2 independent observers. The results were plotted using 5-parameter dose-response curve in GraphPad Prism version 9.1.1. A value of 5 was assigned if cytopathic effect was seen at a dilution of 1:10.

Expression and Purification of RBD

Recombinant RBD (residues 306–543) of SARS-CoV-2 spike protein from the reference sequence Wuhan-Hu-1 (GenBank ID YP_009724390.1) (wild type) or with the mutations N501Y, N501Y-E484K-K417N, L452R-E484Q, E484K, L452R or N439K were expressed and purified in insect cells as we described previously with modifications [20] (Supplementary Figure 1). Briefly, RBD gene sequences were baculovirus-codon-optimized and cloned into pFast dual baculovirus expression vector. The constructs were fused with an N-terminal gp67 signal peptide and C-terminal 6×His tag for secretion and purification. A recombinant bacmid DNA was generated using the Bac-to-Bac system (Thermo Fisher Scientific). Baculovirus was produced by transfecting purified bacmid DNA into Sf9 cells using Cellfectin (Thermo Fisher Scientific) and subsequently used to infect ExpiSf9 cell suspension culture (Thermo Fisher Scientific) at a multiplicity of infection of 1 to 10. Infected ExpiSf9 cells were incubated at 27.5° C with shaking at 125 rpm for 96 hours for protein expression. The supernatant was collected and then concentrated using a 10 kDa MW cutoff Labscale TFF System (Millipore). The RBD protein was purified by Ni-NTA purification system, followed by size exclusion chromatography, and buffer exchanged into 1× phosphate-buffered saline (PBS) pH 7.4. The concentration of purified RBD was determined by using the Bradford Assay Kit (Bio-Rad) according to the manufacturer’s instructions. The purity of recombinant RBD mutants were verified by western blotting (Supplementary Figure 2).

Anti-RBD Assay for Wild Type and Variants

Enzyme immunoassay for anti-RBD IgG antibody was performed as we described previously with modifications [21, 22]. Briefly, 96-well immunoplates (Nunc Immuno modules; Nunc, Denmark) were coated with 100 μL/well (0.2 μg/well) of His-tagged SARS-CoV-2 spike RBD in 0.05 M carbonate bicarbonate buffer (pH 9.6) overnight at 4° C and then followed by incubation with a blocking reagent. After blocking, 100 μL heat-inactivated serum samples at 1:200 dilution were added to the wells and incubated at 37° C for 1 hour. The attached human and mouse antibodies were detected using horseradish-peroxidase-conjugated goat anti-human IgG and anti-mouse IgG antibody respectively (Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts, USA). The reaction was developed by adding diluted 3,3′,5,5′-tetramethylbenzidine single solution and stopped with 0.3 N H₂SO₄. The optical density (OD) was read at 450 and 620 nm. For normalization, mouse monoclonal antibody against His tag (ABclonal, ABclonal, Inc., Woburn, Massachusetts, USA) was diluted in a series of 2-fold dilution from 1:3000, and the EC₅₀ for each RBD was determined using the 5-parameter logistic equation.

Whole Viral Genome Sequencing and Bioinformatics Analysis

Whole viral genome sequencing was performed using nanopore sequencing following the Nanopore protocol PCR tiling of COVID-19 (Version: PTC_9096_v109_revH_06Feb2020) according to the manufacturer’s instructions with minor modifications (Oxford Nanopore Technologies) as we described previously [4, 5, 23]. For bioinformatics analysis, the recommended ARTIC bioinformatics workflow was used with minor modifications applied as described previously [4, 5]. Please refer to Supplementary methods for details.

Statistics

All statistical analysis was performed using PRISM version 9.1.1. The MN titers or normalized OD values were compared between different viruses or RBDs using 1-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test. A P value of <.05 was considered statistically significant.

RESULTS

Neutralizing Antibody Titers Against SARS-CoV-2 Variants

The neutralizing antibody titer of serum specimens was determined using a live virus neutralization assay. The variant viruses included the VOC lineage B.1.351, and the VOI lineages B.1.617.1 and P.3. A D614G virus within lineage B.1.36.27 was used as control. B.1.36.27 is the dominant lineage causing the 4th wave of COVID-19 in Hong Kong between November 2020 and April 2021 [5] and does not have any amino acid mutations at the spike RBD when compared with the reference genome Wuhan/Hu/1 (Supplementary Table 1).

For the 8 patients who had 1 episode of infection with non-VOC/VOI strains (Supplementary Table 2), the MN titer against B.1.36.27 was higher than VOC/VOI strains for 7 patients (Figure 2A). The geometric mean titer (GMT) of B.1.351 and P.3 was 5.7-fold and 5.1-fold lower than that of B.1.36.27. The GMT of B.1.617.1 was 3.4-fold lower than that of B.1.36.27, almost reaching statistical significance (P = .066). The GMT against P.3 was significantly lower than that of B.1.617.1 (P = .0339).

Next, we tested the serum of 12 individuals who have received 2 doses of BNT162b2 and without prior infection. At 21 or 28 days after the first dose, 6 individuals (50%) had detectable levels of MN antibody against B.1.36.27 (Figure 2B), but 3 vaccinees did not have detectable MN antibody against the VOC/VOIs B.1.351, P.3 or B.1.617.1. The GMT of B.1.351, P.3 and B.1.617.1 were lower than that of B.1.36.27 after 1 dose of vaccine, though not statistically significant. For the 12 individuals who had received 2 doses of vaccines, all had detectable MN antibody against B.1.36.27 (Figure 2C). The GMT of B.1.351, P.3 and B.1.617.1 was 7.3-fold, 5.9-fold, and 4.4-fold lower than that of B.1.36.27 after 2 doses of vaccine, all reaching statistical significance (P < .0001). The GMT of B.1.351 was also significantly lower than that of B.1.617.1 (P = .0234).

We also tested serum from 1 patient with reinfection (both episodes due to non-VOC/VOIs) and 2 vaccine recipients who had prior infection. For the patient with reinfection and 1 of the vaccine recipients with prior infection, both B.1.351 and P.3 had lower titer than the 2 other strains (Figure 3A). For the remaining vaccine recipient, the MN antibody titers were similar for all 4 lineages.

For the 4 patients infected with VOCs/VOIs, the 2 patients infected with B.1.351 had highest MN titer against B.1.351, whereas the 2 patients infected with B.1.617.1 had highest titer against B.1.617.1 (Figure 3B). Notably, for 1 of the patients infected with B.1.617.1, the MN titer against B.1.617.1 virus (MN titer: 269) was at least 7-fold higher than other viruses tested (MN titer: 20–38).

RBD Binding

Next, we assessed the binding of convalescent-phase or vaccine serum for recombinant RBDs using enzyme immunoassay (Supplementary Table 3). The recombinant RBD tested include the wild-type virus, the VOCs (N501Y, N501Y-E484K-K417N), a VOI (E484K, representative of VOI B.1.525 [eta variant]; L452R; L452R-E484Q), and a mink-associated mutation, N439K [24].

For sera from non-VOC/VOI patients, wild-type and N501Y RBD had similar OD values, and both had significantly higher OD than all other mutants (P ≤ .0001 for all comparisons) (Figure 4A). L452R and N439K RBD had similar ODs, and both were significantly higher than the RBDs that contain mutation at spike amino acid residue 484 (N501Y-E484K-K417N, E484K, L452R-E484Q). In particular, the mean OD of L452R (1.411; standard error of the mean [SEM], 0.263) was 1.26-fold higher than that of L452R-E484Q (1.141; SEM, 0.263). There was no significant differences between the OD values for 3 lineages that contain mutation at spike amino acid residue 484. The results for BNT162b2 vaccine recipient was similar to that of non-VOC/VOI patients, in that wild-type and N501Y RBD had the highest OD, followed by L452R and N439K, with the E484K, L452R-E484Q, and N501Y-E484K-K417N RBD having the lowest OD (Figure 4B).

We also tested the sera from 9 patients with VOCs/VOIs, including the sera from the 4 patients tested with neutralization assay. For serum from 6 patients infected with B.1.617.1 lineage virus, L452R-E484Q and L452R had similar OD as wild type and N501Y, but L452R had significantly higher OD than N501Y-E484K-K417N (P = .0171) (Figure 4C). For the 3 patients with B.1.351, the OD was higher for N501Y-E484K-K417N RBD than for L452R or L452R-E484K RBD (Figure 4D).

Discussion

SARS-CoV-2 variants with RBD mutations are particularly worrisome as these are more likely to escape humoral immunity induced by natural infection or COVID-19 vaccine. Previous studies suggested that B.1.351 and P.1 variants have reduced susceptibility to neutralization by antibodies induced by vaccine or natural infection due to non-VOC/VOI viruses [25]. However, the data on B.1.617.1, which emerged in India, and P.3, which emerged in the Philippines, are scarce. In this study, we simultaneously compared neutralization susceptibility of B.1.617.1, P.3, and B.1.351 (all carrying mutations at amino acid position 484) with a D614G virus without RBD mutations. We also determined the effect of RBD binding due to different mutations at amino acid residue 417, 439, 452, 484, and 501, which are present alone or in combination among VOCs or VOIs. We showed that both B.1.617.1 and P.3 confer reduced susceptibility to sera from mRNA vaccinee or from COVID-19 patients infected with non-VOC/VOI strains. Notably, the reduction of P.3 is similar to that of B.1.351, while the reduction of B.1.617.1 is slightly less than those of P.3 or B.1.351. Furthermore, we showed that mutation at spike amino acid mutation 484, either alone or in combination with other RBD mutations, confer the greatest reduction in serum IgG binding.

We showed that B.1.617.1 with L452R-E484Q mutation had 3.4-fold and 4.4-fold lower MN titer than B.1.36.27 for non-VOC/VOI patients and vaccine recipients, respectively. Our results concur with recent studies, which showed a reduction of neutralizing antibody titer ranging from 1.94-fold to 7.17-fold for vaccine recipients [26–28], and 1.96- to 6.5-fold reduction for convalescent sera [24, 28]. Our RBD binding assay showed that both the single L452R or the double L452R-E484Q mutations affect IgG binding, but the magnitude of reduction is greater for L452R-E484Q. Previous study also showed that E484K or L452R alone reduced the binding of some monoclonal antibodies [29]. The mutation L452R is also present in the B.1.427/B.1.429 lineages from California, which are more resistant to neutralization by serum than nonvariant viruses [15]. Taken together, both L452R and E484Q are important in contributing to the reduced susceptibility to neutralization.

Our data showed that the neutralization titer against P.3 is similar to that of B.1.351 for both vaccinee and non-VOC/VOI COVID-19 patients, although P.3 lacks the K417N mutation in B.1.351. In addition to the spike RBD mutations, P.3 lineage virus also contains the spike NTD mutation 141-143del, and the furin cleavage site mutation P681H, which is also found in the B.1.1.7 lineage [14]. Whether these mutations help the virus to escape neutralization remain to be determined.

B.1.617.1, B.1.351, and P.3 carry mutation at spike amino residue 484, and our RBD binding assay showed that RBDs with mutation at residue 484 had the greatest reduction in IgG binding from patients or vaccinees. E484K alone reduced the binding of RBD to human ACE2 and affected the binding of some monoclonal antibodies [29]. Previous study on B.1.525 lineage, which carries E484K without other spike RBD mutations, had a slightly lower neutralizing antibody titer than the wild-type virus for BNT162b2 vaccinees or convalescent sera [30, 31].

For patients infected with B.1.617.1 or B.1.351, their serum MN titers against the virus from their respective lineages were much higher than other lineages, including the B.1.36.27 lineage without any RBD mutations. Our result concurs with the study by Cele et al [32], which showed that sera from patients infected with B.1.351 had reduced neutralization against earlier viruses. The reduced cross-neutralization against heterologous viruses, including non-VOC/VOI strains, has implications on reinfection and vaccine design. First, patients infected B.1.617.1 or P.3 lineage viruses are still susceptible to reinfection from viruses without RBD mutations or other RBD mutations. Second, testing for second generation COVID-19 vaccines should include viruses from ancestral D614G lineage and variant lineages with different RBD mutations.

For the patient with reinfection due to D614G and a vaccine recipient with prior COVID-19 infection, their MN titers against B.1.617.1 was similar to those of B.1.36.27, although the MN titers against B.1.351 and P.3 were lower. Because there have not been any locally acquired cases with L452R variants in Hong Kong, it is unlikely that these individuals have prior exposure to B.1.617.1 lineage viruses.

N439K mutation confers reduced serum IgG binding when compared with wild type. N439K is found in different SARS-CoV-2 lineages, including mink-associated human cases [33]. Our findings corroborate with previous studies that showed reduced susceptibility to monoclonal antibodies [33].

There are several limitations in this study. First, we only tested 1 virus from each lineage. There may be differences in neutralization susceptibility among viruses within the same lineage. Second, there are relatively few patients infected with variants. Third, we have not examined the binding to spike N-terminal domain or S2 subunit, which are also targets of monoclonal neutralizing antibodies [6, 34, 35]. Fourth, we currently do not have any clinical data regarding reinfection or vaccine efficacy against B.1.617.1 and P.3. Finally, we did not assess T-cell immunity of these patients, which is also important for protection.

In conclusion, our results suggest that both B.1.617.1 and P.3 are less susceptible to neutralization, which may affect vaccine effectiveness and the risk of reinfection. Although many recovered COVID-19 patients or vaccinees still have relatively high titers of neutralizing antibody against the variants, the decline of antibody level over time may predispose these patients to be more susceptible to variants. Mutations at the spike amino acid residue 484, located at the key region of antibody binding [36], has greatest effect on antibody binding. Antibody induced by natural infection from a variant virus may not protect variants with other RBD mutations. Hence, immunogenicity testing for newer generation COVID-19 vaccines should evaluate viruses from lineages carrying different spike mutations.

Notes

Acknowledgments.The authors gratefully acknowledge the originating and submitting laboratories who contributed sequences to GISAID (Supplementary Table 4).

Financial support.This study was partly supported by Health and Medical Research Fund (HMRF), the Food and Health Bureau, The Government of the Hong Kong Special Administrative Region (Ref No. COVID190124); Consultancy Service for Enhancing Laboratory Surveillance of Emerging Infectious Diseases and Research Capability on Antimicrobial Resistance for Department of Health of the HKSAR; and donations of Richard Yu and Carol Yu, Shaw Foundation Hong Kong, Michael Seak-Kan Tong, May Tam Mak Mei Yin, Lee Wan Keung Charity Foundation Limited, Hong Kong Sanatorium & Hospital, Respiratory Viral Research Foundation Limited, Hui Ming, Hui Hoy and Chow Sin Lan Charity Fund Limited, Chan Yin Chuen Memorial Charitable Foundation, Marina Man-Wai Lee, the Hong Kong Hainan Commercial Association South China Microbiology Research Fund, the Jessie & George Ho Charitable Foundation, Kai Chong Tong, Tse Kam Ming Laurence, Foo Oi Foundation Limited, Betty Hing-Chu Lee, and Ping Cham So. The funding sources had no role in study design, data collection, analysis, interpretation, or writing of the report.

Potential conflicts of interest.The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

References

Author notes