Tracking of Spike Protein Amino Acid Polymorphisms in Portugal
Using publicly available epidemiological surveillance data (https://covidcg.org/?tab=group), single non-synonymous substitutions in the SARS-CoV-2 Spike protein in Portugal were tracked. Figure 1-A shows the percentage of virus strains with polymorphisms among all strains in Portugal. In Portugal, the single amino acid substitutions with the highest frequencies were D614G (91.5%), D839Y (12.8%), A222V (12.3%), P1162R (3.5%), S477N (1.6%), and L176F (1.6%). Therefore, we screened six strains with single-site mutation frequencies greater than 1.5% and further investigated combinations of mutations at these sites.
Tracking of Co-mutations in Strains in Portugal
Using epidemiological data (https://www.epicov.org/epi3/frontend#1bd174), combined mutations in Portugal and the six most frequent single amino acid substitutions were tracked. The mutant strain B.1.1.7 spread rapidly in the United Kingdom in December and to various European countries, including Portugal. Among mutant strains that appeared in Portugal in December, the UK strain B.1.1.7 accounted for 73.1%; this strain had joint L216F and M740V mutations. At the same time, the variant B.1.258 with B.1.258 + L1063F and B.1.258 + N751Y also appeared. Figure 1-B shows the counts and changes in virus strains with a high frequency of naturally occurring joint mutations in Portugal. From December 31, 2020 to March 3, 2021, D614G was still the most frequent mutation (94.9%), followed by A222V + D614G (24.0%), B.1.1.7 (15.5%), D839Y + D614G (7.3%), P1162R + D614G + A222V (5.2%), S477N + D614G (5.0%), and L176F + D614G (1.3%). The frequencies of the four variants D614G, A222V + D614G, B.1.1.7, S477N + D614G, and P1162R + D614G + A222V showed an increasing trend, among which the increase in B.1.1.7 was second only to D614G. The frequencies of D839Y + D614G and L176F + D614G in Portugal showed a decreasing trend. The counts of the B.1.1.7 + L216F, B.1.1.7 + M740V, B.1.258, B.1.258 + L1063F, and B.1.258 + N751Y variants did not increase.
Infectivity analysis
Infectivity of mutant pseudoviruses in various cells
To evaluate the infectivity of natural variants in Portugal, we infected susceptible cells (Huh7, hACE2, Vero, and LLC-MK2 with 12 mutant pseudoviruses. As shown in Fig. 2-A, infectivity (as evaluated by RLU values) for all pseudoviruses was highest in Huh7 cells, followed by hACE2 cells, while Vero and LLC-MK2 cells had relatively low infectivity. Figure 2-B shows the infectivity of mutants in descending order of the mutation frequency in Portugal. For the four types of cells, the RLU values for D839Y + D614G and B.1.258 were lower than those for D614G, and the infectivity of P1162R + D614G + A222V, L216F + B.1.1.7 was about 2 times higher than that of D614G. On the whole, estimates of infectivity for the 11 natural mutants were similar to the infectivity of D614G (setting a 4-fold increase as the threshold for a significant difference).
Infectivity of different mutant pseudoviruses in cells with receptor overexpression
SARS-CoV-2 infection occurs by S protein binding to the host cell ACE2 receptor, thereby infecting cells. To further evaluate the infectivity of the natural mutant strains in Portugal, we infected HEK-293T cells overexpressing ACE2 receptors with 12 natural mutant strains and evaluated RLU.
First, we constructed 14 cell lines expressing ACE2 receptors of different species. After transient transfection, flow cytometry was used to detect the expression of ACE2 in each cell. We used untreated HEK-293T cells as a negative control. The rate of ACE2 protein expression was 20.8–43.8%, with an average of 32.3%.
We also infected these 14 cells overexpressing ACE2 receptors with 12 natural mutant pseudoviruses. As shown in Fig. 3, infectivity did not differ by more than four-fold among cells overexpressing ACE2 receptors of different species. All pseudoviruses showed higher infectivity in 293T cells overexpressing mouse ACE2 receptor than in other cells. In cells overexpressing rabbit ACE2 and dog ACE2, all pseudoviruses showed generally low infectivity.
Neutralizing activity
Neutralizing antibody detection in serum after immunization
The distribution of the 12 natural mutations in the SARS-CoV-2 S protein in Portugal evaluated in the study is shown in Fig. 4-A. Currently, most monoclonal antibodies (mAb) are directed against the receptor binding domain (RBD) of the SARS-CoV-2 S protein. Therefore, the impact of the N439K, S477N, N501Y, and A570D mutations on mAb protection were evaluated, and the effects of mutations at other sites on transmission and infectivity of the virus should be further studied.
First, we synthesized different segments of SARS-CoV-2 S peptides, including RBD, S1, and S2, and use these peptides as a vaccine to immunize BALB/c mice to obtain post-immunization serum. We also used the SARS-CoV-2 S full-length plasmid as a DNA vaccine to immunize BALB/c mice. We used these sera and 12 pseudoviruses to evaluate protective effects of the neutralizing antibody (Fig. 4-B). Using the neutralization ID50 titers for the D614G pseudovirus and different mouse sera for reference, the neutralization activity was evaluated for the other 11 pseudoviruses. The serum neutralized 11 pseudoviruses with no immune escape (defined as a 4-fold reduction in the ID50 value compared with that for D614G).
Monoclonal antibody neutralization activity
To determine the specific sites affecting the antigenicity of the virus in Portugal, we constructed seven pseudoviruses, S477N, D839Y, L176F, L216F + D614G, M740V + D614G, L1063F + D614G, and N751Y + D614G. All pseudoviruses constructed in this paper were treated with 11 mAbs. The IC50 values for the 11 monoclonal antibodies were similar (i.e., approximately 0.03–0.18 µg/mL), with no significant differences (P > 0.05).
As shown in Fig. 5, the epidemic strain S477N + D614G (accounting for 5.0% of strains) and its corresponding single-site mutant in Portugal can escape the monoclonal antibody 09-7B8. Accordingly, escape could be attributed to the S477N mutation (using D614G as the standard and a reduction in neutralizing activity of > 4 times as the threshold for immune escape). We also found that the UK epidemic strain B.1.1.7 and its variants can escape the monoclonal antibodies 03-1F9, 2H10, 03-10D12-1C3, 03-10F9-1A2, 11D12-1, CB6, and HB27. However, no immune escape was found for M740V + D614G and L216F + D614G. Therefore, escape could mainly be explained by the mutation in B.1.1.7. In addition to escape from the monoclonal antibody HB27 by the UK strain, immune escape was also observed for B.1.258 and its variants. Compared with the neutralizing effects of L1063F + D614G and N751Y + D614G, escape can be explained by the mutation in B.1.258.