key: cord-1006821-fwrz779x
authors: Motozono, Chihiro; Toyoda, Mako; Tan, Toong Seng; Hamana, Hiroshi; Aritsu, Yoshiki; Miyashita, Yusuke; Oshiumi, Hiroyuki; Nakamura, Kimitoshi; Okada, Seiji; Udaka, Keiko; Kitamatsu, Mizuki; Kishi, Hiroyuki; Ueno, Takamasa
title: The SARS-CoV-2 Omicron BA.1 spike G446S potentiates HLA-A*24:02-restricted T cell immunity
date: 2022-04-18
journal: bioRxiv
DOI: 10.1101/2022.04.17.488095
sha: ad5668d09a2121eb747839be9eb01b04b2865068
doc_id: 1006821
cord_uid: fwrz779x

Although the Omicron variant of the SARS-CoV-2 virus is resistant to neutralizing antibodies, it retains susceptibility to cellular immunity. Here, we characterized vaccine-induced T cells specific for various SARS-CoV-2 variants and identified HLA-A*24:02-restricted CD8+ T cells that strongly suppressed Omicron BA.1 replication. Mutagenesis analyses revealed that a G446S mutation, located just outside the N-terminus of the cognate epitope, augmented TCR recognition of this variant. In contrast, no enhanced suppression of replication was observed against cells infected with the prototype, Omicron BA.2, and Delta variants that express G446. The enhancing effect of the G446S mutation was lost when target cells were treated with inhibitors of tripeptidyl peptidase II, a protein that mediates antigen processing. These results demonstrate that the G446S mutation in the Omicron BA.1 variant affects antigen processing/presentation and potentiates antiviral activity by vaccine-induced T cells, leading to enhanced T cell immunity towards emerging variants.

) that had been engineered to express 122 spike protein from the prototype, Omicron BA.1 or Delta variants. These target cell 123 lines expressed comparable levels of spike protein, as determined by western blot 124

(Extended Data Fig. 1g) . Interestingly, the level of IFN-γ production by the NF9-125 specific T cell lines from both donors was higher toward target cells expressing 126

Omicron BA.1 spike protein and lower toward cells expressing Delta spike protein, 127 compared to cells expressing the prototype spike protein (Fig. 2a) . On the other 128 hand, the level of IFN-γ production by the QI9-specific T cell lines was comparable 129 against target cells expressing Omicron BA.1, Delta, and prototype spike proteins 130 (Fig. 2b) . The enhanced sensitivity of NF9/A24-specific T cells to Omicron BA.1 131 spike protein, compared to QI9/A24-specific T cells, was maintained at different E:T 132 ratios (Fig. 2c) . Analysis of a further six donors confirmed these results (Fig. 2d of amino acid residues outside the NF9 peptide affect the sensitivity of the target 141 cells to NF9/A24-specific T cells. 142 143

The G446S mutation in the SARS-CoV-2 Omicron BA.1 spike protein is 144 responsible for enhanced NF9/A24-specific TCR recognition 145 We next identified TCR pairs specific for the NF9/A24 and QI9/A24 epitopes by 146 single-cell sorting with NF9/A24 and QI9/A24 tetramers 19 . These studies focused on 147 TCR sequences from 4 donors (Supplementary Table 2 ). Pairs of TCR  and  148 chains were reconstituted in a TCR-deficient NFAT-luciferase reporter cell line. 149

TCRs specific for NF9/A24 (#5-3 and #12-3) and QI9/A24 (#43 and #57) were 150 expressed on the cell surface and bound cognate tetramers at levels comparable to 151 those of T cell lines (Fig. 3a) . Both of the NF9/A24-specific TCR lines responded to 152 cells expressing the Omicron BA.1 spike protein to a greater extent than those 153 expressing the prototype, Alpha (B.1.1.7), Beta (B.1.351), and Gamma (P.1) spike 154 proteins (Fig. 3b) . These data indicate that the level of the NF9 peptide on the cell 155 surface was enhanced in target cells expressing the Omicron BA.1 spike protein. 156 Consistent with this, peptide-titration experiments using the #5-3 TCR indicated that 157 the amount of NF9 peptide expressed on the target cells was almost 3-fold greater 158 on cells expressing Omicron BA.1 spike protein than cells expressing the prototype 159 spike protein (Extended Data Fig. 2a) . Also, the TCR responses were significantly 160 reduced in cells expressing Delta and Lambda (C.37 lineage) spike protein, agreeing 161 with our previous reports that L452R and L452Q in Delta and Lambda, respectively, 162 mediate escape from vaccine-induced NF9/A24-specific T cells 18 . In contrast, both 163 QI9/A24-specific TCRs responded comparably to cells expressing the spike protein 164 from the prototype and variants (Fig. 3b) . 165 It has been shown that mutations adjacent to T cell epitopes in HIV-1 can 166 affect antigen processing and subsequent display to T cells 14, 20, 21, 22 . To 167 determine whether this is also true for SARS-CoV-2 -specific T cells, we introduced 168 mutations adjacent to the NF9 peptide sequence in BA.1 spike protein 1 . There are 169 two mutations, N440K and G446S, located at 8 and 2 amino acids preceding the 170 NF9 peptide sequence in Omicron BA.1 spike protein (Table 2) . Therefore, we 171 sought to examine whether N440K or G446S could enhance epitope recognition by 172 NF9/A24-specific T cells. 173

The introduction of both mutations, or the G446S mutation alone, to the 174 prototype spike protein, resulted in significantly enhanced recognition by NF9/A24-175 specific TCRs. In contrast, the introduction of N440K alone and reversion of the 176 sequence (S446G) in Omicron BA.1 spike protein was recognized by NF9/A24-177 specific TCRs at levels similar to those induced by the prototype spike protein (Fig.  178 3c). 198 QI9-specific T cells from the same donors produced comparable levels of response to stimulation with target cells expressing all spike proteins tested ( Fig.  200 3d). Together, these data suggest that the introduction of a serine at position 446 of 201

Omicron BA.1 spike protein is sufficient to induce enhanced T cell recognition of the 202 NF9 epitope. This enhanced recognition is diminished against Omicron BA.2 spike 203 protein due to the absence of the G446S mutation. 204 205

Tripeptidyl peptidase II (TPPII) inhibitor reduces the enhanced recognition of 206 the NF9 epitope 207

To determine how the G446S mutation affects the antigen processing pathway, we 208 performed a TCR-sensitivity assay using Omicron BA.1 spike protein-expressing 209 target cells pre-treated with a panel of protease inhibitors that are involved in N-210

terminal processing/trimming of the peptide; e.g., bestatin (aminopeptidase inhibitor), 211 butabindide (tripeptidyl peptidase II inhibitor) and ERAP1 inhibitor compound 3 13 . 212

The enhanced sensitivity of NF9/A24-specific TCRs to Omicron BA.1 spike by (GV34 213 #5-3, Vku19 #12-3, and GV34 #2-2) was significantly reduced in the presence of 214 TPPII inhibitor (Fig. 3e, BA.2 (Fig. 4b) .

No suppression was observed against the Delta variant ( Fig. 4b) , 232

presumably due to the T cell escape mutation L452R in the Delta spike protein ( Fig.  233 2c and 2d). On the other hand, QI9-specific T cells comparably suppressed viral 234

replication of prototype and these variants in all three donors tested (Fig. 4c ). We 235 also tested the antiviral activity of T cells from three additional donors and confirmed 236 that NF9-specific T cells have the capacity to inhibit Omicron BA.1, but not BA.2 237 replication to a greater extent than the prototype. These data indicate that vaccine-238 induced T cells can have an enhanced capacity to recognize and suppress emerging 239 SARS-CoV-2 BA.1 variant. 240 241

In this study, we report that vaccine-induced HLA-A*24:02-restricted, NF9-specific T 243 cells efficiently recognize target cells expressing the Omicron BA.1 spike protein and 244

strongly suppress viral replication of the Omicron BA.1 variant compared to the 245 prototype virus. The G446S mutation in Omicron BA.1 spike protein, located 246 adjacent to the N terminus of the NF9 epitope (residues 448-456), is responsible for 247 the efficient generation of the epitope. This is presumably due to enhanced antigen 248 processing and presentation of the epitope. These data indicate that vaccine-249 induced T cells can have an enhanced capacity to cross recognize and suppress 250 emerging SARS-CoV-2 variants. 251

The generation of HLA class I-restricted peptides is profoundly influenced 252 by amino acid variations, not only within, but also around the core epitope. Changes 253

in the epitope-flanking region can result in the inhibition of epitope presentation or a 254 significant increase in the generation of the epitope 22 . In our study, we used a TCR-255

based quantification assay to demonstrate that the presentation of the NF9 epitope 256 on the surface of Omicron BA.1 spike protein-expressing cells was estimated to be 257 almost 3-fold increased relative to that of the prototype. The finding that T cell 258 recognition of NF9 epitope was reduced when the Omicron BA.1 spike protein-259 expressing target cells were pre-treated with butabindide, an inhibitor of TPPII, 260

suggested that generation of NF9 epitope requires TPPII-mediated removal of 2-3 261 amino acids from the N terminus of the peptide as TPPII is known to mediate this 262 process 23 . This finding is consistent with reports that the flanking regions of some 263

HIV-1 epitopes impact proteasomal processing of the epitope 14, 20 . Further studies 264 are needed to clarify how mutations in the spike protein and other proteins in SARS-265

CoV-2 variants affect antigen processing/presentation for T cell recognition, 266

providing better insights for the rational design of vaccine antigens to induce efficient 267 cellular immunity. 268 We and others previously reported that the NF9 is an immunodominant 269 epitope presented by HLA-A*24:02 both in convalescent 15, 24, 25, 26, 27, 28 to the prototype. This is presumably due to the absence of G446S mutation in BA.2 278 and prototype. It will be interesting to see whether vaccine-induced T cells respond 279 comparably, or differently, or control replication of SARS-CoV-2 variants of concern 280

in the context of HLA-A*24:02 in vaccinated donors. 281 IFN-γ ELISpot or AIM (Activation-Induced marker) assays using 282 overlapping peptides are powerful methods to evaluate the breadth of T cell 283 responses to overall viral proteins in vaccinated and COVID-19 convalescent 284

donors 29, 30 . Recent studies using these assays have demonstrated that T cells in 285 vaccinated donors and convalescents can cross-recognize Omicron variants 8, 9, 11, 12 . 286 However, these assays do not reveal antiviral functions of individual T cells against 287 variants of concern, including the Omicron variant and the effect of mutations on 288 antigen processing/presentation in virus-infected cells. Here, we found that a 289 mutation located outside the epitope could enhance the antiviral activity of vaccine-290 induced T cells against the Omicron BA. 

donors are shown. In a, a statistically significant difference versus without T cells (*p 489 < 0.05) was determined by an unpaired two-tailed Student's t-test. In b and c, a 490 statistically significant difference versus prototype (*p < 0.05) was determined by a 491

Mann-Whitney test. ns, no statistical significance. a-c Data are representative of two 492 independent experiments. 493 494 Table 1 Binding of spike-derived epitopes to HLA-A*24:02 495 496 Table 2 Spike-derived HLA-A24-restricted NF9 epitopes and the N-terminal 497 flanking region from the variant 498 Dead cells were stained with 7-aminoactinomycin D (Biolegend, Cat# 420404) . After 586 incubation for 20 min on ice, the cells were fixed with 1% paraformaldehyde (Nacalai 587

Tesque, Cat# 09154-85), and the levels of tetramer + CD8 + T cells were analyzed by 588 flow cytometry using a FACS Canto II (BD Biosciences). The data obtained by flow 589 cytometry were analyzed with FlowJo software (Tree Star). 590 591

Activation-Induced Marker Assay 592

An activation-induced marker assay was performed as previously described 15 

After incubation at 37°C for 24 h, the cells were washed, and surface stained 600 with following antibodies: CD3 FITC (UCHT1)

PerCP/Cy5.5 (HCD14), CD19 PerCP/Cy5.5 (HIB19), CD25 PEcy7 (M-A251) and 602 CD137 APC

Dead cells were stained with

After incubation for 20 min on ice, the cells were fixed 604 with 1% paraformaldehyde (Nacalai Tesque, Cat# 09154-85), and the levels of 605 protein surface expression were analyzed by flow cytometry using a FACS Canto II 606 (BD Biosciences). The data obtained by flow cytometry were analyzed with FlowJo 607 software

Lambda (C.37 612 lineage) and Delta (B.1.617.2), Omicron BA.1 (B.1.1529.1), and Omicron BA.2 613 (B.1.1529.2) variant were prepared in our previous studies 18, 34, 35 . Plasmids 614 expressing the point mutants were generated by site-directed overlap extension 615 PCR using pC-SARS2-spike D614G or SARS2-Omicron-spike as the template and 616 the following primers listed in Supplementary Table 3. Primers for the construction 617 of spike derivatives, related to Fig. 2 and 3. The resulting PCR fragment was 618 digested with KpnI and NotI and inserted into the corresponding site of the pCAGGS 619 vector. Nucleotide sequences were determined by Genetic Analyzer 3500xL 620

Intracellular cytokine staining was performed as previously described 15

5 cells) were transfected with 2 μg of plasmids 626 expressing prototype spike or its derivatives using PEI Max (Polysciences, Cat# 627 24765-1) according to the manufacturer's protocol. At two days post transfection, the 628 transfectants were harvested and mixed with the T cell lines generated from HLA-629

A*24:02 + vaccinated donors (see above) and incubated with RPMI 1640 medium 630 (Thermo Fisher Scientific, Cat# 11875101) containing 10% FBS, 5 μg/ml brefeldin A 631 (Sigma-Aldrich, Cat# B7651) in a 96-well U plate at 37°C for 5 h

CD14 PerCP/Cy5.5 (HCD14), CD19 PerCP/Cy5.5

After incubation at 4°C for 20 min, the cells were fixed and permeabilized 636 with a Cytofix/Cytoperm Fixation/Permeabilization solution kit (BD Biosciences, Cat# 637 554714) and were stained with IFN-γ PE (4S.B3; BD)

The samples for immunoblotting were prepared as described previously 36 with some 644 modifications. Briefly, transfected cells were lysed on ice for 15 min in a buffer (100 645 mM NaCl, 1 mM

The resultant samples were resuspended in 1X Laemmli buffer 648 containing 5% β-Mercaptoethanol (Bio-Rad), boiled for 10 min and subjected to 649 protein separation by SDS-PAGE in 4 -20% Mini-PROTEAN TGX precast gels

The membranes were 651 incubated in a blocking buffer (Nacalai Tesque) for 1 h at room temperature and then 652 mixed with primary antibodies, including rabbit anti-SARS-CoV-2 Spike (S1/S2) 653 polyclonal antibody (1:2,000; Invitrogen) and mouse anti-β-actin monoclonal 654 antibody (1:5,000; Wako), followed by staining with the horseradish peroxidase 655 (HRP)-conjugated anti-rabbit (1:50,000; GE healthcare) and anti-mouse (1:25,000; 656 GE healthcare) IgG secondary antibodies

ImmunoStar LD enhanced chemiluminescence reagents (Wako) and visualized 658 using ImageQuant LAS 400

Jurkat∆-Luc) for functional analysis of TCRs 661 DNA fragment of NFAT-RE-Luc2P-SV40 pro-HygroR was amplified from pGL4.3 662 (Promega) by PCR. The DNA fragment was cloned into Stu I/Sal I site of PiggyBac 663 vector PB530A-2 (SBI) by Gibson assembly method. The resultant vector 664 [PB_NFAT-RE-Luc2P-SV40 pro-HygroR] was electroporated into endogenous TCR 665 knocked-out Jurkat cells 37 with Transposase expression vector

To select Jurkat reporter cell (Jurkat∆-Luc) integrated with NFAT-RE-Luc2P-SV40 667 pro-HygroR, Hygromycin-B selection was performed at 500 ug/ml concentration for 668 14 days

TCR cDNA amplification from single T cells and construction of TCR 671 expression vector 672 The cryopreserved PBMCs were stained with NF9/A24 and QI9/A24 tetramers

) 38 . The amplified TCRα and TCRβ cDNA 680 fragments were connected to the missing constant region and linked to the blasticidin 681 S resistance (BlaR) gene by the Gibson assembly method with P2A ribosomal 682 skipping sequences. Resultant TCRβ-P2A-TCRα-P2A-BlaR DNA was cloned into 683 the PiggyBac vector (SBI, Cat# PB530A-2) by the Gibson assembly method. 684 685 TCR sensitivity assay 686 The plasmid PB TCR-P2A-BlaR was electroporated into Jurkat∆-Luc with 687 Transposase vector (SBI, Cat# PB200PA-1) using Neon® Transfection System 688

cells stably expressing TCRs were selected with RPMI medium 690 containing 10 μg/ml of blasticidin-S for 10-14 days. These cells were cocultured with 691 A549-ACE2-A2402 cells expressing each spike protein an E:T ratio of 2:1 and 692 incubated with RPMI 1640 medium (Thermo Fisher Scientific, Cat# 11875101) 693 containing 10% FBS at 37°C for 6 h. The mixture was measured for luciferase 694 production using a luminescent substrate

Live virus suppression assay 697 A549 cells expressing ACE2/A2402 (1 × 10 4 cells) were infected with each SARS-698

glycerol and 0.4 U/μl recombinant RNase 704 inhibitor (Promega, Cat# N2615) and then incubated at room temperature for 10 min. 705 90 μl of RNase-free water (Nacalai tesque, Cat# 06442-95) was added, and 3 µl of 706 diluted sample was used as the template

For the primer, Primer/Probe N2 710 (2019-nCoV) (Takara, Cat# XD0008) were used as follows: NIID_2019-nCOV_N_ 711 forward, 5-AAATTTTGGGGACCAGGAAC-3

The viral RNA copy number was 714 standardized with a Positive Control RNA Mix (2019-nCoV) (Takara, Cat#XA0142)

Relative viral copy was calculated as viral RNA copy number obtained by virus-716 infected targets without T cells normalized to 1

QUANTIFICATION AND STATISTICAL ANALYSIS 719 Data analyses were performed using Prism 9 (GraphPad Software). Data are 720 presented as median or average with SD