key: cord-0960318-x7ug65i7
authors: Seephetdee, Chotiwat; Bhukhai, Kanit; Buasri, Nattawut; Leelukkanaveera, Puttipatch; Lerdwattanasombat, Pat; Manopwisedjaroen, Suwimon; Phueakphud, Nut; Kuhaudomlarp, Sakonwan; Olmedillas, Eduardo; Saphire, Erica Ollmann; Thitithanyanont, Arunee; Hongeng, Suradej; Wongtrakoongate, Patompon
title: Broad neutralization of SARS-CoV-2 variants by circular mRNA producing VFLIP-X spike in mice
date: 2022-03-28
journal: bioRxiv
DOI: 10.1101/2022.03.17.484759
sha: bd8a698ffe5f3be02e88be210c581a54dd441484
doc_id: 960318
cord_uid: x7ug65i7

Next-generation COVID-19 vaccines are critical due to the ongoing evolution of SARS-CoV-2 virus and waning duration of the neutralizing antibody response against current vaccines. The mRNA vaccines mRNA-1273 and BNT162b2 were developed using linear transcripts encoding the prefusion-stabilized trimers (S-2P) of the wildtype spike, which have shown a reduced neutralizing activity against the variants of concern B.1.617.2 and B.1.1.529. Recently, a new version of spike trimer, termed VFLIP has been suggested to possess native-like glycosylation, and greater pre-fusion trimeric stability as opposed to S-2P. Here, we report that the spike protein VFLIP-X, containing six rationally substituted amino acids to reflect emerging variants (K417N, L452R, T478K, E484K, N501Y and D614G), offers a promising candidate for a next-generation SARS-CoV-2 vaccine. Mice immunized by a circular mRNA (circRNA) vaccine prototype producing VFLIP-X elicited neutralizing antibodies for up to 7 weeks post-boost against SARS-CoV-2 variants of concern (VOCs) and variants of interest (VOIs). In addition, a balance in TH1 and TH2 responses was achieved by immunization with VFLIP-X. Our results indicate that the VFLIP-X delivered by circRNA confers humoral and cellular immune responses, as well as neutralizing activity against broad SARS-CoV-2 variants.

Implementing next-generation vaccines and therapeutics is vital to combating and controlling continuous evolution and ongoing global transmission of SARS-CoV-2 variants of concern (VOCs) and variants of interest (VOIs). The first-generation prefusion-stabilized spike engineered by two proline substitutions (S-2P) has served as an initial immunogen in current SARS-CoV-2 vaccines, presenting one or more RBDs "up" [1] [2] [3] . As serum neutralizing activity is primarily directed to the prefusion spike, structure-based engineering that better stabilizes the prefusion state could elicit more potent neutralizing antibody responses 4 . Importantly, immunodominant neutralizing epitopes are mainly located in the RBD [4] [5] [6] . However, multiple RBD mutations in emerging variants resulted in changes in immunodominance hierarchy and impaired neutralization potency of RBD-directed neutralizing antibodies 7, 8 . Specifically, common mutations, including substitutions at positions 417, 452, 484 and 501, have been identified as major drivers of antigenic differences 8, 9 . Other RBDtargeted neutralizing monoclonal antibodies are unaffected by mutations in the RBD of emerging variants [10] [11] [12] . Moreover, the utilization of RBD as an immunogen in several vaccine platforms has shown promising clinical immunogenicity results [13] [14] [15] [16] . Thus, rationally engineered SARS-CoV-2 RBD and spike proteins could serve as potential immunogens to elicit potent protective immune responses.

Next-generation vaccine design strategies have been suggested to engineer antigens with enhanced stability and native-like quaternary structure, conformational dynamics, and glycosylation profiles 17, 18 . A recently described spike protein, namely VFLIP (five (V) prolines, Flexibly-Linked, Inter-Protomer disulfide), has been engineered by using five proline substitutions in the S2 subunit, a flexible S1/S2 linker, and two cysteine substitutions to introduce an inter-protomer disulfide bond formation 19 . In this work, one proline substitution at residue 986, shared between S-2P and the second-generation HexaPro, was converted back to lysine (K986) restoring native salt bridge formation between K986 and D427. The VFLIP spike displays improved thermostability and native RBD motion. In addition, glycosylation patterns of VFLIP are more similar to the authentic virus than previously engineered spike constructs. Moreover, immunization of VFLIP subunit candidate vaccine in mice results in superior immunogenicity compared to S-2P. When combined with next-generation vaccine platforms, such as mRNA, the antigen design strategy might pave the way for improved, cross-variant neutralization of SARS-CoV-2 20 .

The inherent instability of linear mRNA transcripts impedes their utilization in vaccines as longer half-life of mRNA confers improved immunogenicity. Unlike linear counterparts, circRNAs impart greater stability due to the covalently closed structure, rendering protection from degradation by exonucleases. Efficient circularization of long RNAs is achieved through a permuted intron-exon splicing strategy with other assisting elements, including homology arms and spacer sequences 21 . To produce translatable exogenous circRNAs, these RNAs have been engineered to comprise IRES elements of several viruses (e.g., EMCV, CVB3, PV, etc.) placed prior to coding sequence.

Importantly, circRNAs elicit more durable protein expression as compared to their linear cognates. In this work, we utilized the circRNA platform to engineer the spike protein VFLIP-X, which is VFLIP containing six rationally substituted amino acids. We show in mice that the circRNA vaccine prototype producing VFLIP-X elicited neutralizing antibodies for up to 7 weeks post-boost against SARS-CoV-2 VOCs and VOIs. VFLIP-X delivered by circRNA also induces favorable humoral and cellular immune responses. Together, our work highlights the potential of a SARS-CoV-2 circRNA vaccine expressing VFLIP-X spike as a next-generation COVID-19 vaccine.

We initially confirmed circRNA-driven in vivo protein expression by intramuscularly injecting BALB/c mice with LNP-formulated circRNA encoding firefly luciferase (FLuc). At 24 and 48 hours after the administration, we clearly observed bioluminescence in those mice (Fig. S1 ), suggesting that the circRNA template is applicable for development of an mRNA vaccine prototype.

To develop a circRNA expressing SARS-CoV-2 spike as a vaccine prototype capable of neutralizing broad SARS-CoV-2 variants, the full-length, membrane-bound version of the recently engineered VFLIP spike possessing native-like glycosylation was chosen 19 . Further, a substitution of six amino acids was rationally selected based on the co-mutation (D614G) found in all SARS-CoV-2 variants as well as five mutations (K417N, L452R, T478K, E484K and N501Y) co-identified in several variants of concern (VOCs) and variants of interest (VOIs) (Fig. 1A) . The selected six mutations and the amino acid substitutions pertinent to the originally reported VFLIP spike were structurally presented in the spike trimers ( Fig 1B) . We named this spike construct VFLIP-Cross (VFLIP-X). An unrooted phylogenetic tree was constructed to visualize relationships among amino acid sequences derived from the wildtype isolate containing the six rationally substituted amino acids (wildtype-X) and SARS-CoV-2 VOC and VOI spikes. Compared to VOC and VOI spikes, wildtype-X spike is found at a distinct clade among those spike variants, and is more related to Omicron (B.1.1.529) spike variant (Fig. 1C) . The result suggests that the six rationally substituted amino acids confer a unique sequence feature, yet interrelated to VOC and VOI spikes.

To ascertain whether the VFLIP-X spike can be expressed as a full-length spike in vitro, protein expression was determined in HEK293T transfected with a circRNA prototype encoding the spike protein. We also compared expression of the VFLIP-X spike to a membrane-bound, prefusionstabilized spike containing the six rationally substituted amino acids (HexaPro-X). Similar expression levels of full-length HexaPro-X and VFLIP-X were determined using western blot, immunofluorescence staining and flow cytometry analyses using a polyclonal anti-RBD antibody ( Fig.   1D -1F), indicating that the circRNA prototype is capable of producing the spike antigens.

To elucidate immunogenicity of the circRNA vaccine prototype, we formulated circRNAs expressing HexaPro-X or VFLIP-X with LNPs, and determined the physicochemical properties of the circRNA-LNPs. We confirmed similar average size, polydispersity index, and encapsulation efficiency of circRNA-LNPs producing either HexaPro-X or VFLIP-X (Table S1 ). Next, seven-week-old female BALB/c mice were intramuscularly immunized with 5 μg of the circRNAs using a prime-boost regimen separated by a 3-week interval. Immunization with VFLIP-X induced high B.1.1.529 S-binding serum IgG titers similar to HexaPro-X ( Fig. 2A) . Importantly, we found that only sera from mice vaccinated by VFLIP-X elicits high neutralization against B.1.1.529 pseudovirus. In contrast, HexaPro-X induced little to no neutralizing titers (Fig. 2B) , the result of which is consistent with a previous finding using the membrane-bound version of wildtype HexaPro spike 22 . Therefore, only the circRNA expressing VFLIP-X was thus chosen for subsequent experiments.

To determine whether VFLIP-X provides a cross-variant neutralizing activity, sera samples from mice administered with 1 or 5 μg of VFLIP-X 7 weeks post-boost were collected and tested for neutralizing activity against live SARS-CoV-2 variants (wildtype, B. 

The humoral and cellular immune responses were assessed in BALB/C mice immunized with 1 or 5 μg of VFLIP-X in a series of experiments. Two-dose immunization of VFLIP-X induced high B.1.1.529 S-binding serum IgG titers in a dose-dependent manner (Fig. 4A ). The binding IgG titers were increased one week after the second immunization and sustained up to 7 weeks post-boost.

The potent B.1.1.529 pseudovirus-neutralizing activity was observed in mice receiving 5 μg of VFLIP-X ( Fig. 4B ). Although the binding IgG titers were detected at 2 weeks after the first immunization, the neutralizing titers were detectable after the second immunization. At 5 weeks post-boost, the levels of neutralizing antibodies were increased compared to one-week post-boost and sustained up to 10 weeks after the first immunization. These results demonstrate that the cross-neutralizing VFLIP-X spike expressed under the circRNA vaccine candidate is a potent immunogen.

Next, we characterized B.1.1.529 S specific cellular immune response in mice at 7 weeks post-boost. Enzyme-linked immunospot (ELISpot) assay was performed using isolated splenocytes from mice that received 1 or 5 μg of VFLIP-X restimulated with S peptide mixture to evaluate specific T cell responses. Following restimulation with B.1.1.529 S peptide pool, we found that splenocytes isolated from the immunization groups elicited strong IFN-γ, but low IL-4, responses (Fig. 4C ). This result suggests that the vaccine candidate does not induce T H 2-biased responses. We further elucidated the balance of T H 1 and T H 2 responses by comparing levels of IgG2a and IgG1, which are surrogate markers for T H 1 and T H 2 responses, respectively. Mice receiving 5 μg of VFLIP-X produced robust IgG2a and IgG1 levels, as shown by B.1.1.529 S-specific ELISA (Fig. 4D) . Together, the IgG subclass and ELISpot profiles demonstrate that the immunization with VFLIP-X led to the induction of type 1 antiviral response rather than a T H 2-skewed response.

mRNA vaccines formulated with the first-generation wildtype spike antigen have proven to be highly effective against the ancestral SARS-CoV-2 variant. However, the emergence of variants of concern, especially the heavily mutated B. and E484 are positioned in the "peak" subsection of SARS-CoV-2 spike. K417 and L452 are positioned in the "valley" subsection. N501 is positioned in the "mesa" subsection 7 . They have been shown to either enhance affinity between the spike protein and host ACE2 receptor or escape immunity derived from natural infection, vaccination, and monoclonal antibodies 7, 27 . The mutations K417N/L452R/T478K are found in Delta plus variant. Structurally, the aliphatic-to-positively charged mutation at T478K has been suggested to enhance binding between the Delta variant S and ACE2 via K478-Q24 interaction 28, 29 . K417N, T478K, E484A and N501Y have been demonstrated to participate in Omicron's immune evasion 30 . Even though E484A caused more escape from D614G sera than E484K as determined by a deep mutational scanning study, the effect of single substitution at E484K in significantly poor neutralization is similar to that observed in human influenza viruses; a single amino acid substitution confers a large portion of the immunogenic difference 8 .

Following a prime-boost intramuscular immunization, VFLIP-X appeared to induce strong pseudovirus-neutralizing antibody response against B.1.1.529, whereas mice immunized with HexaPro-X generated little to no neutralizing antibody (Fig. 2) . Thus, antigen expression and binding IgG titers were not predictive of neutralizing titers. Similar results have been observed by vaccination using an mRNA vaccine encoding full-length, HexaPro spike 22 . By using VFLIP-X, potent neutralizing activity against emerging variants was elicited by a 5 μg dose (Fig. 3) . We observed a robust Notably, neutralizing antibodies have been recently shown to mainly target prefusion conformation, suggesting that vaccines utilizing prefusion-stabilized spikes might elicit greater neutralizing titers 4 . In contrast to that notion, spike proteins kept in closed conformation, including VFLIP, have been demonstrated to elicit more potent neutralizing responses than the more opened conformation spikes S-2P and HexaPro 19, 38 . In addition, the VFLIP spike displays more native-like glycosylation profiles than other prefusion-stabilized spikes, presumably better preserving the antigenicity of spike immunogen 19,39,40 . Therefore, a balance in prefusion-stabilized and metastable states, opened and closed structures, as well as glycosylation profiles might be required when revising next-generation vaccines against SARS-CoV-2.

Our VFLIP-X candidate vaccine evaluated in this work elicited a balanced T H 1-T H 2 response, suggesting the protective immunity induction rather than enhanced disease (Fig. 4C-D) . are consistent with other studies on mRNA vaccines against SARS-CoV-2 45, 46 . Altogether, this work highlights the potential of a SARS-CoV-2 circRNA vaccine expressing a highly stable VFLIP-X spike immunogen as a next-generation COVID-19 vaccine preventing emerging SARS-CoV-2 variants.

The sequence encoding SARS-CoV-2 spike with mutations in RBD (K417N, L452R, T478K, E484K, and N501Y), D614G, and structural changes within S1/S2 cleavage site and S2 subunit amino acid sequences derived from wildtype spike, SARS-CoV-2 spike variants of concern (VOCs) and variants of interest (VOIs), and wildtype-X spike (wildtype spike with the six rationally substituted amino acids) was constructed by maximum likelihood using IQ-TREE software.

The circRNAs were produced by T7 RNA polymerase-based in vitro transcription as described previously 47 

Transfected HEK293T cells were harvested by scraping at 24-hour post-transfection and lysed with NP-40 lysis buffer supplemented with protease inhibitor and PMSF. Proteins were collected from supernatant after centrifugation at 12,000 rpm for 20 minutes. The concentration of protein was determined by BCA protein assay kit (Millipore). The heated protein samples were then analyzed by SDS-PAGE using 10% acrylamide/bis-acrylamide and western blot. The proteins were transferred to a PVDF membrane using a wet transfer system. Blotted proteins were detected with a rabbit polyclonal antibody that recognizes SARS-CoV-2 spike RBD (40592-T62, SinoBiological) and a donkey anti-rabbit IgG-HRP antibody (sc-2077, Santa Cruz Biotechnology).

To detect surface protein expression, transfected HEK293T cells were stained with a rabbit polyclonal antibody that recognizes SARS-CoV-2 spike RBD (40592-T62, SinoBiological) and an Alexa Fluor 488-conjugated goat anti-rabbit IgG (A11034, Invitrogen). Cells were then acquired on a BD Accuri C6 plus and analyzed by FlowJo software version 10.6.2.

Transfected HEK293T cells were fixed with 4% paraformaldehyde (PFA). Cells were then blocked with 20% FBS in PBS and incubated with a rabbit polyclonal antibody that recognizes SARS-CoV-2 spike RBD (40592-T62, SinoBiological) followed by an Alexa Fluor 594-conjugated goat anti-rabbit IgG (A11037, Invitrogen). DNA was stained with Hoechst. Images were acquired with a fluorescence microscope.

Mouse 

The production of lentivirus-based SARS-CoV-2 pseudovirus and neutralization assay were performed as described previously 48 . Briefly, the SARS-CoV-2 pseudovirus were produced by cotransfection of plasmids pHAGE-CMV-Luc2-IRES-ZsGreen-W, HDM-Hgpm2, HDM-tat1b, pRC-CMV-Rev1b, and SARS-CoV-2 spike expressing plasmid into HEK293T cells using jetPRIME transfection reagent.

The determination of 50% neutralization titer (pVNT 50 ) of immunized mouse sera were performed in HEK293T-hACE2 cells. Cells were seeded in white wall 96-well plate (12,500 cells/well).

After 24 hours of incubation, the mouse sera were serially 2-fold diluted starting at 1:40, incubated with 4x10 6 relative light unit (RLU)/ml SARS-CoV-2 pseudovirus at 37°C for 1 hour, and added to each well. After 48 hours, luciferase assay was performed using Bright-Glo Luciferase Assay (Promega) and RLU were determined by Cytation7 Cell Imaging Multi-Mode Reader (Bio Tek).

Neutralization titers were defined as the reciprocal serum dilution at which RLU were reduced by 50% compared to the virus control wells after subtraction of background RLU in cell control wells.

The microneutralization assay was performed as described previously with some modifications 49 

Mouse presented as GMT ± geometric SD. Horizontal dotted lines represent assay limits of detection. Control group was compared with 5 μg immunized group by student's t-test (parametric, two-tailed unpaired).

* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. Table S1 . Physicochemical properties of LNP-RNAs (data are presented as mean ±SD, n=3).

Stabilized coronavirus spikes are resistant to conformational changes induced by receptor recognition or proteolysis

Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen

Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation

SARS-CoV-2 spike conformation determines plasma neutralizing activity. bioRxiv

Mapping Neutralizing and Immunodominant Sites on the SARS-CoV-2 Spike Receptor-Binding Domain by Structure-Guided High-Resolution Serology

The antigenic anatomy of SARS-CoV-2 receptor binding domain

Defining variant-resistant epitopes targeted by SARS-CoV-2 antibodies: A global consortium study

Mapping SARS-CoV-2 antigenic relationships and serological responses. bioRxiv

A SARS-CoV-2 variant elicits an antibody response with a shifted immunodominance hierarchy

Antibody evasion by the P.1 strain of SARS-CoV-2

Structural basis for accommodation of emerging B.1.351 and B.1.1.7 variants by two potent SARS-CoV-2 neutralizing antibodies

Broadly neutralizing antibodies overcome SARS-CoV-2 Omicron antigenic shift

Phase I/II study of COVID-19 RNA vaccine BNT162b1 in adults

Safety and immunogenicity of a recombinant tandem-repeat dimeric RBDbased protein subunit vaccine (ZF2001) against COVID-19 in adults: two randomised, double-blind, placebo-controlled, phase 1 and 2 trials. The Lancet Infectious Diseases 21

Phase I interim results of a phase I/II study of the IgG-Fc fusion COVID-19 subunit vaccine, AKS-452

Safety and immunogenicity of the SARS-CoV-2 ARCoV mRNA vaccine in Chinese adults: a randomised, double-blind, placebo-controlled, phase 1 trial

SARS-CoV-2 variants, spike mutations and immune escape

Structure-Based Vaccine Antigen Design. Annual review of medicine 70

Structure-based design of a highly stable, covalently-linked SARS-CoV-2 spike trimer with improved structural properties and immunogenicity. bioRxiv

Stopping pandemics before they start: Lessons learned from SARS-CoV-2

Engineering circular RNA for potent and stable translation in eukaryotic cells

Immunogenicity and efficacy of mRNA COVID-19 vaccine MRT5500 in preclinical animal models. npj Vaccines 6

mRNA-1273 or mRNA-Omicron boost in vaccinated macaques elicits comparable B cell expansion, neutralizing antibodies and protection against Omicron. bioRxiv

Replicating RNA platform enables rapid response to the SARS-CoV-2 Omicron variant and elicits enhanced protection in naïve hamsters compared to ancestral vaccine. bioRxiv

Boosting with Omicron-matched or historical mRNA vaccines increases neutralizing antibody responses and protection against B.1.1.529 infection in mice. bioRxiv

Omicron-specific mRNA vaccine induced potent neutralizing antibody against Omicron but not other SARS-CoV-2 variants. bioRxiv

SARS-CoV-2 variants and effectiveness of vaccines: a review of current evidence

Molecular basis of immune evasion by the Delta and Kappa SARS-CoV-2 variants

A Computational Dissection of Spike protein of SARS-CoV-2 Omicron Variant. bioRxiv

Structural basis of SARS-CoV-2 Omicron immune evasion and receptor engagement

Neutralization of SARS-CoV-2 Variants B.1.429 and B.1.351

SARS-CoV-2 immune evasion by the B.1.427/B.1.429 variant of concern

Neutralization of the SARS-CoV-2 Mu Variant by Convalescent and Vaccine Serum

A cluster of the new SARS-CoV-2 B.1.621 lineage in Italy and sensitivity of the viral isolate to the BNT162b2 vaccine

Low Neutralizing Antibody Titers against the Mu Variant of SARS-CoV-2 in 31 BNT162b2 Vaccinated Individuals in Colombia

SARS-CoV-2 Lambda Variant Remains Susceptible to Neutralization by mRNA Vaccine-elicited Antibodies and Convalescent Serum. bioRxiv

The SARS-CoV-2 Lambda variant exhibits enhanced infectivity and immune resistance

SARS-CoV-2 Spike Protein Stabilized in the Closed State Induces Potent Neutralizing Responses

Glycosylation and Serological Reactivity of an Expression-enhanced SARS-CoV-2 Viral Spike Mimetic

Molecular Architecture of the SARS-CoV-2 Virus

COVID-19 vaccine mRNA-1273 elicits a protective immune profile in mice that is not associated with vaccine-enhanced disease upon SARS-CoV-2 challenge

Immunization with SARS Coronavirus Vaccines Leads to Pulmonary Immunopathology on Challenge with the SARS Virus

RESPIRATORY SYNCYTIAL VIRUS DISEASE IN INFANTS DESPITE PRIOR ADMINISTRATION OF ANTIGENIC INACTIVATED VACCINE12

In contrast to T<sub>H</sub>2-biased approaches, T<sub>H</sub>1 COVID-19 vaccines protect Syrian hamsters from severe disease in the absence of dexamethasone-treatable vaccine-associated enhanced respiratory pathology. bioRxiv

SARS-CoV-2 mRNA vaccine design enabled by prototype pathogen preparedness

BNT162b vaccines protect rhesus macaques from SARS-CoV-2

Engineering circular RNA for potent and stable translation in eukaryotic cells

Protocol and Reagents for Pseudotyping Lentiviral Particles with SARS-CoV-2 Spike Protein for Neutralization Assays

Mice Immunized with the Vaccine Candidate HexaPro Spike Produce Neutralizing Antibodies against SARS-CoV-2. Vaccines (Basel)

Structure-based design of prefusion-stabilized SARS-CoV-2 spikes

This research project was supported by Mahidol University and Ramathibodi Foundation.