Evolution of a Functionally Intact but Antigenically Distinct DENV Fusion Loop

A hallmark of Dengue virus (DENV) pathogenesis is the potential for antibody-dependent enhancement, which is associated with deadly DENV secondary infection, complicates the identification of correlates of protection, and negatively impacts the safety and efficacy of DENV vaccines. ADE is linked to antibodies targeting the fusion loop (FL) motif of the envelope protein, which is completely conserved in mosquito-borne flaviviruses and required for viral entry and fusion. In the current study, we utilized saturation mutagenesis and directed evolution to engineer a functional variant with a mutated FL (D2-FL) which is not neutralized by FL-targeting monoclonal antibodies. The FL mutations were combined with our previously evolved prM cleavage site to create a mature version of D2-FL (D2-FLM), which evades both prM- and FL-Abs but retains sensitivity to other type-specific and quaternary cross-reactive (CR) Abs. CR serum from heterotypic (DENV4) infected non-human primates (NHP) showed lower neutralization titers against D2-FL and D2-FLM than isogenic wildtype DENV2 while similar neutralization titers were observed in serum from homotypic (DENV2) infected NHP. We propose D2-FL and D2-FLM as valuable tools to delineate CR Ab subtypes in serum as well as an exciting platform for safer live attenuated DENV vaccines suitable for naïve individuals and children.

of all four serotypes. However, creating formulations that elicit a balanced response has 44 proven challenging. 8 Additionally, lab-grown strains differ from patient-derived DENVs in 45 both maturation status and antigenicity. 9 In particular, Abs targeting the fusion loop (FL) 46 have been reported to neutralize lab and patient strains with differing strengths and have 47 been observed to facilitate Fcγ-receptor uptake in vitro and therefore ADE. 10-12 Currently, 48 there is a single FDA-approved DENV vaccine, Dengvaxia. However, it is only approved 49 for use in individuals aged 9-16 with previous DENV infection living in endemic areas and 50 is contraindicated for use in naïve individuals and younger children. In naïve children, 51 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint vaccination stimulated non-neutralizing CR Abs that increased the risk of severe disease 52 after DENV infection. 13,14 Other DENV vaccines have been tested or are currently 53 undergoing clinical trial, but thus far none have been approved for use in the United 54 States. 15 The vaccine Qdenga has been approved in the European Union, Indonesia, and 55 Brazil, although vaccine efficacy in adults, naïve individuals, and with all serotypes has 56 not yet been shown. 57 The DENV FL is located in Envelope (E) protein domain II (EDII) and is involved 58 in monomer-monomer contacts with EDIII. 16 During the DENV infection cycle, low pH 59 triggers a conformational change in the E protein. 17 The structure of the virion rearranges,  Figure 1A). Although the extreme conservation and critical role in entry have led 67 to it being considered extremely difficult to change the FL, we successfully tested the 68 hypothesis that massively parallel directed-evolution could produce viable DENV FL 69 mutants that were still capable of fusion and entry, while altering the antigenic footprint.  CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

75
To engineer a virus with a novel antigenic footprint at the FL, we targeted the core 76 conserved FL motif. We generated two different saturation mutagenesis libraries, each 77 with 5 randomized amino acids: DRGXGXGXXXFGK (Library 1; AA 101, 103, 105-107) 78 and DRGXXXXXGLFGK (Library 2 AA 101-105). Library 1 was designed to mutate known 79 residues targeted by FL mAbs while Library 2 focused on a continuous linear peptide that supporting the importance of these residues. 16 When modeled on the pre-fusion DENV2 94 structure, the N103S and G106L mutations are located at the interface with the 95 neighboring monomer EDIII domain, protected from the aqueous environment. In the 96 post-fusion form, the two residues are located between W101 and F108 and form the 97 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint bowl concavity above the chlorine ion in the post-fusion trimer ( Figure 1D). We used 98 reverse genetics to re-derive the FL N103S/G106L/T171A mutant, which we term D2-FL. 99 As enhancing Abs also target prM, 19 we also created a mature version of D2-FL termed  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint by ~2-logs in both variants, 1N5 neutralization is reduced by ~1-log for D2-FL and reduced 121 to background levels for D2-FL, and no neutralization was observed for 1L6 or 4G2 for 122 either variant ( Figure 3A). 18 Focusing on the D2-FLM virus containing both evolved motifs, 123 we then characterized the antigenicity of the whole virion with a panel of mAbs. As 124 expected, D2-FLM was unable to be neutralized by the prM Abs 1E16 and 5M22; the Ab 125 2H2 does not neutralize either DV2-WT or D2-FLM ( Figure 3B). For Abs targeting 126 epitopes in non-mutated regions, including the ED1 and EDE epitopes that target EDII 127 and EDIII, FRNT50 values were generally comparable, although EDE1-C10 shows a 128 moderate but statistically significant reduction between DV2-WT and D2-FLM, indicating 129 that the overall virion structural integrity was intact ( Figure 3B).  (Table 1). However, in two NHPs infected with DENV4, strong 139 neutralization potency (FRNT50 between 1: 100 -1:1,000) was demonstrated against   include strain selection and serotype balance. 42 In the current study we used DENV2 203 S16803, a prototype for DENV2. 43 However, S16803 was isolated several decades ago, 204 and it may be beneficial to utilize more contemporaneous strains. 8,44 Work is currently 205 ongoing to demonstrate the portability of the evolved FL motif on additional DENV2 206 strains and other serotypes, which is essential for tetravalent vaccine production. D2-FLM 207 was highly attenuated in Vero cells, creating a challenge for vaccine production.  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint ACKNOWLEDGEMENTS 215 We thank members of the Tse, Baric, and DeSilva laboratories for helpful discussions.

249
DENV2 S16803 was used in this study. Recombinant viruses were created using a four-    Titers of viral supernatant were determined using a standard DENV focus-forming assay.

289
In brief, cells were seeded at 2x10 4 cells per well of a 96-well plate one day before (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint TrueBlue HRP substrate (SeraCare) and counted using an automated Immunospot 300 analyzer (Cellular Technology).

301
Thermal Stability Assay

302
The indicated viruses were thawed and incubated at temperatures ranging from 4°C to 303 55°C for one hour. Following, viral titers were determined by focus-forming assay as 304 described above.

336
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

362
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint 363 . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 4, 2023. ; https://doi.org/10.1101/2023.03.22.533803 doi: bioRxiv preprint