Protective Zika vaccines engineered to eliminate enhancement of dengue infection via immunodominance switch

Dai, Lianpan; Xu, Kun; Li, Jinhe; Huang, Qingrui; Song, Jian; Han, Yuxuan; Zheng, Tianyi; Gao, Ping; Lu, Xuancheng; Yang, Huabing; Liu, Kefang; Xia, Qianfeng; Wang, Qihui; Chai, Yan; Qi, Jianxun; Yan, Jinghua; Gao, George F.

doi:10.1038/s41590-021-00966-6

Download PDF

Article
Published: 15 July 2021

Protective Zika vaccines engineered to eliminate enhancement of dengue infection via immunodominance switch

Nature Immunology volume 22, pages 958–968 (2021)Cite this article

12k Accesses
23 Citations
59 Altmetric
Metrics details

Subjects

Abstract

Antibody-dependent enhancement (ADE) is an important safety concern for vaccine development against dengue virus (DENV) and its antigenically related Zika virus (ZIKV) because vaccine may prime deleterious antibodies to enhance natural infections. Cross-reactive antibodies targeting the conserved fusion loop epitope (FLE) are known as the main sources of ADE. We design ZIKV immunogens engineered to change the FLE conformation but preserve neutralizing epitopes. Single vaccination conferred sterilizing immunity against ZIKV without ADE of DENV-serotype 1–4 infections and abrogated maternal–neonatal transmission in mice. Unlike the wild-type-based vaccine inducing predominately cross-reactive ADE-prone antibodies, B cell profiling revealed that the engineered vaccines switched immunodominance to dispersed patterns without DENV enhancement. The crystal structure of the engineered immunogen showed the dimeric conformation of the envelope protein with FLE disruption. We provide vaccine candidates that will prevent both ZIKV infection and infection-/vaccination-induced DENV ADE.

Potent Zika and dengue cross-neutralizing antibodies induced by Zika vaccination in a dengue-experienced donor

Article 03 February 2020

Vincent Dussupt, Rajeshwer S. Sankhala, … Shelly J. Krebs

Prior flavivirus immunity skews the yellow fever vaccine response to cross-reactive antibodies with potential to enhance dengue virus infection

Article Open access 24 February 2024

Antonio Santos-Peral, Fabian Luppa, … Simon Rothenfusser

A protective Zika virus E-dimer-based subunit vaccine engineered to abrogate antibody-dependent enhancement of dengue infection

Article 02 September 2019

Jose Luis Slon-Campos, Wanwisa Dejnirattisai, … Gavin R. Screaton

Main

ZIKV is a mosquito-borne pathogen belonging to the genus Flavivirus in the family Flaviviridae¹. The 2015–2016 ZIKV epidemic spread to 84 countries worldwide, including China^2,3. ZIKV infection during pregnancy can lead to miscarriage and catastrophic congenital Zika syndrome in neonates, including microcephaly and neurodevelopmental disorders^4,5,6, and is also associated with Guillain–Barré syndrome in adults⁷. Evidence so far has demonstrated that ZIKV can break through barriers from blood to brain, placenta, testes and eyes to cause pathogenicity⁸. We and others found that ZIKV can cause testis damage and male infertility in a mouse model^9,10. Nevertheless, no prophylactics or therapeutics are available, highlighting the need for vaccine development. With global efforts to develop vaccines, a notable concern has been raised that ZIKV vaccine may prime cross-reactive antibodies to enhance natural infections of the antigenically related DENV via ADE^11,12,13,14. Antibodies that can bind flavivirus particles without neutralizing activities will cause ADE via Fcγ receptor-mediated virus uptake^14,15. Extensive ADE was observed in heterotypic secondary DENV infection and can explain severe dengue disease in human¹⁶. ADE has been the principal hurdle to DENV vaccine development and is implicated as a limitation of the first licensed dengue vaccine, Dengvaxia, which is recommended only for DENV-seropositive, not DENV-seronegative, individuals due to more frequent severe clinical outcomes^11,14,17,18. ADE also complicated ZIKV vaccine development itself. Mounting evidence has demonstrated that pre-existing ZIKV antibodies significantly enhance DENV infection in vitro and aggravated dengue disease in vivo using both mouse and macaque models^12,13,19,20. More importantly, a recent study of pediatric cohorts in Nicaragua clearly shows that previous ZIKV infection enhanced future risk of DENV2 disease and severity, as well as DENV3 severity, in humans^21,22. This highlights that an effective and safe ZIKV vaccine is highly needed to (1) prevent homogenous ZIKV infection; (2) prevent future severe DENV infection caused by natural ZIKV infection and (3) prevent vaccine-induced DENV ADE. In particular, ZIKV infection was found to activate ADE-prone antibodies de novo in DENV-experienced individuals²³. Therefore, the possible risk for DENV ADE should be taken into account for ZIKV vaccine design, especially, given that ZIKV/DENV are transmitted by same vectors (usually Aedes aegypti) and endanger people living in same regions¹¹.

ADE-prone antibodies target mainly epitopes on precursor membrane protein (prM) and fusion loop (FL) of envelope protein (E)^24,25 in DENV. These antibodies are cross-reactive between serotypes or serocomplexes of flavivirus due to epitope conservation, and are usually poorly neutralizing²⁶. More potent neutralizing antibodies (NAbs) recognize other epitopes^{26,27,28,29,30}. For ZIKV, neutralizing monoclonal antibodies (mAbs) have been reported to recognize epitopes on domains I (DI), DII and DIII, or quaternary epitopes on the virus surface^{19,27,28,31,32}. Strategies to prevent induction of ADE-prone antibodies have been described recently for the design of new ZIKV vaccines^11,13. Either mutation¹³ or minimal exposure¹¹ of the FL epitope (FLE) was demonstrated to reduce the ADE of DENV infection in mice. However, after ZIKV challenge, substantial virus burden can still be detected in sera¹³, or in targeted tissues including spleen, brain, placenta and fetal head^11,13. Given the many detrimental effects observed on fetuses and neonates following ZIKV infection in pregnant women, sterilizing immunity may be required for vaccine development to prevent long-term sequelae³³.

Here, we described an antibody-to-vaccine development process. By extracting insights from the structural basis of FLE mAbs, we engineered a ZIKV immunogen to (1) eliminate the FLE while (2) still presenting potent neutralizing epitopes. Introducing mutations to key residues of FLE could easily cause immunogen impairment, and this obstacle was overcome by homologous replacement by naturally occurring variants from arthropod-specific flaviviruses. The resultant immunogens, MutB/C, were developed as ZIKV vaccines. Impressively, a single vaccination was completely protective and sterilizing, with elimination of ADE for DENV in mice, and prevented maternal–fetal transmission against ZIKV challenge. By analysis of vaccine-induced B cell pools at single-cell level, we found that, unlike the wild-type (WT) construct inducing predominantly cross-reactive ADE-prone antibodies, MutB/C-based vaccines switched immunodominance to ZIKV-specific patterns without DENV enhancement. The crystal structure of the immunogen further revealed the molecular basis for the vaccine potency and ADE abrogation.

Results

ZIKV vaccines expressing either prM/E or M/E induce DENV ADE

We previously described a protective ZIKV vaccine using chimpanzee adenovirus type 7 (AdC7) as a vector to express ZIKV-M/E³⁴. Since many of the developing ZIKV vaccines are based on prM/E³³, we also generated AdC7 expressing prM/E for comparison. These two vaccines were given to groups of BALB/c mice. Four weeks later, sera were first assessed for neutralizing activities against ZIKV infection, showing comparable magnitudes of neutralizing titers between these two groups (Extended Data Fig. 1a). Next, sera were evaluated for their ability to enhance DENV infection on K562 cells bearing the Fcγ receptor IIA (CD32A). Notably, dilutions of sera from mice immunized with either AdC7-M/E or AdC7-prM/E dramatically enhanced the infections of all four serotypes of DENV, implicating potential safety concerns (Extended Data Fig. 1b–e).

Rational design of ZIKV immunogens

We previously isolated a group of FLE mAbs from a ZIKV-infected patient³¹. A representative mAb, Z6, was evaluated for ADE of DENV infection. As expected, Z6 promoted all four serotypes of DENV infections as bell-shaped curves (Extended Data Fig. 1f). To better understand the FLE to guide vaccine design, we determined the crystal structure of Z6 bound to ZIKV-E at a resolution of 3.0 Å (Supplementary Table 1) as a glimpse at the first structure of a human FLE mAb. Z6 straddled FL from the top of DII with balanced interactions with both heavy and light chains (Fig. 1a,b and Extended Data Fig. 2a,e). Comparing the three structures of known FLE mAbs (Z6, 2A10G6 and E53)^35,36, we found they approached FL with different angles, contacting residues within W101–K110 with the help from the bc loop (Fig. 1a,b and Extended Data Fig. 2). The structural analyses pinpoint potential residues for engineering to destroy the FLE.

**Fig. 1: Engineering ZIKV immunogens.**

Next, we sought to engineer ZIKV immunogen. M/E was chosen as template because (1) it was previously demonstrated as a protective immunogen^34,37,38, and (2) it avoided the precursor peptide—another main target for ADE-prone antibodies^24,25,39. To characterize the immunogens antigenically, mAbs bound to different regions of ZIKV-E were used as probes. Z6 and 2A10G6 bound to FLE, whereas Z3L1, Z20 and Z23 were ZIKV-specific neutralizing mAbs that bound epitopes on DI, DII and DIII, respectively³¹ (Fig. 1c).

Given that residue W101 in FL was demonstrated previously to be bound by most FLE antibodies^26,40 and contributed most interactions to Z6 and 2A10G6 (Fig. 1b and Extended Data Fig. 2e), we therefore attempted to mutate W101 to reduce FLE. As expected, W101 mutated to the other 19 amino acids, except W101F, abrogated the interactions with FLE mAbs (Extended Data Fig. 3a,b). Nevertheless, these mutations dramatically reduced antigen expression and epitope stability. In particular, the DI/DII epitopes for Z3L1 and Z20 were barely probed (Extended Data Fig. 3b), implying a sacrifice of immunogenicity. Therefore, we investigated other strategies.

Structural analysis showed that, except for W101, other FL residues, including G106, L107 and W108 (Fig. 1b and Extended Data Fig. 2) are also important for FLE. We therefore sought to replace the entire FL. FL amino acids are highly conserved within the genus Flavivirus, but have variations in distantly related species such as arthropod-specific flaviviruses (Fig. 1d and Extended Data Fig. 4a). We assumed these naturally occurring FL variants may be structurally stable due to evolutionary selection. Therefore, we substituted FL residues (D98–G109) of ZIKV with their homologs in arthropod-specific flaviviruses, namely MutA, MutB and MutC (Fig. 1d). Five mutations were introduced in each construct (D98N, N103T, G106F/L, L107E/K and F108W) (Fig. 1d). These mutants exhibited no discernable binding to FLE mAbs, yet substantial antigen expression displaying neutralizing epitopes probed by Z3L1, Z23 and Z20 (Fig. 1e). Antigen expression can also be detected in supernatant from transfected cells by capture enzyme-linked immunosorbent assay (ELISA) (Fig. 1f). Due to the relatively low antigen expression of MutA (Fig. 1e,f), we focused on the other two designs, MutB and MutC, for further immunogen evaluation.

We generated AdC7-M/E-MutB and AdC7-M/E-MutC as vaccines. Both vaccines had substantial antigen expression of correctly folded both intracellular and secreted antigen when infecting cells (Extended Data Fig. 5). We immunized groups of BALB/c mice with AdC7-M/E-MutB or AdC7-M/E-MutC. AdC7-M/E-WT and PBS were given as controls; 4 weeks later, mice sera were collected for measurement. We observed comparable ZIKV-E-reactive serologic responses between groups of mice receiving vaccines based on WT, MutB or MutC, respectively (Fig. 1g). Moreover, sera from the WT group were highly cross-reactive to four serotypes of DENV; however, sera from MutB/C groups showed little or no cross-reactions to DENV1–4 (similar to sham) (Fig. 1g).

MutB/C vaccines induced protective immunity without DENV ADE

To evaluate vaccine efficacy, serum samples were further assessed for neutralizing activities against ZIKV infection (SMGC-1 strain). Impressively, MutB/C-based vaccines elicited magnitudes of neutralizing titers comparable to those elicited by the WT form (Fig. 2a). High levels of ZIKV-E-binding antibodies and NAbs were detected 12- and 24-weeks postvaccination until our last serum sampling (Extended Data Fig. 4b,c). To further explore the prophylactic efficacy of FL-engineered vaccines, IFNAR1-deficient (Ifnar1^−/−) mice were used as a sensitive model for ZIKV challenge^34,41. At 4 weeks after a single-dose immunization, mice receiving either the WT- or MutB/C-based vaccines developed high levels of ZIKV NAbs (Extended Data Fig. 4d). At day 30 postvaccination, they were challenged with a lethal dose (5 × 10⁶ focus-forming units (FFU)) of ZIKV-SMGC-1. As expected, all mice receiving the sham vaccine succumbed to ZIKV challenge by 8 days postinfection (DPI), with severe weight loss starting from 4 DPI (Fig. 2b,c). In contrast, all mice receiving the WT- or MutB/C-based vaccines showed no mortality or morbidity (Fig. 2b,c). To validate the protective efficacy of vaccines against viremia, virus burden in blood was measured at both 3 DPI and 6 DPI. As expected, mice with sham vaccination displayed sustainable viremia before death. In contrast, all mice receiving WT- or MutB/C-based vaccines showed no detectable viremia (Fig. 2d). To further test protection efficacy of the vaccines against virus burden in ZIKV target tissues, a challenge experiment was performed. A single-dose vaccination elicited binding and neutralizing antibodies (Extended Data Fig. 6a,b). At both days 3 and 7 postchallenge, mice were euthanized for quantification of ZIKV RNA in tissues. Like the WT form, the MutB/C-based vaccines protected mice, with undetectable viral RNA in all tissues tested, including brain, spinal cord, testis, spleen, liver and eye, whereas high viral burdens were detected in sham-vaccinated animals (Fig. 2e and Extended Data Fig. 6c). Given that the FL-engineered vaccines completely protected mice against ZIKV viremia and tissue infection, we further tested whether they conferred sterilizing immunity. Sera collected pre- and post-ZIKV challenge were analyzed for neutralizing activities. Mice immunized with the MutB-based vaccine exhibited a lack of anamnestic responses, as did the WT form, indicating sterilizing immunity (Fig. 2f).

**Fig. 2: Neutralizing activities and protective efficacies of FL-engineered vaccines.**

Given that the MutB/C-based vaccines induced antibodies with negligible DENV cross-reactivity (Fig. 1g), we further investigated the ADE potential of these vaccines. Sera from previously vaccinated BABL/c mice were further tested for enhancement of virus infection on K562 cells. As expected, the WT-based vaccine induced antibodies capable of enhancing infections of DENV1–4 (Fig. 3a–d). However, consistent with the markedly reduced cross-reactivity of serum to DENV, the ADE effect was almost absent in sera from MutB/C groups (Fig. 3a–d). Additionally, we also assessed the infection-enhancing activities of sera for ZIKV. Almost all flavivirus antibodies will induce ADE against the homologous virus at subneutralizing concentrations^11,42. Accordingly, we observed either WT- or MutB/C-based vaccines induced serologic responses enhancing ZIKV infection in K562 cells (Fig. 3e). However, sera from MutB-based vaccine showed dramatically decreased ADE of homologous ZIKV infection, with approximately half of the peak enhancement at low dilution (Fig. 3e).

**Fig. 3: Evaluation of ADE effect for FL-engineered vaccines.**

To further interrogate the physiological significance of these results, the vaccine-induced ADE for DENV were tested in vivo by the established passive transfer model in mice deficient in both IFNAR1 and IFNGR (AG6). Donor mice (BALB/c) were immunized with AdC7 expressing M/E-WT, M/E-MutB and M/E-MutC, respectively. PBS was given as the immune-naïve control. AG6 mice were adoptive transferred with equivalent amounts of pooled sera from groups of donor mice, and were further challenged with 5,000 FFU of DENV2 (NGC strain). The sera of donor mice from MutB/C group produced ZIKV-E-specific IgG and NAbs comparable to those of the WT group (Extended Data Fig. 7a,b), but exhibited no obvious DENV enhancement (Extended Data Fig. 7c). Consistent with the ADE tendency in vitro, mice receiving sera from donors vaccinated with the WT form succumbed to DENV2-NGC challenge significantly earlier than those receiving sera from the immune-naïve or MutB/C-vaccinated donors (WT versus Sham: P = 0.0019; WT versus MutB: P = 0.0003; WT versus MutC: P < 0.0001) (Fig. 3f). In contrast, the death curves between mice receiving donor sera from immune-naïve and MutB/C-vaccinated mice exhibited no significant difference, suggesting no in vivo ADE for DENV2 (Fig. 3f). Consistently, mice receiving sera from donors immunized with the MutB/C-based vaccines did not suffer from severe weight loss compared with those from the immune-naïve group (Fig. 3g).

MutB/C vaccines abrogate maternal–neonatal transmission

Maternal infection with ZIKV can cause a severe congenital syndrome in humans. We therefore sought to evaluate vaccine efficacy for fetal protection against ZIKV challenge in the pregnant mouse model. Groups of female Ifnar1^−/− mice were vaccinated with MutB/C-based, WT-based or sham vaccines (Fig. 4a). Both the MutB/C and WT groups elicited robust ZIKV-E-specific binding and NAb titers (Fig. 4b). These mice were then crossed with male Ifnar1^−/− mice overnight for pregnancy. Pregnant mice were then challenged with 1 × 10⁶ FFU ZIKV-SMGC-1 via the intraperitoneal (i.p.) route on embryonic day 5.5 (d_e5.5) after observation of a vaginal plug. On d_e13.5, we performed cesarean (C) sections to obtain maternal and fetal tissues from embryos for evaluation (Fig. 4a). As expected, high levels of viral RNA can be detected in maternal brain, spleen and serum from sham-vaccinated mice. In contrast, no viral loads were detectable in these tissues from the MutB/C and WT groups (Fig. 4c). We further measured virus burden in fetal and placental tissues from these vaccinated dams. High titers of virus RNA were detected in both placenta (mean 7.9 log copies per gram) and fetal head (mean 5.2 log copies per gram) in sham-vaccinated dams. In contrast, no virus load was detected in almost all placentas and heads of pups born to dams vaccinated with MutB/C- or WT-based vaccines (Fig. 4d,e), highlighting the potential of MutB/C-based vaccines to abrogate maternal–neonatal transmission of ZIKV.

**Fig. 4: MutB/C-based vaccines protect placental and fetal tissues following ZIKV challenge.**

MutB/C vaccines switch immunodominance of antibody responses

Given that the MutB/C-based vaccines elicited completely protective immunity without ADE of DENV infection, we further assessed how these vaccines affected antibody responses using B cell profiling at the single-cell level. Three groups of BALB/c mice were immunized with WT- or the MutB/C-based vaccines. At 20 days postvaccination, germinal center B cells (B_GC) (GL-7⁺ B220^hi CD38^lo IgD⁻ CD93⁻ CD138⁻) from the lymph nodes that bound to ZIKV-E were sorted by flow cytometry (Extended Data Fig. 8). We used a protein mixture of E-monomer/dimer conjugated with fluorochrome to probe the ZIKV-E-reactive B_GC, precluding experiment noise from adenovirus-elicited B_GC. To obtain the paired variable (V)-regions, we applied the chromium immune profiling solution on these ZIKV-E-reactive B_GC for single-cell B cell receptor sequencing (scBCR-seq) (Fig. 5a). Sequences covering the full-length V(D)J segments for heavy and light chains were selected for further analyses. As a result, from the WT group, 451 variable regions for heavy chain (HV) and 661 variable regions for light chain (LV) were obtained with 334 in pairs; from the MutB group, 310 HV and 379 LV were obtained with 234 in pairs; from the MutC group, 664 HV and 776 LV were obtained with 515 in pairs (Fig. 5b,c and Supplementary Table 2). Impressively, antigen-reactive B_GC clones elicited by the WT-based vaccine used a narrow pool of V genes, with ~60% utilizing IGHV9-2-1, IGHV1-22 or IGHV7-3 for HV, and also ~60% utilizing IGKV10-96, IGKV14-111 or IGKV6-23 for LV (Fig. 5c; Supplementary Table 2). In contrast, antigen-reactive B_GC clones elicited by the MutB/C-based vaccines displayed more diverse and dispersed V-gene usages for both HV and LV (Fig. 5c and Supplementary Table 2). The frequency of the dominant V genes utilized in the WT group were reduced dramatically or absent in the repertoires of MutB/C groups (Fig. 5c and Supplementary Table 2).

**Fig. 5: Profiling of B cell repertoire by scBCR-seq to dissect the immune responses induced by WT- or MutB/C-based vaccines.**

To profile the full B cell receptor (BCR) usage, we next analyzed the germline genes for the HV and LV in pairs (HV:LV) for these three groups. Consistent with the findings from the respective analyses for HV and LV, we observed the WT-based vaccine expanded B_GC clones using the narrow pool of HV:LV genes, expressing predominately IGHV9-2-1:IGKV10-96 (29.9%), IGHV1-22:IGKV14-111 (14.4%) and IGKV1-22:IGKV6-23 (7.5%) (Fig. 5d and Supplementary Table 3). In contrast, the MutB/C-based vaccines expanded E-reactive B_GC clones using larger pools of HV:LV genes with more dispersed gene frequencies (Fig. 5e,f and Supplementary Table 3). In particular, the high-frequent V-gene pairs used in the WT group were absent or minimized in the MutB/C groups. For instance, the most dominant gene pair, IGHV9-2-1:IGKV10-96, was absent in MutB/C groups (Fig. 5e,f and Supplementary Table 3). Taken together, the repertoire profiling suggested that, compared with the WT form, MutB/C-based vaccines switched immunodominance of ZIKV-E-reactive B cell responses.

Main classes of FLE mAbs to enhance DENV infection

To further characterize the antibodies elicited from these three vaccination groups, a panel of antibody genes was synthesized, representing the top most frequent clusters of genetically similar B_GC clones, for 63.4% of the WT group, 46.6% of the MutB group and 43.9% of the MutC group (Supplementary Table 4). The HV and LV sequences were then cloned into mouse IgG2a and Igκ expression vectors, respectively. The mAbs derived from the WT, MutB and MutC groups were termed ZWT.1-10, ZMutB.1-8 and ZMutC.1-13, respectively (Supplementary Table 4). We first tested their binding properties to ZIKV-E by ELISA. Most mAbs can bind to soluble E (sE)-monomer/dimer, with a frequency of 90% (9/10) in the WT group, 87.5% (7/8) in the MutB group and 76.9% (10/13) in the MutC group (Fig. 5b and Supplementary Table 4), suggesting the reliability of our approach. The 26 E-binding mAbs were further evaluated. Notably, we found that none of the mAbs derived from the WT group interacted with ZIKV-sE-MutC, but were cross-reactive with serotypes of DENV-sE, implying characteristics of FLE mAb (Fig. 6a). However, mAbs derived from the MutB/C groups had almost no-binding to DENV1–4 (Fig. 6a). Next, representative mAbs in each group were tested for their ADE effects for DENV infections on K562 cells. Consistently, the dominant mAbs derived from WT-vaccinated mice largely enhanced DENV1–4 infections, whereas the unrelated mAb did not (Fig. 6b–e). In contrast, the dominant mAbs from MutB/C-vaccinated mice did not show ADE for any of the four serotypes of DENV (Fig. 6f–i).

**Fig. 6: Characterization of representative mAbs elicited in mice at high frequency by WT- or MutB/C-based ZIKV vaccines.**

The FLE mAbs elicited from the WT group used predominantly four classes of HV:LV gene pairs (Fig. 6j). Next, we sought to analyze the publicly available sequence for murine FLE mAbs and found four such mAbs: 6B6C-1 from St. Louis encephalitis virus (SLEV)-immune⁴³, 4G2 and 2A10G6 from DENV-immune^44,45 and E53 from WNV-immune⁴⁶. Interestingly, these four mAbs were genetically close to the four main classes of FLE mAbs in the WT group (Fig. 6j). 6B6C-1 and 4G2 used the same HV:LV genes as ZWT.1-3 and ZWT4-5, respectively (Fig. 6j), with high sequence similarities (Extended Data Figs. 9a–d). 2A10G6 and E53 used the same HV gene as ZWT.6 and ZWT.8, respectively (Fig. 6j). Besides, 2A10G6 shared exactly the same amino acids with ZWT.6 in CDRL3, a region found to determine the FL-binding for light chain of 2A10G6 (Extended Data Fig. 9e)³⁵. Taken together, flavivirus FLE mAbs, isolated from mice with ZIKV-, DENV-, SLEV- and WNV-immune, were found to use almost the same few germline genes.

Structural basis of the FL-engineered immunogen

Since the FL-engineered vaccines were completely protective, with elimination of ADE for DENV in mice, we sought to explore the underlying molecular basis. ZIKV-sE-MutC were expressed and purified as representative for evaluation. A surface plasmon resonance (SPR) experiment was first performed to evaluate the binding affinities of sE-MutC to FLE and non-FLE mAbs with sE-WT tested as a control. Consistently, either the DI or DIII targeting mAb (Z3L1 or Z23) bound to sE-MutC with affinity (equilibrium dissociation constant (K_D)) almost equivalent to its binding to sE-WT, indicating the stable presentation of neutralizing epitopes. In contrast, binding between sE-MutC and FLE mAbs (Z6 and 2A10G6) was undetectable (Fig. 7a and Supplementary Fig. 1), suggesting FLE disruption.

**Fig. 7: Structural characterization of the FL-engineered immunogen.**

To further explore the molecular basis of vaccine potency and ADE abrogation at the atomic level, we determined the crystal structure of sE-MutC bound to Z3L1 at 3.1 Å resolution (Supplementary Table 1). Although five-residue mutations were introduced into FL, sE-MutC was still arranged as an E-dimer to engage Z3L1 in its prefusion state (Extended Data Fig. 10a). When folded, it resembled its WT form (root mean squared (r.m.s.) deviation of 1.04 Å for their C_α positions), showing the normal secondary, tertiary and quaternary structures, indicating the stable display of neutralizing epitopes (Fig. 7b). The sE-MutC dimer presented neutralizing epitopes as E proteins in mature ZIKV virions (Extended Data Fig. 10b). Interface analysis revealed the interaction network used to maintain dimer formation (Supplementary Table 5). In particular, compared with the WT form, FL-MutC built alternative interactions with the adjacent E protomer, introducing four potential hydrogen bonds with the neighboring DI/DIII residues (three by D98N and one by F108W) (Fig. 7c,d and Supplementary Table 5). Superimposing the structure of FL-MutC on its WT form, we found both FLs were arranged as similar conformations (r.m.s. deviation of 0.35 Å for their C_α positions), with only side-chain variations at mutated sites (Fig. 7b). To further interpret the potential elimination of FLE-antibody induction, we superimposed the DII structure of sE-MutC with its homologs from the structure-known complexes of FLE mAb bound to flavivirus-E. Notably, while the FL-WT was well accommodated by cavities of these FLE mAbs, mutations at G106, L107 and F108 would obviously interfere with reciprocal docking (Fig. 7e–g). The long side-chain protruding from G106L causes severe clashes for all these FLE mAbs (Fig. 7e–g). Moreover, L107E generated a charged side-chain and may destroy the local hydrophobic interactions that existed in FL-WT bound to these FLE mAbs (Fig. 7e–g). Besides, F108W also causes steric hindrance for 2A10G6 (Fig. 7f). Taken together, these three mutations contribute synergistically to abrogation of the elicitation of FLE mAbs.

Discussion

Given increasing knowledge of the long-term negative impact on children born with confirmed ZIKV infection during pregnancy in humans⁴⁷, sterilizing immunity may be required to completely prevent viral seeding of the placenta and fetal infection^13,33. Previous studies have described that the introduction of FLE mutations (including W101R) into ZIKV vaccine dramatically reduced ADE for DENV. However, these mutations sacrificed neutralizing titers significantly compared with the unmutated construct and, as a result, protection against viremia and tissue infections was incomplete¹³. Mutations at W101 impaired the antigen’s ability to stably present neutralizing epitopes on E domains (Extended Data Fig. 3), and are also likely to disrupt E-dimer formation, with the E-dimer epitope (EDE) destroyed¹¹. We designed mutations based on naturally occurring FL variants within members of the genus Flavivirus that maintained structural stability. These mutations retained the neutralizing epitopes presented on E domains, and also maintained E-dimer formation to present the potential EDE. Indeed, the MutB/C-based vaccines elicited E-dimer-dependent or -preferred antibodies, represented as mAbs ZMutB.2, ZMutB.3, ZMutB.4, ZMutB.7 and ZMutC.8 (Fig. 6a). Given that EDEs are targeted by potent DENV/ZIKV NAbs^11,26,28,48, keeping the E-dimer conformation may be important for vaccine potency. The strategy to retain EDE is also a goal of DENV vaccine with ADE abrogation⁴⁹.

Whether the antigen is soluble or particulate needs to be further determined. However, we propose that intracellular and secreted antigens can stimulate B/T cell responses. Also, intracellular antigens are certainly targets processed by APC for T cell activation.

The L107K in MutB are proposed to exert a similar repulsive effect to interfering with reciprocal binding by FLE mAbs. Therefore, we propose that MutB and MutC may have similar mechanisms to destroy FLE. The structural analysis provides a basis for the rational design of DENV vaccine devoid of ADE. A previous study with a DENV1 DNA plasmid vaccine reduced the ADE of DENV2 infection⁵⁰. Analogously, in that study, two mutations (a long side-chain G106R and a charged side-chain L107D) were introduced.

Here, we provide effective and safe Zika vaccine candidates to prevent both ZIKV infection and infection/vaccination-induced severe dengue disease. We used a 1.6 × 10¹¹ virus particles (vp) dose vaccine in this proof-of-concept study; however, the minimum dose should be explored as our next step in vaccine development. MutB/C could be applied to other vaccine platforms, such as DNA, messenger RNA, different virus vectors or subunit proteins. This study in a mouse model provided immunological trends for vaccine evaluation in humans and highlighted attractive candidates for further development. The global analysis of analogous amino acid sequences from members of a whole genus affords new insight into vaccine-oriented protein design.

Methods

Cells, viruses and animals

BHK-21 cells (ATCC catalog no. CCL-10), HEK293 cells (ATCC catalog no. CRL-1573), human embryonic kidney cells 293 T (HEK293T, ATCC catalog no. CRL-3216) and African green monkey kidney epithelial cells (Vero cells; ATCC catalog no. CCL-81) were all maintained in complete DMEM (Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco) and incubated at 37 °C under 5% CO₂. K562 cells (ATCC catalog no. CCL-243) were maintained in RPMI-1640 medium (Gibco) containing 10% FBS at 37 °C under 5% CO₂. Aedes albopictus C6/36 cells (ATCC, catalog no. CRL-166) were cultured in RPMI-1640 medium containing 10% FBS at 28 °C.

ZIKV-SMGC-1 was isolated in our laboratory previously³. DENV1 was isolated from an infected patient in Shenzhen Third People’s Hospital, Shenzhen, China. DENV2 strain New Guinea C (NGC strain) (GenBank accession no. KM204118.1) was kindly provided by J. An (Capital Medical University, Beijing, China). DENV3 strain YN02 (GenBank accession no. KF824903) and DENV4 strain GZB5 (GenBank accession no. AF289029) were kindly provided by C.-F. Qin (Beijing Institute of Microbiology and Epidemiology, Beijing, China). ZIKV, DENV1, DENV2, DENV3 and DENV4 were propagated in C6/36 cells and titrated by a focus-forming assay on BHK-21 cells.

BALB/c mice were purchased from Beijing Vital River Laboratory Animal Technology Company (licensed by Charles River), and housed in specific-pathogen-free (SPF) mouse facilities in the Institute of Microbiology, Chinese Academy of Science (IMCAS). Ifnar1^−/− mice were purchased from the Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences and Peking Union Medical College (CAMS&PUMC), and bred in the Laboratory Animal Center, Chinese Center for Disease Control and Prevention (China CDC). AG6 mice deficient in both IFNAR1 and IFNGR were generated by Q. Leng (Institut Pasteur of Shanghai, CAS, China) from C57BL/6 background Ifngr1^−/− mice (Jackson laboratories, catalog no. 003288) and Ifnar1^−/− A129 mice (B&G Company). They were bred in the Laboratory Animal Center, China CDC. All animals were allowed free access to water and a standard chow diet and provided with a 12-h light and dark cycle (temperature, 20–25 °C; humidity, 40–70%).

Ethics statement

All animal experiments were approved by the Committee on the Ethics of Animal Experiments of the IMCAS, and conducted in compliance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the IMCAS Ethics Committee.

Construction and production of recombinant chimpanzee adenovirus

The recombinant chimpanzee type 7 adenovirus (AdC7) encoding Japanese encephalitis virus signal peptide (JEVsp) and full-length membrane and envelope (M/E, defined as amino acids 216–794 of the ZIKV polyprotein) genes from ZIKV-FSS13025 (GenBank accession no. JN860885.1) was generated in our laboratory previously³⁴. The prM/E-WT construct contains the same signal peptide (JEVsp) and full-length precursor membrane and envelope (prM/E, defined as amino acids 123–794 of the ZIKV polyprotein). Plasmid containing JEVsp-M/E-WT was used as the template to generate MutA (D98N, N103T, G106F, L107E, F108W), MutB (D98N, N103T, G106F, L107K, F108W) and MutC (D98N, N103T, G106L, L107E, F108W). The prM/E-WT, M/E-MutB and M/E-MutC cassettes were cloned into pAdC7, forming recombinant adenovirus genomes, which were linearized and transfected into HEK293 cells to rescue the recombinant adenovirus that was then further propagated and purified by cesium chloride density gradient centrifugation as previously described³⁴.

Protein expression and purification of ZIKV/DENV-sE proteins

The coding sequence for ectodomain residues of E proteins from ZIKV (1–409), DENV1 (1–400), DENV2 (1–398), DENV3 (1–400) and DENV4 (1–400) were codon optimized and cloned into pET21a vector (Novagen) with NdeI and XhoI restriction sites. The coding sequence for the ectodomain of E-MutC was generated by introducing mutations D98N, N103T, G106L, L107E and F108W into the WT-ZIKV-E construct. sE proteins were expressed in Escherichia coli strain BL21 (DE3) as inclusion bodies and then refolded in vitro using a previously described method³⁵, with some modifications. Briefly, aliquots of inclusion body were diluted dropwise with stirring into refolding buffer containing 100 mM Tris (pH 8.0), 400 mM l-Arg HCl, 2 mM EDTA, 5 mM reduced glutathione, 0.5 mM oxidized glutathione and 10% glycerol. The refolded proteins were concentrated using an Amicon 400 concentrator with 30 kDa cutoff membrane and then adjusted to 20 mM Tris (pH 8.0), 50 mM NaCl and 5% glycerol. Next, sE proteins were further purified using the ÄKTA Pure System (GE Healthcare) by gel filtration on a HiLoad Superdex 200 prep grade column (GE Healthcare).

Antibody expression and purification

The human monoclonal antibodies (mAbs) Z6, Z3L1, Z20, Z23 and the murine mAbs ZWT.1-10, ZMutB.1-8 and ZMutC.1-13 were expressed and purified as previously described³¹. Briefly, HEK293T cells were cotransfected with pCAGGS plasmids containing coding sequences for Ig heavy chain and light chain. The cell culture was collected on days 3 and 7 post-transfection. The supernatant was mixed with one volume of buffer containing 20 mM sodium phosphate, pH 7.0, and filtered with a 0.22-μm filter. The mixture was passed through the HiTrap Protein A FF (GE Healthcare) column. The bound protein was detached from the column by 0.1 M glycine, pH 3.0. The elution was adjusted to neutral pH by adding 1 M Tris-HCl, pH 9.0 and further purified by gel filtration. The antibody was finally buffered with PBS, concentrated and stored at −80 °C before further use.

Murine mAb 2A10G6 was prepared according to a previously described method³⁵. Briefly, hybridoma cells producing 2A10G6 were injected into mice and the ascites were collected. 2A10G6 was further purified by HiTrap Protein G HP (GE Healthcare) with binding buffer of 20 mM sodium phosphate (pH 7.0) and elution buffer of 0.1 M glycine acid-HCl (pH 3.0). Protein was concentrated and stored at −80 °C until use.

To generated the fragment of antibody-binding (Fab), purified Z6, 2A10G6, Z3L1 or Z23 antibodies were digested with immobilized papain (Thermo Scientific) according to the manufacturer’s instructions. Fab fractions were purified by HiTrap Protein A FF (5 ml, GE Healthcare) and gel filtration.

Z3L1 single-chain variable fragment (scFv) was constructed as V_L-(GGGS)4-V_H and cloned into pET21a. The protein was expressed in E. coli (BL21) as inclusion bodies.

Neutralization assay

The neutralizing activity of mouse serum was assessed using a previously described ZIKV microneutralization (MN) assay³⁴, with some modifications. Briefly, heat-inactivated serum was serially diluted and incubated with ZIKV-SMGC-1 (100 PFU) for 1 h at 37 °C. The virus–serum mixture was added to preplated Vero cell monolayers in 96-well plates. After incubation for 4 days, the culture medium was removed, and the cells were fixed with cold methanol at −20 °C for 20 min and washed with PBS with 0.05% (v/v) Tween-20 (PBST). The following steps were all performed at room temperature (RT). The cells were blocked with 3% skim milk for 30 min and then stained for 2 h with human anti-ZIKV-E mAb Z6, which was diluted to 4 μg ml⁻¹. Cells were washed with PBST, followed by incubation for 2 h with goat anti-human IgG-HRP (Proteintech), diluted 1:1,500. Plates were then washed and developed with 3,3′,5,5′-tetramethylbenzidine (TMB) substrate for 20 min. The reaction was stopped with 2 M hydrochloric acid, and the absorbance was measured at 450 nm using a microplate reader (PerkinElmer). Neutralization titers, MN₅₀, were determined by fitting nonlinear regression curves using GraphPad Prism (v.6 and v.9.0.1) as the reciprocal of the serum dilution required for 50% neutralization of infection. MN₅₀ titers calculated as <20 were reported as half the limit of detection (LOD).

Epitope detection for ZIKV-E by flow cytometry

HEK293T cells were transfected with pCAGGS plasmids encoding WT or mutated ZIKV-M/E proteins, or infected with AdC7 viruses encoding WT or mutated ZIKV-M/E proteins. After 48 h, the cells were digested by trypsin and collected, followed by incubation in Fixation and Permeabilization solution buffer (BD, catalog no. 554722) for 20 min at 4 °C. The cells were then divided into several aliquots and each was stained with mAb (Z6, 2A10G6, Z3L1, Z23 or Z20) for 1 h at 4 °C. The cells were then washed twice with 1× Perm/Wash buffer and stained with Goat Anti-Human IgG-fluorescein isothiocyanate (FITC) (Proteintech) for 1 h at 4 °C. After washing twice with 1× Perm/Wash buffer, the cells were resuspended in PBS. Subsequently, FITC-positive HEK293T cells were detected by FACSCanton flow cytometer (BD Bioscience). The data were further analyzed by Flowjo v.7.6.1.

Capture ELISA

Cell supernatants were collected from HEK293T cells 3 days after plasmid transfection or 2 days after infection. Purified mouse anti-ZIKV IgG were plated at 500 ng per well in 96-well ELISA plates overnight at 4 °C. Plates were blocked with 5% skimmed milk in room temperature for 30 min; 200 μl cell supernatant was added to each well overnight at 4 °C. Antibody Z3L1 conjugated with horseradish peroxidase (HRP) was used to trap the antigen at a concentration of 1 μg ml⁻¹ at RT for 3 h. TMB chromogen solution was then used for color development reaction, which was stopped by addition of 2 M HCl. Reaction color changes were quantified at 450 nm in a microplate reader (PerkinElmer).

Immunization

AdC7 vaccines expressing ZIKV-M/E-WT, -prM/E-WT, -M/E-MutB or -M/E-MutC were diluted in PBS. BALB/c mice or Ifnar1^−/− mice were immunized with vaccine through the intramuscular (i.m.) route. Equivalent volume of PBS was injected as a sham control. The blood samples were collected postvaccination. The sera isolated were used for determination of antibody titer or ADE activity.

Animal protection against virus challenge

To evaluate the protective efficacy of ZIKV vaccines against lethal ZIKV challenge, immunized Ifnar1^−/− mice were infected with ZIKV-SMGC-1 via i.p. injection. After ZIKV challenge, the animals were monitored daily for clinical signs of disease. Sera were collected to determine viral loads by quantitative reverse transcription–polymerase chain reaction (qRT–PCR). Mice that had lost 25% of initial body weight were euthanized on animal welfare grounds.

To measure virus burdens in targeted tissues, immunized Ifnar1^−/− mice were i.p. challenged with ZIKV-SMGC-1. Mice were euthanized and necropsied. Organs were removed, weighed and homogenized in PBS with a tissue grinder. Samples were further centrifuged, and the supernatants were stored at −80 °C until virus titration by qRT–PCR.

qRT–PCR

Virus RNA was isolated from 200 μl supernatants of homogenized tissues or 50 μl of mouse serum using a GenePure Plus nucleic acid extraction instrument (Hangzhou Bioer Technology). ZIKV-specific qRT–PCR assays were performed using a FastKing One Step Probe RT–qPCR kit (Tiangen Biotech) on a CFX96 Touch real-time PCR detection system (Bio-Rad) according to the manufacturer’s protocol. Primers and probe were designed to amplify a region of the NS3 gene from ZIKV-FSS13025 and ZIKV-SMGC-1, with sequences as follows: forward primer, TTGGCTGGCCTATCAGGTTG; reverse primer, CACCTCGGTTTGAGCACTCT; the probe FAM-CCACACCTCTGCCGGCACAC-TAMRA (where FAM is 6-carboxyfluorescein, and TAMRA is 6-carboxytetramethylrhodamine). Viral loads were expressed on a log₁₀ scale as viral copies per gram or copies per milliliter after calculation with a standard curve. Viral copy numbers below the LOD were set as half of the LOD.

ELISA

Binding properties of murine sera and the isolated mAbs to ZIKV-sE (monomer), ZIKV-sE (dimer), ZIKV-sE-MutC or DENV-sE (serotype 1–4) were determined by ELISA. First, 96-well plates (catalog no. 3590; Corning) were coated overnight with 3 μg ml⁻¹ of ZIKV/DENV-E or -E-mutant in 0.05 M carbonate-bicarbonate buffer, pH 9.6, and blocked in 5% skim milk in PBS. Sera were serially diluted and added to each well. For mAbs expressed from HEK293T cells, culture supernatant was added. Plates were incubated with goat anti-mouse IgG-HRP antibody (murine serum and mAbs isolated from mice) or goat anti-human IgG-HRP antibody (mAb Z6, Z3L1 and Z23), and subsequently developed with TMB substrate. Reactions were stopped with 2 M hydrochloric acid, and the absorbance was measured at 450 nm using a microplate reader (PerkinElmer). The endpoint titers were defined as the highest reciprocal dilution of serum to give an absorbance greater than fivefold background values.

In vitro evaluation of ADE of virus infection

Heat-inactivated serum was serially diluted with RPMI-1640 medium supplemented with 1% FBS and incubated with ZIKV or DENV1, 2, 3 or 4 for 1 h at 37 °C. The serum–virus mixture was then mixed with 3 × 10⁴ K562 cells and incubated for 2 h at 37 °C. RPMI-1640 medium supplemented with 2% FBS were then added. After incubation for 4 days, the K562 cells were centrifuged and washed once with PBS, followed by incubation with Fixation and Permeabilization solution (BD, catalog no. 554722) for 20 min at 4 °C. The cells were washed twice with 1× Perm/Wash buffer (BD, 554723) and stained with FITC-conjugated human mAb Z6 for 1 h at 4 °C. After washing twice with 1× Perm/Wash buffer, the cells were resuspended in PBS. Virus infected K562 cells were detected by FACSCanton flow cytometer (BD Bioscience) and further analyzed by Flowjo v.7.6.1.

In vivo ADE of DENV2 evaluation

AG6 mice were used as the infection model. BALB/c mice were immunized with vaccines or sham as sera donor. At 4 weeks postvaccination, blood samples were collected, and sera were isolated, heat-inactivated, mixed together for each vaccine group, aliquoted and stored at –80 °C until use. Before adoptive transfer, the sera were first diluted tenfold with PBS. AG6 mice received diluted sera by passive transfer. AG6 mice were challenged with DENV2-NGC and monitored daily for survival and weight loss postchallenge. Mice that had lost 25% of initial body weight were euthanized on animal welfare grounds.

Mouse pregnancy experiment

Female Ifnar1^−/− mice were vaccinated with vaccines or sham. Serum samples were collected on day 28, heat inactivated (56 °C for 30 min) for detection of ZIKV-E-specific IgG and MN₅₀. Ifnar1^−/− female mice were then mated with Ifnar1^−/− males in individual cages. The vaginal plug was checked daily. Once the plug was identified, the mice were defined at starting day embryotic stage (d_e0.5). The mice were infected with ZIKV-SMGC-1 at d_e5.5. All animals were sacrificed at d_e13.5. Blood, brain, spleen of maternal mice and placentas and heads of fetuses were harvested for measuring viral loads by qRT–PCR.

SPR assay

SPR binding experiments were carried out using a BIAcore 8000 device (GE Healthcare) at RT. The buffers for all proteins used for kinetic analyses were exchanged for BIAcore buffer consisting of 10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA and 0.005% (v/v) Tween-20 by gel filtration. Purified ZIKV-sE or ZIKV-sE-MutC proteins were immobilized on a CM5 chip with the standard EDC/NHS coupling method at about 6,000 response units (RU). Serial dilutions of Fabs were prepared and used to flow over the chip surface. Data were collected over time. After each cycle, the sensor surface was regenerated via a short treatment using 30 mM NaOH (Z6 mAb) or 3 M MgCl₂ (2A10G6 antibody). The apparent equilibrium dissociation constants (apparent binding affinity, K_D) for each sE-antibody were calculated using BIAcore 8000 analysis software (BIAevaluation v.4.1). Each set of equilibrium binding responses was fitted to the single binding site model.

Isolation of antigen-specific B_GC cells

Single-cell suspensions were obtained from the lymph nodes of immunized mice. The cells from each group of mice were pooled together. Antibodies used for staining were anti-mouse GL-7-FITC, CD138-PE, CD38-PE/Cy7, CD93-APC, B220-BV421 and IgD⁻BV510. All antibodies were used according to manufacturer’s instructions. To stain the ZIKV-E-binding B cells, ZIKV-sE monomer (described above) and dimer (a kind gift from R. Gong, Wuhan Institute of Virology, CAS, China)⁵¹ were used as bait by biotinylation via Biotin–Protein Ligase/BirA Enzyme (GeneCopoeia). BV711–Streptavidin was used to probe the biotinylated-ZIKV-E proteins. Antigen-specific B_GC gated as GL-7⁺, B220^hi, CD38^lo, IgD⁻, CD93⁻, CD138⁻ and ZIKV-E⁺ were sorted using a BD FACSAria II flow cytometer (BD Biosciences).

Single-cell library construction and sequencing

Samples for single BCR V(D)J clonotype were processed using a Chromium Single Cell 5′ Library and Gel Bead Kit following the manufacturer’s user guide (10x Genomics). After FACS sorting, cells were centrifuged, resuspended in 3% FBS (Sigma-Aldrich)/phosphate buffer solution (Thermo Fisher Scientific) and subjected to cell quality control using a cell counter. All processed B cells had cell viability >70%. After determining cell density, cells were injected into three channels, aiming to achieve around 1,000–3,000 cells per channel. Gel beads-in-emulsion (GEMs) were formed in channels of a chip in the 10x Chromium instrument, and then collected in an Eppendorf plate for the GEM-reverse transcription (GEM-RT) reaction. After GEM clean up, GEM-RT products were subjected to two rounds of 9 and 15 PCR cycles using custom primers, followed by SPRIselect (Beckman Coulter) bead clean up. Single-cell BCR V(D)J libraries were prepared following the manufacturer’s user guide, profiled using the Bioanalyzer High Sensitivity DNA kit (Agilent Technologies) and quantified with a Kapa Library Quantification Kit (Kapa Biosystems). Libraries were sequenced by paired-end sequencing (2 × 150 bp) on an Illumina NovaSeq (Illumina). BCL data were converted to demultiplexed FASTQ files using Illumina bcl2fastq v.2.20.

Cellranger V(D)J data processing

We retained the first 26 bases for read 1 covering the 16-nt cell barcode and 10-nt unique molecular identifier (UMI). The FASTQ files were analyzed subsequently. The Cell Ranger Single-Cell Software Suite (v.3.1.0) was used to perform barcode processing and single-cell V(D)J analysis. The FASTQ files were processed using the Cell Ranger V(D)J pipeline. First, reads were filtered for valid cell barcodes and UMIs. Cell barcodes with 1-Hamming-distance from a list of known barcodes were considered. A UMI with 1-Hamming-distance from another UMI with more reads for a same barcode was corrected to this UMI with more reads. The filtered reads were used to assembly contigs by alignment with the GRCm38 V(D)J reference genome, then defined alignments of V, D and J segments to a contig, identifying CDR3 sequences, and, from these data, determine whether a contig is productive, meaning that it is likely to correspond to a functional B cell receptor. The V(D)J sequences and clonotypes from single-cell 5′ data were analyzed, searched and visualized using the LoupeTM V(D)J browser v.3.0.0. Sequences were further annotated and analyzed using IgBLAST v1.6.1 (ref. ⁵²) to identify variable region gene segments and somatic mutations.

Crystallization, data collection and structure determination

For the ZIKV-sE/Z6 complex, ZIKV-sE was mixed with Z6-Fab at a stoichiometry of 1.3:1 and incubated on ice for 2 h. The mixture was then subjected to gel filtration. The survival peak containing the indicated Fab and sE was pooled, buffered in 20 mM Tris, 50 mM NaCl and 5% glycerol (pH 8.0). It was concentrated to 10 or 5 mg ml⁻¹ for crystallization.

To obtain ZIKV-sE-MutC/Z3L1 complex, the inclusion bodies of ZIKV-sE-MutC and Z3L1-scFv were diluted into refolding buffer containing 100 mM Tris (pH 8.0), 400 mM l-Arg HCl, 2 mM EDTA, 5 mM reduced glutathione, 0.5 mM oxidized glutathione and 10% glycerol. After 12 h of slow stirring at 4 °C, the refolded ZIKV-sE-MutC/Z3L1-scFv complex was concentrated and further purified in an ÄKTA Pure System (GE Healthcare) by gel filtration on a HiLoadTM SuperdexTM 200 preparation grade column (GE Healthcare). The survival peak containing the indicated ZIKV-sE-MutC/Z3L1-scFv was pooled and further polished using an ion exchange column (Source 15Q; GE Healthcare). The indicated complex was further buffered in 20 mM Tris, 50 mM NaCl and 5% glycerol (pH 8.0), and concentrated to 10 or 5 mg ml⁻¹ for crystallization.

Crystallization trials were performed in sitting drops of 600 nl. All crystals were obtained by mixing equal volumes of protein and reservoir solution, using a Mosquito LCP robot (TTP LabTech). Diffraction-quality crystals of ZIKV-sE/Z6-Fab were obtained at 18 °C in condition consisting of 0.1 M amino acids, 0.1 M buffer system 3 at pH 8.5 (Morpheus MD1-46) and 30% v/v Precipitant Mix 3 (Morpheus MD1-46). Diffraction-quality crystals of ZIKV-sE-MutC/Z3L1-scFv were obtained at 18 °C in conditions consisting of 0.12 M ethylene glycol, 0.1 M buffer system 3 (Morpheus MD1-46) at pH 8.5 and 30% v/v Precipitant Mix 2 (Morpheus MD1-46).

All diffraction data were collected at the Shanghai Synchrotron Radiation Facility and processed with HKL2000 (ref. ⁵³). The ZIKV-sE/Z6-Fab dataset was collected at BL17U1, whereas the ZIKV-sE-MutC/Z3L1-scFv dataset was collected at BL19U1. The complex structure was solved by the molecular replacement module of Phaser⁵⁴ from the CCP4 program suite⁵⁵, with the previously reported structure of sE from ZIKV (PDB 5JHM), ZIKV-sE/2A10G6-Fab (PDB: 5JHL) and ZIKV-sE/Z3L1-Fab (PDB 5GZN) as the search models. Restrained rigid-body refinement with REFMAC5 (CCP4 suite) and manual model building with Coot⁵⁶ were then performed. Further rounds of refinement were then done with phenix.refine⁵⁷. During model building and refinement, the program of MolProbity⁵⁸ was used to validate the stereochemistry of the structure. Structural data were analyzed by PyMOL and the figures were generated using Chimera⁵⁹.

Statistical analysis

GraphPad Prism v.6 or v.9.01 software was used for data analysis. Data are represented as mean ± s.e.m. Statistical significance was determined by unpaired, two-sided Student’s t-test for two-group comparisons, one-way analysis of variance (ANOVA) for comparisons of more than two groups, two-way ANOVA for comparisons of more than two groups with two or more items, or log-rank test for survival experiments. P values <0.05 were considered to be statistically significant, where *P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001.

Reporting Summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The crystal structures of ZIKV-sE-Z6 and ZIKV-sE-MutC-Z3L1 have been deposited in the Protein Data Bank (PDB) under accession codes 7BQ5 and 7BPK, respectively. The raw data for scBCR-seq have been deposited at https://www.scidb.cn/en/datalist?tag=1. Source data are provided with this paper. All other data supporting the findings of this study are available within the paper or from the corresponding author upon reasonable request.

References

Knipe, D. M., Howley, P. M. (eds) Fields Virology 6th edn (Lippincott Williams & Wilkins Health, 2013).
Baud, D., Gubler, D. J., Schaub, B., Lanteri, M. C. & Musso, D. An update on Zika virus infection. Lancet 390, 2099–2109 (2017).
Article PubMed Google Scholar
Wang, Q. et al. Genetic and biological characterization of Zika virus from human cases imported through Shenzhen Port. Chin. Sci. Bull. 61, 2463–2474 (2016).
Article Google Scholar
Brasil, P. et al. Zika virus infection in pregnant women in Rio de Janeiro. N. Engl. J. Med. 375, 2321–2334 (2016).
Article PubMed PubMed Central Google Scholar
Rasmussen, S. A., Jamieson, D. J., Honein, M. A. & Petersen, L. R. Zika virus and birth defects—reviewing the evidence for causality. N. Engl. J. Med. 374, 1981–1987 (2016).
Article CAS PubMed Google Scholar
van der Eijk, A. A. et al. Miscarriage associated with Zika virus infection. N. Engl. J. Med. 375, 1002–1004 (2016).
Article PubMed Google Scholar
Cao-Lormeau, V. M. et al. Guillain–Barre Syndrome outbreak associated with Zika virus infection in French Polynesia: a case-control study. Lancet 387, 1531–1539 (2016).
Article PubMed PubMed Central Google Scholar
Miner, J. J. & Diamond, M. S. Zika virus pathogenesis and tissue tropism. Cell Host Microbe 21, 134–142 (2017).
Article CAS PubMed PubMed Central Google Scholar
Ma, W. et al. Zika virus causes testis damage and leads to male infertility in mice. Cell 167, 1511–1524.e10 (2016).
Article CAS PubMed Google Scholar
Govero, J. et al. Zika virus infection damages the testes in mice. Nature 540, 438–442 (2016).
Article CAS PubMed PubMed Central Google Scholar
Slon-Campos, J. L. et al. A protective Zika virus E-dimer-based subunit vaccine engineered to abrogate antibody-dependent enhancement of dengue infection. Nat. Immunol. 20, 1291–1298 (2019).
Article CAS PubMed PubMed Central Google Scholar
Fowler, A. M. et al. Maternally acquired Zika antibodies enhance dengue disease severity in mice. Cell Host Microbe 24, 743–750.e5 (2018).
Article CAS PubMed PubMed Central Google Scholar
Richner, J. M. et al. Modified mRNA vaccines protect against Zika virus infection. Cell 168, 1114–1125.e10 (2017).
Article CAS PubMed PubMed Central Google Scholar
Rey, F. A., Stiasny, K., Vaney, M. C., Dellarole, M. & Heinz, F. X. The bright and the dark side of human antibody responses to flaviviruses: lessons for vaccine design. EMBO Rep. 19, 206–224 (2018).
Article CAS PubMed Google Scholar
Halstead, S. B. & O’Rourke, E. J. Dengue viruses and mononuclear phagocytes. I. Infection enhancement by non-neutralizing antibody. J. Exp. Med. 146, 201–217 (1977).
Article CAS PubMed PubMed Central Google Scholar
Katzelnick, L. C. et al. Antibody-dependent enhancement of severe dengue disease in humans. Science 358, 929–932 (2017).
Article CAS PubMed PubMed Central Google Scholar
Halstead, S. B. Dengvaxia sensitizes seronegatives to vaccine enhanced disease regardless of age. Vaccine 35, 6355–6358 (2017).
Article CAS PubMed Google Scholar
Sridhar, S. et al. Effect of dengue serostatus on dengue vaccine safety and efficacy. N. Engl. J. Med. 379, 327–340 (2018).
Article PubMed Google Scholar
Stettler, K. et al. Specificity, cross-reactivity, and function of antibodies elicited by Zika virus infection. Science 353, 823–826 (2016).
Article CAS PubMed Google Scholar
George, J. et al. Prior exposure to Zika virus significantly enhances peak Dengue-2 viremia in rhesus macaques. Sci. Rep. 7, 10498 (2017).
Article PubMed PubMed Central CAS Google Scholar
Katzelnick, L. C. et al. Zika virus infection enhances future risk of severe dengue disease. Science 369, 1123–1128 (2020).
Article CAS PubMed PubMed Central Google Scholar
York, A. Zika virus enhances dengue risk. Nat. Rev. Microbiol. 18, 605 (2020).
PubMed PubMed Central Google Scholar
Rogers, T. F. et al. Zika virus activates de novo and cross-reactive memory B cell responses in dengue-experienced donors. Sci. Immunol. 2, eaan6809 (2017).
Article PubMed PubMed Central Google Scholar
Dejnirattisai, W. et al. Cross-reacting antibodies enhance dengue virus infection in humans. Science 328, 745–748 (2010).
Article CAS PubMed Google Scholar
Beltramello, M. et al. The human immune response to Dengue virus is dominated by highly cross-reactive antibodies endowed with neutralizing and enhancing activity. Cell Host Microbe 8, 271–283 (2010).
Article CAS PubMed Google Scholar
Dejnirattisai, W. et al. A new class of highly potent, broadly neutralizing antibodies isolated from viremic patients infected with dengue virus. Nat. Immunol. 16, 170–177 (2015).
Article CAS PubMed Google Scholar
Dai, L., Wang, Q., Song, H. & Gao, G. F. Zika virus envelope protein and antibody complexes. Subcell. Biochem. 88, 147–168 (2018).
Article CAS PubMed Google Scholar
Barba-Spaeth, G. et al. Structural basis of potent Zika–dengue virus antibody cross-neutralization. Nature 536, 48–53 (2016).
Article CAS PubMed Google Scholar
Rouvinski, A. et al. Recognition determinants of broadly neutralizing human antibodies against dengue viruses. Nature 520, 109–113 (2015).
Article CAS PubMed Google Scholar
Fibriansah, G. & Lok, S. M. The development of therapeutic antibodies against dengue virus. Antivir. Res. 128, 7–19 (2016).
Article CAS PubMed Google Scholar
Wang, Q. et al. Molecular determinants of human neutralizing antibodies isolated from a patient infected with Zika virus. Sci. Transl. Med. 8, 369ra179 (2016).
Article PubMed CAS Google Scholar
Zhao, H. et al. Structural basis of Zika virus-specific antibody protection. Cell 166, 1016–1027 (2016).
Article CAS PubMed PubMed Central Google Scholar
Abbink, P., Stephenson, K. E. & Barouch, D. H. Zika virus vaccines. Nat. Rev. Microbiol. 16, 594–600 (2018).
Article CAS PubMed PubMed Central Google Scholar
Xu, K. et al. Recombinant chimpanzee adenovirus vaccine AdC7-M/E protects against Zika virus infection and testis damage. J. Virol. 92, e01722-17 (2018).
PubMed PubMed Central Google Scholar
Dai, L. et al. Structures of the Zika virus envelope protein and its complex with a flavivirus broadly protective antibody. Cell Host Microbe 19, 696–704 (2016).
Article CAS PubMed Google Scholar
Cherrier, M. V. et al. Structural basis for the preferential recognition of immature flaviviruses by a fusion-loop antibody. EMBO J. 28, 3269–3276 (2009).
Article CAS PubMed PubMed Central Google Scholar
Larocca, R. A. et al. Vaccine protection against Zika virus from Brazil. Nature 536, 474–478 (2016).
Article CAS PubMed PubMed Central Google Scholar
Abbink, P. et al. Durability and correlates of vaccine protection against Zika virus in rhesus monkeys. Sci. Transl. Med. 9, eaao4163 (2017).
Article PubMed PubMed Central CAS Google Scholar
Smith, S. A. et al. Dengue virus prM-specific human monoclonal antibodies with virus replication-enhancing properties recognize a single immunodominant antigenic site. J. Virol. 90, 780–789 (2016).
Article CAS PubMed Google Scholar
Tsai, W. Y. et al. High-avidity and potently neutralizing cross-reactive human monoclonal antibodies derived from secondary dengue virus infection. J. Virol. 87, 12562–12575 (2013).
Article CAS PubMed PubMed Central Google Scholar
Lazear, H. M. et al. A mouse model of Zika virus pathogenesis. Cell Host Microbe 19, 720–730 (2016).
Article CAS PubMed PubMed Central Google Scholar
Pierson, T. C. et al. The stoichiometry of antibody-mediated neutralization and enhancement of West Nile virus infection. Cell Host Microbe 1, 135–145 (2007).
Article CAS PubMed PubMed Central Google Scholar
Crill, W. D. & Chang, G. J. Localization and characterization of flavivirus envelope glycoprotein cross-reactive epitopes. J. Virol. 78, 13975–13986 (2004).
Article CAS PubMed PubMed Central Google Scholar
Bennett, K. M. et al. Hybrid flagellin as a T cell independent vaccine scaffold. BMC Biotechnol. 15, 71 (2015).
Article PubMed PubMed Central CAS Google Scholar
Deng, Y. Q. et al. A broadly flavivirus cross-neutralizing monoclonal antibody that recognizes a novel epitope within the fusion loop of E protein. PLoS ONE 6, e16059 (2011).
Article CAS PubMed PubMed Central Google Scholar
Oliphant, T. et al. Antibody recognition and neutralization determinants on domains I and II of West Nile Virus envelope protein. J. Virol. 80, 12149–12159 (2006).
Article CAS PubMed PubMed Central Google Scholar
van der Linden, V. et al. Description of 13 infants born during October 2015–January 2016 with congenital Zika virus infection without microcephaly at birth—Brazil. MMWR Morb. Mortal. Wkly Rep. 65, 1343–1348 (2016).
Article PubMed Google Scholar
Fernandez, E. et al. Human antibodies to the dengue virus E-dimer epitope have therapeutic activity against Zika virus infection. Nat. Immunol. 18, 1261–1269 (2017).
Article CAS PubMed PubMed Central Google Scholar
Rouvinski, A. et al. Covalently linked dengue virus envelope glycoprotein dimers reduce exposure of the immunodominant fusion loop epitope. Nat. Commun. 8, 15411 (2017).
Article CAS PubMed PubMed Central Google Scholar
Crill, W. D. et al. Sculpting humoral immunity through dengue vaccination to enhance protective immunity. Front. Immunol. 3, 334 (2012).
Article CAS PubMed PubMed Central Google Scholar
Yang, C. et al. Characterization of two engineered dimeric Zika virus envelope proteins as immunogens for neutralizing antibody selection and vaccine design. J. Biol. Chem. 294, 10638–10648 (2019).
Article CAS PubMed PubMed Central Google Scholar
Ye, J., Ma, N., Madden, T. L. & Ostell, J. M. IgBLAST: an immunoglobulin variable domain sequence analysis tool. Nucleic Acids Res. 41, W34–W40 (2013).
Article PubMed PubMed Central Google Scholar
Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326 (1997).
Article CAS PubMed Google Scholar
Read, R. J. Pushing the boundaries of molecular replacement with maximum likelihood. Acta Crystallogr. D Biol. Crystallogr. 57, 1373–1382 (2001).
Article CAS PubMed Google Scholar
Collaborative Computational Project, Number 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D Biol. Crystallogr. 50, 760–763 (1994).
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
Article PubMed CAS Google Scholar
Aad, G. et al. Search for new particles in two-jet final states in 7 TeV proton-proton collisions with the ATLAS detector at the LHC. Phys. Rev. Lett. 105, 161801 (2010).
Article CAS PubMed Google Scholar
Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12–21 (2010).
Article CAS PubMed Google Scholar
Pettersen, E. F. et al. UCSF Chimera—a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Article CAS PubMed Google Scholar

Download references

Acknowledgements

We thank the staff of BL17U and BL19U beamlines at National Center for Protein Sciences Shanghai and Shanghai Synchrotron Radiation Facility (Shanghai, China) for assistance during data collection. We are grateful to J. Jia (Institute of Biophysics, Chinese Academy of Sciences (CAS), China) for technical support during BD FACSAria II manipulation. We thank Y. Chen (Institute of Biophysics, CAS, China) for technical support with BIAcore experiments. We thank D. Zhou (Institut Pasteur of Shanghai, CAS, China) for providing us with the AdC7 vector. We thank G. Cheng (Tsinghua University, Beijing, China) and Q. Leng (Institut Pasteur of Shanghai, CAS, China) for providing AG6 mice for breeding. We thank R. Gong (Wuhan Institute of Virology, CAS) for providing ZIKV-sE dimer protein. We thank Y. Chen (China CDC) for breeding of AG6 and Ifnar1^−/− mice. We thank C. Liu (Guangxi University, China) for technical support. This work is supported by the National Science and Technology Major Project (2016YFE0205800 and 2020YFA0907100) (L.D. and G.F.G.), the National Natural Science Foundation of China (NSFC) (81991494 and 81801648) (L.D.) and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB29010202) (G.F.G). L.D. is supported by Youth Innovation Promotion Association CAS (2018113).

Author information

These authors contributed equally: Lianpan Dai, Kun Xu, Jinhe Li.

Authors and Affiliations

CAS Key Laboratory of Pathogen Microbiology and Immunology, Institute of Microbiology, Chinese Academy of Sciences (CAS), Beijing, China
Lianpan Dai, Jian Song, Qihui Wang, Yan Chai, Jianxun Qi & George F. Gao
University of Chinese Academy of Sciences, Beijing, China
Lianpan Dai, Jinhe Li, Yuxuan Han, Tianyi Zheng, Ping Gao & George F. Gao
Key Laboratory of Tropical Translational Medicine of Ministry of Education, School of Tropical Medicine and Laboratory Medicine, The First Affiliated Hospital, Hainan Medical University, Hainan, China
Lianpan Dai, Kun Xu & Qianfeng Xia
Research Network of Immunity and Health (RNIH), Beijing Institutes of Life Science, Chinese Academy of Sciences, Beijing, China
Jinhe Li, Tianyi Zheng, Ping Gao & George F. Gao
CAS Key Laboratory of Microbial Physiological and Metabolic Engineering, Institute of Microbiology, Chinese Academy of Sciences, Beijing, China
Qingrui Huang, Huabing Yang & Jinghua Yan
Laboratory Animal Center, Chinese Center for Disease Control and Prevention (China CDC), Beijing, China
Xuancheng Lu
Faculty of Health Sciences, University of Macau, Macau SAR, China
Kefang Liu

Authors

Lianpan Dai
View author publications
You can also search for this author in PubMed Google Scholar
Kun Xu
View author publications
You can also search for this author in PubMed Google Scholar
Jinhe Li
View author publications
You can also search for this author in PubMed Google Scholar
Qingrui Huang
View author publications
You can also search for this author in PubMed Google Scholar
Jian Song
View author publications
You can also search for this author in PubMed Google Scholar
Yuxuan Han
View author publications
You can also search for this author in PubMed Google Scholar
Tianyi Zheng
View author publications
You can also search for this author in PubMed Google Scholar
Ping Gao
View author publications
You can also search for this author in PubMed Google Scholar
Xuancheng Lu
View author publications
You can also search for this author in PubMed Google Scholar
Huabing Yang
View author publications
You can also search for this author in PubMed Google Scholar
Kefang Liu
View author publications
You can also search for this author in PubMed Google Scholar
Qianfeng Xia
View author publications
You can also search for this author in PubMed Google Scholar
Qihui Wang
View author publications
You can also search for this author in PubMed Google Scholar
Yan Chai
View author publications
You can also search for this author in PubMed Google Scholar
Jianxun Qi
View author publications
You can also search for this author in PubMed Google Scholar
Jinghua Yan
View author publications
You can also search for this author in PubMed Google Scholar
George F. Gao
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

L.D., J.Y. and G.F.G. designed the experiments. L.D., K.X., J.L., Q.H., J.S., Y.H., T.Z., P.G., X.L., H.Y., K.L., Q.W., Y.C. and J.Q. performed the investigations and assays. L.D., K.X., J.Y. and G.F.G. analyzed the data. L.D. and K.X. wrote the manuscript. L.D., K.X., Q.X., Q.W., J.Y. and G.F.G. discussed and edited the manuscript.

Corresponding authors

Correspondence to Lianpan Dai, Jinghua Yan or George F. Gao.

Ethics declarations

Competing interests

G.F.G, L.D., J.Y., K.X., Y.H., Q.W., Q.H. and J.L. are listed as the coinventors for the pending patent for the ZIKV vaccines described in this study. The other authors declare no competing interests.

Additional information

Peer review Information Nature Immunology thanks R. Rappuoli and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. J. D. K. Wilson was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Antibodies induced by ZIKV vaccines enhance DENV infections.

a, Neutralizing activities of ZIKV vaccines. 6-8-week old female BALB/c mice (n = 5 biologically independent samples) were immunized with a single dose of AdC7-M/E or AdC7-prM/E (1.6 × 10¹¹ vp) via i.m. route. PBS was given as a sham vaccine. Sera were collected from blood at 4 weeks postvaccination. Magnitudes of NAbs were measured as MN₅₀ titers. Data are means ±standard errors of means (SEM). P values were analyzed with One-way ANOVA test (ns, not significant). The dashed line indicates the LOD. b-e, ADE activities of the immune sera (n = 3 biologically independent samples). Enhancement curves of K562 cells infected with DENV1 (b), DENV2 (c), DENV3 (d), and DENV4 (e) in the presence of serially diluted mouse immune sera as indicated. Infected cells were quantified by flow cytometry and were normalized to the peak infectivity of FLE mAb Z6. Data are means ± SEM. f, ADE activities of the human mAb Z6. Enhancement curves of K562 cells infected with DENV1, 2, 3 and 4 in the presence of serially diluted Z6 as indicated.

Source data

Extended Data Fig. 2 Comparison of FLE mAbs bound to flavivirus-E.

a, Z6 bound to ZIKV-sE. Z6 Fab is colored in green and ZIKV-sE is colored in grey. FL is highlighted in magenta. b, 2A10G6 bound to ZIKV-sE (PDB: 5JHL). 2A10G6-Fab is colored in cyan and ZIKV-sE is colored in grey. FL is highlighted in magenta. c, E53 bound to WNV-sE (PDB: 3I50). E53 Fab is colored in orange and ZIKV-sE is colored in grey. FL is highlighted in magenta. d, The Z6, 2A10G6 and E53 bound to E are overlapped, showing different approaching angles. e, The E residues participating in interactions of antibodies are listed. The numbers of contacts are highlighted for the main contributors in each sE.

Extended Data Fig. 3 Flow cytometry detection of antigen expression and epitope display.

a, Gating strategy to monitor ZIKV antigen expression by FACS analysis, related to Fig. 1e, Extended Data Figs. 3b and 5a. b, Plasmid expressing M/E-WT and 19 E-W101 mutants as indicated were probed by FLE mAbs or neutralizing mAbs targeting DI, DII or DIII.

Extended Data Fig. 4 Phylogenetic trees flaviviruses, neutralizing titers in Ifnar^−/− mice and ADE activities of the human mAb Z5.

a, Phylogenetic trees of 73 flaviviruses based on the amino acid sequences of the E. b-c, Groups of 5-week-old female Ifnar^−/− mice (n = 7) were immunized with a single dose (1.6 × 10¹¹ vp) of AdC7-M/E-WT, -M/E-MutB or -M/E-MutC via i.m. route. PBS was given as the sham vaccine. Sera were collected at week 4, 12 and 24. Serologic binding and neutralizing antibodies were detected. Data are means ± SEM. d, Groups of 5-week-old Female Ifnar^−/− mice (n = 8-9) were immunized with a single dose (1.6 × 10¹¹ vp) of AdC7-M/E-WT, -M/E-MutB or -M/E-MutC via i.m. route. PBS was given as the sham vaccine. Sera were collected at 4 weeks postimmunization. MN₅₀ titers of NAbs were measured. Results were pooled from two independent experiments (related to Fig. 2b–e). Data are means ± SEM. P values were analyzed with two-tailed Student’s t test (ns, not significant; *, P < 0.05; P values are available in source data). The dashed line indicates the LOD. e, ADE activities of the ZIKV FLE mAb Z5. Enhancement curves of K562 cells infected with DENV1, 2, 3 and 4 in the presence of serially diluted Z5 as indicated. Infected cells were quantified by flow cytometry.

Source data

Extended Data Fig. 5 Antigen characterization of AdC7 vectored vaccines.

HEK293 cells were infected with AdC7 expressing ZIKV-M/E WT, MutB or MutC. Non-infected cells were used as the negative control. a, Flow cytometry detection of antigen expression and epitope display probed by FLE mAbs or neutralizing mAbs targeting epitopes on DI, DII or DIII. b, Capture ELISA quantifying the secretion of ZIKV antigen from AdC7 virus infected cells. Data are means of triplicates.

Source data

Extended Data Fig. 6 Protective efficacy of MutB/C-based vaccine against ZIKV infection at 3 DPI.

a-c, Groups of Ifnar1^−/− mice (n = 4) (Female:male of 2:2 in both WT and MutB groups; 3:1 in MutC group; 4:0 in sham group) received a single immunization of 1.6 × 10¹¹ vp of AdC7-M/E-WT, AdC7-M/E-MutB, AdC7-M/E-MutC via the i.m. route. PBS was given as the sham vaccine. Serum samples were collected for detection of ZIKV-E-specific IgG (a) and NAb titers (b). Thirty-days after vaccination, mice were challenged with 5 × 10⁶ FFU of ZIKV-SMGC-1 via the i.p. route. At 3 DPI, tissues from the brain, spinal cord, testis, spleen, liver and eye were harvested for measurement of ZIKV loads (c). Data are means ± SEM. The dashed line indicates the LOD.

Source data

Extended Data Fig. 7 Evaluation of the sera from donor mice, related to Figs. 3f,g.

a,b, The pooled serum from donors were tested for ZIKV-E-specific IgG (a) and NAb titers (b). The dashed line indicates the LOD. c, ADE activities of the pooled donor sera in cell culture. Shown are the enhancement curves of K562 cells infected with DENV1, DENV2, DENV3, and DENV4, respectively, in the presence of serially diluted immunized mice sera as indicated. Infected cells were quantified by flow cytometry. Relative infectivity for each sample were normalized to the peak infectivity of mAb Z6.

Source data

Extended Data Fig. 8 FACS plots showing the gating strategy to isolate single ZIKV-E+ B_GC.

The ZIKV-E-binding B_GC (GL-7⁺ B220^hi CD38^lo IgD⁻CD93⁻CD138⁻) from the lymph nodes of a group of BALB/c mice primed with ZIKV vaccines was sorted by flow cytometry.

Extended Data Fig. 9 Amino acid sequences and alignment of FLE mAbs.

a, Amino acid sequence alignment of VH-region for mAb ZWT.1-3 and 6B6C-1. b, Amino acid sequence alignment of VL-region for mAb ZWT.1-3 and 6B6C-1. c, Amino acid sequence alignment of VH-region for mAb ZWT.4-6, 4G2 and 2A10G6. d, Amino acid sequence alignment of VL-region for mAb ZWT.4-5 and 4G2. e, Amino acid sequence alignment of VL-region for mAb ZWT.6 and 2A10G6.

Extended Data Fig. 10 Overview of E-MutC bound to Z3L1_scFv and model on virus particle.

a, Overall structure of complex of sE-MutC/Z3L1_scFv. The sE-MutC is shown as ribbon, Z3L1_scFvs are showed as surface. b, Left: the cryo-electron microscopy structure of the mature ZIKV particle (PDB: 5IZ7). Middle: ZIKV-E-MutC/Z3L1-scFv docked onto the the mature ZIKV particle (PDB: 5IZ7), Right: ZIKV-E-MutC docked onto the the mature ZIKV particle (PDB: 5IZ7). The Z3L1 footprint are highlighted in green.

Supplementary information

Supplementary Information

Supplementary Tables 1, 2, 4 and 5, Fig. 1, and PDB reports for 7BPK and 7BQ5.

Reporting Summary

Peer Review Information

Supplementary Table 3

Paired V gene usage.

Source data

Source Data Fig. 1

Statistical source data.

Source Data Fig. 2

Statistical source data.

Source Data Fig. 3

Statistical source data.

Source Data Fig. 4

Statistical source data.

Source Data Fig. 6

Statistical source data.

Source Data Extended Data Fig. 1

Statistical source data.

Source Data Extended Data Fig. 4

Statistical source data.

Source Data Extended Data Fig. 5

Statistical source data.

Source Data Extended Data Fig. 6

Statistical source data.

Source Data Extended Data Fig. 7

Statistical source data.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Dai, L., Xu, K., Li, J. et al. Protective Zika vaccines engineered to eliminate enhancement of dengue infection via immunodominance switch. Nat Immunol 22, 958–968 (2021). https://doi.org/10.1038/s41590-021-00966-6

Download citation

Received: 24 February 2021
Accepted: 26 May 2021
Published: 15 July 2021
Issue Date: August 2021
DOI: https://doi.org/10.1038/s41590-021-00966-6

This article is cited by

Development of flavivirus subviral particles with low cross-reactivity by mutations of a distinct antigenic domain
- Koshiro Tabata
- Yukari Itakura
- Yasuko Orba
Applied Microbiology and Biotechnology (2023)
Current status and perspectives on vaccine development against dengue virus infection
- Jisang Park
- Ju Kim
- Yong-Suk Jang
Journal of Microbiology (2022)

Subjects

Abstract

Similar content being viewed by others

Main

Results

ZIKV vaccines expressing either prM/E or M/E induce DENV ADE

Rational design of ZIKV immunogens

MutB/C vaccines induced protective immunity without DENV ADE

MutB/C vaccines abrogate maternal–neonatal transmission

MutB/C vaccines switch immunodominance of antibody responses

Main classes of FLE mAbs to enhance DENV infection

Structural basis of the FL-engineered immunogen

Discussion

Methods

Cells, viruses and animals

Ethics statement

Construction and production of recombinant chimpanzee adenovirus

Protein expression and purification of ZIKV/DENV-sE proteins

Antibody expression and purification

Neutralization assay

Epitope detection for ZIKV-E by flow cytometry

Capture ELISA

Immunization

Animal protection against virus challenge

qRT–PCR

ELISA

In vitro evaluation of ADE of virus infection

In vivo ADE of DENV2 evaluation

Mouse pregnancy experiment

SPR assay

Isolation of antigen-specific BGC cells

Single-cell library construction and sequencing

Cellranger V(D)J data processing

Crystallization, data collection and structure determination

Statistical analysis

Reporting Summary

Data availability

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Additional information

Extended data

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Search

Quick links

Isolation of antigen-specific B_GC cells