Phylogenetic classification and conservation analysis of HPV L1 and its surface loops

We obtained 231 representative HPV types L1 sequences from the International Human Papillomavirus Reference Center [33]. According to the IARC carcinogenicity classification, all identified carcinogenic or possibly carcinogenic HPV types belong to the Alpha genus [4]. Consequently, we analyzed the evolutionary relationships of L1 protein sequences across all 65 reported HPV types within the Alpha genus, focusing specifically on the prevalence of 25 potentially carcinogenic HPV types and the low-risk types HPV6 and HPV11 among women with invasive cervical cancer in China and globally [2, 56]. As shown in Fig. 1a, HPV53, HPV56, and HPV66 belong to the Alpha 6 group, and their L1 proteins are evolutionarily closer to each other than to other high-risk or low-risk types identified by the IARC.

Fig. 1

Evolutionary relationships and biophysical characterization of HPV53, HPV56, and HPV66 L1 protein constructs. a Phylogenetic analysis of 65 alpha-genus HPV types based on L1 protein sequences, with HPV53, HPV56, and HPV66 highlighted within the α6 group, indicating their close evolutionary relationship within this clade. b Comparative structural analysis of the BC, DE, EF, FG, and HI surface loops in the L1 proteins of HPV53 (blue), HPV56 (red), and HPV66 (yellow), as predicted by the AlphaFold3 server. c SDS-PAGE and WB analysis of purified N-terminally truncated WT L1 proteins. Western blot analysis was performed under reducing conditions using SDS-PAGE, with the broad-spectrum linear monoclonal antibody 4B3 for detection. d TEM images of the VLPs, with a scale bar representing 100 nm. e Biophysical characterization of the WT VLPs, including DLS for particle size distribution, HPSEC for elution profiles, AUC for sedimentation coefficients, and DSC for thermal stability profiles, providing comprehensive insights into the structural integrity and stability of the VLPs

Using Clustal Omega software, we assessed the amino acid sequence homology of HPV53, HPV56, and HPV66 L1 proteins, revealing a substantial average homology of 75.70%. Notably, HPV56 and HPV66 exhibited a high sequence similarity of 87.87%, whereas HPV53 showed slightly lower homology with the other two, at 79.13% with HPV56 and 78.73% with HPV66 (Supplementary Fig. 1). To assess conservation within five key surface-exposed loops (BC, DE, EF, FG, and HI) of the HPV L1 protein, aligned sequences for each loop were uploaded to WebLogo (version 3.7.12) to generate conservation profiles. The analysis indicated that while the loop regions of 65 Alpha-genus HPV L1 proteins were generally less conserved (Supplementary Fig. 2), the loop structures of HPV53, HPV56, and HPV66—each belonging to the Alpha 6 group—displayed greater similarity. As shown in Fig. 1b, structural predictions generated by AlphaFold3 demonstrate substantial overlap in the FG, and HI loops of HPV53, HPV56, and HPV66, suggesting these regions as promising targets for vaccine design due to their immunogenic potential.

Expression, purification, and characterization of HPV53, 56 and 66 WT VLPs

Gene encoding for the HPV53, HPV56 and HPV66 L1 proteins were prepared as constructs with various N-terminal truncations (5-, 10-, 15-aa), and cloned into the pTO-T7 vector for protein expression in E. coli system. This led to the purification of highly pure HPV53, HPV56, and HPV66 L1 proteins, each with a molecular weight of approximately 55 kDa. These purified proteins were isolated through a combination of cation exchange column chromatography and Superdex 200 column chromatography, as illustrated in Fig. 1c. Subsequent in vitro assays facilitated the self-assembly of HPV VLPs. To comprehensively characterize the physical and chemical attributes of these VLPs, an array of analytical techniques was deployed, including TEM, HPSEC, AUC, DLS, and DSC.

TEM analysis provided critical insights into the VLPs' morphological characteristics (Fig. 1d). Majorly, the HPV53 VLPs displayed a mix of hollow spherules, including some irregularly shaped particles, whereas HPV56 and HPV66 VLPs presented a homogeneous population of spherules in size and shape. DLS data highlighted differences in particle size among the VLPs, with mean radii of 57.73 nm for HPV53, 43.86 nm for HPV56, and 38.10 nm for HPV66 VLPs. This trend was mirrored in HPSEC results, where the retention times inversely correlated with particle size—12.36 min for HPV53, 13.11 min for HPV56, and 13.44 min for HPV66. Sedimentation coefficients derived from AUC analysis were 130 S for HPV53, 134 S for HPV56, and 141 S for HPV66 VLPs, reinforcing the high purity and uniformity of the assembled VLPs. Significantly, thermal stability assessments revealed Tm values of 76.76 °C for HPV53, 83.40 °C for HPV56, and 84.59 °C for HPV66 VLPs, indicating notable thermal stability across all three types (Fig. 1e).

WT HPV53, 56, and 66 VLPs exhibit strong type-specific immunogenicity with limited cross-neutralization reactivity

To investigate the immunogenicity and cross-reactivity of WT HPV53, HPV56, and HPV66 VLPs, SPF BALB/c female mice aged 6 weeks were randomly assigned into groups (n = 5) and were subsequently immunized via intraperitoneal injection, utilizing aluminum adjuvant at doses of 5.0 μg, 1.0 μg, and 0.2 μg per mouse. The immunization regimen spanned 0, 2, and 4 weeks, with orbital venous blood collection conducted at the 6-week mark.

Utilizing ELISA, we quantified the elicited titer levels of type-specific antibodies and those capable of cross-type binding against HPV53, HPV56, and HPV66 WT VLPs, in addition to a combined VLP preparation. Remarkably, as illustrated in Fig. 2a, the induced type-specific binding antibody titers for HPV53, HPV56, and HPV66 spanned from 10^5 to 10^7 across the administered dosages, indicating potent immunogenicity. Moreover, these antibodies demonstrated significant levels of cross-reactivity; especially notable was the similarity in titer levels against HPV53 and HPV56 when induced by HPV66, approaching the autologous titer against HPV66 itself (p > 0.05 for both the 5.0 and 0.2 μg groups).

Fig. 2

Vaccination with HPV53, HPV56, and HPV66 WT VLPs induces specific and cross-reactive binding and neutralizing antibodies. a Quantification of binding antibody titers in vaccinated BALB/c mice using ELISA. b Measurement of neutralizing antibody titers in serum samples through a PBNA. BALB/c mice (n = 5) were intraperitoneally inoculated with high (5.0 μg/dose), middle (1.0 μg/dose), or low (0.2 μg/dose) doses of WT HPV53, HPV56, and HPV66 VLPs at weeks 0, 2, and 4. Binding and neutralization antibody levels were assessed at week 6 after the initial vaccination. All data were subjected to two-way ANOVA and are presented as mean ± standard deviation (SD); *P < 0.05, **P < 0.01; ***P < 0.001; ****P < 0.0001. The error bars represent the SD, and symbols denote individual mice

Further analysis of the immune responses was conducted via a PBNA, assessing both type-specific and cross-type neutralizing antibody titers against HPV53, HPV56, and HPV66. Presented in Fig. 2b, the findings revealed that each VLP type successfully elicited strong type-specific neutralizing antibodies, with titers ranging from 10^4 to 10^5 among the dosing groups. However, the induction of cross-neutralizing antibodies between the HPV types was more nuanced. The analysis indicated a variable efficacy in cross-neutralization, with generally higher antibody titers observed among genetically closer HPV types. Specifically, HPV66 elicited moderate cross-neutralizing antibody titers against HPV56, yet it was notably less effective in generating cross-neutralizing responses towards the more genetically distant HPV53 (p < 0.0001). This finding is corroborated by the results of the ED50 assay, where the ED50 metric denotes the minimum effective dose required to achieve seroconversion in 50% of the subjects within a specified experimental group. As depicted in Table 1 and Supplementary Table 1, for their respective variants, HPV53, HPV56, and HPV66 demonstrated ED50 values of 0.004 μg, 0.019 μg, and 0.043 μg, respectively. Notably, while HPV53 and HPV56 were ineffective at inducing seroconversion in at least half of the mice when attempting cross-protection against other types, HPV66 uniquely succeeded in facilitating seropositive conversion in 5 mice against the HPV56.

Table 1 Half-effective dosage (ED50) of HPV 56/66/53 WT and double-type chimeras with aluminum adjuvant in miceImpact of backbone type selection on assembly and particle uniformity in HPV53/56/66 cross-type immunogen designed

The construction of triple type chimeric molecules across HPV 53, HPV56, and HPV66 represents a formidable endeavor in our study. Preliminary estimates suggest that through surface loop swapping, at least 60 unique tripartite chimeric VLP configurations could theoretically be realized. However, the practicality of such a task is daunting, underscored by the immense workload involved. A more streamlined approach, favoring the initial combination of two HPV types before integrating the third, stands to substantially minimize the verification workload of chimeric constructs. The study systematically examines the influence of backbone types selections on the assembly efficiency and immunogenicity of chimeric molecules, an aspect hitherto unexplored in previous research. Focusing on HPV53, HPV56, and HPV66 as backbone types, the investigation meticulously compares the effects of different immunogenic epitope reshaping strategies. Specifically, ten chimeric proteins employing HPV56 and HPV66 as backbone types (defined as H56-66 BC/DE/EF/FG/HI and H66-56 BC/DE/EF/FG/HI, respectively) were designed. Parallel efforts saw the use of HPV53 as a backbone, with homologous loop regions successively swapped with those from HPV56 and HPV66, resulting in a further ten chimeric constructs (notably, H53-56 BC/DE/EF/FG/HI and H53-66 BC/DE/EF/FG/HI). Expression of all double type chimeric proteins was achieved in E. coli. Subsequently, the engineered L1 proteins underwent purification to high purity via a meticulously optimized two-step column chromatography process (illustrated in Fig. 3a) before being subjected to in vitro self-assembly.

Fig. 3

Comprehensive analysis of the double-type chimeric VLPs intertype HPV53/56/66. Both the WT and chimeric L1 proteins were subjected to reducing SDS-PAGE (a) and western blotting (b) with a wide-spectrum linear mAb 4B3. TEM imaging (scale bar: 100 nm) (c) and HPSEC profiles (d) reveal the morphology and size distribution of H56-66, H66-56, H53-56, and H53-66 chimeric VLPs. (e) AUC profiles highlight the sedimentation coefficients, aiding in understanding the complex dynamics of H56-66 and H66-56 VLPs in solution. These analyses provide insights into the physical and chemical properties of the double-type chimeric VLPs, as compared to the WT VLPs

Analytical scrutiny via TEM (demonstrated in Fig. 3b and Supplementary Fig. 3), and HPSEC (Fig. 3c), revealed distinctive particulate attributes—specifically, morphology and uniformity—across the double-type chimeric VLPs contingent on the type of backbone employed. It was observed that chimeric proteins with HPV56 and HPV66 serving as the backbone were predominantly competent in assembling in vitro into VLPs bearing close resemblance to the wild-type, entity, characterized by HPSEC column retention times spanning 12–14 min. In contrast, chimeric VLPs with HPV53 as the foundational backbone exhibited considerable heterogeneity in size and demonstrably inferior particulate morphology, evidenced by protracted HPSEC column retention times of 14–16 min. Further AUC of the ten most proficiently assembled chimeric VLPs featuring HPV56 and HPV66 as backbone types (H56-66 and H66-56) delineated sedimentation coefficients ranging from 125–140 S, analogous to those observed in the wild-type, with the singular exception of H56-66EF which manifested signs of suboptimal assembly. These compellingly suggest that chimeric constructs between the genetically closer HPV56 and HPV66 facilitate a preservation of physicochemical characteristics akin to the wild-type, portraying HPV53—owing to its marginal genetic dissimilarity—as a less favorable candidate for the skeletal backbone in chimeric molecule design.

Backbone types and their unique impact on cross-type immunogenicity in chimeric VLPs

To assess the immunogenicity and potential cross-neutralization capabilities of chimeric VLPs based on different backbone types, we formulated all 20 double-type chimeric VLPs with aluminum adjuvant and administered them to female BALB/c mice via three intraperitoneal injections. The elicited type-specific and cross-type neutralizing antibody titers were assiduously quantified using the PBNA, yielding insightful data into the molecular intricacies of VLP-induced immunity.

As illustrated in Fig. 4a, the investigation into five chimeric VLPs employing HPV56 as their foundational backbone resulted in varied neutralizing antibody titers against HPV56. Notably, the swapping of EF, FG, and HI loops with those from HPV66 (H56-66EF/FG/HI) led to a subtle reduction in titers compared to the WT HPV56, especially pronounced with the HI loop replacement. Despite these variations, the neutralization potency against HPV56 sustained a high and dose-responsive tier, implying a pivotal role of the EF, FG, and HI loops in steering type-specific neutralizing responses towards HPV56. Additionally, the comparative analysis of neutralizing antibody titers against HPV66, induced by H56-66 chimeric VLPs, underscored the efficacy of DE and HI loop substitutions, with H56-66DE and H56-66HI eliciting titers peaking at 10^4 against HPV66.

Fig. 4

Cross-immunogenicity induction by double-type chimeric VLPs based on diverse backbone type. Analysis of the immunogenic response induced by HPV56/66 (a), H53-56 and H53-66 (b) chimeric VLPs in BALB/c mice subjected to a range of immunization dosages. Groups of five BALB/c mice received intraperitoneal injections of either high (5.0 μg/dose), medium (1.0 μg/dose), or low (0.2 μg/dose) concentrations of WT HPV56, HPV66, and HPV56/66 chimeric VLPs on a schedule at weeks 0, 2, and 4. The elicitation of neutralizing antibodies was quantitatively assessed six weeks following the initial immunization, employing neutralization assays with the assay’s sensitivity threshold delineated by a dotted line. All data were subjected to one-way ANOVA and are presented as mean ± standard deviation (SD); *P < 0.05, **P < 0.01; ***P < 0.001; ****P < 0.0001. The error bars represent the SD, and symbols denote individual mice. c Heatmap representation of logarithmic fold changes in neutralizing antibody titers targeting both backbone and loop types, elicited by high-dose administrations of double-type chimeric VLPs alongside those induced by WT VLP. (d) Heatmap illustrating the sequence consistency across homologous loop regions of HPV56, HPV66, and HPV53. Darker shades indicate higher levels of consistency

Parallel outcomes were observed with chimeric VLPs having HPV66 as the backbone (H66-56), where the substitution of the DE loop significantly impacted the immunogenicity against HPV66, pointing to the DE loop's role as a potent neutralization epitope of HPV66. The strategic reconfiguration of HPV56's EF, FG, and HI loops onto the HPV66 backbone markedly elevated type-specific neutralizing antibody levels against HPV56 over those induced by WT HPV66 VLPs, especially pronounced with the HI loop replacement. Specifically, the H66-56HI chimera invoking a noteworthy increase in anti-HPV56 neutralizing titers, which ranged from 20,480 to 81,920 (5.0 μg group, P < 0.05), 1,280 to 10,240 (1.0 μg group) and 1,280 to 2,560 (0.2 μg group, P < 0.01); these were all significantly higher than those elicited by the WT HPV66 VLPs (anti-HPV56 titers of 80 to 5,120). Taken together, most HPV56/66 chimeric VLP not only enhanced titers against the heterologous type but also maintained their potency in eliciting neutralizing antibodies against their analogous types, demonstrating excellent cross-neutralization capacity.

Conversely, chimeric VLPs with HPV53 as the backbone exhibited suboptimal particle assembly and diminished cross-neutralization capacity (Fig. 4b). Notwithstanding, notable findings emerged: among the H53-56 VLP variants, only H53-56HI achieved neutralizing antibody titers against HPV56 comparable to those of WT HPV56, affirming the immunodominant nature of the HI loop for HPV56.In the case of the H53-66 VLP, the DE loop of HPV66 exhibited superior neutralizing antibody induction on the heterologous framework, mirroring the findings with HPV56 as the foundation. However, chimeras incorporating all five HPV66 loops into an HPV53 backbone proved to be less effective, highlighting the inadequacy of HPV53 as a scaffold for chimeric constructions. Additionally, regarding changes in immunogenicity specific to HPV53 itself, both H53-56 and H53-66 chimeric VLPs revealed a consistent pattern whereby substitutions of HPV53's DE, FG, and HI loops impacted its immunogenicity to varying degrees, suggesting these loop regions potentially hold more advantageous positions in the induction of neutralizing antibodies against HPV53.

We organized the logarithmic fold changes of the neutralizing antibody titers, which targeted both the backbone and loop types, elicited by high-dose administrations of double-type chimeric VLPs, in conjunction with those induced by WT VLP, into a heatmap (Fig. 4c). This visual representation effectively underscored that the HI loop of HPV56 constitutes its principal immunogenic loop, modifications in the DE loop of HPV66 are pivotal for adjusting the HPV66 titer, and replacements within the DE, FG, HI loops of HPV53 significantly influence its immunogenicity. Furthermore, we performed a detailed analysis of the sequence homology across the loop regions of these three HPV types, as depicted in Fig. 4d. Remarkably, HPV56 and HPV66 not only demonstrate a higher level of full-length sequence homology but also exhibit greater amino acid sequence consistency within their respective loop regions compared to those of HPV53, with the consistency of the FG loop between HPV53 and HPV66 comparable to that observed in the HPV56/66 FG loop. Taken together, these observations indicate that the phylogenetic closeness among HPV types is inversely related to the variability in loop regions, thus facilitating the successful execution of homologous loop substitutions.

Among the array of twenty double-type chimeric VLPs explored, H56-66DE, H66-56FG, and H66-56HI emerged as frontrunners in cross-type neutralization efficacy, prompting further evaluation of their ED50 values. As illustrated in Table 1 and Supplementary Table 2, HPV56 exhibited an ED50 of 0.021 μg against its own variant, while for HPV66, the ED50 exceeded 0.900 μg. Conversely, HPV66 demonstrated an ED50 of 0.057 μg against itself and 0.747 μg against HPV56, aligning with the previously noted elevation of HPV66 to HPV56 cross-neutralizing antibody titers. Notably, H66-56HI's ED50 values signified a remarkable enhancement in cross-immunogenicity versus the individual WT VLP, ED50 values within the discernible range against both HPV56 and HPV66, measuring 0.300 μg and 0.100 μg, respectively, highlighting its promise as a potential vaccine candidate for cross-protection studies. Thus, H66-56HI is earmarked for extended cross-immunization research, reflecting its superior immunogenic and cross-neutralization profile.

Triple-type cross-neutralization achieved by incorporating HPV53 and HPV56’s immunodominant loops into HPV66 VLP

Through homologous loop swapping between different types, we identified distinct immunodominant loop regions for HPV53, HPV56, and HPV66, providing an opportunity for the integration of immunodominant epitopes. We have demonstrated that using the HPV66 scaffold can functionally remodel the immunodominant loop HI of HPV56, enabling cross-neutralization between the two types. Cryo-EM confirmed the structural integrity of the H66-56HI capsid and determined the resolutions of HPV53, HPV56, HPV66, and H66-56HI capsids to be 10.12 Å, 10.20 Å, 18.00 Å, and 9.20 Å, respectively. These measurements followed the established FSC standard of 0.143 (see Fig. 5a–d and Supplementary Fig. 4). The structural properties observed in H66-56HI were consistent with those of the WT VLPs, with DLS quantifying an average particle radius of 32.41 nm (Fig. 5e). Importantly, H66-56HI exhibited greater thermal stability compared to WT HPV66, with its Tm peaking at 86 °C (Fig. 5f). This enhanced stability underscores its potential as a robust vaccine candidate.

Fig. 5

Cryo-EM structural analysis and biophysical characterization of WT and H66-56HI VLPs (a–d) Cryo-EM elucidates the intricate structural details of WT HPV53 (a), 56 (b), 66 (c), and H66-56HI chimeric VLPs (d), with resolutions of 10.12 Å, 10.20 Å, 18.00 Å, and 9.20 Å, respectively. Radial color gradients from 200 Å to 300 Å accentuate conformational subtleties. e DLS measurements provide particle size distribution, and (f) DSC analyzes the thermal stability of the H66-56HI chimeric VLP

Extending epitope integration across multiple HPV types resulted in the development of the tri-type chimeric H66-56HI-53FG VLP. Following standard expression, purification, and in vitro assembly protocols, well-assembled chimeric VLPs were obtained (Fig. 6a–b). Despite undergoing dual epitope remodeling, these VLPs retained structural characteristics similar to the wild-type counterparts. Cryo-EM analysis revealed that H66-56HI-53FG adopts a regular icosahedral symmetry (T = 7), comparable to HPV53 and HPV66 WT VLPs, with a resolution of 14.22 Å (FSC = 0.143), further demonstrating structural consistency with H66-56HI (Fig. 6c). Immunogenicity evaluation in a mouse model (Fig. 6d) demonstrated that H66-56HI-53FG significantly improved neutralizing antibody titers against exogenous loop types (P < 0.05) while maintaining its ability to induce neutralizing antibodies against its own type. This improvement was especially evident in the high-dose group and remained significant in the low-dose group. Therefore, our integration strategy combining dominant epitopes with optimized scaffold VLPs successfully generated cross-immunogens capable of neutralizing multiple HPV types. This approach enhances our understanding of multi-type cross-immunogen design and highlights the potential of complex antigen design to broaden the protective scope of HPV vaccines.

Fig. 6

Comprehensive analysis of H66-56HI-53FG VLPs. a Schematic of recombinant chimeric HPV53/56/66 L1 proteins, highlighting the immunodominant loops of WT HPV53 (pink) and HPV56 (yellow) on an HPV66 scaffold (gray), with TEM confirmation (scale bar: 100 nm). b Integrity of the H66-56HI-53FG construct assessed via reduced SDS-PAGE and western blot using mAb 4B3, with particle uniformity examined by HPSEC elution profiles and AUC sedimentation coefficients. c TEM image of H66-56HI-53FG VLPs, showing well-formed particles (scale bar: 100 nm, left panel). Cryo-EM reconstruction of H66-56HI-53FG VLPs reveals an icosahedral structure with radial color gradients ranging from 180 Å to 300 Å, highlighting key structural features (middle panel). The resolution of the VLPs was determined to be 14.22 Å, based on the FSC threshold of 0.143 (right panel). d Immunogenic potential was assessed in BALB/c mice (n = 5) with intraperitoneal injections of VLPs at 5.0 or 1.0 μg/dose at weeks 0, 2, and 4. Neutralizing antibody responses were measured at week 6 post-initial dose using specific neutralization assays, with sensitivity thresholds indicated by dotted lines. ANOVA analysis of results, presented as mean ± SD, shows statistical significance (*P < 0.05, **P < 0.01; ***P < 0.001; ****P < 0.0001), with error bars and individual data points indicating immune response variability