Enhancing Vaccine Immunogenicity of H9N2 Influenza HA by Locking Its Pre-Fusion Conformation via Cleavage Site Engineering

Simple Summary

The hemagglutinin (HA) protein of influenza A virus is cleaved into HA1 and HA2 subunits during maturation. However, whether the subsequent conformational change alters HA immunogenicity remains unclear. In this study, we identified a circular loop structure flanking the cleavage site that is critical for HA expression (P7~P1, P′1~P′12). We generated a panel of mutants that either increased or decreased the cleavage efficiency of the HA0 precursor. To investigate the impact of conformational change on immunogenicity, we assessed these HA variants in a mouse model. Specifically, the Dlt5 (P6~P1, P′1~P′13) mutant reduced cleavage efficiency, resulting in elicited immunogenicity and significantly higher antibody titers in mice. In contrast, all other cleavable mutants induced similar antibody levels, suggesting the stabilized, conformationally locked HA protein possesses greater immunogenicity.

Abstract

Avian influenza (AI) significantly threatens poultry health and causes major economic losses in the poultry industry. Vaccination remains crucial for AI prevention and control. The major protective epitopes of influenza viruses are located on hemagglutinin (HA), a surface glycoprotein essential for viral infection. Most influenza vaccines induce neutralizing antibodies against HA to block viral entry. HA maturation requires the HA0 precursor to be proteolytically cleaved at a conserved site by host proteases to yield HA1 and HA2 subunits. A subsequent acidic condition triggers HA conformational changes, enabling viral–host membrane fusion. However, whether HA conformational variations affect immunogenicity remains unclear. In this study, the cleavage site of the HA gene from an H9N2 avian influenza virus was modified to block the proteolytic cleavage of the HA protein. Our results revealed distinct proteolytic patterns of certain mutants, which exhibited either increased or decreased cleavage efficiencies compared to the wild-type (WT) HA. However, none of the mutants exhibited completely abolished HA0 cleavage. To assess the immunogenicity of these variants, BALB/c mice were immunized with DNA vaccines expressing either WT or mutant HA proteins. Strikingly, the mutant HA protein with a 19-amino-acid deletion Dlt5 (P6~P1, P1’~P′13) at the cleavage site exhibited reduced cleavage efficiency and induced significantly higher HI antibody titers compared to the WT. These results offer valuable perspectives for enhancing avian influenza vaccine efficacy through strategic modification of HA cleavage properties.

1. Introduction

Avian influenza is an acute respiratory disease, affecting poultry and waterfowl, caused by avian influenza viruses (AIVs) [1]. The avian influenza viruses pose a threat to public health as well [1]. The influenza A virus genome comprises eight distinct segments that encode 10–16 viral proteins. Among these, HA and neuraminidase (NA) genes encode the two principal glycoproteins that form the viral surface spikes [2]. HA contains the major neutralizing epitopes, which are attractive targets for vaccine and drug development [3]. The HA glycoprotein exhibits a bipartite structure, which consists of a globular head domain positioned distally from the viral envelope and a stem-like domain proximal to the viral membrane [4]. The influenza virus genome encodes the hemagglutinin precursor HA0, which undergoes post-translational glycosylation in the endoplasmic reticulum and Golgi apparatus of infected cells, subsequently forming homotrimers through the action of molecular chaperones. Following this maturation process, HA0 undergoes proteolytic cleavage by extracellular or intracellular proteases, generating the disulfide-linked HA1 and HA2 subunits [5]. This cleavage event represents a critical step for both the maturation of HA and the production of infectious influenza virus particles [6]. During the infection, influenza viruses enter cells through receptor-mediated endocytosis. Following receptor binding, HA undergoes proteolytic cleavage by host proteases (e.g., TMPRSS2 and furin), which primes the protein for pH-dependent conformational changes in late endosomes. The structural rearrangements expose the fusion peptide, enabling fusion between the viral envelope and endosomal membrane [6], to release viral ribonucleoprotein complexes into the cytoplasm. However, how the proteolytic cleavage affects the immunogenicity of the HA protein remains largely unclear.

HA is the dominant protective antigen of the influenza virus, presenting key neutralizing epitopes on its surface that elicit potent antibody responses. A critical determinant of its antigenic characteristics is its conformational structures. For instance, Wei Wang et al. demonstrated that the conformational structure of HA is correlated with the stimulation of broadly neutralizing monoclonal antibodies targeting the stem region [7]. A complementary finding revealed that acid-induced conformational changes in HA can abolish the binding of HA-specific monoclonal antibodies, indicating that the native pre-fusion structure is essential for effective antibody recognition [8]. These findings collectively underscore a direct link between HA conformational structure and immunogenicity. Although proteolytic cleavage and the subsequent low-pH-induced conformational rearrangement of HA are essential steps in the life cycle of the influenza virus, how the proteolytic cleavage of the HA protein affects the antigenicity of the HA protein remains largely unclear.

Proteolytic cleavage is the prerequisite of conformational change in the HA protein. To demonstrate the relationship between proteolytic cleavage and immunogenicity of the HA protein, the cleavage site was systematically modified to create a series of HA mutants. The cleavage efficiency and antigenicity of HA mutants were evaluated in this study. The results showed that the mutant HA protein with a 19-amino-acid deletion (Dlt5: P6~P1, P′1~P′13) at the cleavage site exhibited reduced cleavage efficiency and induced significantly higher HI antibody titers compared to the WT. This finding correlates the conformational structure of the HA protein with its antigenicity and provides a rationale for engineering optimized vaccine antigens.

2. Materials and Methods

2.1. Site-Directed Mutagenesis

In this study, the hemagglutinin (HA) gene from an H9N2 avian influenza virus (A/Avian/Shanghai/SH04/2014) was cloned into the expression vector PRE-3FLAG to generate a recombinant plasmid 3FLAG-H9HA-WT. This plasmid then served as the template for PCR-mediated site-directed mutagenesis, which was performed using specifically designed overlapping primers containing the desired nucleotide substitutions, following a standard quick-change mutagenesis protocol. Briefly, a linear, double-stranded DNA fragment incorporating the mutation was amplified using a high-fidelity DNA polymerase (Vazyme Biomed, Nanjing, China). After PCR, the reaction mixture was treated with DpnI endonuclease (TransGen Biotech, Beijing, China) to selectively digest methylated parental DNA templates, leaving the newly synthesized, unmethylated mutant DNA strands intact. The resulting linear dsDNA product bearing homologous ends was subsequently transformed into competent Escherichia coli cells for circularization and propagation [9]. Finally, all mutant plasmid constructs were confirmed by DNA sequencing (Tsingke Biotech, Beijing, China). Amino acids along the cleavage site were systematically substituted one by one, with all mutations detailed in

Table 1. Summary of clones with mutated cleavage site sequences.

Name	Amino Acid Sequence of the Cleavage Sites	Mutagenesis Site
Wild-type HA (WT)	SRSSR↓GLF	A334S
M1	SRSST↓GLF	R338T
M2	SRSST↓ELF	G339E
M3	SRSIT↓ELF	S337I
M4	SRSST↓EGF	L340G
M5	SRYIT↓ELF	S336Y
M6	SRSST↓EGS	F341S
M7	STYIT↓ELF	R335T
M8	SRSIT↓EGS	S337I
M9	STYIT↓EGF	L340G
M10	SRYIT↓EGS	S336Y
M11	STYIT↓EGS	F341S
R344V	SRSSV↓GLF	R338V
R338Q	SRSSQ↓GLF	R338Q
MC	SRKKR↓GLF	S336K, S337K
MC5	SRRRKKR↓GLF	Insert RR

(“↓” denotes the cleavage site. Green letter represents mutations for moderate change on physicochemical properties of amino acids, red letter represents mutations for change on basic amino acids).

. Recombinant plasmids containing either deletions, insertions, or peptide replacements at the cleavage site were generated using a similar strategy, with the specific construction details provided in

Table 2. Summary of mutagenesis approaches for generating deletion mutants, and strategies for screening residues near the circular loop critical to HA expression.

Name	Deletion/Substitution/Insertion
Dlt1	Delete SRSSR↓GLF (P5~P1, P′1~P′3)
Dlt2	Delete P5~P1, P′1~P′3, then insert AAAAAAAA
Dlt3	Delete P5~P1, P′1~P′3, then insert GSGGSGGGG
Dlt4	Delete P11~P1, P′1~P′20, then insert 4 α-helices (15 aa)
Dlt5	Delete P6~P1, P′1~P′13
Dlt6	Delete P7~P1, P′1~P′12
Dlt7	Delete P11~P1, P′1~P′20
Dlt8	Delete P6~P1, P′1~P′11
Dlt9	Delete P9~P1, P′1~P′20
Dlt10	Delete P9~P1, P′1~P′20, then substitute P15~P10, P′21~P′26 with GGGSGGGSGGGS
Dlt11	Delete P9~P1, P′1~P′20, then substitute P19~P10, P′21~P′31 with GSGGSGGGSGGGSGGGSGGSG
Dlt12	Delete P19~P1, P′1~P′31
Dlt13	Delete P19~P1, P′1~P′31, then insert GGGSGGGSGGGS
Dlt7+3	Delete P7~P1, P′1~P′3
Dlt7+6	Delete P7~P1, P′1~P′6
Dlt7+12	Delete P7~P1, P′1~P′12
Dlt11+3	Delete P11~P1, P′1~P′3
Dlt5+16	Delete P5~P1, P′1~P′16
Dlt5+20	Delete P5~P1, P′1~P′20

“↓” denotes the cleavage site.

. More deletions were performed at both sides of the cleavage site to define the enzyme recognition site, with all variants listed in Table 2. All primers are available upon request.

The conserved cleavage site was designated using standard protease nomenclature. The arginine (R) residue was defined as the P1 position (toward the HA1 N-terminus, with preceding residues as P2, P3, etc.), while the glycine (G) residue was defined as the P′1 position (toward the HA2 C-terminus, with following residues as P′2, P′3, etc.). Throughout this article, “↓” denotes the cleavage site within the amino acid sequence. The structures of WT and circle loop deletion HA were predicted by SWISS-MODEL (SWISS-MODEL (expasy.org)).

2.2. Western-Blotting

The constructed plasmids were transfected into HEK293T cells at 90% confluency. After 48 h, cells were harvested and lysed in RIPA buffer (Yeasen, Shanghai, China) supplemented with protease inhibitor cocktail. Whole-cell lysates (WCL) were subjected to Western blot analysis to assess HA protein expression. Equal amounts of protein were separated by SDS-PAGE and subsequently transferred onto a PVDF membrane (GE Healthcare, Chicago, IL, USA). The membranes were blocked with 5% skim milk in TBST overnight at 4 °C. HA0 and HA1 bearing an N-terminal FLAG tag were probed using an anti-FLAG primary antibody (Yeasen, Shanghai, China). Following incubation with an HRP-conjugated secondary antibody, target proteins were visualized by enhanced chemiluminescence (ECL) detection (Yeasen, Shanghai, China). HA cleavage efficiency was quantified by calculating the ratio of HA1 to HA0 band intensities. Densitometric analysis was performed using ImageJ (version 154) software to estimate the gradient with three independent experiments.

2.3. Animal Study

Thirty-four 4-week-old female BALB/c mice were randomly divided into six experimental groups (as outlined in

Table 3. The HA mutants selected for animal study.

Groups	Plasmids	Number of Mice	Cleavage Levels
1	WT	6	Moderate
2	MC5	6	Increased
3	M4	6	Decreased
4	M10	6	Decreased
5	Dlt5	6	Decreased
6	Mock	4	N/A

N/A: Not applicable.

). Based on previous results, four mutant plasmids with distinct cleavage efficiencies and a WT HA expressing plasmid were selected: the WT (moderate cleavage, serving as a baseline), mutant MC5 (slightly enhanced cleavage, high expression), mutant M4 (reduced cleavage), mutant M10 (reduced cleavage), and mutant Dlt5 (reduced cleavage).

Each mouse received intramuscular injections of plasmid DNA (5 μg per dose) according to a two-dose schedule: a primary immunization on Day 0, followed by a booster at week two. Terminal blood collection was performed two weeks after the booster immunization to assess the immune response. Briefly, blood was collected from anesthetized mice (using Zoletil^®50, Virbac, Carros, France) via the retro-orbital plexus with capillary tubes. The samples were incubated at 37 °C to allow for clotting, stored at 4 °C overnight, and subsequently centrifuged to isolate the serum.

2.4. Hemagglutination Inhibition Assay (HI)

Serum antibody levels were evaluated using a standard HI assay. Briefly, a homologous H9N2 virus (A/Avian/Shanghai/SH04/2014) was standardized to 4 hemagglutination (HA) per 25 µL units for the assay. Mouse blood samples were collected and allowed to clot at 37 °C for 1 h, followed by overnight incubation at 4 °C. After centrifugation at 2000× g to remove red blood cells (RBCs), serum was isolated and treated with receptor-destroying enzyme (RDE; Denka Seiken, Tokyo, Japan) to remove nonspecific inhibitors. HI titers were determined by serial two-fold dilutions of the treated sera in V-bottom 96-well plates. An equal volume of the standardized virus solution was added to each serum dilution and incubated at room temperature for 30–60 min, followed by the addition of a 0.5–1% suspension of chicken RBCs. The reciprocal of the highest serum dilution that completely inhibited hemagglutination was recorded as the HI titer.

2.5. Statistical Analysis

All experiments included internal controls (e.g., β-actin) and were performed in at least three independent replicates. Statistical analysis was performed using one-way ANOVA in GraphPad Prism Version 5.0 (GraphPad software Inc., La Jolla, CA, USA). Significance levels were defined as *: p < 0.05, **: p < 0.01, and ***: p < 0.001.

3. Results

3.1. Site-Directed Mutagenesis of Single Amino Acids

To alter the cleavage stability of the HA protein, mutations were introduced into the amino acids at the conserved HA cleavage site. Using 3FLAG-H9HA-WT as a template, site-directed mutagenesis was performed to generate mutants by substituting, inserting, or deleting residues at or near the HA cleavage site. In this manuscript, the symbol “↓” in the protein sequence denotes the cleavage site, with arginine and glycine in the R/G motif designated as P1 and P′1, respectively. Amino acid substitutions were selected based on side-chain similarity to minimize potential disruption to the overall structure of the HA protein.

To investigate the cleavage site, amino acids were substituted on both sides one by one. Figure 1A shows that none of the single-site mutations completely eliminated hemagglutinin cleavage. The cleavage efficiency was assessed based on the width ratio of the HA0 and HA1 bands. Statistical analysis illustrated that the prototype WT, with nearly equal HA0 and HA1 band widths, served as the reference for medium cleavage, whilst M1, M2, M3 and M4 cleaved with reducing efficiency. This trend was, however, reversed in M5, M6 and M7. Further, the cleavage started to reduce from M8, M9, M10 and M11. All the mutants showed reduced cleavage efficiency compared to the prototype, among which M4 and M10 showed the lowest cleavage efficiency, and M10 cleaved significantly less than WT.

The mutants H9-R344V and H9-R338Q were constructed to introduce single-point mutations R338V and R338Q at the P1 position. Previous reports suggested that these mutations would abolish HA protein cleavage. In contrast, our results demonstrated that the hemagglutinin proteins carrying the R338V and R338Q mutations remained cleavable (Figure 1B). Furthermore, the R338Q mutant resulted in an almost undetectable protein expression level.

To enhance HA protein cleavage, the serine residues at P2 and P3 were mutated to lysine (mutant MC), and a double-arginine was inserted between P3 and P4 (mutant MC5). The results showed that only MC5 mutants exhibited slightly increased cleavage efficiency, with a higher expression level (Figure 1B).

3.2. Deletion or Substitution of the Cleavage Motif

Since single-site mutagenesis failed to eliminate HA0 cleavage, a strategy involving deletion or substitution of the flanking cleavage site sequence was employed. More mutants with the cleavage site deletion were constructed. Dlt1, Dlt2, Dlt3, Dlt4, Dlt5 and Dlt7 modifications reduced the cleavage efficiency compared to the WT protein. Dlt6 is a mutant containing a 19-amino-acid deletion, wherein a circular loop encompassing P7~P1 and P′1~P′12 was deleted, leading to a loss of HA expression, as shown in Figure 2A. Notably, the results reveal the importance of the circle loop in HA stability.

To further verify the necessary motif required for cleavage, more amino acids were deleted or replaced. Deleting P6~P′11, P9~P′20 and P19~P′31, which deleted 17, 29 and 50 amino acids, respectively, resulted in Dlt8, Dlt9 and Dlt12. The results in Figure 2A showed that Dlt8 and Dlt9 showed significantly higher cleavage efficiency of HA0, whilst Dlt12 showed a very low expression level. Based on Dlt8, Dlt9 and Dlt12, we further employed additional GS peptides to replace the original sequences. Specifically, Dlt10 was generated by replacing P15~P10, P′21~P′26 of Dlt9 with 12GS; Dlt11 was produced by replacing P19~P10 and P′21~P′31 of Dlt9 with 21GS; and Dlt13 was constructed by inserting additional 12GS into Dlt12. The results showed that Dlt10 was still able to cleave. Although Dlt11 and Dlt12 could not be expressed, Dlt13, which contained a 50-amino-acid deletion and 12GS insertion, was able to be expressed and to be cleaved as well (Figure 2A).

To investigate the key amino acids near the circle loop that affected the expression and cleavage efficiency of the HA protein, the residues along the cleavage site were mutated with two or three amino acids at a time. Detailed fragment deletions with P7~P′3, P7~P′6, P7~P′12, P11~P′3, P5~P′16 and P5~P′20 were performed. Figure 2B shows that, except for the P7~P′12 deleting mutant (Dlt6), all other HA proteins can be expressed successfully. This result indicates that the amino acids in P7~P′12, which form a circular structure, are important for the successful expression of the HA protein. The predicted structures of the HA P7~P′12 circular loop (WT) and Dlt6 are shown in Figure 3A,B.

3.3. HA Proteins with Different Cleavage Efficiencies Induce Varying Levels of Antibodies in Mice

To determine whether the antigenicity of HA differs between pre-fusion and post-fusion states, antibody titers induced by HAs with varying cleavage efficiencies were measured. Serum HI titers were analyzed after immunization with DNA vaccines expressing WT or mutant HA proteins. As shown in Figure 4, the WT inoculated group exhibited a serum HI titer of 3.5 Log2, which was significantly higher than the mock control, confirming the induction of specific antibodies. Among the mutants, the titer in the M4 inoculated group (reduced cleavage) was 3.5 Log2, the same as that in the WT group, which was slightly higher than those in the MC5 and M10 groups (HI titers were 3 Log2). Notably, immunization with the Dlt5 mutant, which carries a 19-amino-acid deletion, resulted in a significantly elevated HI titer of 6.67 Log2. This suggests that the HA protein with cleavage site (P6~P1, P1′~P′13) deletion exhibits enhanced immunogenicity, which may be due to the low cleavage efficiency in vivo.

4. Discussion

While research on HA cleavage sites is fundamental to virology and vaccinology, this study provides mechanistic insights and practical validation. Specifically, our investigation focused on how cleavage site modifications alter HA processing efficiency and immunogenicity.

To study the immunogenicity of epitopes on the pre-fusion conformation of the HA protein, an uncleavable form of HA is required. Although mutations at H9-R344V and H9-R338Q have been reported (R343V [10] and R329Q [11], respectively) to block HA cleavage, the same mutations in this study failed to inhibit cleavage. According to the structure of the HA protein cleavage site, a total of 19 amino acids P7~P1 and P′1~P′12 formed a circular structure [11]. Dlt6 was designed to delete these 19 amino acids. Notably, deletion of this circular structure resulted in the failure of HA expression. The mutation has been constructed twice with different methods. The results infer that the circular structure is important for HA expression or stability.

Studies showed that increasing the number of basic amino acids at the cleavage site promotes HA cleavage [12,13,14,15]; therefore, MC and MC5 variants were designed. As expected, the HA containing more basic amino acids at the cleavage site exhibited slightly enhanced cleavage efficiency than the WT HA, but not significantly. It may partially be because there are other proteases in vivo that recognize this cleavage site other than furin protease [16]. Furthermore, the MC5 variant showed elevated expression efficiency that has not been reported before.

Surprisingly, HA proteins with deletions of up to 38 amino acids (Dlt13) remained cleavable. However, the specific host proteases responsible for HA protein activation remain poorly characterized, and only a limited number have been reported so far. Stieneke Gröber et al. first showed that influenza virus HA proteins with polybasic cleavage sites undergo activation through cleavage by furin protease [17]. A key molecular determinant of highly pathogenic avian influenza viruses is a polybasic cleavage site in the HA protein, which enables intracellular cleavage by furin. AIVs are classified into two categories, highly pathogenic avian influenza viruses (HPAIVs) and low pathogenic avian influenza viruses (LPAIVs), based on pathogenicity and HA cleavage sites [18]. The proteolytic cleavage of HA protein is an essential prerequisite for its conformational transformation in acidic intracellular environments [5], which can be cleaved by furin and proprotein convertase 5/6 (PC5/6) [16], supporting the systemic infection. However, the proteases involved in the cleavage of low-pathogenic avian influenza virus HA proteins have not yet been well understood [19]. In addition, the cleavage recognition site is not clear. LPAIVs require trypsin-like serine proteases present exclusively in the respiratory tract [20], thereby confining the viral tropism. Most of the influenza virus HA can be cleaved at the site of infection by host cell surface proteases, such as TMPRSS2, TMPRSS4, human respiratory trypsin-like proteases, kallikrein-related peptidases 5 (KLK5) and 12 (KLK12), and Matrix/ST 14 [10,21]. In the present study, the HA protein remained cleavable even when its canonical cleavage site was completely deleted. This finding suggests that the site may not be absolutely specific and that host proteases might recognize and cleave to the HA protein based on structural features through alternative mechanisms.

Our study investigated the HI antibody titers induced by HA cleavage site mutants. A key finding was that deleting a 19-amino-acid region near the cleavage site (Dlt5) significantly enhanced immunogenicity. The structure of the loop region is conservative among H1, H3, H5 and H9 subtypes, which allows the recognition of the structure [22]. We propose that this deletion impairs intracellular cleavage in vivo, preventing the HA protein from undergoing its conformational change. As a result, HA remains locked in its pre-fusion state, which presents potent neutralizing epitopes and ultimately leads to the induction of higher antibody titers. Consistent with this principle, a separate investigation on H5N1 demonstrated that mutating the HA cleavage site (RRKRR to T) yielded a more immunogenic vaccine candidate [23]. Other studies on yeast expression systems have further shown that removing the polybasic motif (e.g., RRKRR in H5N1 HA) can improve immunogenicity by promoting the formation of stable, higher-order oligomers [24]. An alternative and possible explanation is that the Dlt5 mutant may exhibit enhanced expression or stability in vivo, which would be consistent with its increased immunogenicity despite having comparable in vitro levels. Therefore, future studies employing purified proteins are warranted to address this. It is important to note that while HI titers served as a practical and well-correlated initial screening for immunogenicity in this study, they do not directly measure virus-neutralizing capacity. The enhanced HI responses observed for the Dlt5 mutant strongly suggest improved antigenicity and predicted the potential for eliciting functional antibodies. Future studies employing microneutralization assays and animal challenge experiments will be essential to confirm the direct protective efficacy conferred by this stabilized HA conformation. Consequently, cleavage site engineering represents a versatile approach to optimizing vaccine antigens, primarily through conformational stabilization as demonstrated, and potentially through mechanisms such as enhanced oligomerization.

In summary, this study systematically evaluated the impact of HA cleavage site modifications. A critical circular loop was identified, whose deletion abrogates HA expression in 293T cells. More importantly, the results demonstrated that deleting a 19-amino-acid region near the cleavage site generates potent, stabilized vaccine antigens, thereby eliciting significantly higher HI antibody titers in immunized mice. The enhanced immunogenicity may be a consequence of stabilization of the mutant HA in the pre-fusion conformation in vivo, which optimally exposes neutralizing epitopes, thereby eliciting a superior antibody response compared to the post-fusion state. Further study is warranted to verify the protective efficacy of the engineered antigen. Our work links conformational structures via cleavage site engineering to enhanced immunogenicity, paving the way for rationally designed next-generation influenza vaccines.