Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2022 May 20;119(21):e2200821119. doi: 10.1073/pnas.2200821119

Influenza chimeric hemagglutinin structures in complex with broadly protective antibodies to the stem and trimer interface

Xueyong Zhu a,1, Julianna Han a,1, Weina Sun b, Eduard Puente-Massaguer b, Wenli Yu a, Peter Palese b,c, Florian Krammer b,d, Andrew B Ward a,2, Ian A Wilson a,e,2
PMCID: PMC9173763  PMID: 35594401

Significance

Sequential vaccination with chimeric hemagglutinins (cHAs) that contain the same stem, but different head domains, can direct the antibody immune response to the conserved stem region. Structural information on cHAs is limited to one example containing a group 1 stem. Here, X-ray and electron microscopy structures revealed that cHAs cH4/3 and cH15/3, which contain a group 2 stem, with broadly protective stem antibody 31.a.83, have intact, native-like stem regions. Furthermore, the cHA head domains were in an open conformation, which exposes the conserved trimer interface region. Structure determination of cH15/3 HA with 31.a.83 and head interface antibody FluA-20 suggests that broadly protective antibodies could be elicited against both the conserved HA stem and the head interface.

Keywords: universal vaccine design, chimeric influenza hemagglutinin, X-ray crystallography, negative-stain electron microscopy, HA trimer interface and stem

Abstract

Influenza virus hemagglutinin (HA) has been the primary target for influenza vaccine development. Broadly protective antibodies targeting conserved regions of the HA unlock the possibility of generating universal influenza immunity. Two group 2 influenza A chimeric HAs, cH4/3 and cH15/3, were previously designed to elicit antibodies to the conserved HA stem. Here, we show by X-ray crystallography and negative-stain electron microscopy that a broadly protective antistem antibody can stably bind to cH4/3 and cH15/3 HAs, thereby validating their potential as universal vaccine immunogens. Furthermore, flexibility was observed in the head domain of the chimeric HA structures, suggesting that antibodies could also potentially interact with the head interface epitope. Our structural and binding studies demonstrated that a broadly protective antihead trimeric interface antibody could indeed target the more open head domain of the cH15/3 HA trimer. Thus, in addition to inducing broadly protective antibodies against the conserved HA stem, chimeric HAs may also be able to elicit antibodies against the conserved trimer interface in the HA head domain, thereby increasing the vaccine efficacy.

Influenza virus infection continues to be a constant threat to public health worldwide, although the frequency of infections has dramatically decreased since the emergence of COVID-19 and the ensuing countermeasures (e.g., masks, social distancing). Influenza viruses evolve and escape host immune responses by antigenic drift through amino acid mutation and increased glycosylation in the hemagglutinin (HA) and neuraminidase (NA) surface antigens (seasonal viruses), and antigenic shift through acquisition of zoonotic HAs and NAs (pandemic viruses). Annual seasonal influenza vaccines are generally effective against the viruses circulating that year in the human population, but they are suboptimal or ineffective against drifted seasonal or novel pandemic strains mainly due to sequence variation in the HA major surface glycoprotein. Efforts are currently ongoing for development of more broadly protective or universal influenza vaccines (1–3). One such strategy focuses on the HA, which is composed of an immunodominant, highly variable, head domain and an immunosubdominant, but more conserved, stem (or stalk) domain (4). Antibodies targeting the conserved stem region of HA are capable of conferring protection against a broad range of virus strains and subtypes (5–7).

Many approaches for stem-based universal influenza vaccines are under development, such as design of headless HAs by removing the globular head to create modified HAs that encompass the HA1 and HA2 components of the stem domain (8–10), and through design of a hyperglycosylated HA head domain to mask the more variable regions of the globular head (11). We developed a different approach through generating chimeric HAs (cHA) to focus the immune response on the HA stem (12). Indeed, the 2009 pandemic H1N1 virus was somewhat akin to this situation, as the H1 head differed substantially from the HAs in prior seasonal H1N1 viruses, and boosted the titer of antibodies against the conserved HA stem region (13). The cHA molecules contain either conserved H1 or H3 stem domains, representing influenza A virus group 1 and group 2 HAs, respectively, which are attached to “exotic” head domains from HA subtypes to which humans would not normally have been exposed (14). Sequential vaccination of cHA proteins with the same stem domain, but varied head domains, directs the immune response to the subdominant stem region (12, 13). Recently, some of the authors here reported proof-of-concept phase I clinical trial results of a group 1 cHA-based universal influenza virus vaccine (15). Vaccination with cH8/1 (H8 head and H1 stem) and cH5/1 (H5 head and H1 stem) viruses was found to be safe and induced a broad, strong, durable, and functional immune response to the HA stem region, suggesting cHAs have the potential to be developed as universal vaccines.

Even though the stem domain remains the main target for broadly protective vaccine design, we and others recently reported epitopes in the HA head protomer–protomer interface (i.e., trimer interface) that were not originally thought to be accessible in the HA prefusion conformation (16–18). Some interface-targeting antibodies can react with most influenza A virus subtypes and confer in vivo protection. Among them, antibody FluA-20 was isolated from a vaccinated patient who received inactivated seasonal influenza vaccines for over two decades and participated in clinical trials of experimental H5N1 and H7N9 vaccines (16). FluA-20 has the largest breadth among such antibodies and was able to bind 14 of 15 tested subtypes in both group 1 and group 2 HAs (16). Crystal structures of FluA-20 Fab in complex with HA head domains of H1 and H3 revealed FluA-20 binds to an epitope largely buried in the HA trimer interface that consists of conserved residues in the 220-loop and the 90-loop. Electron microscopy studies revealed that FluA-20 can induce disassociation of native HA trimers, consistent with requirement for structural flexibility of the heads in the HA trimer to gain accessibility to this epitope (16).

The structural information of cHA has been limited so far to only one cryoelectron tomography structure for a group 1 cH5/1 HA displayed on the surface of influenza virus (19). The cH5/1 HA structurally differed from native HAs where the cH5/1 head was rotated about 60° relative to the stem when compared to native H1 or H5 HA structures, leading to a much more open configuration of the cH5/1 trimers.

Here, we report on crystal and negative-stain electron microscopy (nsEM) structures of two group 2 cHAs cH4/3 (H4 head and H3 stem) and cH15/3 (H15 head and H3 stem) and complexes with broadly protective antistem antibodies 31.a.83 (20) and CR9114 (21) that neutralize group 1 and group 2 influenza A viruses, and with FluA-20, a broad HA head trimer interface antibody.

Results

Chimeric cH4/3 HA in Complex with Broadly Protective Antistem Antibody 31.a.83 Has an Intact Stem but Open Head.

The ectodomain of cH4/3 HA was designed to contain the head domain (residues 53 to 276a of HA1 with an egg-passage adaptive mutation G225E, H3 numbering) of H4 HA from the H4N6 virus of A/duck/Czechoslovakia/1956 and the stem domain (HA1 residues 11 to 52, 277 to 329 and HA2 residues 1 to 176) of H3 HA from the H3N2 virus of A/Hong Kong/4801/2014 (HK14). The cH4/3 HA was overexpressed in a baculovirus expression system, as previously described (22). After trypsin cleavage to remove the trimerization domain and 6xHis-tag, the purified cH4/3 HA protein was recovered in mostly trimeric form.

To elucidate the structural basis for the cH4/3 HA elicitation of broadly protective antibodies against influenza virus, we determined the crystal structure of cH4/3 HA in complex with 31.a.83 Fab to 4.9 Å resolution (Fig. 1 and SI Appendix, Table S1). The chimeric cH4/3 HA retains its ability to bind the antistem broadly protective antibody 31.a.83 with one Fab per HA protomer (three Fabs per trimer) (Fig. 1A), indicating that the stem epitope is faithfully preserved in the cH4/3 construct. However, comparisons with native H4 or H3 HA structures (Fig. 1B) show that, while the cH4/3 HA stem domain is intact, the cH4/3 HA globular head is more open. Compared with native H4 HA, the cH4/3 HA head domain is rotated away from the HA trimer center by about 10° around the head/stem domain hinge (HA1 residues 53 and 276a) and also about 5° counterclockwise around the HA trimer threefold axis (Fig. 1B).

Fig. 1.

Fig. 1.

Open in a new tab

Crystal structure of cH4/3 HA in complex with 31.a.83 Fab and nsEM of cH4/3 HA. (A) Overall structure of cH4/3 HA in complex with 31.a.83 Fab. One HA–Fab protomer is colored green for H4 HA head domain and pink for H3 HA stem domain, yellow for the Fab light chain, and cyan for the Fab heavy chain. The other HA and Fab protomers are in light gray. (B) Surface representation of cH4/3 HA and comparison with the native H4 structure from A/duck/Czechoslovakia/1956 (H4N6) (PDB ID code 5XL1) and H3 structure A/Victoria/361/2011 (H3N2) (PDB ID code 4WE8) after alignment of the corresponding stem domains in standard side view (Upper) and top view (Lower). The stem domains are colored pink, and the three head domains of each HA trimer are colored green, blue, and magenta. The same coloring scheme is used in Figs. 2B and 3C. (C) Representative 3D reconstruction (gray) of cH4/3 HA in nsEM. Upper depicts an averaged reconstruction of all cH4/3 HA particles. Lower depicts subclassified cH4/3 HA reconstructions showing a variety of flexible states from tight trimer (light pink) to flexible and disassembling trimer (maroon).

Furthermore, we assessed the conformation and stability of apo (unliganded) cH4/3 HA using nsEM. A variety of flexible states of intact cH4/3 HA were observed in the two-dimensional (2D) class averages ranging from a more closed trimer to more open forms where the heads were widely separated resulting in loosely associated but still intact trimers (SI Appendix, Fig. S1). Three-dimensional (3D) classification and reconstruction of cH4/3 HA particles yielded density maps of HA with the head domains in different states of openness (Fig. 1C), consistent with the 2D class averages (SI Appendix, Fig. S1).

Chimeric cH15/3 HA in Complex with 31.a.83 Fab also Has an Intact Stem but Widely Open Head.

The structural integrity of cHA was further explored for another group 2 cH15/3 HA. The ectodomain of cH15/3 HA consists of an HA head domain (residues 53 to 276a of HA1 with egg-passage adaptive mutation I282F, H3 numbering) of H15 HA from the H15N9 virus of A/shearwater/West Australia/2576/79 and an HA stem domain (HA1 residues 11 to 52, 277 to 329 and HA2 residues 1 to 176) of HK14 H3 HA. The chimeric cH15/3 HA contained the same stem domain as cH4/3 HA and was overexpressed in a baculovirus expression system. The purified cH15/3 HA protein was recovered in mostly trimeric form after thrombin cleavage to remove the trimerization domain and 6xHis-tag.

The crystal structure of cH15/3 HA in complex with 31.a.83 Fab was determined at 4.6 Å resolution (Fig. 2 and SI Appendix, Table S1). One Fab was bound per HA protomer (three per trimer), also indicating structural integrity of the 31.a.83 epitope in cH15/3 HA. Similar to cH4/3 HA, the cH15/3 HA stem domain was intact, but the head domain of cH15/3 HA was open to a much larger extent than cH4/3 and was rotated away from HA trimer center about 15° around the head/stem domain hinge (HA1 residues 53 and 276a) (Fig. 2B). Due to the openness of the HA head domain, the more extended 260-loop of the cH15/3 HA head domain as well as the 50-loop was able to contact the light-chain variable domain of 31.a.83.

Fig. 2.

Fig. 2.

Open in a new tab

Crystal structure of cH15/3 HA in complex with 31.a.83 Fab and nsEM of cH15/3. (A) Overall structure of cH15/3 HA in complex with 31.a.83 Fab. One HA–Fab protomer is colored green for H15 HA head domain and pink for H3 HA stem domain, yellow for the Fab light chain, and cyan for the Fab heavy chain. (B) Surface representation of cH15/3 HA and comparison with the native H4 HA structure from A/shearwater/West Australia/2576/79 (H15N9) (PDB ID code 5TG8) and native H3 HA structure from A/Victoria/361/2011 (H3N2) (PDB ID code 4WE8) after alignment of the corresponding stem domains in standard side view (Upper) and top view (Lower). The coloring scheme is the same as in Fig. 1B. (C) Representative 3D reconstruction of cH15/3 HA in nsEM. The Upper panel depicts an averaged reconstruction of all cH15/3 HA particles. The Lower panel depicts subclassified c15/3 HA reconstructions showing a variety of flexible states from tight trimer (sky blue) to flexible and disassembling trimer (dark blue).

For apo cH15/3 HA, we also observed flexibility in both the head and stem domains by nsEM. In the 2D class averages, the intact cH15/3 HA ranged from a relatively tight trimer to widely open head and stem domains (SI Appendix, Fig. S2). Three-dimensional classification and reconstruction resulted in HA density maps with a substantial amount of separation between protomers in both the head and stem domains (Fig. 2C). When classified in 3D, a range of flexible states was observed, with the most open state being visualized as disassembled head and stem domains (Fig. 2C and SI Appendix, Fig. S2).

Chimeric cH15/3 HA in Ternary Complex with 31.a.83 Fab and HA Head Trimer Interface Antibody FluA-20 Fab.

As the head domains of cH4/3 and cH15/3 HAs displayed more open conformations, both in the unliganded state and when bound to 31.a.83 (Figs. 1 and 2), we considered whether antibodies that target the trimer interface in the globular head, such as FluA-20, could bind these cHAs. Bio-layer interferometry (BLI) assays revealed strong binding of FluA-20 to cH15/3 HA, but weak binding to cH4/3 HA (Fig. 3A), consistent with our findings that the HA head domains of cH15/3 HA were more open than cH4/3 HA (Figs. 1B and 2B). In the nsEM analysis, the cH4/3 HA trimer was found to form some complexes with FluA-20 Fab, although most HA trimers remained unbound (SI Appendix, Fig. S3). However, cH15/3 HA trimer was fully bound with FluA-20 Fabs (three Fabs per trimer) (SI Appendix, Fig. S4), which is consistent with the BLI results.

Fig. 3.

Fig. 3.

Open in a new tab

Crystal structure of cH15/3 HA in complex with 31.a.83 and FluA-20 Fabs. (A) The binding traces of cH15/3 and cH4/3 HAs (at a concentration of 1.065 μM) to immobilized FluA-20 Fab in a BLI assay. (B) Overall structure of cH15/3 HA in complex with 31.a.83 Fab and FluA-20 Fab. One HA–Fab protomer is colored green for H15 HA head domain and pink for H3 HA stem domain, yellow for the 31.a.83 Fab light chain and cyan for the 31.a.83 Fab heavy chain, brown for the FluA-20 Fab light chain, and magenta for the FluA-20 Fab heavy chain. The other HA and Fab protomers are in light gray. (C) Surface representation of cH15/3 HA from its complex with 31.a.83 Fab, and cH15/3 HA from its complex with 31.a.83 Fab and FluA-20 Fab, after alignment of the corresponding stem domains in standard side view (Upper) and top view (Lower). The coloring scheme is the same as in Fig. 1B.

To provide further structural characterization of FluA-20 interactions with cH15/3 HA, we determined a crystal structure of cH15/3 HA in complex with FluA-20 and 31.a.83 Fabs at 5.4 Å resolution (Fig. 3 and SI Appendix, Table S1). In this ternary complex, the H3 stem domain remained intact and similar to the native H3 HA structure; the 31.a.83 Fab epitope on cH15/3 HA stem domain had an estimated buried surface area (BSA) (23) of 620 Å2. The more open head allows FluA-20 to bind to its epitope in the trimer interface of cH15/3 HA head with an estimated BSA of 450 Å2. Unlike the cH4/3 HA complex, cH15/3 HA 260-loop exhibited no contacts with the 31.a.83 Fab light chain, but instead interacted with FluA-20 Fab from a neighboring cH15/3 HA protomer with an estimated BSA of 160 Å2. The 50-loop remained in close proximity to the 31.a.83 light chain variable domain (Fig. 3B) as in cH4/3 HA complex (Fig. 2A). In this ternary complex, the two antibody Fabs contacted each other with the 31.a.83 Fab interacting with FluA-20 Fab from a neighboring HA protomer with an estimated BSA of 260 Å2.

These findings further indicate that the flexibility of the head domain allows cH15/3 HA to adopt more than one open conformation as revealed by comparison of cH15/3 HA in the binary complex with 31.a.83 Fab and in the ternary complex with 31.a.83 and FluA-20 Fabs (Fig. 3C). Upon FluA-20 Fab binding, the HA head rotates about 15° back toward the trimer center around the head/stem hinge region, but about 30° clockwise around the trimer threefold axis.

The cH15/3 HA epitopes of 31.a.83 and FluA-20 are located in the HA central stem and HA head trimer interface, respectively (Fig. 4A). FluA-20 is bound to the 220-loop and 90-loop epitope residues in the HA head trimer interface (Fig. 4B and SI Appendix, Fig. S5A), which is usually buried in the native HA trimer (16). In our previous study on FluA-20 (16), the HA epitope residues on 220-loop and 90-loop were found to be generally conserved in both group 1 and group 2 HAs, consistent with the breadth of protection by FluA-20. Moreover, FluA-20 also contacts the 260-loop residues from a neighboring cH15/3 HA protomer (Figs. 3B and 4B and SI Appendix, Fig. S5A); in H15 HA, the 260-loop has a long insertion in comparison with other subtypes (SI Appendix, Fig. S5B). The four potential FluA-20 contacting residues HA1 Pro261D-Ser-Gly-Ile261G in the 260-loop are from this insertion and do not have bulky side chains, which suggests that only the main chain of the 260-loop is primarily involved in interaction with FluA-20. It would not be expected that the 260-loop would form such interactions with other HA subtypes.

Fig. 4.

Fig. 4.

Open in a new tab

FluA-20 contacts the open HA heads of the cH15/3 HA trimer. (A) Surface representation of cH15/3 HA showing the epitopes for 31.a.83 (in brown) and FluA-20 (in magenta). One cH15/3 protomer is colored in green for the HA head and cyan for the HA stem. For clarity, the other two protomers in cH15/3 HA trimer are colored in gray. (B) Surface representation of cH15/3 HA with FluA-20 Fab in the ternary complex of cH15/3 HA with 31.a.83 and FluA-20 Fabs. FluA-20 heavy and light chains are colored in yellow and light gray, respectively.

Since cH15/3 HA was observed to also have some flexibility in the stem domain in nsEM, we tested binding of the broadly protective antistem antibody CR9114 (21) to the canonical stem epitope of cH15/3 HA (SI Appendix, Fig. S6). We observed CR9114 Fab bound cH15/3 HA stem in full stoichiometry (three Fabs per trimer) by nsEM that appeared to rigidify not only the stem domain but also the head domain. When FluA-20 Fab was added to this complex, FluA-20 bound cH15/3 HA in full 3:1 stoichiometry (SI Appendix, Fig. S7), as in the X-ray structure of the ternary complex with FluA-20 and 31.a.83 Fabs.

FluA-20 Neutralizes cH4/3- and cH15/3-Expressing Influenza Viruses.

It has been shown that FluA-20 and other antibodies targeting the same interface epitope have no neutralizing activity against wild-type influenza viruses in vitro because their epitope is cryptic and typically not accessible on intact prefusion HA. However, the slightly more open structure of the cH4/3 HA head and the wide-open structure of the cH15/3 HA head suggested that FluA-20 may be able to neutralize these viruses. We performed plaque reduction neutralization assays with a wild-type HK14 virus as well as cH4/3HK14N2HK14 and cH15/3HK14N2HK14 viruses that shared the HA stem domain, as well as NA with the HK14 virus. An antistem antibody, mAb 9H10, was used as control, which should neutralize all three viruses. Indeed, 9H10 neutralized these viruses with 50% inhibitory concentrations (IC50) of 0.54 μg/mL, 3.86 μg/mL, and 6.32 μg/mL, respectively (Fig. 5). As expected, FluA-20 had negligible neutralizing activity against wild-type HK14 virus [n.b., did not neutralize but still protected (16)]. However, the cH4/3HK14N2HK14 and cH15/3HK14N2HK14 viruses were efficiently neutralized by FluA-20 with IC50s of 0.16 μg/mL and 0.29 μg/mL, respectively.

Fig. 5.

Fig. 5.

Open in a new tab

Neutralizing activity of FluA-20 against wild type H3N2, cH4/3N2 and cH15/3N2 viruses. (A) Neutralization of H3N2 virus of A/Hong Kong/4801/2014 (HK14) by FluA-20 and stem-targeting mAb 9H10. Neutralization activity of the same antibodies against viruses cH4/3HK14N2HK14 (B) and cH15/3HK14N2HK14 (C), which both possess the stem domain (and NA) from HK14 H3N2 virus.

Discussion

Recently, the results from a randomized, placebo-controlled phase 1 trial (NCT03300050) on group 1 cHA-based influenza virus vaccine candidates containing cH5/1 and cH8/1 HAs were disclosed (15). This study revealed that immunizations with influenza virus vaccines containing these two group 1 cHAs were safe and induced strong, broad, durable, and functional antistem antibodies. Here, to aid in development of group 2 cHA vaccine candidates, we determined crystal structures of cH4/3 HA and cH15/3 HA in complex with the cross-reactive influenza A (group 1 and group 2) antistem antibody 31.a.83. The crystal structures revealed intact, native-like cH4/3 HA and cH15/3 HA stem regions that were recognized by the 31.a.83 antibody, demonstrating their potential to induce broad protective antibodies to the central stem region and perhaps other stem epitopes (24). Furthermore, the HA heads were in a more open conformation compared to their individual homotypic structures (H4 and H15 HAs). EM studies of cH4/3 or cH15/3 HAs also demonstrated substantial structural flexibility in the head, as well as some flexibility in the stem region. The more open cHA heads in cH4/3 and cH15/3 are most likely due to the mismatch of their heterologous head domains attached to the H3 stem (SI Appendix, Fig. S8).

As the cHAs adopted a more open head configuration, particularly for cH15/3 HA, we tested and demonstrated that antibody FluA-20, which recognizes the HA head trimer interface, could bind cHAs, enabling structural studies of cH15/3 HA with 31.a.83 and FluA-20 in a ternary complex. While 31.a.83 bound to a conserved epitope in the HA stem domain (20), FluA-20 recognized the open HA head trimer interface where the critical epitope residues are conserved across most subtypes of influenza A virus (16). In previous studies, FluA-20 binding to the HA trimer destabilized the trimeric interface of native HA, and EM studies revealed that the HA0 trimer dissociated to Fab-bound monomeric HA upon FluA-20 binding (16). This structural plasticity or “breathing” of the native HA head domain can therefore allow antibodies to access cryptic epitopes that are only transiently exposed. For cH15/3 HA, due to the more open HA head domain, FluA-20 is able to bind the trimer interface epitope while the HA remains in a trimeric state. It should be noted that the vaccine cHA construct differs from the cHA construct here, as it is membrane-embedded (15) compared to the soluble ectodomain in the X-ray, EM, and binding studies. However, efficient neutralization of live cH4/3HK14N2HK14 and cH15/3HK14N2HK14 virus by FluA-20 provides evidence that cH4/3 and cH15/3 HAs likely also display a more open head conformation in their membrane-bound form on virions. While antistem antibodies have been the focus of universal vaccine design, the unexpected finding that FluA-20 can stably bind cH15/3 HA suggests that two distinct types of protective antibodies could be elicited by some cHA vaccine candidates (15).

Structural differences of cHAs can be seen compared to native HAs as in this study of cH4/3 and cH15/3 HAs, and in a previous study of cH5/1 HA (19), which likely arise from fitting an HA head from a different subtype on the HA stem of another subtype (e.g., H1 or H3). Some diversity has also been found in the natural configuration (or relative disposition) of HA heads on HA stems, depending on the HA subtype (25). Thus, cHA constructs where a mismatched head is attached to a particular stem may also be subject to different dispositions (rotation and translation) of their heads relative to the stem. Notwithstanding, influenza viruses expressing cHA constructs grow normally, and can infect cells and induce production of broadly protective antibodies in clinical trials and animal studies (12, 15). One key criterion for vaccine design is that the candidate construct should maintain a native-like conformation of the epitopes targeted for engagement by protective antibodies. The structural integrity of HA stem domains, or at least the conserved central stem epitopes, such as in cH4/3 HA and cH15/3 HA, indicates that the cHAs are suitable for elicitation of broadly protective antibodies against these conformational epitopes in the stem.

Thus, the structural and binding results here suggest that the cHAs could potentially elicit protective antibodies to multiple highly conserved epitopes that would further enhance their efficacy as a universal vaccine, providing a foundation for future human clinical trials of group 2 cHAs.

Materials and Methods

Cloning, Baculovirus Expression, and Purification of cHAs.

The ectodomains of influenza cH4/3 and cH15/3 HAs were expressed essentially as previously described (22). Chimeric H4/3 HA consists of an H4 head domain (HA1 residues 53 to 276a; H3 numbering) from A/duck/Czechoslovakia/1956 (H4N6) with a cell passage adaptive mutation G225E, and an H3 stem domain (HA1 residues 11 to 52, 277 to 329 and H3 HA2 residues 1 to 176) from A/Hong Kong/4801/2014 (H3N2). Chimeric H15/3 HA consists of an H15 head domain (HA1 residues 53 to 276a) from A/shearwater/West Australia/2576/79 (H15N9) with a cell passage adaptive mutation I282F, and an H3 stem domain (HA1 residues 11 to 52, 277 to 329 and H3 HA2 residues 1 to 176) from A/Hong Kong/4801/2014 (H3N2). Both cH4/3 and cH15/3 HAs were expressed in a baculovirus system for structural and functional analyses. The cDNAs corresponding to the ectodomain of cH4/3 and cH15/3 HAs were incorporated into a baculovirus transfer vector, pFastbacHT-A (Invitrogen), with an N-terminal gp67 signal peptide, a C-terminal foldon trimerization domain and His6-tag, with a trypsin cleavage site and thrombin cleavage site incorporated to enable cleavage between the HA ectodomain and the foldon trimerization domain/His6-tag. The constructed plasmids were used to transform DH10bac competent bacterial cells by site-specific transposition (Tn-7 mediated) to form a recombinant bacmid with β-galactosidase blue–white receptor selection. The purified recombinant bacmids were used to transfect Sf9 insect cells for overexpression. HA protein was produced in suspension cultures of High Five cells with recombinant baculovirus at a multiplicity of infection of 5 to 10 and incubated at 28 °C shaking at 110 RPM. After 72 h, High Five cells were removed by centrifugation and supernatants containing secreted, soluble HA protein was concentrated and buffer-exchanged into 20 mM Tris pH 8.0, 150 mM NaCl, and then further purified by metal affinity chromatography using Ni-nitrilotriacetic acid (NTA) resin (Qiagen). For crystal structure determination, the foldon trimerization domain and His6-tag were cleaved from the cHA ectodomain with thrombin for cH15/3 HA (n.b. trypsin cleaves c15/3 HA into pieces, likely due to wide open HA head) and with trypsin for cH4/3 HA (which gave more uniform cleavage to HA1/HA2), and purified further by size-exclusion chromatography on a Hiload 16/90 Superdex 200 column (GE Healthcare) in 20 mM Tris pH 8.0, 150 mM NaCl, and 0.02% NaN3. The purified HAs were quantified by optical absorbance at 280 nm, and purity and integrity were analyzed by reducing and nonreducing sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS/PAGE).

Expression of Recombinant Antibody Fabs in Mammalian Cells.

The synthesized DNA fragments of light- and heavy-chain variable regions (VL and VH) of human antibodies 31.a.83 (20) and FluA-20 (16) were cloned into a mammalian cell expression vector phCMV containing the corresponding human λ or κ CL region, and the CH1 region of human IgG1, respectively. The Fabs were expressed by transient cotransfection of the Fab heavy- and light-chain expression vectors into ExpiCHO cells (Thermo Fisher Scientific) following methods that we reported previously (26). Affinity chromatography using CaptureSelect CH1-XL affinity matrix (Thermo Fisher Scientific) was used to purify recombinant Fabs from culture supernatant and followed by size-exclusion chromatography using a Superdex 200 column (GE Healthcare). Purified Fabs were quantified by optical absorbance at 280 nm and buffered in 20 mM Tris pH 8.0, 150 mM NaCl, and 0.02% NaN3. The purity and integrity of recombinant Fabs were analyzed by reducing and nonreducing SDS/PAGE.

Plaque Reduction Neutralization Assay.

Dilutions of mAbs 9H10 (27) and FluA-20 (16) IgGs in 1× minimum essential medium (MEM) were incubated with 50 μL of H3N2 viruses of A/Hong Kong/4801/2014 (HK14), cH4/3HK14N2HK14, cH15/3HK14N2HK14 (14) at 30 (targeted) plaque forming units (PFU) per well of a 12-well plate for 1 h at room temperature on a shaker. cH4/3HK14N2HK14 and cH15/3HK14N2HK14 viruses possess the stem domain (and NA) from HK14 H3N2 strain. The incubated solution was then transferred onto a cultured monolayer of Madin–Darby canine kidney (MDCK) cells in 12-well cell culture plates, followed by 1-h incubation and intermittent rocking every 15 min. The agar overlay applied was supplemented with the corresponding mAb dilutions, and plates were incubated at 37 °C for 48 h. 4% (vol/vol) p-formaldehyde in phosphate buffered saline at pH7.4 (PBS) was added to each well to fix cells at 4 °C for 1 h. The plaques were stained with antistalk mAbs 9H10 and 12D1 at a final concentration of 10 μg/mL in PBS for 1 h at room temperature with shaking. Following washes with PBS, anti-mouse IgG horseradish peroxidase (HRP) secondary antibody (Sigma-Aldrich) diluted to 1:3,000 was added to wells. The plates were developed using TrueBlue reagent (KPL), and the number of plaques was counted. Percent inhibition or plaque reduction was calculated in comparison to an irrelevant antibody control (anti-Ebola virus mAb 2E5), and the IC50, defined as the concentration of the mAb at which there is 50% plaque reduction, was calculated on Prism (GraphPad) using a nonlinear regression model.

Crystal Structure Determination.

The sitting-drop vapor diffusion method was used for crystallization experiments on our robotic CrystalMation system (Rigaku). Diffraction quality crystals for the cH4/3 HA ectodomain trimer in complex with 31.a.83 Fab were obtained by mixing 0.1 µL of the HA protein at 6.2 mg/mL in 20 mM Tris pH 8.0, 150 mM NaCl and 0.02% (vol/vol) NaN3 with 0.1 µL of the well solution in 0.1 M imidazole pH 8.0 and 10% (wt/vol) polyethylene glycol 8,000 at 20 °C. The cocrystals of cH4/3 HA with 31.a.83 Fab were cryoprotected in mother liquor with addition of 20% (wt/vol) polyethylene glycol 200 before being flash cooled at 100 K. The complex of cH15/3 HA and 31.a.83 Fab at 6.3 mg/mL was crystallized in 0.1 M Tris, pH 7.0, 40% (wt/vol) polyethylene glycol 300 and 5% (wt/vol) polyethylene glycol 1,000 at 20 °C. The cocrystals of cH15/3 HA with 31.a.83 Fab were flashed-cooled at 100 K without additional cryoprotectant. The ternary complex of cH15/3 HA with 31.a.83 Fab and FluA-20 Fab at 6.1 mg/mL was crystallized in 0.1 M sodium cacodylate, pH 6.5, 0.2 LiSO4, and 30% (wt/vol) polyethylene glycol 400 at 20 °C. The cocrystals of cH15/3 HA with 31.a.83 Fab and FluA-20 Fab were flashed-cooled at 100 K without additional cryo-protectant. We also tried extensively to crystallize cH4/3 and cH15/3 with different antibody Fabs to achieve higher resolution, but could not obtain higher quality crystals.

Diffraction data were collected at synchrotron beamlines, and integrated and scaled with HKL2000 (28). Statistics of the data collection are summarized in SI Appendix, Table S1.

Crystal structures were determined by molecular replacement (MR) using the program Phaser (29). The cH4/3 HA and 31.a.83 Fab complex structure was determined using the H4 head domain structure from A/duck/Czechoslovakia/1956 (H4N6) (PDB ID code 5XL1), the H3 stem domain structure from A/Victoria/361/2011 (H3N2) (PDB ID code 4WE8), which shares 98% amino acid identity with H3 from A/Hong Kong/4801/2014 (H3N2), and the 31.a.83 Fab structure (PDB ID code 5KAQ) as input MR models. The structure of cH15/3 HA with 31.a.83 Fab was determined using the H15 head domain structure from A/shearwater/West Australia/2576/79 (H15N9) (PDB ID code 5TG8), the H3 stem domain structure from A/Victoria/361/2011 (H3N2) (PDB ID code 4WE8), and 31.a.83 Fab structure (PDB ID code 5KAQ) as input MR models. The cH15/3 HA structure in complex with Fabs 31.a.83 and FluA-20 was determined using the H15 head domain structure from A/shearwater/West Australia/2576/79 (H15N9) (PDB ID code 5TG8), the H3 stem domain structure from A/Victoria/361/2011 (H3N2) (PDB ID code 4WE8), and the 31.a.83 Fab structure (PDB ID code 5KAQ) and FluA-20 Fab structure (PDB ID code 6OCB) as input MR models. Initial rigid-body refinement was performed in REFMAC5 (30), and subsequently rigid-body and group ADP refinement was carried out in Phenix (31). Models were examined and modified with the program Coot to assess model fit and any potential clashes (32). Final refinement statistics for these moderate resolution structures are summarized in SI Appendix, Table S1. The quality of the structures was analyzed using the QC-Check program initially developed at JCSG (qc-check.usc.edu) and MolProbity (33). All figures were generated with PyMol (www.pymol.org).

Bio-Layer Interferometry.

cHAs binding with antibody Fabs were evaluated by BLI using an Octet Red instrument (ForteBio). For binding of cH15/3 HA and cH4/3 HA with FluA-20 Fab, the FluA-20 Fab with an His6-tag at 50 µg/mL in 1× kinetics buffer (1× PBS, pH 7.4, 0.01% BSA, and 0.002% Tween 20) was loaded onto Ni-NTA biosensors and incubated with 1.065 μM of cH4/3 HA or cH15/3 HA in 1× kinetics buffer. Experimental data were fit with a 1:1 binding model. All binding data were collected at 28 °C.

Immune Complexes for EM Analysis.

CR9114 Fab expressed and purified from High Five insect cell culture (21) was incubated with cH15/3 HA in a greater than 3:1 molar ratio for 30 min at room temperature prior to mounting on grids for EM analysis. To make FluA-20 Fab–cHA complexes, FluA-20 Fab was added to cH4/3 HA, cH15/3 HA, or cH15/3 HA complexed with CR9114 Fab immediately before mounting the samples on grids. The total complexing time with FluA-20 Fab was about 90 s, as the negative-stain process takes 90 s after applying the sample.

nsEM.

Protein samples were diluted to 10 μg/mL in TBS buffer (20 mM Tris, pH 7.6, 150 mM NaCl) prior to mounting on glow-discharged 400 mesh copper grids (Electron Microscopy Sciences) previously coated with a carbon evaporator (Cressington). Excess sample was wicked off the grid using filter paper, then the grid was stained with 2% (wt/vol) uranyl formate for 30 s After wicking off the first application of stain with filter paper, the grid was stained a second time with uranyl formate for 45 s. After removing the second application of uranyl formate, grids were air dried for 1 min before storage.

Protein samples were imaged on a T20 electron microscope operating at 200kV with an Eagle CCD 4k camera (FEI). Micrographs were collected using Leginon, particles were picked using difference of Gaussian picker and Appion, and particles were classified and reconstructed in Relion (34–37). Figures were made using University of California, San Francisco Chimera (38) and Photoshop CS6.

Cryo-EM studies were also attempted on the cHA-Fab complexes but without success, which may be due in part to the inherent flexibility of these complexes.

Supplementary Material

Supplementary File

Acknowledgments

We thank Henry Tien for automated robotic crystal screening, and Bill Anderson, Hannah Turner, and Charles Bowman for running the electron microscopy suite computational resources at The Scripps Research Institute. The work was supported in part by NIH National Institute of Allergy and Infectious Diseases (NIAID) Collaborative Influenza Vaccine Innovation Centers Contract 75N93109C00051 (to I.A.W., A.B.W., P.P., and F.K.). The work was also partially funded by NIH Centers of Excellence for Influenza Research and Response 75N93021C00014 (to P.P. and F.K.), and NIAID Grants P01 AI097092-07 (to P.P.) and AI145870-03 (to P.P.). J.H. was funded by NIAID 2 T32 AI007244-36. X-ray diffraction data were collected at the Advanced Photon Source (APS) beamline 23ID-B (GM/CA CAT) and the Stanford Synchrotron Radiation Lightsource (SSRL) beamline 12-2. GM/CA CAT is funded in whole or in part with federal funds from the National Cancer Institute (Y1-CO-1020) and the National Institute of General Medical Sciences (NIGMS) (Y1-GM-1104). Use of the APS was supported by the US Department of Energy (DOE), Basic Energy Sciences, Office of Science, under Contract DE-AC02-06CH11357. The SSRL is a Directorate of SLAC National Accelerator Laboratory and an Office of Science User Facility operated for the US DOE of Science by Stanford University. The SSRL Structural Molecular Biology Program is supported by the DOE Office of Biological and Environmental Research, and by the NIH, NIGMS (including P41GM103393), and the National Center for Research Resources (NCRR) (P41RR001209). The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIAID, NIGMS, NCRR, or NIH.

Footnotes

Reviewers: K.L., University of Washington; and S.S., University of British Columbia.

Competing interest statement: The Icahn School of Medicine at Mount Sinai has filed patent applications regarding chimeric HA influenza virus vaccines that list F.K. and P.P. as inventors. F.K. and Dr. Kelly Lee (reviewer) were middle coauthors on a 39-author paper (https://rupress.org/jem/article/215/6/1571/42397/Nucleoside-modified-mRNA-vaccines-induce-potent-T) in May 2018, but did not directly collaborate.

This article contains supporting information online at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.2200821119/-/DCSupplemental.

Data Availability

The atomic coordinates and structure factors have been deposited in the Protein Data Bank (PDB) (accession codes 7U8J for cH4/3 HA in complex with 31.a.83 Fab, 7U8L for cH15/3 HA in complex with 31a.83 Fab, and 7U8M for cH15/3 HA in complex with 31.a.83 and FluA-20 Fabs). nsEM 3D maps have been deposited in the Electron Microscopy Databank (EMDB) under accession codes EMD-26928–EMD-26932 for apo cH4/3 HA maps, EMD-26933–EMD-26935 and EMD-26937–EMD-26938 for apo cH15/3 HA maps, and EMD-26939 for the cH15/3 HA in complex with CR9114 Fab map. All other study data are included in the main text and/or SI Appendix.

References

  • 1.Henry C., Palm A. E., Krammer F., Wilson P. C., From original antigenic sin to the universal influenza virus vaccine. Trends Immunol. 39, 70–79 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Nachbagauer R., Palese P., Is a universal influenza virus vaccine possible? Annu. Rev. Med. 71, 315–327 (2020). [DOI] [PubMed] [Google Scholar]
  • 3.Sangesland M., Lingwood D., Antibody focusing to conserved sites of vulnerability: The immunological pathways for ‘universal’ influenza vaccines. Vaccines (Basel) 9, 125 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Julien J. P., Lee P. S., Wilson I. A., Structural insights into key sites of vulnerability on HIV-1 Env and influenza HA. Immunol. Rev. 250, 180–198 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Krammer F., Palese P., Influenza virus hemagglutinin stalk-based antibodies and vaccines. Curr. Opin. Virol. 3, 521–530 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Wu N. C., Wilson I. A., Structural biology of influenza hemagglutinin: An amaranthine adventure. Viruses 12, 1053 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Nath Neerukonda S., Vassell R., Weiss C. D., Neutralizing antibodies targeting the conserved stem region of influenza hemagglutinin. Vaccines (Basel) 8, 382 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Impagliazzo A., et al. , A stable trimeric influenza hemagglutinin stem as a broadly protective immunogen. Science 349, 1301–1306 (2015). [DOI] [PubMed] [Google Scholar]
  • 9.Yassine H. M., et al. , Hemagglutinin-stem nanoparticles generate heterosubtypic influenza protection. Nat. Med. 21, 1065–1070 (2015). [DOI] [PubMed] [Google Scholar]
  • 10.Valkenburg S. A., et al. , Stalking influenza by vaccination with pre-fusion headless HA mini-stem. Sci. Rep. 6, 22666 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Eggink D., Goff P. H., Palese P., Guiding the immune response against influenza virus hemagglutinin toward the conserved stalk domain by hyperglycosylation of the globular head domain. J. Virol. 88, 699–704 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Krammer F., Pica N., Hai R., Margine I., Palese P., Chimeric hemagglutinin influenza virus vaccine constructs elicit broadly protective stalk-specific antibodies. J. Virol. 87, 6542–6550 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Pica N., et al. , Hemagglutinin stalk antibodies elicited by the 2009 pandemic influenza virus as a mechanism for the extinction of seasonal H1N1 viruses. Proc. Natl. Acad. Sci. U.S.A. 109, 2573–2578 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Chen C. J., et al. , Influenza A viruses expressing intra- or intergroup chimeric hemagglutinins. J. Virol. 90, 3789–3793 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Nachbagauer R., et al. , A chimeric hemagglutinin-based universal influenza virus vaccine approach induces broad and long-lasting immunity in a randomized, placebo-controlled phase I trial. Nat. Med. 27, 106–114 (2021). [DOI] [PubMed] [Google Scholar]
  • 16.Bangaru S., et al. , A site of vulnerability on the influenza virus hemagglutinin head domain trimer interface. Cell 177, 1136–1152.e18 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Watanabe A., et al. , Antibodies to a conserved influenza head interface epitope protect by an IgG subtype-dependent mechanism. Cell 177, 1124–1135.e16 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Bajic G., et al. , Influenza antigen engineering focuses immune responses to a subdominant but broadly protective viral epitope. Cell Host Microbe 25, 827–835.e6 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Tran E. E., et al. , Cryo-electron microscopy structures of chimeric hemagglutinin displayed on a universal influenza vaccine candidate. MBio 7, e00257 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Joyce M. G., et al. ; NISC Comparative Sequencing Program, Vaccine-induced antibodies that neutralize group 1 and group 2 influenza A viruses. Cell 166, 609–623 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Dreyfus C., et al. , Highly conserved protective epitopes on influenza B viruses. Science 337, 1343–1348 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Xu X., Zhu X., Dwek R. A., Stevens J., Wilson I. A., Structural characterization of the 1918 influenza virus H1N1 neuraminidase. J. Virol. 82, 10493–10501 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Connolly M. L., Analytical molecular surface calculation. J. Appl. Cryst. 16, 548–558 (1983). [Google Scholar]
  • 24.Guthmiller J. J., et al. , Broadly neutralizing antibodies target a hemagglutinin anchor epitope. Nature 602, 314–320 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gamblin S. J., Skehel J. J., Influenza hemagglutinin and neuraminidase membrane glycoproteins. J. Biol. Chem. 285, 28403–28409 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zhu X., et al. , Structural basis of protection against H7N9 influenza virus by human anti-N9 neuraminidase antibodies. Cell Host Microbe 26, 729–738.e4 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Tan G. S., et al. , Characterization of a broadly neutralizing monoclonal antibody that targets the fusion domain of group 2 influenza A virus hemagglutinin. J. Virol. 88, 13580–13592 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Otwinowski Z., Minor W., Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326 (1997). [DOI] [PubMed] [Google Scholar]
  • 29.McCoy A. J., Grosse-Kunstleve R. W., Storoni L. C., Read R. J., Likelihood-enhanced fast translation functions. Acta Crystallogr. D Biol. Crystallogr. 61, 458–464 (2005). [DOI] [PubMed] [Google Scholar]
  • 30.Murshudov G. N., Vagin A. A., Dodson E. J., Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D Biol. Crystallogr. 53, 240–255 (1997). [DOI] [PubMed] [Google Scholar]
  • 31.Adams P. D., et al. , PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66, 213–221 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Emsley P., Lohkamp B., Scott W. G., Cowtan K., Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Chen V. B., et al. , MolProbity: All-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12–21 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Suloway C., et al. , Automated molecular microscopy: The new Leginon system. J. Struct. Biol. 151, 41–60 (2005). [DOI] [PubMed] [Google Scholar]
  • 35.Lander G. C., et al. , Appion: An integrated, database-driven pipeline to facilitate EM image processing. J. Struct. Biol. 166, 95–102 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Scheres S. H., RELION: Implementation of a Bayesian approach to cryo-EM structure determination. J. Struct. Biol. 180, 519–530 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Zivanov J., et al. , New tools for automated high-resolution cryo-EM structure determination in RELION-3. eLife 7, e42166 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Pettersen E. F., et al. , UCSF Chimera—A visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary File
pnas.2200821119.sapp.pdf (5.1MB, pdf)

Data Availability Statement

The atomic coordinates and structure factors have been deposited in the Protein Data Bank (PDB) (accession codes 7U8J for cH4/3 HA in complex with 31.a.83 Fab, 7U8L for cH15/3 HA in complex with 31a.83 Fab, and 7U8M for cH15/3 HA in complex with 31.a.83 and FluA-20 Fabs). nsEM 3D maps have been deposited in the Electron Microscopy Databank (EMDB) under accession codes EMD-26928–EMD-26932 for apo cH4/3 HA maps, EMD-26933–EMD-26935 and EMD-26937–EMD-26938 for apo cH15/3 HA maps, and EMD-26939 for the cH15/3 HA in complex with CR9114 Fab map. All other study data are included in the main text and/or SI Appendix.

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES