
==== Front
Proc Natl Acad Sci U S A
Proc Natl Acad Sci U S A
PNAS
Proceedings of the National Academy of Sciences of the United States of America
0027-8424
1091-6490
National Academy of Sciences

39231203
202404250
10.1073/pnas.2404250121
research-articleResearch ArticlemicrobioMicrobiology423
Biological Sciences
Microbiology
Structure-based design of a soluble human cytomegalovirus glycoprotein B antigen stabilized in a prefusion-like conformation
Sponholtz Madeline R. a 1 https://orcid.org/0000-0002-0671-670X

Byrne Patrick O. a 1 2 https://orcid.org/0000-0002-6462-8951

Lee Alison G. a https://orcid.org/0000-0001-8305-3332

Ramamohan Ajit R. a https://orcid.org/0009-0005-9778-5987

Goldsmith Jory A. a
McCool Ryan S. a https://orcid.org/0000-0001-8833-8847

Zhou Ling a https://orcid.org/0000-0003-3402-5719

Johnson Nicole V. a https://orcid.org/0000-0003-4351-125X

Hsieh Ching-Lin a
Connors Megan b https://orcid.org/0009-0007-6691-3151

Karthigeyan Krithika P. b
Crooks Chelsea M. b https://orcid.org/0000-0003-3874-4577

Fuller Adelaide S. b
Campbell John D. c
Permar Sallie R. b
Maynard Jennifer A. d https://orcid.org/0000-0002-0363-8486

Yu Dong c
Bottomley Matthew J. c https://orcid.org/0000-0001-9734-5694

McLellan Jason S. jmclellan@austin.utexas.edu
a 3 https://orcid.org/0000-0003-3991-542X

aDepartment of Molecular Biosciences, The University of Texas at Austin, Austin, TX 78712
bDivision of Infectious Diseases, Department of Pediatrics, Weill Cornell Medicine, New York, NY 10065
cDynavax Technologies Corporation, Emeryville, CA 94608
dDepartment of Chemical Engineering, The University of Texas at Austin, Austin, TX 78712
3To whom correspondence may be addressed. Email: jmclellan@austin.utexas.edu.
Edited by Thomas Shenk, Princeton University, Princeton, NJ; received February 28, 2024; accepted July 31, 2024

1M.R.S. and P.O.B. contributed equally to this work.

2Present address: GSK Vaccines Institute for Global Health, Siena 53100, Italy.

4 9 2024
10 9 2024
4 9 2024
121 37 e240425012128 2 2024
31 7 2024
Copyright © 2024 the Author(s). Published by PNAS.
2024
https://creativecommons.org/licenses/by/4.0/ This open access article is distributed under Creative Commons Attribution License 4.0 (CC BY).

Significance

Prefusion-stabilized class I viral fusion proteins have generally been shown to elicit higher neutralizing antibody titers relative to non-prefusion-stabilized proteins when used as vaccine antigens. However, whether this concept extends to class III viral fusion proteins, like herpesvirus glycoprotein B (gB), remains unexplored, partly due to difficulties in stabilizing these antigens. Here, we describe results of a protein engineering campaign that identified prefusion-stabilizing substitutions that maintain human cytomegalovirus (HCMV) gB in a prefusion-like conformation, as confirmed by a 2.8 Å resolution cryo-EM structure. This stabilized gB did not, however, elicit superior neutralizing antibody responses in mice compared to postfusion gB, challenging the longstanding hypothesis that prefusion-stabilized class III fusion proteins elicit superior immune responses.

Human cytomegalovirus (HCMV) glycoprotein B (gB) is a class III membrane fusion protein required for viral entry. HCMV vaccine candidates containing gB have demonstrated moderate clinical efficacy, but no HCMV vaccine has been approved. Here, we used structure-based design to identify and characterize amino acid substitutions that stabilize gB in its metastable prefusion conformation. One variant containing two engineered interprotomer disulfide bonds and two cavity-filling substitutions (gB-C7), displayed increased expression and thermostability. A 2.8 Å resolution cryoelectron microscopy structure shows that gB-C7 adopts a prefusion-like conformation, revealing additional structural elements at the membrane-distal apex. Unlike previous observations for several class I viral fusion proteins, mice immunized with postfusion or prefusion-stabilized forms of soluble gB protein displayed similar neutralizing antibody titers, here specifically against an HCMV laboratory strain on fibroblasts. Collectively, these results identify initial strategies to stabilize class III viral fusion proteins and provide tools to probe gB-directed antibody responses.

cytomegalovirus
herpesvirus
vaccine
cryo-EM
==== Body
pmcHuman cytomegalovirus (HCMV), also known as human betaherpesvirus 5 (1), establishes lifelong latency in infected individuals (2). HCMV infects 60 to 90% of adults worldwide and can be spread through transplacental transmission from mother to fetus or through contact with bodily fluids (2, 3). Primary infection, reinfection, and reactivation of a latent infection all pose risks for the fetus during pregnancy (2). HCMV is the leading infectious cause of birth defects worldwide (4), with about 1 in 200 newborns having congenital infections in the US (5). HCMV is also a common and serious opportunistic infection following solid-organ and stem-cell transplants (2). In healthy patients, HCMV infection can be controlled by a robust immune response, which may come at the cost of decreased immune function over time (6). Despite the considerable disease burden associated with HCMV and the variety of vaccine formulations investigated over the last five decades, no FDA-approved vaccine is available for prevention or treatment (7).

As a member of the Herpesviridae family, HCMV is an enveloped, double-stranded DNA virus (8). Viruses within this family enter host cells through a conserved mechanism, relying on the coordinated actions of multiple surface glycoproteins for receptor binding and membrane fusion (9). The specific receptors and glycoproteins involved vary among different herpesviruses, although glycoprotein B (gB), which mediates membrane fusion, is highly conserved, and required for entry (10, 11). For HCMV, receptor binding is mediated by glycoprotein complexes referred to as Trimer (gH, gL, and gO) and Pentamer (gH, gL, UL128, UL130, and UL131A) (12–14). Receptor binding to Trimer and Pentamer complexes is thought to trigger the irreversible transition of gB from a metastable prefusion conformation to a highly stable postfusion conformation, thereby facilitating fusion of the viral and host-cell membranes. Given its essential role in viral entry, gB is typically included in HCMV vaccine candidate formulations (15). Of the vaccine candidates tested in humans, recombinant gB delivered with the MF59 adjuvant has shown promise in phase II clinical trials (NCT00125502, NCT00133497), achieving 40 to 50% short-lived efficacy in preventing HCMV infection in both adolescent and postpartum cohorts (16, 17), as well as reducing viremia and antiviral prophylaxis in renal transplant patients (18). The immune correlate of protection against HCMV acquisition from these studies was plasma immunoglobulin G (IgG) binding to native gB expressed on the surface of a cell (19), suggesting that the conformation of gB is critical to vaccine efficacy.

HCMV gB is a class III fusion protein encoded by open reading frame UL55. Class III fusion proteins share features of both class I and class II fusion proteins, including trimeric central helices arranged in a coiled-coil and internal fusion loops (9, 20). Class III fusion proteins are found in herpesviruses, rhabdoviruses, thogotoviruses, and baculoviruses (9). Upon translation of the monocistronic gB messenger RNA (mRNA), each protomer undergoes extensive glycosylation, with high occupancy at the 17 to 18 predicted N-linked glycosylation sites per protomer (21) (Fig. 1A). Additionally, two O-linked glycosylation sites are located near the N terminus of each protomer (22). Three gB protomers associate to form a metastable prefusion trimer, which is processed by furin (457-R-X-K/R-R-460) and incorporated into the HCMV virion. Each protomer has two hydrophobic fusion loops located at the membrane-proximal tip of structural domain I (DI) that pack against the hydrophobic membrane-proximal region (MPR) of a neighboring protomer in the prefusion trimer (Fig. 1B and SI Appendix, Fig. S1) (23). The metastable prefusion gB trimer undergoes significant conformational rearrangement to facilitate membrane fusion. This involves transitioning to an extended intermediate state, in which DI disengages from the MPR, rotates with DII approximately 180° relative to DIII, and embeds the fusion loops into the host-cell membrane (23). This intermediate, which bridges the viral and host-cell membranes, then collapses into the highly stable postfusion conformation, bringing the viral and host-cell membranes together to create a fusion pore (9, 23).

Fig. 1. Exemplary substitutions for HCMV gB stabilization. (A) Schematic of WT Towne strain HCMV gB and ectodomain base construct (gB Base). Native disulfide bonds are shown as connecting black lines. N-linked glycosylation sites are shown as branched lines. The native furin cleavage site is shown as a thick black line. The N-terminal signal sequence (SS) and C-terminal domain (C-term) are shown as white boxes. The first and second regions expected to move during the conformational rearrangement from pre-to-postfusion gB are colored blue and pink, respectively, and the region that does not undergo substantial rearrangement is colored yellow. The MPR and TM domain are shown in gray, with the TM demarcated by dashed blue and black lines. The soluble gB Base construct consists of the first 704 residues of WT gB with substitutions C246S, R457S, and R460S, of which the latter two remove the native furin cleavage site. In gB Base, residue 704 is followed by the foldon (Fd) domain and C-terminal tags. (B) Side view of trimeric prefusion HCMV gB (PDB ID: 7KDP) (23). One protomer is colored as in (A) and shown as a ribbon diagram, one protomer is colored gray and shown as a cartoon trace of the α-carbon backbone, and one protomer is shown as a transparent surface. Exemplary substitutions are shown as spheres with sulfur atoms in yellow, nitrogen atoms in blue, and oxygen atoms in red. The approximate location of the viral membrane is shown as dashed blue lines. Insets show exemplary substitutions as sticks. Anticipated hydrogen bonds are shown as black dashed lines and anticipated disulfide bonds are shown as yellow dashed lines. (C) Absolute and relative expression levels of individual variants determined by quantitative BLI. Variants are colored by substitution type.

Two 3.6 Å resolution crystal structures of postfusion HCMV gB were published in 2015 (24, 25) and these closely resembled the postfusion structures of gB homologs from herpes simplex virus type 1 (HSV-1) and Epstein–Barr virus (EBV) (26, 27). Subsequent cryoelectron tomography (cryo-ET) reconstructions of full-length, membrane-bound prefusion HCMV gB (28–30) revealed a shorter and more compact structure than observed for postfusion gB and suggested a domain architecture similar to prefusion G from vesicular stomatitis virus (VSV) (31). Recently, a 3.6 Å resolution cryoelectron microscopy (cryo-EM) structure of prefusion HCMV gB was determined using detergent-solubilized full-length gB purified from virions, complexed with the neutralizing human antibody SM5-1 (32), and stabilized in the prefusion conformation with a thiourea fusion inhibitor and a chemical cross-linker (23). This high-resolution structure of a prefusion herpesvirus gB protein provided the necessary data to guide the structure-based design of a stabilized prefusion immunogen.

In evaluating the humoral immune response to HCMV gB, antigenic mapping has identified six antigenic domains (AD-1–6), of which all but AD-3 and AD-6 are capable of eliciting neutralizing antibodies (32, 33). AD-1, located on structural domain IV (DIV) (SI Appendix, Fig. S1), is considered the immunodominant region of gB and elicits primarily nonneutralizing antibodies (32, 34). AD-1 is partially obscured by DI, DII, and DV in the prefusion conformation but is highly accessible in the postfusion conformation (23–25). AD-2 consists of the first 85 N-terminal residues of gB, which are flexible and have not been resolved in any HCMV gB structure but can elicit potently neutralizing antibodies that bind linear epitopes, including the human antibodies TRL345 and 3-25 (35, 36). AD-3 corresponds to the C-terminal cytoplasmic tail that is presumably inaccessible on intact virions and elicits exclusively nonneutralizing antibodies (32, 37). AD-4, located on DII, contains the epitope for broadly neutralizing antibodies 7H3 and SM5-1 (32, 38). AD-5, located on DI, contains the epitope for the neutralizing antibody 1G2 (32) and can elicit comparatively high titers of neutralizing antibodies (39). An additional antigenic region corresponding to DV, AD-6, was recently identified in a study assessing the antibody response elicited by the recombinant gB vaccine tested in clinical trial NCT00299260 and shown to elicit antibodies that are nonneutralizing but limit cell-to-cell spread of HCMV in both fibroblasts and epithelial cells (33). Although it has been shown that some prefusion-stabilized class I fusion proteins elicit higher quality immune responses relative to immunization with postfusion or nonprefusion-stabilized variants (40–42), this has not yet been reported for class III herpesvirus fusion proteins, primarily due to the difficulty of producing prefusion gB.

Here, we used the published structure of full-length, detergent-solubilized prefusion HCMV gB (PDB ID: 7KDP) (23) to guide the engineering of a soluble HCMV gB ectodomain construct stabilized in a prefusion-like conformation. We designed and biochemically characterized a soluble postfusion ectodomain base construct (gB Base) and numerous amino acid substitution variants. As no antibodies that bind exclusively to the prefusion conformation of HCMV gB have been described, we analyzed protein expression, thermal stability, dispersity, and conformational homogeneity to evaluate gB variants. This led us to identify gB-C7, a variant with improved expression and thermostability relative to gB Base. A 2.8 Å resolution cryo-EM structure of gB-C7 in complex with neutralizing antibodies 1G2 and 7H3 (32, 38) reveals that gB-C7 folds into a prefusion-like conformation, even though it lacks the MPR and transmembrane domain (TM). Mice were immunized with gB-C7 or the gB Base postfusion construct to assess the resulting gB-specific and HCMV-neutralizing antibody titers in fibroblasts. This work identifies a strategy for stabilizing class III viral fusion proteins, provides structural insights into the prefusion conformation of HCMV gB when bound by neutralizing antibodies, and creates reagents for the isolation of gB-directed antibodies and assessment of antibody responses from infected or vaccinated individuals.

Results

Design and Initial Characterization of Single Substitution HCMV gB Variants.

We first designed a soluble base construct (gB Base) comprising the HCMV gB ectodomain (Towne strain, residues 1 to 704) followed by a C-terminal T4 fibritin (foldon) trimerization motif, an octa-histidine tag, and a Twin-Strep affinity tag (Fig. 1A and SI Appendix, Fig. S1). We also substituted an unpaired cysteine at position 246 for serine (C246S) and eliminated the furin cleavage site with two serine substitutions (R457S, R460S), as described previously (24). Approximately 3 mg/L of gB Base was routinely purified from transiently transfected FreeStyle 293 cell cultures. To structurally characterize the base construct, we determined a 3.4 Å resolution cryo-EM structure of gB Base in complex with the fragments of antigen binding (Fabs) from high affinity, neutralizing antibodies 1G2 and 7H3 (32, 38) (SI Appendix, Fig. S2 and Table S1). As expected, the structure revealed that gB Base adopts the postfusion conformation and closely resembles previously determined crystal and cryo-EM structures of postfusion HCMV gB (SI Appendix, Fig. S3) (23–25).

To stabilize the soluble gB ectodomain in the prefusion conformation, we designed and characterized 24 gB variants, each containing one or two amino acid substitutions (Fig. 1B). Types of substitutions included: engineered disulfide bonds to covalently link regions that separate during the conformational rearrangement from prefusion to postfusion (disulfide), substitutions to neutralize internal charge imbalances (polar), hydrophobic residues to fill internal cavities (cavity-filling), and proline residues to disfavor refolding of secondary structure (proline) (43). To evaluate the expression of the variants, we performed small-scale (4 mL) transient transfections of FreeStyle 293 cells followed by quantification of the expressed protein yield in clarified media by bio-layer interferometry (BLI, Fig. 1C). The gB Base construct yielded 2.9 µg/mL of medium on average and the variants yielded between 0.5 and 7.7 µg/mL. Of the 24 variants tested, 16 increased HCMV gB expression relative to gB Base (six disulfide, two polar, five cavity-filling, and three proline).

We next purified variants from 40 mL cultures of FreeStyle 293 cells to conduct further characterization. When purified and analyzed by reducing SDS-PAGE, gB Base and its variants yielded a predominant band around 130 kDa (Fig. 2A and SI Appendix, Fig. S4), consistent with the molecular weight of glycosylated monomeric HCMV gB ectodomain. When analyzed by nonreducing SDS-PAGE, gB samples migrated as three distinct species, likely corresponding to monomers and disulfide-linked dimers and trimers (Fig. 2A). Five interprotomer disulfide variants (Q98C/N658C, V134C/I653C, N220C/E657C, H222C/E657C, and S674C/E698C) exhibited the greatest proportion of the trimer species (bands >460 kDa), indicating the formation of intermolecular disulfide bonds between protomers. As expected, disulfide variants designed to form intramolecular bonds predominately ran as monomers (Fig. 2A).

Fig. 2. Characterization of single disulfide gB variants. (A) Reducing and non-reducing SDS-PAGE of gB variants. Molecular weight standards are indicated on the left in kDa. White, gray, and black triangles correspond to monomeric, dimeric, and trimeric gB molecular weights, respectively. (B) SEC traces of purified gB variants. The approximate location of the void volume, HMW peak, and LMW peak are identified with black arrows. (C) DSF analysis of gB variant thermostability colored as in (B).

All single disulfide variants were then analyzed by size-exclusion chromatography (SEC) and differential scanning fluorimetry (DSF) to assess their solution behavior and thermal stability. When separated by SEC, the gB Base construct yielded two main peaks: one high molecular weight (HMW) peak with a retention volume of ~12 mL and one low molecular weight (LMW) peak with a retention volume of ~14 mL (Fig. 2B). The LMW peak was consistent with gB ectodomain trimers, whereas the HMW peak was consistent with oligomers of trimers, which are known to form via the exposed hydrophobic residues in the fusion loops (25, 44). Since the hydrophobic fusion loops are farther apart in the prefusion conformation of gB (23), we hypothesized that stabilizing gB in its prefusion conformation would disrupt the hydrophobic oligomer interface. This, in turn, would disfavor higher-order oligomerization and favor monodisperse trimers, quantified as an increase in the ratio of the area under the curve (AUC) of the LMW peak relative to the HMW peak (AUCLMW/HMW). The gB Base construct displayed a low AUCLMW/HMW ratio, indicating fewer monodisperse trimers, whereas a subset of interprotomer disulfide variants exhibited relatively high AUCLMW/HMW ratios, indicating more monodisperse trimers (SI Appendix, Table S2). As evaluated by DSF, five disulfide variants displayed increases in melting temperature (Tm) compared to the base construct (Tm of 59 °C): Variants V134C/I653C, I356C/A500C, H222C/E657C, Q98C/N568C, and N220C/E657C exhibited Tm values of 61, 62, 67, 69, and 71 °C, respectively (Fig. 2C). Of these, all but I356C/A500C were interprotomer disulfide variants. Overall, interprotomer disulfide variants V134C/I653C, N220C/E657C, and H222C/E657C displayed pronounced increases in expressed protein yield (Fig. 1C), AUCLMW/HMW ratio (Fig. 2B), and thermal stability (Fig. 2C) relative to gB Base.

Biochemical and Structural Characterization of Combination Variants.

We next engineered seven gB combination variants containing two or three beneficial single substitutions and evaluated them for additive effects (listed in Table 1). We first selected V134C/I653C, N220C/E657C, and H222C/E657C based on their increased expression, favorable AUCLMW/HMW ratios, and Tm values (Figs. 1C and 2 B and C). Notably, all three disulfides were designed to covalently link DI and DV, which are adjacent in the prefusion conformation but separated in the postfusion conformation (Fig. 1B and SI Appendix, Figs. S1 and S3). We selected four additional variants (T100L/A267I, K130Y, K260W, and V273F) based on their increased expression relative to gB Base (Fig. 1C). As judged by quantitative BLI, all seven gB combination variants (gB-C1 through gB-C7) increased expression relative to gB Base (Fig. 3A). Variants gB-C1 (N220C/E657C, T100L/A267I), gB-C6 (H222C/E657C, V134C/I653C), and gB-C7 (H222C/E657C, V134C/I653C, T100L/A267I) displayed additive increases in expression relative to gB Base and their constituent single substitution variants (Figs. 1C and 3A). In contrast, gB-C4 (N220C/E657C, K130Y) exhibited slightly reduced expression relative to the N220C/E657C variant, suggesting these substitutions may interfere with each other (Figs. 1C and 3A). When analyzed by nonreducing SDS-PAGE, all combination variants displayed an increased proportion of the trimer species relative to gB Base (Fig. 3B). DSF analysis revealed that five combination variants exhibited increases in thermal stability relative to their parental single disulfide variants; gB-C6, gB-C5, and gB-C7 exhibited Tm values of 71, 72, and 72 °C, respectively (Fig. 3C). In addition, gB-C4 and gB-C1 exhibited biphasic DSF profiles with dominant peaks at 73 and 75 °C, respectively, and shoulders around 68 °C. Biphasic DSF profiles may reflect a heterogeneous population with two distinct unfolding events (45), or a two-state transition, such as the dissociation of a higher-order oligomer followed by the denaturation of the trimeric fusion protein (46, 47), neither of which were considered favorable.

Table 1. HCMV gB combination variants

Combo variant name	Substitution 1	Substitution 2	Substitution 3	
gB-C1	N220C/E657C	T100L/A267I	–	
gB-C2	H222C/E657C	K260W	–	
gB-C3	H222C/E657C	V273F	–	
gB-C4	N220C/E657C	K130Y	–	
gB-C5	N220C/E657C	V134C/I653C	–	
gB-C6	H222C/E657C	V134C/I653C	–	
gB-C7	H222C/E657C	V134C/I653C	T100L/A267I	

Fig. 3. Characterization of combination gB variants. (A) Absolute and relative expression levels of individual variants determined by quantitative BLI. Dashed lines denote onefold and fivefold increases in expression relative to gB Base. (B) Reducing and non-reducing SDS-PAGE gels of gB variants. Molecular weight standards are indicated at the Left in kDa. White, gray, and black triangles correspond to monomeric, dimeric, and trimeric gB molecular weights, respectively. (C) DSF analysis of gB variant thermostability colored as in (A). (D) Negative-stain EM 2D class averages of six variants.

To assess the effect of single and multiple substitutions on the conformation of the soluble HCMV gB ectodomain, we conducted structural studies on a subset of variants by negative-stain EM (ns-EM) (Fig. 3D and SI Appendix, Fig. S5). We used two high-affinity, neutralizing antibodies to aid in conformational characterization, 1G2 and 7H3, which bind DI and DII, respectively (32, 38). Complexes of the gB Base construct with 1G2 and 7H3 yielded 2D class averages consistent with gB in the postfusion conformation bound to three 1G2 Fabs and either zero, one, or two 7H3 Fabs (Fig. 3D and SI Appendix, Fig. S5). Oligomers of postfusion gB (also known as rosettes) were observed for gB Base by ns-EM, consistent with the HMW peaks observed by SEC and previous findings on the spontaneous oligomerization of soluble postfusion gB (SI Appendix, Fig. S5) (25, 44, 48). The single disulfide variants V134C/I653C, N220C/E657C, and N222C/E657C each appeared more compact than gB Base, although a minority of the 2D class averages corresponded with side views of gB in the postfusion conformation (Fig. 3D and SI Appendix, Fig. S5). We hypothesized that these more compact particles corresponded to gB in prefusion-like conformations. Although single substitution variants T100L/A267I and S367C/A503C appeared promising based on their increased expression relative to gB Base (Fig. 1C), both appeared to be in the postfusion conformation by ns-EM (SI Appendix, Fig. S5). The combination variants gB-C6 and gB-C7, both of which contain the V134C/I653C substitution, appeared to be conformationally similar to the single disulfide V134C/I653C variant (Fig. 3D). Additionally, gB-C7 displayed distinct side views compared to gB Base (Fig. 3D). The gB-C7 variant contains two disulfide substitutions along with a paired cavity-filling substitution (V134C/I653C, H222C/E657C, T100L/A267I). Overall, gB-C7 displayed higher expression than gB Base by a factor of 4.2, exhibited the greatest trimeric fraction when analyzed by nonreducing SDS-PAGE, had an 11 °C increase in Tm relative to gB Base, and resembled prefusion gB by ns-EM (Fig. 3 A–D). Given these properties, we focused on this construct for additional characterization.

HCMV gB-C7 Maintains a Prefusion-Like Conformation.

We determined a cryo-EM structure of gB-C7 in complex with 1G2 and 7H3 Fabs that reached a global resolution of 2.8 Å when refined with C3 symmetry (Fig. 4 and SI Appendix, Figs. S6 and S7 and Table S1). We performed local refinement to account for relative motion between DI bound by 1G2 and the rest of the gB-C7 complex, yielding a 3.1 Å resolution local reconstruction (SI Appendix, Fig. S6). We then combined the 2.8 Å global map with the 3.1 Å local map in Phenix to generate a composite map (49), which we used to model the gB-C7 complex (Fig. 4A). With this high-resolution map, we were able to model previously unresolved residues at the gB membrane-distal apex (residues 437 to 447 and 474 to 482, Fig. 4C and SI Appendix, Fig. S8). These residues adopt extended α-helices and are in agreement with an AlphaFold2 (AF2) predicted model (SI Appendix, Fig. S9) (50). Consistent with the previously reported prefusion HCMV gB structure (23), N-terminal residues 1 to 78 and a portion of the apex (residues 448 to 473) are unresolved in our map, likely due to the intrinsic flexibility of these regions. This is also consistent with the AF2 model, in which these regions are predicted to be unstructured. However, unlike the previously reported structure, we were unable to resolve the fusion loops of DI (residues 149 to 163, 192 to 197, and 233 to 246) or the majority of DV (residues 662 to 704). This is likely due to the omission of the MPR in our soluble ectodomain construct, which appears to be critical for interacting with the fusion loops and stabilizing DI (23). Further, our structure reveals that in the absence of the MPR, DI is shifted outward relative to its more compact positioning in the full-length prefusion HCMV gB structure reported previously (Fig. 4C). Thus, the MPR or a soluble mimic likely needs to be present to maintain a compact prefusion conformation of gB.

Fig. 4. Cryo-EM structure of gB-C7 bound to 1G2 and 7H3 Fabs. (A) Side (Top, Left) and Top (Top, Right) views of the composite EM map (global plus local refinement) of HCMV gB-C7 complexed with 1G2 and 7H3 Fabs shown above the gB-C7 complex model (Bottom, Left). One protomer of the model is colored as in Fig. 1A and shown as a ribbon diagram, the second is colored gray and shown as a cartoon tube trace of the α-carbon backbone, and the third is shown as a transparent surface. The Inset (Bottom, Right) shows a zoomed view of the substitutions that comprise design gB-C7 with side chains shown as sticks. (B) Side view of HCMV gB-C7 bound to 1G2 and 7H3 Fabs shown as a ribbon diagram (Top) above a zoomed view of the binding interface between 7H3 and gB-C7 (Bottom). 7H3 HC is pink, 7H3 light chain (LC) is pale pink, and gB-C7 is yellow. In the zoomed interface, key residues are shown as sticks, and hydrogen bonds are shown as black dashed lines. (C) The structure of gB-C7 (yellow) is superimposed with the previously determined structure of prefusion HCMV gB (blue, PDB ID: 7KDP) (23), both shown as cartoon tube traces of the α-carbon backbones. Side and Top views are shown for the superimposition of HCMV gB both as a trimer (Left) and as a single protomer (Right). Shifts in domain arrangement are highlighted with arrows.

All stabilizing substitutions in gB-C7 were designed to maintain DI in its prefusion position. The two disulfide bonds (V134C/I653C and H222C/E657C) were designed to bridge DI and DV (Fig. 4A and SI Appendix, Fig. S10), whereas the cavity-filling substitutions T100L/A267I were designed to stabilize the interface between DI and DIV via hydrophobic packing. We observed good map features for all side chains at these sites of substitution (SI Appendix, Fig. S10), although the map was not contiguous between the cysteines of the H222C/E657C disulfide. This may indicate that the H222C/E657C disulfides formed incompletely under these conditions, or they may have suffered damage from the electron beam (51).

The 1G2 and 7H3 interfaces were well resolved in the local and global maps, respectively, revealing extensive hydrogen bond networks that are maintained in both the prefusion and postfusion conformations (Fig. 4B and SI Appendix, Fig. S11). In line with the previously determined 1G2-bound postfusion HCMV gB structure (PDB ID: 5C6T), 1G2 binds a hydrophobic patch on DI, which corresponds to AD-5 of HCMV gB (24) (SI Appendix, Fig. S11). Consistent with a previously reported ns-EM reconstruction of postfusion HCMV gB in complex with 7H3 Fab (52), 7H3 binds the same DII epitope as the high-affinity, neutralizing antibody SM5-1 (32), which corresponds to AD-4 of HCMV gB (23, 53). The structures of gB Base and gB-C7 reported here are the first high-resolution structures of HCMV gB in complex with 7H3 Fab, providing a detailed characterization of the gB:7H3 interface, which includes 13 hydrogen bonds and 808 Å2 of buried surface area on gB-C7 (Fig. 4B). Like SM5-1, binding appears to be largely mediated by an unusually long heavy chain (HC) CDR3 (23, 53).

Unlike gB-C7, which exhibits full 1G2 and 7H3 Fab occupancy in the 2.8 Å cryo-EM structure, 7H3 Fab bound to only one of the three protomers of gB Base (SI Appendix, Fig. S3). Comparison between occupied and unoccupied protomers reveals that 7H3 Fab displaces a flexible loop between DII and DIII (residues 466 to 475) that localizes to the apex in prefusion gB and the trunk in postfusion gB (SI Appendix, Fig. S3B). When 7H3 is not bound, 10 additional residues of this DII/DIII loop are resolved, packing against the unoccupied 7H3 epitope. In a recently determined structure of postfusion HCMV gB in complex with SM5-1 Fab (PDB ID: 7KDD), SM5-1 binding also appears to displace the DII/DIII loop resolved in our gB Base EM map (23). Our structural studies suggest that 7H3 and the DII/DIII loop compete to occupy a similar position in the postfusion conformation, explaining the partial occupancy of 7H3 Fab when complexed with gB Base observed by ns-EM and cryo-EM (Fig. 3D and SI Appendix, Figs. S3 and S5). This partial occupancy, which has not been recognized in binding studies or reported previously, occurred reproducibly in our structural studies and suggests that 7H3 occupancy could be used to evaluate the prefusion character of gB proteins.

Stabilized Prefusion-Like gB and Postfusion gB Are Similarly Immunogenic in Mice.

To assess the immunogenicity of the stabilized prefusion-like variants, three groups of eight BALB/c mice were immunized with gB antigens at weeks 0, 3, and 6. Each group was immunized with 2.5 µg of one of the following gB variants, which were selected to span a range of conformations: gB Base (postfusion), gB V134C/I653C (partially prefusion-stabilized), and gB-C7 (the most stabilized prefusion-like combination variant) (Fig. 5A and SI Appendix, Fig. S12A). To potentiate immune responses, the gB antigens were combined with CpG 1018 adjuvant plus alum (54). A fourth group of mice was immunized with gB-C7 without adjuvant to assess the impact of adjuvant on the quality of the immune response. Sera were collected 2 wk after the third immunization (Fig. 5A). To quantify antibody titers, we performed serum enzyme-linked immunosorbent assays (ELISAs) using antibody SM5-1 (32) expressed with murine IgG2a fragment crystallizable (mFc) domains as a standard. This neutralizing human antibody has a similar binding affinity for gB Base and gB-C7 (SI Appendix, Fig. S12B), consistent with previous findings that the SM5-1 Fab binds both postfusion and prefusion conformations (23). All gB variants were highly immunogenic, with serum antigen-binding IgG titers (1/ED50) ranging between 103 and 106 (SI Appendix, Fig. S12 C and D). Antibodies elicited by immunization with adjuvanted gB Base, gB V134C/I653C, or gB-C7 exhibited similar binding to gB Base with SM5-1-binding-equivalent concentrations of 1.50, 1.45, and 1.22 mg/mL, respectively (Fig. 5B and SI Appendix, Table S3). Similar trends were observed when measuring antibody binding against gB-C7, where mice immunized with adjuvanted gB Base, gB V134C/I653C, or gB-C7 had similar responses, with SM5-1-binding-equivalent concentrations of 0.25, 0.25, and 0.22 mg/mL, respectively (Fig. 5C). Predictably, mice immunized with gB-C7 without an adjuvant had significantly lower antibody titers against gB Base (0.11 mg/mL) and gB-C7 (0.02 mg/mL), indicating the immune response to gB-C7 can be enhanced by an adjuvant. Interestingly, immunization with postfusion or stabilized prefusion-like gB did not appear to bias antibody binding toward specific gB conformations (Fig. 5 B and C) and anti-gB Base titers correlated strongly with anti-gB-C7 titers (SI Appendix, Fig. S13A). These data are consistent with previous observations that many, if not all, known immunogenic gB epitopes are shared between conformations (23). Antibody binding to gB-C7 was lower for all treatment groups compared to gB Base (Fig. 5 B and C), suggesting there may be reduced availability of some prefusion epitopes. This is consistent with the differential exposure of certain regions of gB in the two conformations (23), including DIV and DV, which correspond to AD-1 and AD-6, respectively (32, 33, 34, 55). However, when we performed IgG mapping with a Luminex-based binding antibody multiplex assay (BAMA) to assess variation in IgG binding by structural domain (56), we found no significant differences in binding among sera from each group of mice immunized with adjuvanted gB antigens (SI Appendix, Fig. S14).

Fig. 5. Engineered gB variants are immunogenic in mice. (A) Schematic of mouse immunization. 6-to-8-week-old female BALB/c mice (n = 8/group) were immunized at weeks 0, 3, and 6 with 2.5 μg of gB Base (orange), gB V134C/I653C (red), or gB-C7 (green), all adjuvanted with CpG 1018 plus alum or unadjuvanted gB-C7 (purple). Blood samples were collected from mice 2 wk after the week 6 injection. SM5-1-equivalent antibody concentrations of immunized mouse sera that bind to (B) gB Base or (C) gB-C7. Plots represent ELISA measurements relative to an SM5-1 mFc standard from two experimental replicates. AD169-GFP infection of human MRC-5 fibroblasts. (D) Neutralization by Cytogam is enhanced 3.6-fold in the presence of 12.5% guinea pig complement. Data are averaged across 16 plates with SD shown. Neutralization by mouse sera in the (E) absence and (F) presence of 12.5% guinea pig complement. Plots represent averages from two independent experiments. In violin plots, horizontal lines represent the first quartile, median, and third quartile. Statistical significance was determined by one-way ANOVA followed by Tukey’s HSD test in GraphPad Prism: *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

Prefusion-Like gB Elicits Weakly Neutralizing, Complement-Enhanced Antibodies.

To test the hypothesis that immunization with prefusion-stabilized gB would produce superior neutralization titers, as is the case for multiple class I fusion proteins (40, 57–59), we measured the capacity of immunized mouse sera (with or without guinea pig complement) to neutralize HCMV laboratory strain AD169 infection of MRC-5 fibroblasts. We used Cytogam, a human hyperimmune globulin containing high titers of anti-HCMV polyclonal antibodies, as our positive control (60). In contrast to previous reports (61), but consistent with observations that complement can play a key role in neutralization by anti-gB monoclonal antibodies (62), we found that Cytogam-mediated neutralization of HCMV AD169 was enhanced 3.6-fold by the addition of 12.5% exogenous guinea pig complement (Fig. 5D). The immunized mouse sera displayed weak neutralizing activity in the absence of complement (Fig. 5E) and the addition of guinea pig complement also enhanced neutralization activity of sera from mice immunized with adjuvanted antigens (Fig. 5F). No gB variants in this study consistently elicited complement-independent neutralizing antibodies in mice, although sera from several mice from different treatment groups exhibited measurable neutralization activity in the absence of complement (Fig. 5E and SI Appendix, Fig. S12E and Table S3). Notably, the two mice with the highest complement-independent responses were immunized with gB Base. Prior gB vaccination studies in humans (63) and rabbits (64) have also shown poor complement-independent neutralization of HCMV despite high anti-gB titers, consistent with the observation that gB vaccination often stimulates antibody responses directed toward nonneutralizing epitopes (65). Despite similar anti-gB binding titers across gB variants, we observed weaker complement-dependent neutralization responses against HCMV AD169 in fibroblasts from immunization with increasingly prefusion-like gB variants: gB Base appeared to be the most potent, followed by gB V134C/I653C and gB-C7 (Fig. 5F and SI Appendix, Fig. S12 and Table S3). Anti-gB antibody titers correlated positively with complement-dependent neutralization but not complement-independent neutralization (SI Appendix, Fig. S13), suggesting gB binding and opsonization may be key factors in complement-dependent neutralization. Both gB Base and gB-C7 bound similarly to a panel of HCMV gB AD-2 antibodies, although the AD-2 site 1 (AD-2S1) TRL345 unmutated common ancestor (UCA) antibody (66) bound better to gB-C7 than gB Base (SI Appendix, Fig. S15). Collectively, the immunization data with the constructs described here suggest that prefusion-stabilized HCMV gB proteins are not superior immunogens to postfusion gB for the elicitation of potent neutralizing antibodies, at least against HCMV laboratory strain AD169 in fibroblasts.

Discussion

The stabilization of fusion proteins in their prefusion conformations has been shown to be a beneficial approach to optimizing subunit vaccine antigens against a variety of viruses (43, 67). This strategy works well when the fusion protein contains neutralization-sensitive antigenic sites found exclusively on the prefusion conformation. This is true for the class I fusion protein respiratory syncytial virus (RSV) F, which displays two highly neutralization-sensitive sites exclusively at the apex of prefusion F (68). However, to the best of our knowledge, no class III fusion proteins have been sufficiently stabilized in their prefusion conformations to enable testing for improved immunogenicity. Aiming to stabilize a soluble HCMV gB construct in its prefusion conformation, we engineered and biochemically characterized a base construct, gB Base, along with 24 single substitution variants and seven combination variants (Figs. 1–3). With ns-EM, we found that certain constructs appeared to be partially stabilized in the prefusion conformation (Fig. 3D and SI Appendix, Fig. S5). We determined a 2.8 Å resolution cryo-EM structure of our most promising combination variant, gB-C7, confirming that it is stabilized in a prefusion-like conformation. We then immunized mice with gB Base, gB V134C/I653C (a single substitution variant), and gB-C7, and found that immunization with prefusion-like gB did not elicit a superior immune response as measured by the neutralizing antibody response against HCMV AD169 in fibroblasts (Fig. 5).

Following the recently proposed model of HCMV gB rearrangement in which DI and DII are the first domains to undergo rearrangement from prefusion to postfusion (23), we hypothesized that substitutions aiming to tether DI in its prefusion position might globally maintain gB in the prefusion conformation (Fig. 1B). In line with this hypothesis, single substitutions to DI displayed better expression and thermostability relative to gB Base and other domain substitutions (Figs. 1C and 2). The membrane-distal apex of HCMV gB (DII and DIII) was stabilized in the prefusion conformation in the gB-C7 structure— in which all stabilizing substitutions were made to DI, DIV, and DV—demonstrating that stabilizing DI positioning can indeed stabilize gB in a prefusion-like conformation. However, compared to the recently published structure of wild-type (WT) HCMV gB in the prefusion conformation (PDB ID: 7KDP) (23), DI in the gB-C7 structure is shifted distally and only partially resolved by local refinement (Fig. 4C), suggesting that DI is highly flexible and primed to rearrange in the absence of the MPR. Similar domain flexibility is displayed by the prefusion conformation of VSV G, a fellow class III fusion protein, that, in contrast to HCMV gB, does not possess an MPR and instead anchors its fusion loops directly into the viral membrane (31, 69). In comparison to the first crystal structure determined for VSV G (31), a second crystal structure published more recently displayed an 11° tilt of the fusion domain relative to the rest of the protein (69), underscoring the flexibility of this domain in ectodomain constructs of class III fusion proteins.

In designing single substitutions aimed at stabilizing HCMV in the prefusion conformation, we hypothesized that structure-based vaccine design strategies that have been successfully applied to class I fusion proteins would translate to class III fusion proteins. For severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and Middle East respiratory syndrome coronavirus (MERS-CoV) spikes as well as RSV F and many other class I fusion proteins, proline substitutions—particularly helix capping proline substitutions—have been an effective method to stabilize the prefusion conformation (57, 59, 70). These substitutions generally target helix–turn–helix motifs in the prefusion conformation that rearrange to form extended helices in the postfusion conformation—by introducing prolines at these turns, the extended coiled-coil of the postfusion conformation is disfavored (43). Recently, proline substitutions aimed at capping a helix in DIII of HSV-1 gB and varicella-zoster virus (VZV) gB were shown to bias gB toward a prefusion-like conformation in the membrane (30, 71). However, the proline substitution was not sufficient to maintain the HSV-1 gB prefusion conformation in an ectodomain construct (30). Considering the recently published structure of prefusion HCMV gB (PDB ID: 7KDP) (23), as well as our structure of gB-C7, we hypothesize that a similarly located proline substitution in HCMV gB may distort the native architecture of the membrane-distal apex of the prefusion conformation. Of the five proline single substitution variants we tested, V645P, D646P, and L484P achieved around 50% increases in expression (Fig. 1C). While L484P likely disrupts the central helix, V645P and D646P localize to a coil between α-helices in the prefusion conformation (Figs. 1B and 4). This coil rearranges to an extended form in the postfusion conformation that packs between the central helices of DIII, a conserved motif seen in many postfusion herpesvirus fusion proteins (SI Appendix, Figs. S1 and S3) (24–27). As such, proline substitutions targeted to this region may translate well to gB homologs. Overall, as most class III fusion proteins do not possess a prefusion helix–turn–helix motif that rearranges to an extended coiled-coil in the postfusion conformation, the strategy of proline capping to maintain the prefusion conformation may have limited potential in the context of herpesvirus fusion proteins. However, well-placed proline substitutions, such as V645P and D646P, may still have a positive impact on expression and stability.

We found that interprotomer disulfide substitutions, particularly those designed to bond DI to DV (H222C/E657C, N220C/E657C, and V134C/I653C), were the most effective method to stabilize HCMV gB in a prefusion-like conformation (Figs. 1C, 3D, and 4A and SI Appendix, Fig. S5). As both DI and DV undergo conformational changes, these disulfide bonds stabilize the prefusion conformation by preventing rearrangement of the two domains. Interprotomer disulfide bonds have been used to stabilize several class I fusion proteins (58, 72–75), but often these disulfides reduce protein expression. However, here we found that interprotomer disulfide substitutions in HCMV gB generally had a neutral-to-positive impact on expression (Fig. 1C) in addition to increasing thermostability and stabilizing gB in a prefusion-like conformation (Figs. 2C, 3C, and 4A and SI Appendix, Fig. S5). As such, targeted interprotomer disulfide substitutions may be a general approach to stabilizing class III viral fusion proteins.

Since a vaccine candidate consisting of recombinant, structurally undefined HCMV gB paired with MF59 adjuvant achieved ~50% efficacy in preventing HCMV infection in phase II clinical trials (NCT00125502, NCT00133497) (16, 17), development of a prefusion-stabilized HCMV gB antigen holds the potential to produce a more efficacious HCMV vaccine candidate. However, in our mouse immunization study, we found that gB-C7 was similarly immunogenic to gB Base (Fig. 5 B and C), and did not elicit higher titers of antibodies capable of neutralizing HCMV AD169 in fibroblasts. In fact, neutralization by sera from mice immunized with gB-C7 was dampened in comparison to gB Base (Fig. 5 E and F). There are several possibilities for the observed result: 1) insufficient stabilization of the prefusion-like gB construct, 2) the oligomerization of gB Base (SI Appendix, Fig. S5) may increase its immunogenicity, and 3) prefusion gB naturally lacks neutralization-sensitive epitopes found exclusively in this conformation, supported by the lack of prefusion-specific gB antibodies isolated to date. Indeed, as HCMV gB conformational rearrangement seems to be largely dominated by rigid body movements rather than reorganization of secondary structural elements, shared epitopes between prefusion and postfusion predominate (23). Nevertheless, the differing domain arrangement between prefusion and postfusion conformations suggests that it might be feasible to elicit antibodies that bind across domain interfaces to target potentially vulnerable, prefusion-specific epitopes. In future studies, our stabilized prefusion-like soluble HCMV gB-C7 immunogen should be assessed for its ability to elicit neutralizing antibodies against more clinically relevant HCMV isolates (76) and cell types (epithelial, endothelial cells) (77, 78). Further, nonneutralizing antibody functions that are associated with gB vaccine efficacy (63, 79) and decreased risk of congenital CMV transmission (80, 81) should be explored. HCMV gB-C7 should also facilitate isolation of prefusion-specific antibodies, particularly those that target regions at the well-stabilized membrane-distal apex, which may prove useful for therapeutic applications and as reagents for vaccine research and development.

Methods

Design Scheme for Prefusion-Stabilized HCMV gB Variants.

The HCMV gB base construct (gB Base) comprises ectodomain residues 1 to 704 of HCMV gB Towne Strain (UniProtKB: P13201) with serine substituted at residues 246, 457, and 460 as previously described (24), followed by a C-terminal T4 fibritin (foldon) trimerization motif, an HRV3C protease recognition site, an octa-histidine tag, and a Twin-Strep affinity tag cloned into the mammalian expression vector pαH (82) and verified by DNA sequencing (Fig. 1A). All HCMV gB variants were constructed into this plasmid by Gibson assembly and verified by DNA sequencing. Based on the HCMV gB prefusion structure (PDB ID: 7KDP) and postfusion structures (PDB IDs: 5CXF, 5C6T, 7KDD) (23–25), residues were considered for disulfide bond, polar, cavity-filling, and proline designs. Combinations were chosen to test whether pairing of designs could result in additive effects.

Quantification of HCMV gB Expression by BLI.

Plasmids encoding HCMV gB variants were transfected into FreeStyle 293-F cells (Thermo Fisher) using polyethyleneimine (PEI). Culture medium was harvested 6 d after transfection by centrifugation. The clarified medium was diluted fivefold with 1× HBS EP+ (10 mM HEPES pH 8.0, 150 mM NaCl, 3 mM EDTA, 0.005% (w/v) Tween 20, 0.01% (w/v) sodium azide), then pipetted into black-walled 96-well plates (Greiner Bio-One) and loaded into a bio-layer interferometer (Octet RED96, ForteBio). Anti-human IgG Fc capture (AHC) biosensor tips (Sartorius) were loaded with anti-foldon IgG MF5, provided by Vicente Mas from Instituto de Salud Carlos III, then dipped into wells containing HCMV gB variants. The amount of HCMV gB in each sample was quantified by fitting the response as a function of time to the linear portion of the binding curve. We used a dilution series of purified HCMV gB in conditioned FreeStyle 293 medium (10 mg/mL to 0.156 mg/mL) to generate a standard curve, which allowed us to report the absolute concentration of HCMV gB in the conditioned media. Data were plotted as an average of three independent biological replicates.

Purification of HCMV gB Variants by Affinity and SEC.

Plasmids encoding HCMV gB variants were transiently transfected into 40 mL FreeStyle 293-F cell cultures using PEI. After 5 to 6 d, medium was harvested by centrifugation, 0.22 µm filtered, then passed over Strep-Tactin Sepharose resin (IBA Lifesciences) by gravity, washed with 3 column volumes of 1× PBS, and eluted with Strep-Tactin elution buffer (100 mM Tris-Cl pH 8.0, 150 mM NaCl, 1 mM EDTA, and 2.5 mM desthiobiotin) (IBA Lifesciences). Elution fractions were analyzed by SDS-PAGE. Fractions containing HCMV gB were pooled, concentrated with Amicon Ultra centrifugal filters (MilliporeSigma), and flash-frozen in liquid nitrogen. Samples were thawed in a room temperature water bath just before injection onto a Superose 6 Increase 10/300 GL column (Cytiva). The SEC running buffer was composed of 2 mM Tris pH 8.0, 200 mM NaCl, and 0.02% (w/v) sodium azide. Desired fractions were pooled, concentrated, aliquoted, and flash-frozen in liquid nitrogen for further analysis. HCMV gB protein was prepared in a similar fashion for cryo-EM, with the following exceptions: 5 μM kifunensine was added approximately 4 h after transfection for both gB Base and gB-C7. For gB-C7, concentrated eluate was treated with 5% (w/w) HRV3C protease at 4 °C overnight to remove affinity tags prior to SEC. For additional details, see SI Appendix, Extended Methods.

Negative-Stain Electron Microscopy.

Concentrated, purified HCMV gB constructs were mixed with Fabs (1G2, 7H3, or both 1G2 and 7H3) at a ratio of 1.2:1 (Fab:gB), incubated for 15 to 30 min at room temperature, then diluted to a working concentration of 0.02 to 0.1 mg/mL in SEC running buffer. Diluted samples were immediately applied to glow-discharged copper-supported carbon grids (Formvar, 400 mesh) and stained with 2% methylamine tungstate (Nano-W, Nanoprobes). Grids were loaded onto one of two transmission electron microscopes (TEMs): i) a Japan Electron Optics Laboratory (JEOL) 2010F TEM or ii) a JEOL NEOARM. The nominal magnifications for the 2010F and NEOARM were 60,000× (pixel size = 3.6 Å) and 50,000× (pixel size = 2.16 Å), respectively. Both microscopes operated at 200 kV and were equipped with OneView cameras (Gatan). Micrographs were acquired in 2 k × 2 k mode for the 2010F and 4 k × 4 k mode for the NEOARM using Digital Micrograph (Gatan), then exported to cryoSPARC (Structura Biotechnology) for contrast transfer function (CTF) correction, particle picking, and 2D classification (83). 3D volumes were generated using ab initio reconstruction, and data were further processed through heterogeneous and homogeneous refinements. Structural figures were produced using ChimeraX (84).

Cryo-EM Sample Preparation and Data Collection.

CF-400 1.2/1.3 grids (Electron Microscopy Sciences) were glow discharged for 60 s at 15 mAmps (PELCO easiGlow™ Glow Discharge Cleaning System) prior to sample application. Samples were prepared in EM buffer composed of 2 mM Tris pH 8.0, 200 mM NaCl, 0.02% (w/v) sodium azide, 3% (v/v) glycerol, and 0.01% (w/v) amphipol A8-35.

A 3.0 mg/mL solution of gB Base complex was prepared by incubating roughly equimolar concentrations of gB Base and 1G2 Fab and a twofold molar excess of 7H3 Fab in EM buffer. The complex was incubated for 30 min at 4 °C before adding 10× CMC CHAPS (VitroEase™ Buffer Screening Kit, ThermoFisher Scientific) to a final concentration of 0.5× CMC. Immediately after, 3 µL of the gB Base complex solution was applied onto grids, which were double blotted and subsequently plunge-frozen.

A 4.0 mg/mL solution of gB-C7 complex was prepared by incubating roughly equimolar concentrations of gB-C7, 1G2 Fab, and 7H3 Fab in EM buffer. The complex was incubated for 30 min at 4 °C before adding 10× CMC CHAPS (VitroEase™ Buffer Screening Kit, ThermoFisher Scientific) to a final concentration of 0.25× CMC. Immediately after, 3 µL of the gB-C7 complex solution was applied onto grids, which were blotted and subsequently plunge-frozen.

Grids were plunge-frozen in liquid ethane using a Vitrobot Mark IV (ThermoFisher Scientific) set to 100% humidity and 4 °C with a blot time of 5 s, a blot force of -1, and a wait time of 5 s. Cryo-EM datasets were collected at 105,000× magnification corresponding to a calibrated pixel size of 0.83 Å on an FEI Titan Krios operating at 300 kV and equipped with a K3 direct electron detector (Gatan) (SI Appendix, Table S1). A total of 8,679 exposures were collected for the gB Base complex dataset, 1,308 exposures of which were collected without tilt and 7,371 exposures of which were collected with a 30° tilt. A total of 12,524 exposures were collected for the gB-C7 complex dataset, all without tilt. Data were collected using SerialEM 3.9.0 beta (85).

Cryo-EM Data Processing, Model Building, and Refinement.

Gain reference correction was performed before the micrographs were imported into cryoSPARC Live. Motion correction, patch CTF estimation, defocus estimation, micrograph curation, particle picking, and 2D classification were initially performed in cryoSPARC Live (83). Micrographs and selected particles were then exported into CryoSPARC for 2D classification, ab initio reconstruction, heterogeneous refinement, homogenous refinement, and subsequent nonuniform homogeneous refinement of final classes. For the gB-C7 complex, a mask was created around domain I and 1G2 using ChimeraX and imported to cryoSPARC for local refinement. The resulting local map was combined with the global map using PHENIX combine_focused_maps (49). The EM processing pipelines for the gB Base and gB-C7 datasets are summarized in SI Appendix, Figs. S1 and S3, respectively. Initial models of gB Base, 1G2 Fab, and 7H3 Fab were predicted using AlphaFold2 (AF2) (50) while an initial model of gB-C7 was built with ModelAngelo (86). Models were then fit into the experimental cryo-EM maps in ChimeraX (84). Iterative model building and refinement were performed using PHENIX (49), COOT (87), and ISOLDE (88). The protein interfaces, surfaces, and assemblies (PISA) service at the European Bioinformatics Institute was used to determine buried surface area and interacting residues (89), and structural figures were produced using ChimeraX (84).

Animal Experiment.

In vivo immunogenicity of gB protein in combination with adjuvants was evaluated in 6-to-8-week-old female BALB/c mice (Charles River) at Aragen Bioscience (Morgan Hill, CA). All procedures were carried out under institutional IACUC-approved protocols. Groups of 8 mice were immunized by the intramuscular route at Days 0, 21, and 42 with 2.5 µg/dose of gB protein (gB Base, gB V134C/I653C, or gB-C7) in the presence or absence (gB-C7 only) of CpG 1018 adjuvant (10 µg; Dynavax) and aluminum hydroxide (alum; Alhydrogel®; 50 µg; InvivoGen). Immunogens were prepared by first mixing gB protein with alum for 30 min, followed by addition of CpG 1018 for an additional 5 min of mixing. Mice were injected within 1 h of immunogen preparations. Blood samples were collected by the submandibular route at Day 0 and Day 42 (post second immunization) and by cardiac puncture at Day 56 (day of killing) for serum harvests.

HCMV Preparation.

Human MRC-5 fibroblasts (ATCC, CCL-171) were cultured in Dulbecco’s modified Eagle medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C, 5% CO2. Virus stocks were produced as described from AD169-GFP BAC (90), which was a gift from Thomas Shenk (Princeton). In brief, MRC-5 cells were transfected with the BAC to generate a P0 stock. MRC-5 cells grown to 80% confluent monolayers were then infected using the P0 stock at an MOI of 0.01 and cultured for ~2 wk. Virus was concentrated from the supernatant using centrifugation with a 17% sorbitol cushion to generate the P1 virus used in all experiments. The titer of viral stocks was measured by infecting MRC-5 cells grown to 80% confluence on 24-well culture plates with serially diluted virus in a 500 μL volume. After 1 h, cells were washed once with cell culture medium and incubated overnight. About 17 h postinfection, cells were detached from the plate and %GFP-positive cells were measured on an Attune flow cytometer (ThermoFisher Scientific). Titer was quantified by the following calculation and reported as infectious units (IU) per mL:Virus concentration=(%GFPtest-%GFPuninfected)×#cells×dilution factorVolume.

In Vitro Neutralization Assay.

One day prior to the neutralization assay, MRC-5 cells were seeded at 5 to 10 × 103 cells/well in a cell-culture treated black, clear-bottom 96-well plate. Mouse sera were heat-inactivated (56 °C for 30 min) and √10 serially diluted (4 × 10−2 to 1.3 × 10−4). For assays including complement, guinea pig complement (Cedarlane, CL4051) was mixed with cell culture media prior to addition to virus and sera at a final concentration of 12.5%. Cytogam hyperimmune globulin was used as a positive control and was √10-fold serially diluted (100 to 0.32 μg/mL). Sera or antibody was mixed with 2500 IU AD169-GFP in culture medium and coincubated with sera for 2 h at 37 °C, 5% CO2. Cells were coincubated with virus/antibody mixture for 1 h at 37 °C, 5% CO2. Cells were then washed, overlaid with culture medium, and incubated for 17 to 20 h at 37 °C, 5% CO2. Wells were then stained with 2 μM Hoechst 33342 nuclear stain (Invitrogen H1399) and imaged on a Cytation C10 confocal plate reader (BioTek). All experimental conditions were performed in duplicate wells. Uninfected and infected, no-antibody controls were performed in quadruplicate on every plate. Nuclei and GFP+ cells were counted using Gen5 imaging software (v3.13, BioTek) and used to calculate the ratio of infected cells to number of nuclei (R). %Infection was calculated using the following equation:%Infection=Rtest-Rno virusRno Ab-Rno virus.

Averages of two experimental replicates were analyzed as ID50 and IC50 curves calculated by fitting to a 4PL curve (inhibitor, normalized response). The distribution of IC50 data was determined to be log-normal by the Shapiro–Wilk test for normality.

Data Processing and Statistical Analysis.

EM processing was performed using CryoSPARC v4.2.0 and subsequent versions (83). Statistical analyses were performed using GraphPad Prism 9.5.1 and subsequent versions (GraphPad Software). We also used GraphPad Prism 9.5.1 and subsequent versions to plot the data. Information about the statistical tests performed can be found in the figure legends.

Supplementary Material

Appendix 01 (PDF)

We thank members of the Texas Materials Institute EM facility (Karalee Jarvis, Raluca Gearba, and Xun Zhan) and The University of Texas at Austin Center for Biomedical Research Support (Michelle Mikesh) for technical assistance with negative-stain EM. We thank members of the Sauer Structural Biology Lab (Evan Schwartz, Axel Brilot) for technical assistance with cryo-EM data collection. We thank Kaci Erwin and James Guerra for technical assistance with mammalian cell culture, and we thank Justine Meccio for administrative support. We thank Emily J. Rundlet for providing helpful comments on the manuscript. Funding for these studies was provided by Dynavax Technologies Corporation. The Sauer Structural Biology Laboratory is supported by the University of Texas College of Natural Sciences and by award RR160023 from the Cancer Prevention and Research Institute of Texas.

Author contributions

M.R.S., P.O.B., A.G.L., A.R.R., J.A.G., R.S.M., L.Z., N.V.J., C.-L.H., K.P.K., J.D.C., S.R.P., J.A.M., D.Y., M.J.B., and J.S.M. designed research; M.R.S., P.O.B., A.G.L., A.R.R., J.A.G., R.S.M., L.Z., N.V.J., C.-L.H., M.C., K.P.K., C.M.C., and A.S.F. performed research; M.R.S., P.O.B., A.G.L., A.R.R., J.A.G., R.S.M., L.Z., N.V.J., C.-L.H., K.P.K., J.D.C., S.R.P., J.A.M., D.Y., M.J.B., and J.S.M. analyzed data; and M.R.S., P.O.B., A.G.L., J.A.M., and J.S.M. wrote the paper.

Competing interests

M.R.S., P.O.B., A.R.R., J.A.G., R.S.M., L.Z., N.V.J., C.-L.H., and J.S.M. are inventors on a patent application entitled “Prefusion-Stabilized CMV gB Proteins” (PCT/US2023/073369). J.D.C., D.Y., and M.J.B. are employees of Dynavax and hold Dynavax stock. The other authors declare that they have no competing interests.

Data, Materials, and Software Availability

Atomic coordinates have been deposited with the Protein Data Bank under accession codes 8VYM (91) and 8VYN (92). Cryo-EM maps have been deposited with the Electron Microscopy Data Bank under accession numbers EMD-43667, EMD-43670, 43671, and 43672 (93–96). All other data are included in the manuscript and/or SI Appendix.

Supporting Information

This article is a PNAS Direct Submission.

Preprint servers: This manuscript was deposited as a preprint to bioRxiv.
==== Refs
1 D. Gatherer , ICTV virus taxonomy profile: Herpesviridae 2021. J. Gen. Virol. 102 , 001673 (2021).34704922
2 P. Griffiths, M. Reeves, Pathogenesis of human cytomegalovirus in the immunocompromised host. Nat. Rev. Microbiol. 19 , 759–773 (2021).34168328
3 S. Sapuan, A. A. Theodosiou, B. L. Strang, P. T. Heath, C. E. Jones, A systematic review and meta-analysis of the prevalence of human cytomegalovirus shedding in seropositive pregnant women. Rev. Med. Virol. 32 , e2399 (2022).36196755
4 S. Manicklal, V. C. Emery, T. Lazzarotto, S. B. Boppana, R. K. Gupta, The “silent” global burden of congenital cytomegalovirus. Clin. Microbiol. Rev. 26 , 86–102 (2013).23297260
5 Anonymous, “Babies born with congenital CMV” in Cytomegalovirus (CMV) and Congenital CMV Infection (Centers for Disease Control and Prevention, 2022).
6 G. Picarda, C. A. Benedict, Cytomegalovirus: Shape-shifting the immune system. J. Immunol. 200 , 3881–3889 (2018).29866770
7 C. S. Nelson, B. C. Herold, S. R. Permar, A new era in cytomegalovirus vaccinology: Considerations for rational design of next-generation vaccines to prevent congenital cytomegalovirus infection. NPJ Vaccines 3 , 38 (2018).30275984
8 S. Scarpini , Development of a vaccine against human cytomegalovirus: Advances, barriers, and implications for the clinical practice. Vaccines (Basel) 9 , 551 (2021).34070277
9 J. M. White, A. E. Ward, L. Odongo, L. K. Tamm, Viral membrane fusion: A dance between proteins and lipids. Annu. Rev. Virol. 10 , 139–161 (2023).37774128
10 S. A. Connolly, T. S. Jardetzky, R. Longnecker, The structural basis of herpesvirus entry. Nat. Rev. Microbiol. 19 , 110–121 (2021).33087881
11 D. Navarro , Glycoprotein B of human cytomegalovirus promotes virion penetration into cells, transmission of infection from cell to cell, and fusion of infected cells. Virology 197 , 143–158 (1993).7692667
12 A. L. Vanarsdall, M. C. Chase, D. C. Johnson, Human cytomegalovirus glycoprotein gO complexes with gH/gL, promoting interference with viral entry into human fibroblasts but not entry into epithelial cells. J. Virol. 85 , 11638–11645 (2011).21880752
13 B. Adler , Role of human cytomegalovirus UL131A in cell type-specific virus entry and release. J. Gen. Virol. 87 , 2451–2460 (2006).16894182
14 G. Gerna , Dendritic-cell infection by human cytomegalovirus is restricted to strains carrying functional UL131-128 genes and mediates efficient viral antigen presentation to CD8+ T cells. J. Gen. Virol. 86 , 275–284 (2005).15659746
15 E. B. Struble, H. Murata, T. Komatsu, D. Scott, Immune prophylaxis and therapy for human cytomegalovirus infection. Int. J. Mol. Sci. 22 , 8728 (2021).34445434
16 R. F. Pass , Vaccine prevention of maternal cytomegalovirus infection. N. Engl. J. Med. 360 , 1191–1199 (2009).19297572
17 D. I. Bernstein , Safety and efficacy of a cytomegalovirus glycoprotein B (gB) vaccine in adolescent girls: A randomized clinical trial. Vaccine 34 , 313–319 (2016).26657184
18 P. D. Griffiths , Cytomegalovirus glycoprotein-B vaccine with MF59 adjuvant in transplant recipients: A phase 2 randomised placebo-controlled trial. Lancet 377 , 1256–1263 (2011).21481708
19 J. A. Jenks , Antibody binding to native cytomegalovirus glycoprotein B predicts efficacy of the gB/MF59 vaccine in humans. Sci. Transl. Med. 12 , eabb3611 (2020).33148624
20 M. Backovic, T. S. Jardetzky, Class III viral membrane fusion proteins. Curr. Opin. Struct. Biol. 19 , 189–196 (2009).19356922
21 N. Smargiasso , Exploring the N-glycosylation profile of glycoprotein B from human cytomegalovirus expressed in CHO and Nicotiana tabacum BY-2 Cells. Int. J. Mol. Sci. 20 , 3741 (2019).31370181
22 I. Bagdonaite , Global mapping of O-glycosylation of varicella zoster virus, human cytomegalovirus, and Epstein-Barr virus. J. Biol. Chem. 291 , 12014–12028 (2016).27129252
23 Y. Liu , Prefusion structure of human cytomegalovirus glycoprotein B and structural basis for membrane fusion. Sci. Adv. 7 , eabf3178 (2021).33674318
24 S. Chandramouli , Structure of HCMV glycoprotein B in the postfusion conformation bound to a neutralizing human antibody. Nat. Commun. 6 , 8176 (2015).26365435
25 H. G. Burke, E. E. Heldwein, Crystal structure of the human cytomegalovirus glycoprotein B. PLoS Pathog. 11 , e1005227 (2015).26484870
26 E. E. Heldwein , Crystal structure of glycoprotein B from herpes simplex virus 1. Science 313 , 217–220 (2006).16840698
27 M. Backovic, R. Longnecker, T. S. Jardetzky, Structure of a trimeric variant of the Epstein-Barr virus glycoprotein B. Proc. Natl. Acad. Sci. U.S.A. 106 , 2880–2885 (2009).19196955
28 T. Zeev-Ben-Mordehai , Two distinct trimeric conformations of natively membrane-anchored full-length herpes simplex virus 1 glycoprotein B. Proc. Natl. Acad. Sci. U.S.A. 113 , 4176–4181 (2016).27035968
29 Z. Si , Different functional states of fusion protein gB revealed on human cytomegalovirus by cryo electron tomography with Volta phase plate. PLoS Pathog. 14 , e1007452 (2018).30507948
30 B. Vollmer , The prefusion structure of herpes simplex virus glycoprotein B. Sci. Adv. 6 , eabc1726 (2020).32978151
31 S. Roche, F. A. Rey, Y. Gaudin, S. Bressanelli, Structure of the prefusion form of the vesicular stomatitis virus glycoprotein G. Science 315 , 843–848 (2007).17289996
32 S. Potzsch , B cell repertoire analysis identifies new antigenic domains on glycoprotein B of human cytomegalovirus which are target of neutralizing antibodies. PLoS Pathog. 7 , e1002172 (2011).21852946
33 A. C. Gomes , The cytomegalovirus gB/MF59 vaccine candidate induces antibodies against an antigenic domain controlling cell-to-cell spread. Nat. Commun. 14 , 1041 (2023).36823200
34 H. Meyer, V. A. Sundqvist, L. Pereira, M. Mach, Glycoprotein gp116 of human cytomegalovirus contains epitopes for strain-common and strain-specific antibodies. J. Gen. Virol. 73 , 2375–2383 (1992).1383409
35 L. M. Kauvar , A high-affinity native human antibody neutralizes human cytomegalovirus infection of diverse cell types. Antimicrob. Agents Chemother. 59 , 1558–1568 (2015).25534746
36 X. Ye , Recognition of a highly conserved glycoprotein B epitope by a bivalent antibody neutralizing HCMV at a post-attachment step. PLoS Pathog. 16 , e1008736 (2020).32745149
37 N. Kniess, M. Mach, J. Fay, W. J. Britt, Distribution of linear antigenic sites on glycoprotein gp55 of human cytomegalovirus. J. Virol. 65 , 138–146 (1991).1702157
38 A. Macagno , Isolation of human monoclonal antibodies that potently neutralize human cytomegalovirus infection by targeting different epitopes on the gH/gL/UL128-131A complex. J. Virol. 84 , 1005–1013 (2010).19889756
39 M. Perotti, J. Marcandalli, D. Demurtas, F. Sallusto, L. Perez, Rationally designed human cytomegalovirus gB nanoparticle vaccine with improved immunogenicity. PLoS Pathog. 16 , e1009169 (2020).33370407
40 J. S. McLellan , Structure-based design of a fusion glycoprotein vaccine for respiratory syncytial virus. Science 342 , 592–598 (2013).24179220
41 G. B. E. Stewart-Jones , Structure-based design of a quadrivalent fusion glycoprotein vaccine for human parainfluenza virus types 1–4. Proc. Natl. Acad. Sci. U.S.A. 115 , 12265–12270 (2018).30420505
42 C. L. Hsieh , Structure-based design of prefusion-stabilized human metapneumovirus fusion proteins. Nat. Commun. 13 , 1299 (2022).35288548
43 P. O. Byrne, J. S. McLellan, Principles and practical applications of structure-based vaccine design. Curr. Opin. Immunol. 77 , 102209 (2022).35598506
44 S. Sharma, T. W. Wisner, D. C. Johnson, E. E. Heldwein, HCMV gB shares structural and functional properties with gB proteins from other herpesviruses. Virology 435 , 239–249 (2013).23089254
45 T. W. Geders, K. Gustafson, B. C. Finzel, Use of differential scanning fluorimetry to optimize the purification and crystallization of PLP-dependent enzymes. Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 68 , 596–600 (2012).
46 S. Boivin, S. Kozak, R. Meijers, Optimization of protein purification and characterization using Thermofluor screens. Protein Expr. Purif. 91 , 192–206 (2013).23948764
47 A. Mera, M. Z. T. de Lima, A. Bernardes, W. Garcia, J. R. C. Muniz, Low-resolution structure, oligomerization and its role on the enzymatic activity of a sucrose-6-phosphate hydrolase from Bacillus licheniformis. Amino Acids 51 , 599–610 (2019).30673850
48 M. Backovic, G. P. Leser, R. A. Lamb, R. Longnecker, T. S. Jardetzky, Characterization of EBV gB indicates properties of both class I and class II viral fusion proteins. Virology 368 , 102–113 (2007).17655906
49 P. D. Adams , PHENIX: A comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D Biol. Crystallogr. 66 , 213–221 (2010).20124702
50 J. Jumper , Highly accurate protein structure prediction with AlphaFold. Nature 596 , 583–589 (2021).34265844
51 J. Hattne , Analysis of global and site-specific radiation damage in Cryo-EM. Structure 26 , 759–766.e4 (2018).29706530
52 H. D. Patel , In vitro characterization of human cytomegalovirus-targeting therapeutic monoclonal antibodies LJP538 and LJP539. Antimicrob. Agents Chemother. 60 , 4961–4971 (2016).27270290
53 N. Spindler , Structural basis for the recognition of human cytomegalovirus glycoprotein B by a neutralizing human antibody. PLoS Pathog. 10 , e1004377 (2014).25299639
54 J. D. Campbell, Development of the CpG adjuvant 1018: A case study. Methods Mol. Biol. 1494 , 15–27 (2017).27718183
55 G. Gerna , Human cytomegalovirus serum neutralizing antibodies block virus infection of endothelial/epithelial cells, but not fibroblasts, early during primary infection. J. Gen. Virol. 89 , 853–865 (2008).18343825
56 C. S. Nelson , Human cytomegalovirus glycoprotein B nucleoside-modified mRNA vaccine elicits antibody responses with greater durability and breadth than MF59-adjuvanted gB protein immunization. J. Virol. 94 , e00186-20 (2020).32051265
57 A. Krarup , A highly stable prefusion RSV F vaccine derived from structural analysis of the fusion mechanism. Nat. Commun. 6 , 8143 (2015).
58 M. G. Joyce , Iterative structure-based improvement of a fusion-glycoprotein vaccine against RSV. Nat. Struct. Mol. Biol. 23 , 811–820 (2016).27478931
59 J. Pallesen , Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen. Proc. Natl. Acad. Sci. U.S.A. 114 , E7348–E7357 (2017).28807998
60 M. Germer, P. Herbener, J. Schuttrumpf, Functional properties of human cytomegalovirus hyperimmunoglobulin and standard immunoglobulin preparations. Ann. Transplant 21 , 558–564 (2016).27595792
61 F. Li , Complement enhances in vitro neutralizing potency of antibodies to human cytomegalovirus glycoprotein B (gB) and immune sera induced by gB/MF59 vaccination. NPJ Vaccines 2 , 36 (2017).29263890
62 A. Ohta , Recombinant human monoclonal antibodies to human cytomegalovirus glycoprotein B neutralize virus in a complement-dependent manner. Microbes Infect. 11 , 1029–1036 (2009).19651232
63 C. S. Nelson , HCMV glycoprotein B subunit vaccine efficacy mediated by nonneutralizing antibody effector functions. Proc. Natl. Acad. Sci. U.S.A. 115 , 6267–6272 (2018).29712861
64 A. C. Finnefrock , Preclinical evaluations of peptide-conjugate vaccines targeting the antigenic domain-2 of glycoprotein B of human cytomegalovirus. Hum. Vaccin. Immunother. 12 , 2106–2112 (2016).26986197
65 L. Li , A conditionally replication-defective cytomegalovirus vaccine elicits potent and diverse functional monoclonal antibodies in a phase I clinical trial. NPJ Vaccines 6 , 79 (2021).34078915
66 J. A. Jenks , A single, improbable B cell receptor mutation confers potent neutralization against cytomegalovirus. PLoS Pathog. 19 , e1011107 (2023).36662906
67 W. C. Koff, R. Rappuoli, S. A. Plotkin, Historical advances in structural and molecular biology and how they impacted vaccine development. J. Mol. Biol. 435 , 168113 (2023).37080423
68 M. S. Gilman , Rapid profiling of RSV antibody repertoires from the memory B cells of naturally infected adult donors. Sci. Immunol. 1 , eaaj1879 (2016).28111638
69 F. Beilstein , Identification of a pH-sensitive switch in VSV-G and a crystal structure of the G pre-fusion state highlight the VSV-G structural transition pathway. Cell Rep. 32 , 108042 (2020).32814045
70 L. Rutten , Structure-based design of prefusion-stabilized filovirus glycoprotein trimers. Cell Rep. 30 , 4540–4550.e3 (2020).32234486
71 M. Zhou , Targeted mutagenesis of the herpesvirus fusogen central helix captures transition states. Nat. Commun. 14 , 7958 (2023).38042814
72 J. M. Binley , A recombinant human immunodeficiency virus type 1 envelope glycoprotein complex stabilized by an intermolecular disulfide bond between the gp120 and gp41 subunits is an antigenic mimic of the trimeric virion-associated structure. J. Virol. 74 , 627–643 (2000).10623724
73 J. K. Lee, A. Prussia, J. P. Snyder, R. K. Plemper, Reversible inhibition of the fusion activity of measles virus F protein by an engineered intersubunit disulfide bridge. J. Virol. 81 , 8821–8826 (2007).17553889
74 P. S. Lee, X. Zhu, W. Yu, I. A. Wilson, Design and structure of an engineered disulfide-stabilized influenza virus hemagglutinin trimer. J. Virol. 89 , 7417–7420 (2015).25926650
75 G. B. E. Stewart-Jones , Interprotomer disulfide-stabilized variants of the human metapneumovirus fusion glycoprotein induce high titer-neutralizing responses. Proc. Natl. Acad. Sci. U.S.A. 118 , e2106196118 (2021).34551978
76 G. W. Wilkinson , Human cytomegalovirus: Taking the strain. Med. Microbiol. Immunol. 204 , 273–284 (2015).25894764
77 M. G. Revello, G. Gerna, Human cytomegalovirus tropism for endothelial/epithelial cells: Scientific background and clinical implications. Rev. Med. Virol. 20 , 136–155 (2010).20084641
78 G. Gerna, A. Kabanova, D. Lilleri, Human cytomegalovirus cell tropism and host cell receptors. Vaccines (Basel) 7 , 70 (2019).31336680
79 I. Baraniak , Protection from cytomegalovirus viremia following glycoprotein B vaccination is not dependent on neutralizing antibodies. Proc. Natl. Acad. Sci. U.S.A. 115 , 6273–6278 (2018).29686064
80 E. C. Semmes , Maternal Fc-mediated non-neutralizing antibody responses correlate with protection against congenital human cytomegalovirus infection. J. Clin. Invest. 132 , e156827 (2022).35763348
81 E. C. Semmes , ADCC-activating antibodies correlate with decreased risk of congenital human cytomegalovirus transmission. JCI Insight 8 , e167768 (2023).37427588
82 C. L. Hsieh , Structure-based design of prefusion-stabilized SARS-CoV-2 spikes. Science 369 , 1501–1505 (2020).32703906
83 A. Punjani, J. L. Rubinstein, D. J. Fleet, M. A. Brubaker, cryoSPARC: Algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14 , 290–296 (2017).28165473
84 T. D. Goddard , UCSF ChimeraX: Meeting modern challenges in visualization and analysis. Protein Sci. 27 , 14–25 (2018).28710774
85 D. N. Mastronarde, Automated electron microscope tomography using robust prediction of specimen movements. J. Struct. Biol. 152 , 36–51 (2005).16182563
86 K. Jamali , Automated model building and protein identification in cryo-EM maps. Nature 628 , 450–457 (2024).38408488
87 P. Emsley, K. Cowtan, Coot: Model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60 , 2126–2132 (2004).15572765
88 T. I. Croll, ISOLDE: A physically realistic environment for model building into low-resolution electron-density maps. Acta Crystallogr. D Struct. Biol. 74 , 519–530 (2018).29872003
89 E. Krissinel, K. Henrick, Inference of macromolecular assemblies from crystalline state. J. Mol. Biol. 372 , 774–797 (2007).17681537
90 S. Terhune , Human cytomegalovirus UL38 protein blocks apoptosis. J. Virol. 81 , 3109–3123 (2007).17202209
91 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, Data from “Soluble ectodomain of humancytomegalovirus (HCMV) glycoprotein B (gB) in the postfusion conformation in complex with1G2 and 7H3 Fabs.” Protein Data Bank. 10.2210/pdb8VYM/pdb. Deposited 9 February 2024.
92 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, Data from “Soluble ectodomain of humancytomegalovirus (HCMV) glycoprotein B (gB) stabilized in a prefusion-like conformation incomplex with 1G2 and 7H3, composite map (global and local) and model.” Protein Data Bank. 10.2210/pdb8VYN/pdb. Deposited 9 February 2024.
93 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, EMD-43667. Electron Microscopy Data Bank (EMDB). https://www.ebi.ac.uk/emdb/EMD-43667. Deposited 9 February 2024.
94 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, EMD-43670. Electron Microscopy Data Bank (EMDB). https://www.ebi.ac.uk/emdb/EMD-43670. Deposited 9 February 2024.
95 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, EMD-43671. Electron Microscopy Data Bank (EMDB). https://www.ebi.ac.uk/emdb/EMD-43671. Deposited 9 February 2024.
96 M. R. Sponholtz, P. O. Byrne, J. S. McLellan, EMD-43672. Electron Microscopy Data Bank (EMDB). https://www.ebi.ac.uk/emdb/EMD-43672. Deposited 9 February 2024.
