A Novel Strain-Specific Neutralizing Epitope on Glycoprotein H of Human Cytomegalovirus

ABSTRACT

Human cytomegalovirus (HCMV) is a ubiquitous pathogen that causes severe clinical disease in immunosuppressed patients and congenitally infected newborn infants. Viral envelope glycoproteins represent attractive targets for vaccination or passive immunotherapy. To extend the knowledge of mechanisms of virus neutralization, monoclonal antibodies (MAbs) were generated following immunization of mice with HCMV virions. Hybridoma supernatants were screened for in vitro neutralization activity, yielding three potent MAbs, 6E3, 3C11, and 2B10. MAbs 6E3 and 3C11 blocked infection of all viral strains that were tested, while MAb 2B10 neutralized only 50% of the HCMV strains analyzed. Characterization of the MAbs using indirect immunofluorescence analyses demonstrated their reactivity with recombinantly derived gH. While MAbs 6E3 and 3C11 reacted with gH when expressed alone, 2B10 detected gH only when it was coexpressed with gB and gL. Recognition of gH by 3C11 was dependent on the expression of the entire ectodomain of gH, whereas 6E3 required residues 1 to 629 of gH. The strain-specific determinant for neutralization by Mab 2B10 was identified as a single Met→Ile amino acid polymorphism within gH, located within the central part of the protein. The polymorphism is evenly distributed among described HCMV strains. The 2B10 epitope thus represents a novel strain-specific antibody target site on gH of HCMV. The dependence of the reactivity of 2B10 on the simultaneous presence of gB/gH/gL will be of value in the structural definition of this tripartite complex. The 2B10 epitope may also represent a valuable tool for diagnostics to monitor infections/reinfections with different HCMV strains during pregnancy or after transplantation.

IMPORTANCE HCMV infections are life threatening to people with compromised or immature immune systems. Understanding the antiviral antibody repertoire induced during HCMV infection is a necessary prerequisite to define protective antibody responses. Here, we report three novel anti-gH MAbs that potently neutralized HCMV infectivity. One of these MAbs (2B10) targets a novel strain-specific conformational epitope on gH that only becomes accessible upon coexpression of the minimal fusion machinery gB/gH/gL. Strain specificity is dependent on a single amino acid polymorphism within gH. Our data highlight the importance of strain-specific neutralizing antibody responses against HCMV. The 2B10 epitope may also represent a valuable tool for diagnostics to monitor infections/reinfections with different HCMV strains during pregnancy or after transplantation. In addition, the dependence of the reactivity of 2B10 on the simultaneous presence of gB/gH/gL will be of value in the structural definition of this tripartite complex.

INTRODUCTION

It is estimated that human cytomegalovirus (HCMV) infects more than 80% of the world’s population (1, 2). Within immunocompetent hosts, HCMV is well controlled by the immune system and the infection only occasionally progresses to disease. In contrast, HCMV infections can cause life-threatening disease in immunocompromised individuals and newborn infants. Thus, HCMV is one of the most common opportunistic pathogens affecting immunosuppressed patients and is a major source of increased morbidity and mortality in patients undergoing solid organ or stem cell transplantation (3, 4). In addition, HCMV is not infrequently transmitted to the developing fetus, resulting in HCMV being a leading infectious cause of congenital abnormalities and long-term neurodevelopmental sequelae in infected infants and children worldwide (5). Because of the clinical importance of infections with HCMV, passive and/or active immunization strategies against HCMV are urgently needed and the development of a preventive vaccine has been identified as a high priority by health authorities.

In immunocompetent individuals, infection with HCMV is well controlled by a multilayered immune response, including innate and adaptive immune effector functions. Within the adaptive arm of the immune response, the development of antiviral antibodies contributes to effective control of HCMV infection. In cases of congenital HCMV infection, naturally acquired antiviral antibodies have been considered to be important components of the maternal immune response, since they have been associated with a decreased frequency of transplacental transmission of the virus (6). In addition, several reports suggest that passive transfer of HCMV hyperimmune globulin (HIG) can be beneficial for the prevention and treatment of congenital CMV infections (7–9). Similarly, the administration of HCMV HIG in certain transplant settings can modify posttransplant HCMV disease (reviewed in reference 10). However, it should be also noted that preexisting antibodies to HCMV do not protect seropositive mothers from HCMV infections and/or intrauterine transmission to their offspring. The failure of existing immunity to HCMV to prevent maternal infections and congenital infections has been most readily demonstrated in maternal populations with high HCMV seroprevalence. In such maternal populations, nonprimary maternal infections account for up to 90% of all congenital CMV infections and are responsible for the majority of infected infants with neurodevelopmental sequelae (11). It has been proposed that the source of such nonprimary infections in immune mothers is exposure to an antigenically distinct virus that escapes the HCMV-specific immune response that has been established following primary infection (11, 12).

Neutralizing antibodies (NtAbs) that interfere in vitro with envelope glycoprotein-mediated entry of the virus into host cells are thought to play an important role in the protection against HCMV infection in vivo (13). Within the envelope of HCMV, two proteins or protein complexes have been identified as being the most important targets for the NtAb response: glycoprotein B (gB) and the gH-containing complexes gH/gL/gO (trimeric complex) and gH/gL/unique long 128 (UL128)/UL130/UL131A (pentameric complex [PC]) (14, 15). Glycoprotein B represents the HCMV fusion protein (16), which is essential for infection of all types of target cells. Activation of gB’s fusogenic activity is an open question for HCMV but likely requires its association with the gH/gL/gO complex (17–20). The PC is required in addition to the trimeric complex for efficient targeting of HCMV to epithelial and endothelial cells (19, 21–24). The PC has been shown to induce antibodies with high neutralizing potency that are demonstrable only in cell types for which it is required for entry. In contrast to the PC, HCMV gB or gH/gL proteins elicit serum HCMV NtAbs that block entry into both fibroblasts and other cell types, such as epithelial/endothelial cells (reviewed in reference 14).

Following natural infection with HCMV, gH-specific antibodies have been detected in convalescent-phase serum samples with frequencies between 95% and 100% (25, 26). In some human serum samples, anti-gH antibodies constitute the majority of the neutralizing activity when assayed on fibroblasts (26). Several anti-gH NtAbs have been isolated from HCMV-seropositive individuals (27, 28) and were initially considered for prophylaxis and treatment of HCMV infections nearly 2 decades ago (29). Recently an affinity-matured version of anti-gH antibody MSL-109, in combination with an anti-UL131A antibody, reduced the incidence of HCMV infection in the post-transplant period, delayed viremia, and reduced HCMV disease in high-risk kidney transplant recipients when given prophylactically (30).

Antibodies against gH have also been isolated following immunization of mice and rabbits (31–34). Notably, the neutralization capacity for some of these antibodies was demonstrated to be strain-specific (32, 33). The immunodominant linear epitope targeted by these strain-specific antibodies was mapped to the N terminus of gH and has been proven as a valuable tool to monitor reinfections in mothers with preconceptional immunity and transplant recipients (35–37).

In our continuing efforts to increase knowledge on HCMV-specific antibody targets, we have generated murine monoclonal antibodies following immunization with intact virions. Antibodies obtained by this immunization strategy were subsequently screened for their capacity to block infection of human fibroblast cells. We identified three monoclonal antibodies (MAbs), termed 2B10, 6E3, and 3C11, which neutralized HCMV at nanomolar concentrations. While 6E3 and 3C11 were effective at blocking a broad group of distinct HCMV strains, 2B10 neutralization of HCMV infectivity was strictly strain-specific. The three MAbs were found to be directed against gH, but in contrast to 6E3 and 3C11, MAb 2B10 required coexpression of gB/gH/gL for its reactivity. The strain-specific epitope of MAb 2B10 was mapped to a single residue on the surface of gH, further emphasizing the potential importance of an immune response directed against polymorphic sites expressed by gH.

RESULTS

Isolation and primary characterization of three neutralizing MAbs.

Following immunization of mice with virions from HCMV strain AD169, three neutralizing MAbs, termed 3C11, 6E3, and 2B10, were isolated. While 6E3 is an IgG2b, 2B10 and 3C11 are both IgG2c subtypes. As shown by the results in Fig. 1, MAb 6E3 and 3C11 exhibited neutralizing activity in the nanomolar range (0.07 to 4.5 μg/ml) when tested against the HCMV strains AD169, TB40/E, and Towne in fibroblast cultures (Fig. 1A to C). The neutralizing activities of these MAbs were comparable to that of the gB-specific human MAb C23, which was used throughout this study as a reference antibody (38, 39). In contrast, MAb 2B10 had potent neutralizing activity (50% inhibitory concentration [IC₅₀] between 0.07 to 1.8 μg/ml) against strains AD169 and TB40/E (Fig. 1A and B) but was completely inactive against strain Towne (Fig. 1C).

FIG 1.

Neutralization activities of MAbs 6E3, 3C11, and 2B10 against AD169, TB40/E, and Towne strains. Log₂ dilutions of antibodies were incubated for 1 h with comparable infectious units of the indicated strains before infection of fibroblasts. Forty-eight hours later, infection was quantified by measuring luciferase activity (A and B) or IE1-positive cells (C). The percentage of neutralization (%nt) was calculated relative to the results for the no-virus and no-antibody controls. Neutralizing activity was assessed in duplicate, and the mean values (±standard errors of the means [SEM]) are given. The neutralizing human anti-gB antibody C23 served as an internal control.

To analyze whether the lack of neutralizing activity of MAb 2B10 against strain Towne was unique to this strain of HCMV, additional virus strains were tested for their susceptibility to the neutralizing activity of MAb 2B10. Human foreskin fibroblasts (HFF) were infected with comparable PFU of the HCMV strains Toledo, PAN23, VR1814, Davis, Merlin, or TR and used to determine the neutralizing activities of MAb 6E3, 3C11, and 2B10 (Fig. 2). The neutralization capacities of 6E3 and 3C11 ranged from 0.07 to 4.5 μg/ml; the results from the assay using 6E3 are shown in Fig. 2 as an example. Antibody 2B10 neutralized strains Toledo, PAN23, and VR1814 with IC₅₀s of 0.05 to 0.65 μg/ml (Fig. 2A to C) but failed to neutralize strain Davis, Merlin, or TR even at concentrations of 10 μg/ml (Fig. 2D to F). As expected, MAb C23 neutralized all strains with comparable efficiencies, i.e., IC₅₀s between 0.07 to 0.4 μg/ml. We concluded from these results that MAbs 6E3 and 3C11 were capable of neutralizing a wider spectrum of HCMV strains, while the activity of 2B10 was restricted to a subset of virus strains.

FIG 2.

Neutralization capacities of MAbs 6E3 and 2B10 against various HCMV strains. MAb 2B10 or 6E3 or anti-gB antibody C23 was incubated for 1 h with comparable infectious units of the indicated HCMV strains. HFF cells were infected with the antibody-virus mixtures, and IE1-positive cells counted 48 h later. The percentage of neutralization (%nt) was calculated relative to the results for the no-virus and no-antibody controls. Neutralizing activity was assessed in duplicate, and the mean values (±SEM) are given.

The relative conservation of serological reactivity of envelope glycoproteins gB, gH, and gL and most components of the PC that has been described in the literature suggested that gO could be a likely candidate for the strain-specific reactivity of MAb 2B10. Glycoprotein O is an extremely polymorphic protein, with sequence differences of about 40% between strains, and has been shown to be a target of neutralizing antibodies (Fig. S1 in the supplemental material). In fact, eight distinct gO genotypes have been described (40–42). To determine if MAb 2B10 is directed against gO, we utilized isogenic TB40/E recombinant viruses in which the authentic gO GT1c sequence was replaced by Towne gO GT4 (43). Both 2B10 and 6E3 potently neutralized recombinant TB40/E-gO-Towne strains, with IC₅₀s between 0.16 and 0.6 μg/ml (data not shown), indicating that gO did not account for the strain-specific neutralization by MAb 2B10.

gH as a target of MAbs 6E3, 3C11, and 2B10.

Having excluded gO as a target for MAb 2B10, the MAbs were tested for recognition of candidate envelope glycoproteins gB, gL, and gH by using indirect immunofluorescence analyses of cells transiently expressing the respective protein as previously described (44). To this end, HEK293T cells were transfected with plasmids encoding gB, gL-myc, or gH (Fig. 3A to C) or combinations thereof (Fig. 3D and E). Two days later, cells were stained with antibody to gB (C23), gL (anti-myc antibody), or gH (MSL-109) in combination with one of the murine MAbs 3C11, 6E3, and 2B10 as indicated. These analyses identified gH as the target of all three MAbs (Fig. 3C). None of the antibodies reacted with cells expressing gB or gL (Fig. 3A and B). On gH-expressing cells, MAbs 6E3 and 3C11 showed signals similar to that of MSL-109. Thus, we identified gH as the target for MAbs 6E3 and 3C11.

FIG 3.

Target identification for MAbs 6E3, 3C11, and 2B10. (A to C) Indirect immunofluorescence analyses of HEK293T cells transfected with gB, gL, or gH alone. (D) Linear representation of C-terminal truncation mutants of gH that were coexpressed with gL-myc. Structural domains of gH as deduced from Chandramouli et al. (46) are indicated by different colors and Roman numerals I to VI. Reactivity of 6E3 or 3C11, respectively, as determined by immunofluorescence analyses, is displayed as positive (+) and negative (−) beside each construct. (E) AD169 gH was expressed in combination with gL-myc (c, d, l, m and t, u), gB (e, f, n, o, and v, w), or gB and gL-myc (g, h, p, q, and x, y). (A to E) Two days after transfection, indirect immunofluorescence analyses were performed using primary antibodies human anti-gB MAb C23 (A), rabbit anti-myc antibody (B and D), human anti-gH antibody MSL-109 (C) or anti-His antibody (D) and one of the murine MAbs 6E3, 3C11, and 2B10 as indicated. Proteins were visualized by using appropriate secondary antibodies coupled to Alexa Fluor 488 (green), Alexa Fluor 647 (cyan), and Alexa Fluor 555 (red). For visualization of cell nuclei, costaining with DAPI was carried out.

To gain insights into the antigenic binding sites of the MAbs 6E3 and 3C11, gH truncation constructs that had deletion mutations of individual or combinatorial structural domains of the gH ectodomain fused to a C-terminal His₆ tag were generated (Fig. 3D, scheme). Following transient transfection, the expression of all gH mutants was confirmed via indirect immunofluorescence analyses by using an anti-His antibody. Moreover, signals from gH proteins colocalized with their heterodimeric partner gL, as illustrated by staining with an anti-myc antibody (Fig. 3D). MAb 3C11 interacted only with the protein representing the entire ectodomain of gH (Fig. 3D, a to o). In contrast, MAb 6E3 tolerated truncation of amino acids (aa) 630 to 718 and even showed residual reactivity with mutant gH-aa1-495 (Fig. 3D, p to t). From this result, we concluded that MAbs 3C11 and 6E3 have different requirements for interaction with the antigen. Interestingly, both MAbs were reactive with cells expressing gH in the absence of gL, indicating that formation of the gH/gL complex was not required for the correct presentation of the respective epitopes (Fig. 3C).

In contrast, the immunofluorescence signal obtained with 2B10 differed significantly from the signal obtained with 6E3 or 3C11. While we observed complete congruence of the signals between 6E3 or 3C11 and MSL-109 (Fig. 3C, a to f), only a fraction of cells that expressed gH were positive for recognition of 2B10 (Fig. 3C, g to i). A potential explanation for the low signal could have been the lack of gL in the assay and, consequently, a nonnative conformation of gH, since gH is conformationally stabilized by its heterodimeric partner gL (45). Coexpression of gL, however, did not significantly affect the reactivity of 2B10 (Fig. 3E, a to d). The expression level of gH was not significantly affected upon coexpression with gL and/or gB as determined by the reactivity of 6E3 or 3C11 (Fig. 3E, i to y). The signal from 2B10 was comparable to that of 6E3 or 3C11 only when gH/gL was coexpressed with gB (Fig. 3E, g and h).

To confirm the strain-specific reactivity of 2B10, we performed indirect immunofluorescence analyses of cells expressing gB and gL together with gH from strain AD169, Towne, or Merlin. As shown by the results in Fig. 4, MAb 2B10 did not react with cells expressing gB/gL together with gH of Towne or Merlin, while MAb 3C11 was reactive, confirming the expression of the gH variants. A potential, although unlikely explanation for the lack of reactivity for gH from strains Towne and Merlin was that gB/gL from strain AD169 failed to form a complex with gH from Towne or Merlin, despite the fact that the sequences of gH of Towne and Merlin are 96% homologous to the sequence of AD169 gH at the protein level. In addition, syncytia could be observed in these samples, indicating that a functional gB/gH/gL complex was formed with the gH from Merlin and Towne (Fig. 4, see magnifications of insets). Thus, syncytium formation as observed for the gB/gL and Merlin/Towne gH complexes was considered an internal control for the correct formation and functionality of the gB/gH/gL core fusion machinery (20). Taken together, the results demonstrated that MAbs 6E3 and 3C11 were reactive with gH from a variety of different HCMV strains, while MAb 2B10 only recognized gH from a subset of strains. Finally, efficient reactivity of 2B10 required a gH-containing glycoprotein complex or a conformation of gH that was only present following coexpression of gB/gH/gL.

FIG 4.

Strain-specific immunoreactivity of MAb 2B10. Indirect immunofluorescence analyses of HEK293T cells that were transfected with plasmids coding for gB and gL-myc together with AD169 gH (a to d and n to q), Merlin gH (e to h and r to u), or Towne gH (i to m and v to y). Immunofluorescence analyses were performed as described above by using primary antibodies human anti-gB MAb C23, rabbit anti-myc antibody, and murine MAb 3C11 (a to m) or MAb 2B10 (n to y), respectively.

Epitope mapping of MAb 2B10.

Mapping the epitope of 2B10 was of interest as it might reveal a previously unidentified strain-specific antibody binding site on gH. Since our previous results highlighted that 2B10 reactivity strongly depends on complex formation of gB/gH/gL, we generated a set of AD169/Merlin gH chimeric proteins to map the binding site(s) of MAb 2B10 as an alternative to the generation of gH truncation mutants. Aided by the recently published structure of Merlin gH, chimeric proteins were constructed that were composed of varying lengths of AD169 gH fused to Merlin gH (46). When sequence polymorphisms between gH of AD169 and Merlin were taken into account, the resulting chimeric proteins contained various numbers of the eight gH polymorphic sites (Fig. 5A and Fig. S1). The expression of the newly generated chimeric proteins was confirmed via indirect immunofluorescence analyses using 3C11, which recognized each construct (Fig. 5B). In contrast, 2B10 bound specifically to a chimera containing aa 192 to 296 of AD169 gH, while the results for the remaining chimeric proteins were negative (Fig. 5C, a to y). As illustrated in Fig. 5A, the region comprising aa 192 to 296 contains two sequence polymorphisms that differ between AD169 and Merlin, as follows: (i) the positional homologs at residue 220/219 are Met→IIe, and (ii) those at residue 285/284 comprise an Ala/Thr→Asp exchange (Fig. S1). For further fine mapping of the residues mediating binding of 2B10, the respective residues in AD169 gH were substituted individually for the corresponding amino acid present in Merlin gH, resulting in proteins with mutations M220I (a change of Met to Ile at position 220) and A285D, respectively. Indirect immunofluorescence analyses revealed that MAb 2B10 reacted with AD169-gH-A285D but failed to react with AD169-gH-M220I (Fig. 5C, z to VII), indicating an essential role of M220/219 for binding of 2B10. To confirm M220/219 as the critical amino acid for MAb 2B10, a mutant of Merlin gH was generated, in which the positional homolog isoleucine was exchanged for methionine. This mutant Merlin-gH-I219M was reactive with MAb 2B10 (Fig. 5C, VIII to XI). In summary, the polymorphism M220/219I within glycoprotein H was identified as the residue that determined the virus strain-specific binding of MAb 2B10.

FIG 5.

Epitope mapping of MAb 2B10. (A) Scheme of gH chimeras and gH point mutants used for epitope mapping. Eight gH polymorphisms identified by the alignment as shown in Fig. S1 are indicated as black bars within gH with their respective amino acid sequences at the top. The reactivity of 2B10 as determined by immunofluorescence analyses as shown in panel C is displayed as positive (+) or negative (−) beside each construct. (B, C) Immunofluorescence analyses of HEK293T cells that were transfected with gB and gL-myc together with either of the chimeric proteins composed from gH of AD169 and Merlin (a to y) or point mutations thereof, as indicated (z to XI). Two days later, indirect immunofluorescence analysis was performed exactly as described before using MAb 3C11 (B) or MAb 2B10 (C) for detection of gH.

Further characterization of the 2B10 epitope.

Epitope-paratope interactions in general involve more than a single residue, and the mean area covered by an antibody is estimated to be on the order of 700 Ų (47). In addition, the fact that binding of 2B10 to its target required the simultaneous presence of gB, gH, and gL indicated that a more complex protein structure might be required for proper recognition by 2B10. We therefore attempted to map additional residues on gH that might be involved in 2B10 binding. The available three-dimensional (3-D) protein structure of gH localized the strain-determining polymorphism M220/219I to an exposed loop that is formed by aa 218 to 225 (Fig. 6A, arbitrarily termed loop 1, red). In close proximity to this loop, i.e., approximately 10 to 25 Å distal to this loop 1, are loop 2 (aa 239 to 248, green), loop 3 (aa 329 to 335, magenta), loop 4 (aa 386 to 391, orange), and the more distal loop 5 (aa 284 to 286, cyan), all of which could theoretically contribute binding residues to 2B10 (Fig. 6A). The results of the previous experiment demonstrated that the strain variation situated in loop 5 did not significantly contribute to the 2B10 epitope, since the A285D substitution did not abrogate binding of MAb 2B10 (Fig. 5C, z to III). We therefore concentrated on the remaining loops 1 to 4 to further define the epitope targeted by MAb 2B10. A set of surface-exposed residues that are in proximity to M220 within loop 1 or one of the adjacent loops 2, 3, or 4 of AD169 gH were selected for alanine scanning mutagenesis (Fig. 6A, underlined residues). According to an energetic analysis with MAESTROweb, all of these exchanges cause only minor (<0.6 kcal/mol) protein destabilization (48). Consequently, the mutations were not predicted to cause a significant disturbance of the gH structure.

FIG 6.

Coepitope mapping of MAb 2B10. (A) Strain-determining polymorphism M220/219I (highlighted in yellow) is located within a loop (1, red) as deduced from the gH structure from Chandramouli et al. (46) (PDB ID 5VOB). Putative neighboring coepitopes (loops 2 to 5) are indicated in different colors with their respective amino acid sequences and positions within full length AD169 gH. Mutated residues are underlined. TM, transmembrane domain. (B, C) HEK293T cells were cotransfected with gB, gL-myc, and the indicated mutant of AD169 gH. Two days later, indirect immunofluorescence analysis was performed using the anti-gH antibody 3C11 (B) or 2B10 (C).

Following transient transfection, expression of the new mutants was confirmed by indirect immunofluorescence analysis using 3C11, which detected every transiently expressed protein (Fig. 6B). In contrast, MAb 2B10 interacted exclusively with mutant gH-221DE/AA222, while substitution of any other residue in loop 1 abrogated MAb 2B10 binding (Fig. 6C, a to q). Similarly, amino acid substitutions within loop 2 (gH-242DD/AA243) or loop 3 (gH-332QML/AAA334) also abolished binding of MAb 2B10 (Fig. 6C, r to y). Point mutagenesis within loop 4 (aa 386 to 391) did not affect the reactivity of MAb 2B10 (Fig. 6C, z to XI), suggesting that this loop is not involved in 2B10 binding. Thus, in addition to the essential residue M220/219I of gH, the epitope of 2B10 likely involves additional residues within loop 1 (aa 218 to 225), loop 2 (aa 239 to 248), and loop 3 (aa 329 to 335) of gH, as evidenced by the findings that substitution of two or three surface-exposed residues within these loops abrogated MAb 2B10 interaction.

Neutralization capacity of 2B10 against recombinant Towne-I219M.

As a final proof for the importance of the polymorphism M220/219I for binding and neutralization of 2B10, a recombinant Towne virus encoding the positional substitution I219M in Towne gH was generated. We applied bacterial artificial chromosome (BAC) recombination using a Towne-GFP-BACmid (49) to exchange the isoleucine at position 219 of gH for methionine (Fig. 7A). Nucleotide sequencing of the UL75 open reading frame confirmed correct insertion of the mutation. After reconstitution of infectious virus, HFF cells were infected with comparable PFU of Towne-GFP-gH-I219M or its parental strain and analyzed for neutralization by 2B10. MAb 2B10 failed to neutralize the parental Towne strain (compare with Fig. 1C) but did neutralize the recombinant virus Towne-I219M with IC₅₀s of approximately 0.47 μg/ml (Fig. 7B). Anti-gB MAb C23 neutralized both analyzed strains with comparable IC₅₀s of about 0.3 to 0.45 μg/ml (Fig. 7B). Of note, MAb 2B10’s neutralization capacity toward the laboratory-adapted recombinant strain was in the same range as for the clinical strains Toledo, PAN23, and VR1814 (Fig. 2A to C). This result confirmed the polymorphism M220/219I as the determinant that dictates the neutralizing capacity of MAb 2B10.

FIG 7.

Neutralization capacities of C23 and 2B10 against recombinant Towne-I219M. (A) Schematic representation of recombinant HCMV BACmids generated via homologous recombination in E. coli cells. Top, genomic region UL75 (gH) of the BACmid Towne-GFP; numbers indicate nucleotide positions after reverse complementation of strain Towne (accession number GQ121041). Bottom, expected binding of MAb 2B10 to either Towne wild type (I219) or Towne-gH-I219M. (B) Antibody C23 (left) or 2B10 (right) was incubated with similar PFU of either HCMV strain Towne or its recombinant derivative Towne-gH-I219M before HFF cells were infected with the antibody-virus mixtures. GFP-positive cells were counted 48 h later. The percentage of neutralization was calculated relative to the results for the no-virus and no-antibody controls. Neutralizing activity was assessed in duplicate, and the mean values (±SEM) are given.

DISCUSSION

Immunization of mice with unadjuvanted HCMV particles resulted in the identification of three neutralizing MAbs, 6E3, 3C11, and 2B10, that were specific for gH. Each of these MAbs neutralized infectious virus in the nanomolar range (0.07 to 4.5 μg/ml), activity that is comparable to those of previously described potent human antibodies directed at either gB or gH (27, 28, 39, 50, 51). The IC₅₀s of anti-gB MAb C23 and anti-gH MAbs 6E3, 3C11, and 2B10 varied depending on the HCMV strain used in the neutralization assay. Such an observation is not without precedent and was observed for a variety of antibodies against different viral glycoproteins (i.e., gB, gH, gN, or gO), but whether this variation in IC₅₀s for neutralization is of in vivo relevance remains to be elucidated (50–56). While MAbs 6E3 and 3C11 exhibited comparable reactivities against a number of unrelated HCMV strains, 2B10’s reactivity was strictly strain-specific, showing an “all or nothing” neutralization phenotype (Fig. 1, 2, and 7), and to our knowledge, 2B10 is the first MAb with robust HCMV-neutralizing activity that is absolutely strain-specific. A single amino acid polymorphism in the ectodomain of gH that varied between different HCMV strains provided a structural explanation for the strict strain specificity of 2B10.

Even though the exact epitopes of these MAbs have not been finely mapped yet, indirect immunofluorescence analyses revealed that 3C11 required the entire ectodomain, while for recognition of 6E3, a truncated molecule encompassing aa 1 to 629 was sufficient, indicating that different target structures on gH were required for binding of these MAbs (Fig. 3D). The C terminus of gH has been previously identified as containing binding sites for murine (14-4b) and human (3G16) antibodies, and further mapping studies will determine whether 6E3 or 3C11 binds to similar regions on gH. The epitope bound by 2B10 was more precisely mapped. It involves M220/219 as the crucial anchor residue (Fig. 5, 7, and 8C). The residue is exposed in a loop (arbitrarily designated loop 1) that is located in the central part of the protein that contains AD169 gH aa 218 to 225 within structural domain III of gH (Fig. 6A and 8C) (46). In general, epitopes recognized by antibodies are 3-D structures involving more than a single amino acid, such that neighboring loops will contribute to the formation of the complete epitope. Alanine scans of loop residues that are located within close proximity in the 3-D structure of gH indicated that the neighboring loops 2 and 3 (D241 and D242 or Q331, M332, and L333) may also contribute to binding of MAb 2B10 (Fig. 6 and Fig. 8C). As the exchange of the original residues to alanine most probably resulted in only minimal destabilization of the respective loops, it seems highly likely that the residues identified within these loops contributed to binding of 2B10. Definitive evidence of this assumption will require a 3-D structure of 2B10 bound to gH.

FIG 8.

Antibody binding sites on gH, including the novel antigenic determinant. (A) Linear representation of Merlin gH. Structural domains are indicated by different colors and Roman numerals I to VI. TM, transmembrane domain. Arabic numerals indicate the positions of the bordering amino acids. Known binding sites of gH-specific antibodies isolated from mouse (AP86-SA4, 5C3, 10C10, and 14-4b, as well as 6E3, 3C11, and 2B10), rabbit (15.1, 58.8, 223.4, and 347.4), and human (13H11, 11B12, MSL-109, and 3G16) are indicated with their respective epitopes. (B) Ribbon model of gH (gray) complexed with gL (wheat) as deduced from Chandramouli et al. (46) (PDB ID 5VOB), with antibody binding sites from panel A depicted as spheres. (C) Detailed presentation of the 2B10 epitope, rotated 90 degrees around the vertical axis relative to its orientation in panel B. Amino acids crucial for 2B10 binding are highlighted in red, and Met in yellow.

Our findings raise the possibility that the 2B10 epitope defines a new antigenic domain on gH. As shown in Fig. 8, a number of binding sites of anti-gH antibodies isolated from mice (AP86-SA4, 14-4b, 5C3, 10C10, 6E3, 3C11, and 2B10), rabbits (15.1, 58.8, 223.4, and 347.4), and humans (13H11, 11B12, MSL-109, and 3G16) (27, 28, 31, 33, 34, 57–60) have been mapped. Only MAbs 13H11 and 11B12 bind to epitopes in proximity to the 2B10 binding domain, and for 13H11, aa 238 to 247 (loop 2) have been described as a binding site (57). There are, however, several crucial differences in binding requirements between 2B10 and 13H11/11B12. First, Macagno and colleagues showed that 13H11 binds to gH when expressed alone, which, importantly, is in clear contrast to 2B10, which requires coexpression of gB and gL for optimal reactivity (Fig. 3E) (28). Second, 13H11 binds to recombinant Merlin gH, as demonstrated by native gel shift, enzyme-linked immunosorbent assay (ELISA), and hydrogen deuterium exchange mass spectrometry (HDX-MS) (57), which is in contrast to MAb 2B10, which failed to interact with Merlin gH in indirect immunofluorescence analyses (Fig. 4). Third, MAb 2B10 does not bind to Merlin gH incorporated into the viral particle, as it fails to neutralize the Merlin strain (Fig. 2E). Thus, our findings strongly argue that MAb 2B10 identifies a novel antigenic determinant on gH of HCMV.

In infected cells, gH forms different complexes, of which the disulfide-linked gH/gL complex represents the core complex (61, 62). This core protein complex is essential for proper folding, intracellular transport, and incorporation of trimeric (gH/gL/gO) and pentameric (gH/gL/UL128/UL130/UL131A) complexes into the viral envelope (63, 64). The fact that 6E3 and 3C11 bind to gH when expressed without complex partners indicates that some folding of neutralization-relevant epitopes can take place in the absence of gL, similar to the reactivity described for MAb 14-4b (45). On the other hand, formation of the 2B10 epitope required not only coexpression of gL but also of gB, indicating the formation of additional neutralization-relevant epitopes after the assembly of higher-order gH complexes or, perhaps, even after the interaction of gH/gL with gB. In fact, gB/gH/gL-containing complexes have been described in HCMV-infected cells and on viral particles, where up to 50% of gH/gL is complexed with gB (65). Interestingly, the binding sites for MAbs 2B10 and human antibody 13H11 are located in proximity to a protein domain in gH that is homologous to the domain of herpes simplex virus (HSV) gH that has been proposed to mediate gB binding of HSV-2 gH (57, 66). The gH-specific MAb LP11 has been shown to be directed against an equivalent site within HSV-2 gH and likely interferes with the interaction of gH with gB, as LP11 efficiently blocked syncytium formation as well as cell-cell spread of HSV-2 (66). However, in contrast to LP11, which interferes with the gH/gB interaction, 2B10 appears to bind to gH only when both gB and gL are present.

The target region of 2B10 most probably encompasses three loops within the protein domain comprising residues 218 to 335 of gH that are well conserved in various HCMV strains (Fig. 8 and Fig. S2). As illustrated by the multiple sequence alignment of gH, loop 2 (aa 239 to 248) and loop 3 (aa 329 to 335) are 100% conserved throughout all strains, whereas loop 1 (aa 218 to 225) has only two predominant polymorphisms, with either an isoleucine or methionine at position 220/219 (Fig. S2). About 40% of the available gH sequences are similar to AD169 or TB40/E and encode a methionine at position 220/219, while the remaining 60% contain an isoleucine at position 220/219, as is present in Towne or Merlin. Based on the results of the neutralization assays, we are able to predict that strains containing M220/219 will be neutralized by 2B10 (Fig. S2, boxed in green), while those containing an isoleucine at this position will not (Fig. S2, boxed in red). The protein sequence alignment further revealed that M220/219I is not exclusively restricted to laboratory-adapted strains like AD169 or TB40/E but is also present in clinical HCMV strains isolated from various body fluids (e.g., blood, bronchoalveolar lavage fluid, amniotic fluid, urine, or cervical secretions) that were collected in different parts of the world. From these results, one can argue that (i) gH-M220/219I is a naturally occurring, globally distributed gH polymorphism that (ii) has no obvious tropism for specific cell types and (iii) likely is not associated with any specific HCMV-related clinical phenotype.

Polymorphisms among glycoproteins of HCMV isolates are a well described phenomenon (67). Strain-specific epitopes, e.g., the gH epitope bound by MAb SA4, have been shown to be invaluable for the serological diagnosis of reinfections in mothers with preconceptional immunity or in renal transplant recipients (35–37). In a preclinical mouse model, it was demonstrated that CMV reactivation after bone marrow transplantation could be prevented by therapy with a serum that matched the infecting strain (68). Whether the strain-specific epitope defined in this study could represent an additional polymorphism-associated marker for the immune response during natural infection remains an unanswered question. Our initial attempts to produce the 2B10 binding domain in isolation have been unsuccessful, and the fact that gB is required for proper expression of the epitope indicates that this will likely be a difficult task. However, the fact that neighboring residues are targets for human MAbs like 13H11 and 11B12 indicates that this protein region is immunogenic during infection (28) and, thus, may be relevant to the overall neutralizing antibody response against gH. Thus, the polymorphism defined by the 2B10 epitope may be clinically important, since it could represent another component of the immune evasion strategy of HCMV that contributes to the phenomenon of superinfections (67, 91, 92).

Nevertheless, the polymorphism M220/219I could serve as a novel marker for nucleic acid-based diagnostics. There is an increasing interest in establishing associations between the gH genotype and the pathogenicity of the corresponding herpesvirus strain. In fact, recent publications suggest that gH genotypes are irrelevant for congenital CMV infection but that the gH1 genotype is associated with neurological dysfunctions in congenitally infected children (69–71). Similarly, in transplant recipients undergoing immunosuppressive therapy, diagnosis of more than one gH genotype is a risk factor for worse clinical outcomes (72). Of note, all of these studies rely on a 30-year-old publication that grouped gH into two predominant genotypes: gH1 and gH2 (73). Based on this study, several PCR-based methods were developed that utilize primers and/or TaqMan probes that bind to the heterogeneous N terminus of the corresponding gH genotype (69, 71, 72, 74–77). The newly identified strain-specific epitope on gH could likewise serve for sequence-based PCR diagnostics. According to protein alignments (Fig. S1 and S2), strains like TR, Davis, or BE/21/2010 that share the SA4 epitope with AD169-like strains could be discriminated from them via their M220/219I polymorphism. Thus, M220/219I might be an additional diagnostic marker to distinguish gH genotypes.

In conclusion, this study expands the list of potent neutralizing anti-gH antibodies. MAbs 6E3 and 3C11 bound to the C terminus of gH and neutralized all tested HCMV strains, while 2B10 was strictly strain-specific. The binding site of 2B10 was mapped to a novel conformational epitope on gH that requires coexpression with both gB and gL for its reactivity, an aspect that in the future may allow different conformations of gH complexes to be distinguished.

MATERIALS AND METHODS

Immunization of mice and monoclonal antibody preparation.

BALB/c mice were obtained from in-house breeding based on mice from Charles River Laboratories, maintained under specific-pathogen-free conditions, and used between 10 and 24 weeks of age. Immortalized antibody-producing B-cell lines were generated from the spleens of immunized mice by conventional hybridoma technology as described before (78). Briefly, BALB/c mice were immunized by intravenous injection of 200 μl DPBS (Pan Biotech, Aidenbach, Germany) containing 10 μg of AD169 particles purified using sucrose gradient ultracentrifugation. Mice were boosted twice with 5 μg HCMV particles at days 35 and 52 after primary immunization before a final boost with 10 μg HCMV particles at day 67. Five days later, mice were sacrificed, and spleens were removed and dissected. Splenic cells were separated by using a 70-μm cell strainer and fused with SP2/0 cells. Hybridoma cells were seeded in 96-well F-bottom cell culture microplates containing 200 μl R10-HAT selection medium (RPMI 1640 containing 10% fetal calf serum and hypoxanthine, aminopterin, and thymidine [HAT]; 50× HAT was purchased from Sigma-Aldrich) supplemented with 2% culture supernatant from an interleukin-6 (IL-6)-producing transfectant. About 8 to 10 days later, hybridoma supernatants were tested for their neutralization capacity. Clones of interest were subcloned using a Beckman Coulter MoFlo high-speed cell sorter. Following additional rounds of subcloning, MAbs were purified from the hybridoma supernatants by in-house protein A chromatography and subsequently used in neutralization and immunofluorescence assays.

Ethics statement.

All experiments were conducted in accordance with institutional guidelines for animal care and use. The experiments were approved by the Regierung von Mittelfranken (Government of Middle Franconia) under approval 55.2-2532.2-3/08 and adhered to EC Council Directive 2010/63/EU (79).

Plasmids and bacterial artificial chromosomes (BACmids).

Oligonucleotide primers used for this study were purchased from biomers.net GmbH (Ulm, Germany), and their sequences are listed in Table 1.

TABLE 1.

Primers used for PCR cloning, BACmid recombination, and DNA sequencing

Oligonucleotide no.	Name	Sequence
0-2	5Bam-UL115/164575	AAGCGGATCCCCAACTGGCTCCTTACC
0-3	3Hind3-UL115/163697	CCCGAAGCTTGCGAGCATCCACTGCTT
0-11	5Xba1-gH	GCATTCTAGAATGCGGCCCGGCCTCCCCCCC
0-15	3Sal1-STOP-gH	GCATGTCGACTCAGCATGTCTTGAGCATGCGGTAGAGCAG
0-91	3Hind3-Merlin gHaa295	GCATAAGCTTGRGTCTTTGAGATARGAGTGACGG
0-92	5Hind3-ADgHaa297	GCATAAGCTTTCTYGACGCCGCACTYGACTTC
0-95	3Hind3-Merlin gHaa457	GCATAAGCTTAGAGGCCAGGTGCGTTTTGTGTARTTKTAGG
0-96	5Hind3-ADgHaa460	GCATAAGCTTTCAGCCTTCGCRCGYCAAGAACTCTACC
0-99	3Hind3-Merlin gHaa190	GCATAAGCTTGGYCGRTGTAGTCCYGAGGTGG
0-100	5Hind3-ADgHaa191	GCATAAGCTTTAACCAGACCTGTATCCTCTTTGATGG
1-23	5EcoR1-HindIII-gHaa1	GCATGAATTCAAGCTTATGCGGCCMGGCCTCCCCYCCTACC
1-24	3Xho1-Stop-NotI-gHaa718	GCATCTCGAGTCAGCGGCCGCACGACTGTCGGTRGCGTCCAC
1-31	3Xho1-Stop-Not1gHaa295	GCATCTCGAGTCAGCGGCCGCGTCTTTGAGATARGAGTGACG
1-73	3Xho1-Stop-Not1gHaa434	GCATCTCGAGTCAGCGGCCGCATGTTGCTGATTCTGTTTAG
1-74	3Xho1-Stop-Not1gHaa495	GCATCTCGAGTCAGCGGCCGCGAGGCCCGTTTCTACGATGAAG
1-75	3Xho1-Stop-Not1gHaa629	GCATCTCGAGTCAGCGGCCGCGCGCGTYAGTTCGCATTTAGTTTG
1-42	5upstrUL75To	TCGAATCAGCGTCGTCCCCAC
1-43	3downstrUL75To	GCCCGATATGTAACCAGACCC
1-44	5Towne-I219M	CAGCACCGTCACACCTTGTTTGCACCAAGGCTTTTACCTCATGGACGAACTACGTTACGTTAAAGGATGACGACGATAAGTAGGG
1-45	3Towne-I219M	GAAGTCCTCGGTCAGTGTTATTTTAACGTAACGTAGTTCGTCCATGAGGTAAAAGCCTTGGTGCAACCAATTAACCAATTCTGATTAG
1-46	5Towne gH short	CAGCACCGTCACACCTTGTTTG
7-08	UniversalKANArev	CAACCAATTAACCAATTCTGA
6-36	T7	TAATACGACTCACTATAGGG
18-44	Sp6/BGHrev	GAGGGGCAAACAACAGATGGC

(i) Eukaryotic expression constructs.

The plasmid expressing AD169 gB (pc58) was described before (80). Vector pcUL115-myc was generated via PCR amplification of the ORF encoding gL from the HB5 BACmid using primers 0-2 and 0-3 (Table 1), followed by ligation into a pcDNA3.1(−)-based vector coding for a C-terminal myc/His tag. AD169 gH (pMN156), Merlin gH (pMN157), and Towne gH (pMN158) were generated via PCR of DNA isolated from AD169-, Merlin-, or Towne-infected HFF cells by using oligonucleotides 0-11 and 0-15 (Table 1). The purified PCR products were digested with XbaI and SalI and ligated into pcDNA3.1(+) digested with NheI and XhoI. The ectodomain of AD169 gH, as well as C-terminally truncated versions thereof, were generated via PCR using pMN156 as a template together with a pair of corresponding primers, 1-23 and 1-75 (Table 1). PCR products were digested with HindIII and NotI and subsequently ligated into a pcDNA3.1(+)-based vector that yielded C-terminally His₆-tagged proteins.

Chimeras composed from varying proportions of AD169 gH and Merlin gH were generated via PCR using pMN156 or pMN157 as templates and primer 0-11 or 0-15 in combination with either of the primers 0-91 and 0-100 (Table 1). The resulting gH chimera-encoding plasmids are AD1-190-Merlin191-742 (pMN166), AD1-296-Merlin297-742 (pMN159), AD1-458-Merlin459-742 (pMN161), Merlin1-189-AD192-743 (pMN167), Merlin1-295-AD298-743 (pMN162), and Merlin1-457-AD460-743 (pMN164).

Site-directed mutagenesis was performed using primer pairs listed in Table 2 to generate AD169-gH-M220I (pMN168), AD169-gH-A285D (pMN169), AD169-gH-218YL/AA219 (pMN174), AD169-gH-221DE/AA222 (pMN170), AD169-gH-223LRY/AAA225 (pMN175), AD169-gH-242DD/AA243 (pMN171), AD169-gH-332QML/AAA334 (pMN176), AD169-gH-386SQ/AA387 (pMN172), AD169-gH-388TP/AA389 (pMN177), AD169-gH-390PR/AA391 (pMN178), and Merlin-gH-I219M (pMN173). The integrity of all newly generated plasmids was confirmed by automated DNA sequence analysis (Macrogen, Amsterdam, Netherlands).

TABLE 2.

Primers used for site-directed mutagenesis

Oligonucleotide no.	Name	Sequence
1-01	cADgH-LMD/LID	CCAGGGCTTTTACCTCATCGACGAACTACGTTACG
1-02	ncADgH-LMD/LID	CGTAACGTAGTTCGTCGATGAGGTAAAAGCCCTGG
1-03	cADgH-KAQ/KDQ	GGTACTAGTTAAGAAAGATCAACTAAACCGTCACTCC
1-04	ncADgH-KAQ/KDQ	GGAGTGACGGTTTAGTTGATCTTTCTTAACTAGTACC
1-05	cADgH-221DE/AA222	GGGCTTTTACCTCATGGCCGCACTACGTTACGTTAAAATCAC
1-06	ncADgH-221DE/AA222	GTGATTTTAACGTAACGTAGTGCGGCCATGAGGTAAAAGCCC
1-07	cADgH-242DD/AA243	CGTAGTTACGGTATCTATAGCCGCCGACACACCCATGCTGC
1-08	ncADgH-242DD/AA243	GCAGCATGGGTGTGTCGGCGGCTATAGATACCGTAACTACG
1-09	cADgH-386SQ/AA387	GAATTTATGATCACCTGCCTCGCAGCAACACCACCACGCACC
1-10	ncADgH-386SQ/AA387	GGTGCGTGGTGGTGTTGCTGCGAGGCAGGTGATCATAAATTC
1-11	cMerlin gH-I219M	GCACCAAGGCTTTTACCTCATGGACGAACTACGTTACG
1-12	ncMerlin gH-I219M	CGTAACGTAGTTCGTCCATGAGGTAAAAGCCTTGGTGC
1-13	cADgH-218YL/AA219	CTGTCTGCACCAGGGCTTTGCCGCCATGGACGAACTACGTTACG
1-14	ncADgH-218YL/AA219	CGTAACGTAGTTCGTCCATGGCGGCAAAGCCCTGGTGCAGACAG
1-15	cADgH-223LRY/AAA225	GCTTTTACCTCATGGACGAAGCAGCCGCCGTTAAAATCACACTGACC
1-16	ncADgH-223LRY/AAA225	GGTCAGTGTGATTTTAACGGCGGCTGCTTCGTCCATGAGGTAAAAGC
1-17	cADgH-332QML/AAA334	CTCAAAAGCGGTCGATGTGCAGCGGCGGACCGCCGCACGGTAG
1-18	ncADgH-332QML/AAA334	CTACCGTGCGGCGGTCCGCCGCTGCACATCGACCGCTTTTGAG
1-19	cADgH-388TP/AA389	CACCTGCCTCTCACAAGCAGCACCACGCACCACATTGC
1-20	ncADgH-388TP/AA389	GCAATGTGGTGCGTGGTGCTGCTTGTGAGAGGCAGGTG
1-21	cADgH-390PR/AA391	GCCTCTCACAAACACCAGCAGCCACCACATTGCTGCTATATC
1-22	ncADgH-390PR/AA391	GATATAGCAGCAATGTGGTGGCTGCTGGTGTTTGTGAGAGGC

(ii) BACmid recombination.

BACmid Towne-GFP was generously provided by H. Zhu and colleagues (49). Linear recombination cassettes were utilized to mutate this BACmid by a two-step recombination strategy (81). For this, the kanamycin selection marker gene (aphaI) was amplified by PCR using pEPkan-S (kindly provided by C. Sinzger, Ulm, Germany) as a template and oligonucleotides 1-44 and 7-08, followed by a second PCR with oligonucleotides 1-45 and 1-46 (Table 1). The linear recombination fragment was purified and subsequently transformed into chemically competent E. coli strain GS1783 (gift of C. Sinzger, Ulm, Germany) before homologous recombination was induced. In a second recombination step, the kanamycin cassette was removed as verified by growth selection. Finally, PCR analyses, restriction fragment length polymorphism analysis, and nucleotide sequencing confirmed that the newly generated constructs contained the desired gH amino acid exchange I219M.

Cells and virus strains.

Human embryonic kidney cells (HEK293T), primary human foreskin fibroblasts (HFF), and human fetal lung fibroblasts (MRC-5) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% fetal calf serum (FCS) (Sigma-Aldrich, St. Louis, MO, USA), glutamine (100 μg/ml), and gentamicin (350 μg/ml).

The virus strains used in this study encompassed PAN23, as well as the luciferase-expressing HCMV strains AD169 (HB15luc) and TB40/E (TB40luc) and the recombinant derivative TB40luc-GT4-2 (43, 44, 82, 83). HCMV strains Davis, Merlin, Toledo, Towne, and TR were generously provided by our colleagues and are described elsewhere (84–88). Viral titers were determined by titration of virus stocks in MRC-5 cells, either by measurement of luciferase activity or via an indirect immunofluorescence assay with MAb p63-27, which is directed against the HCMV immediate early 1 (IE1) protein (89).

Antibodies.

The following antibodies were used: human anti-gH antibody MSL-109 (27), human anti-gB MAb C23 (39), and rabbit polyclonal anti-myc antibody ab9106 (Abcam, Cambridge, UK). The Alexa Fluor 488-, 555- and 647-conjugated secondary antibodies for indirect immunofluorescence experiments were purchased from Molecular Probes (Karlsruhe, Germany).

Indirect immunofluorescence analysis.

Indirect immunofluorescence analyses were performed as described before (90). Briefly, HEK293T cells grown on coverslips were transfected via calcium phosphate precipitation. At approximately 36 h posttransfection, the cells were washed three times with phosphate-buffered saline (PBS), followed by fixation with 3% paraformaldehyde for 10 min at room temperature. Then, the cells were permeabilized with PBS–0.2% Triton X-100 on ice for 20 min and blocked in PBS–1% bovine serum albumin (BSA) at room temperature for 30 min. After this, the cells were incubated with the respective primary or secondary antibodies for 30 min at 37°C. Finally, the cells were mounted by using Vectashield mounting medium plus DAPI (4′,6-diamidino-2-phenylindole) (Vector Laboratories, Burlingame, CA). The samples were examined by using a Leica TCS SP5 confocal microscope with 488-nm, 543-nm, or 633-nm laser lines, scanning each channel separately under image capture conditions that eliminated channel overlap. The images were then exported, processed with Adobe Photoshop CS5, and assembled by using CorelDraw X6.

Virus neutralization assay.

HFF or MRC-5 cells were seeded in 96-well plates (1 × 10⁴ cells/well). Comparable infectious units of luciferase-expressing HB15luc and TB40luc, Towne wild type (wt) and recombinant Towne-GFP, TB40luc-gO-GT4 or PAN23, Merlin, TR, Davis, and Toledo strains were preincubated with serial log₂ dilutions of anti-gB MAb C23 or the different anti-gH MAbs (2B10, 6E3, and 3C11) for 1 h at 37°C before the mixtures were added to the cells for 4 h. The inoculum was replaced with fresh medium, and the cells were incubated at 37°C for 48 h. Cells infected with the luciferase-expressing viruses HB15luc and TB40luc were lysed using 100 μl Glo lysis buffer (Promega) per well. Thirty microliters of each cell lysate was transferred to white 96-well LIA plates (Costar), and 50 μl assay buffer (15 mM KH₂PO₄, 25 mM glycylglycine, 15 mM MgSO₄, 4 mM EGTA, 5 mM ATP, 1 mM dithiothreitol [DTT]) was added to each well. Luciferase activity was measured by injection of 50 μl d-luciferin (P.J.K. GmbH, Germany) solution (25 mM glycylglycine, 15 mM MgSO₄, 4 mM EGTA, 2 mM DTT, and 0.05 mM d-luciferin) per well, and detection of chemiluminescence was performed by use of an Orion microplate luminometer (Berthold Technologies). Towne-GFP-infected cells were imaged with a CTL ImmunoSpot S6 analyzer (Cellular Technology Limited, Bonn, Germany), and GFP-positive cells were quantified using the ImmunoSpot version 6.0.0.2 software (Cellular Technology Limited, Bonn, Germany). Cells infected with strain Davis, Merlin, Toledo, Towne wild type, TR, or PAN23 were analyzed by indirect immunofluorescence analysis using MAb p63-27 directed against the HCMV immediate early 1 (IE1) protein as described elsewhere (89). Data derived from these assays were plotted as the percentage of neutralization versus the neutralization in control wells, to which no antibody was added, and the 50% inhibitory concentrations (IC₅₀s) were determined. Neutralization assays were repeated at least two times, and one representative result of technical side-by-side duplicates is shown.