Fig 1.
Expression, purification and characterization of recombinant KSHV gH/gL.
(A) Plasmids encoding the gH ectodomain and gL were co-transfected into 293 cells and recombinant gH/gL was purified from the supernatant. Made with BioRender.com. McGuire, A. (2025). https://BioRender.com/ks0xtn4. (B) Recombinant gH/gL was subjected to size exclusion chromatography on a BioRad ENrichSEC 650 10 x 300 column. Three fractions were collected as indicated. (C) An aliquot of the gH/gL preparation pre-SEC, as well as equal amounts of the three fractions collected from B were analyzed by reducing SDS-PAGE followed by Coomassie staining. (D) Coomassie stained reducing SDS-PAGE gel of recombinant gH/gL treated, or untreated with PNGaseF as indicated. (E) Binding of recombinant gH/gL or recombinant Epstein-Barr virus gp350 to recombinant ephrin receptor A2 (EphA2) was measured by biolayer interferometry.
Fig 2.
gH/gL binding and neutralizing activity in a prospective KSHV-infection cohort.
(A) Plasma collected from a prospective cohort of mothers and their children was evaluated for binding to gH/gL by ELISA. Reciprocal endpoint titers are shown, each dot represents an independent measurement done in duplicate and the bars represent the means. The dashed line indicates the lowest serum dilution tested. Serum from healthy Seattle-area adults that are presumed to be KSHV-negative are included as controls (gray dots). (B) Serum samples from A were evaluated for their ability to neutralize KSHV infection of Vero cells at a 1:10 dilution. Dashed and dotted lines indicate 50% and 80% neutralization respectively. Samples denoted with (+) tested positive for KSHV DNA in contemporaneously collected oral swabs or blood by PCR. “M” indicates that the sample was collected from a mother and C1, C2, and C3 indicate that the sample was collected from a child. The numbers indicate the family household. (C) Correlation analysis of neutralizing activity vs gH/gL binding titer. The mean neutralizing activity and binding titers were plotted on the Y and X axis and analyzed by linear regression.
Fig 3.
Isolation of gH/gL-specific mAbs.
(A) Schematic of memory B cell staining with phycoerythrin (PE) and allophycocyanin (APC) conjugated gH/gL tetramers. Created with BioRender.com. McGuire, A. (2025) https://BioRender.com/bh8ls7o. (B) Staining of memory B cells (live, CD3-, CD14-, CD19+, IgD- IgG+) from a KSHV gH/gL seropositive donor, U35-022-MO (from Fig 2), with APC- and PE-labeled gH/gL. (C) gH/gL staining of memory B cells from a presumed KSHV negative donor (Presumed Negative A, from Fig 2). (D) Biolayer interferometry traces of gH/gL mAbs isolated from single-cell sorted gH/gL-labeled B cells binding to a 500 nM solution of KSHV gH/gL. (E) Biolayer interferometry traces of gH/gL mAbs in D binding to a 500 nM solution of EBV gH/gL. The anti-EBV gH/gL mAb AMMO1 was used as a control, and the dashed line demarcates the association and dissociation steps in D and E. (F) Affinity of the indicated mAbs to serially diluted KSHV gH/gL were measured by biolayer interferometry. (G) The number of amino acid mutations found in the antibody variable heavy (VH) and variable light chain (VL) chain genes of the indicated mAbs.
Fig 4.
Epitope binning of gH/gL-specific mAbs.
Heat map depicting the competition of KSHV mAbs binding to gH/gL as determined by BLI. gHgL was biotinylated and immobilized on a streptavidin biosensor and then immersed in buffer containing the indicated mAb until saturable binding was achieved (y-axis), the biosensor was then immersed in buffer containing the second indicated mAb (x-axis). The percent of binding inhibition is shown as a percentage of residual binding in the presence of the second mAb relative to the binding of the first mAb to gH/gL alone (scale at right). Colored outlines represent epitope clusters as determined by competition BLIs.
Fig 5.
Neutralization by gH/gL-specific mAbs.
(A) Recombinant mAbs were evaluated for their ability to neutralize KSHV infection of Vero cells at a single concentration of 100 µg/mL cells as indicated. Coloring of mAbs represent epitope cluster determined in Fig 4. (B-H) mAbs with neutralizing activity from A were serially diluted and tested for neutralization against KSHV infection of Vero cells as indicated. Each data point represents a technical replicate from one representative experiment. (I) The half-maximal inhibitory concentration (IC50) of the indicated mAbs was calculated from the curves in B-H. Each dot represents the IC50 calculated from an independent experiment carried out in duplicate, the bar represents the mean. (J) The indicated recombinant mAbs were evaluated for their ability to inhibit syncytia formation in a virus-free fusion assay as indicated. Each dot represents the luciferase activity in the presence of each mAb relative to the fusion measured in the absence of mAb. Each dot represents the mean of an independent experiment measured in triplicate. Bars represent the mean and error bars represent standard deviation. (K) The indicated recombinant mAbs were evaluated for their ability to neutralize KSHV infection of RPMI8226 cells as indicated.
Fig 6.
Potent neutralizing mAbs target the EphA2 binding site.
(A) gH/gL was incubated with, or without the indicated mAbs and the binding to immobilized EphA2 was measured by biolayer interferometry. The dashed line represents 100% binding set to the gH/gL alone control. (B-C) MLKH1/gH/gL (B) and MLKH5/gH/gL (C) complexes were purified and visualized by negative stain EM (nsEM). A representative 2D class average (top) and 3D reconstruction (bottom) are shown for each complex. The ribbon structure of gH/gL (PDB entry 7CZF, gH in black, gL in yellow) with each Fab (modeled using Alphafold) was fitted into the nsEM maps. (D) The crystal structure of the EphA2/gH/gL complex (PDB entry 7CZE) is shown for comparison.
Fig 7.
Neutralizing anti-KSHV gH/gL mAbs target multiple epitopes on gH/gL.
(A-C) gH/gL was incubated with MLKH5, MLK10 and MLK3 Fabs (A), MLKH5, MLKH10 and MLKH6 (B) or MLKH5, MLKH10 and MLKH12 (C), and complexes were visualized by nsEM. A representative three-dimensional reconstruction is shown for each complex. A comparison of the different complex is also shown (D). The ribbon structure of gH/gL (PDB entry 7CZF, gH in black, gL in yellow) along with the different Fabs models predicted with Alphafold (MLKH3 in teal, MLKH5 in blue, MLKH6 in red, and MLKH10 in orange and MLKH12 in purple) were fitted into the nsEM maps.
Fig 8.
Cryo-EM structure of MLKH3 and MLKH10 Fab in complex with KSHV gHgL.
(A) Left, cryo-EM map is shown with MLKH10 Fab (VH shown in orange, VL in salmon), MLKH3 Fab (VH shown in teal, VL in light teal), and gHgL in shades of gray and yellow, respectively. Right, 90° view rotation. MLKH5 Fab is not shown in the map. (B) Structure of the complex with the Fabs shown in ribbons and gHgL as a surface in the left panel; in the right panel, the same complex is shown rotated by 90°, (C–G) BSA plots for the residues of each protein involved in the complex. Hydrogen bonds are marked with an H on top of the bar, and salt bridges are marked with an S. (C) BSA plot for gH and gL residues involved with MLKH10 interaction; (D) BSA plot for gH residues involved with MLKH3 interaction; (E) BSA plot for the MLKH3 heavy chain residues only, as the MLKH3 light chain makes no contact with gH; (F) BSA plot for the residues of the MLKH10 heavy chain; (G) BSA plot for the residues of the MLKH10 light chain.