Thermography

Example 9

The TGA thermograms for the crystalline forms 1, 3, 5 and 8 were collected on TGA equipment (TA Instruments). The gases recovered during each run were analyzed by head space mass spectroscopy (Agilent GS system). The measurement allowed to register the temperature at which rapid evaporation of Cl ion (disproportionation) started.

The onset temperature of disproportionation for crystalline forms 1, 3, 5 and 8 was determined based on TGA measurements. The results are shown in Table 3. Crystalline form 8 showed highest thermal stability against disproportionation.

Example 3

We hypothesized that HR1_C is essential to EBOV GP metastability. Since HR1_C in wildtype EBOV GP is equivalent in length (8 aa) to a truncated HR1_N in the prefusion-optimized HIV-1 Env, metastability in EBOV GP may not be sensitive to the HR1_C length and likely requires a different solution. We thus hypothesized that a proline mutation in HR1_C, termed P^1-8, may rigidify the HR1_C bend and improve the EBOV GP trimer stability.

To examine this possibility, eight GPΔmuc-W615L variants, each bearing a proline mutation in HR1_C but without the L extension and foldon at the C terminus, were validated experimentally. All constructs were transiently expressed in 250-ml 293 F cells and purified using an mAb114 column, which captures all GP species. The proline mutation at most positions in HR1_C showed little effect on the composition of GP species except for T577P (P²) and L579P (P⁴), which displayed notable trimer peaks at ˜11 ml in the SEC profiles. In a separate experiment, all eight constructs were transiently expressed in 250-ml 293 F cells and purified using an mAb100 column. Only P² and P⁴ showed any measurable trimer yield, with a notably high SEC peak observed for P⁴ that corresponds to well-formed trimers. The mAb100-purified GP was also analyzed by BN-PAGE, which showed a trimer band for P² and P⁴. Overall, the T577P mutation, P², can substantially increase trimer yield, while the L579P mutation, P⁴, exhibited a less pronounced effect.

Next, the T577P mutation (P²) was incorporated into the GPΔmuc-WL²-foldon construct, resulting in a construct named GPΔmuc-WL²P²-foldon. This construct was expressed transiently in 1-liter 293 F cells and purified using an mAb100 column for SEC characterization on a HiLoad Superdex 200 16/600 GL column. In three production runs, GPΔmuc-WL²P²-foldon generated a trimer peak that was two- and four-fold higher than GPΔmuc-WL²-foldon and wildtype GPΔmuc-foldon, respectively, with an average yield of 2.6 mg after SEC. Protein collected in the SEC range of 55.5-62.0 ml was analyzed by BN-PAGE, which displayed a trimer band across all fractions without any hint of impurity. The thermostability of GPΔmuc-WL²P²-foldon was determined by DSC after mAb100 and SEC purification.

Unexpectedly, two transition peaks were observed in the thermogram, one registered at a lower Tm of 61.6° C. and the other at a higher Tm of 68.2° C. To this end, a second construct bearing the L579P mutation (P⁴), termed GPΔmuc-WL²P⁴-foldon, was also assessed by DSC. Although only one peak was observed in the thermogram with a Tm of 67.0° C., a slight widening at the onset of the peak suggested a similar unfolding behavior upon heating. DSC thus revealed the complexity associated with a proline-rigidified HR1_C bend, which may increase the trimer yield at the cost of reducing GP thermostability. The antigenicity of GPΔmuc-WL²P²-foldon was assessed using the same panel of 10 antibodies by ELISA (FIG. 3F-G) and bio-layer interferometry (BLI). The T577P mutation (P²) appeared to improve GP binding to most antibodies with respect to GPΔmuc-WL²-foldon (FIG. 3G), with a 40% reduction in EC₅₀ observed for bNAb BDBV223, which targets HR2-MPER. Although BLI profiles were almost indistinguishable between wildtype and redesigned GPΔmuc-foldon trimers—all with fast on-rates and flattened dissociation curves, a two-fold higher signal at the lowest concentration (12.5 nM) was observed for GPΔmuc-WL²P²-foldon binding to bNAb BDBV223, consistent with the ELISA data.

Our results thus demonstrated the importance of HR1_C to EBOV GP metastability and an unexpected sensitivity of HR1_C to proline mutation. Recently, Rutten et al. tested proline mutations in HR1_C along with a K588F mutation to stabilize filovirus GP trimers (Cell Rep. 30, 4540-50, 2020). While a similar pattern of increased trimer yield was observed for the T577P mutant, the reported thermostability data appeared to be inconsistent with our DSC measurement. Further investigation is warranted to fully understand the role of HR1_C in filovirus-cell fusion and its impact on GP stability.