(743h) Abstract: Simulated Mutagenesis of VHH Domain In Explicit Solvent: Free-Energy Landscape and Folding Kinetics for the Mutated Hypervariable H1 Loop

In Camelids, a subset of antibodies is fully functional, naturally occurring, single-domain binding agents called VHH or “nanobodies” capable of an extensive antigen-binding repertoire. Conventional antibodies are typically made of two large heavy chains and two small light chains. The antigen binding domain is highly variable and consists of 3 hypervariable loops each in the heavy and light chains (VH and VL). The nanobody lacks the light chain and the region that joins the VH region to the VL region in a conventional antibody [1]. Nanobodies combine many desirable features including a small size, extensive antigen-binding repertoire, high stability at high temperatures, and good solubility; these have fueled a growing interest for their use in many potential applications, including therapeutic and clinical studies to target specific ligands, hidden antigen sites (inside crevices), or tumor cells in which diffusion limits the performance of conventional antibodies due to their large size [2].

This work is an extension of a previous atomistic Molecular Dynamics study [3] that simulated 1-point, 2-point and 3-point mutants of the anti-hCG llama VHH antibody to identify the key mutations that would “humanize” the structure of its hypervariable H1 loop into the type-1 H1 (such “humanized” structures are believed to make nanobodies more compatible in applications involving human therapeutics). In that study, an implicit solvent was used to model the water and a temperature replica exchange method (TREM) was used to accelerate the equilibration of the loops' conformations; it was found there that 2 mutants (having 3 point mutations relative to the wildtype) yielded a stable type-1 H1 loop. In the current work, we use atomistic Molecular Dynamic simulations to perform the following:

1)      Identifying equilibrium structures, via simulations of mutated hypervariable regions in explicit solvent using a generalized hamiltonian replica exchange method (HREM) and the TREM. In contrast to implicit solvent studies, TREM was found to be of limited use in enhancing conformational sampling over brute-force molecular dynamics and did not yield a stable “humanized” loop during our simulations. HREM [4] was applied by targeting different components of the Hamiltonian.

2)      Quantifying thermodynamic driving forces: The free-energy landscape of the system was investigated along various order-parameters critical to the transition through Umbrella Sampling techniques and Metadynamics.

3)      Characterizing kinetics: A forward-flux sampling method was used to monitor the mechanism of loop rearrangement from an initial metastable state to the final stable structure.

Our results regarding the relative stability of the type-1 H1 loop in the different VHH mutants are consistent with those from the solvent-implicit simulations, but they also reveal the existence of multiple metastable basins near the most stable state. We are also able to estimate the rate of selected transitions among conformers and to characterize the transition state (and the role of water molecules) associated with the burial of hydrophobic residues needed to attain the folded loop structure from unfolded states.

[1] Hamers-Casterman, C., Atarhouch, T., Muyldemans, S., Robinson, G., Hamers, C., Bajyana Songa, E., Bendahman, N, & Hamers, R., “Naturally occurring antibodies devoid of light chains”, Nature, 363, 446-448 (1993).

[2] Harmsen, M. M. and De Haard H. J., “Properties production, and applications of camelid single-domain antibody fragments”, App. Microbiol. Biotechnol., 77, 13-22 (2007).

[3] Velez-Vega, C., Fenwick, M. K., and Escobedo, F. A., “Simulated Mutagenesis of the hypervariable loops of llama VHH domain for the recovery of canonical conformations”, J. Phys. Chem., 113, 1785-95 (2009).

[4] Fenwick, M. K. and Escobedo, F. A., “Hybrid Monte Carlo with multidimensional replica exchanges: Conformational equilibria of the hypervariable regions of a Llama VHH antibody domain”, Biopolymers, 68, 160 (2001).