The design and development of novel drugs could benefit considerably from more reliable atomistic modeling schemes able to deal with the appropriate level of theoretical description on size and time scales relevant for complex life-science systems. This has not been possible, mainly due to limitations of the scaling and performance of quantum chemical algorithms on large biomolecules. Recently, we presented a novel approach that allows us to apply quantum simulations to the kinds of biological domains that are typical for classical Hamiltonians. For this we exploited a novel and efficient large-scale parallel sparse matrix-matrix multiplication algorithm. The scheme is sufficiently general to deal with any type of quantum scheme, and we opted our algorithm to an easily parametrizable semiempirical Neglect of Diatomic Differential Orbitals (NDDO) Hamiltonian. In this paper, we present a novel NDDO parameterization specifically tailored for treating intra- and inter-molecular interactions in proteins. We include a validation of static and dynamic properties for individual amino acids as well as for small polypeptides using both experimental and density functional theory benchmarks for reference. We also demonstrate the preeminence of this approach for describing bond-breaking reactions in terms of time-to-solution compared to traditional quantum simulations.