Amyloid fibrils represent a stable form of many misfolded proteins associated with numerous diseases. Among these are Parkinson's disease (alpha-synuclein), Type II diabetes (islet amyloid polypeptide), and Alzheimer's disease (amyloid beta-peptide, Abeta). The appearance of Abeta fibrils in neural tissue is a hallmark of Alzheimer's disease, and many studies have been conducted to determine and analyze the structure of these protein aggregates. The principal toxic species in Alzheimer's disease are believed to be soluble, oligomeric aggregates of Abeta, but numerous studies have found that the insoluble fibrillated form of the peptide also contributes to neurotoxicity. Thus, to design therapeutic agents to combat the progression of Alzheimer's disease, it is worthwhile to understand the thermodynamics of destabilizing these aggregates and the features that contribute to their stability. In this work, we present a systematic study of several factors that influence the stability of Abeta(42) fibrils following in silico mutation. We have employed standard molecular dynamics, as well as center-of-mass pulling and umbrella sampling, to study the thermodynamics of peptide dissociation from the core of a model protofibril at physiological temperature. Results indicate that a finite level of hydration around the Asp23-Lys28 salt bridge is crucial to protofibril stability, while mutation of Phe19 to glycine has no effect on the binding free energy of the terminal peptide. Packing between Ile32 and the aliphatic portion of the Lys28 side chain serves to regulate the level of hydration in the core of the protofibril and thus rigidify the Asp23-Lys28 salt bridge. These observations are important for designing compounds that target Abeta aggregates; interrupting these native interactions may destabilize these assemblies and ameliorate their toxicity.