Biopharmaceutical proteins are often formulated and freeze dried in agents that protect them from deleterious reactions that can compromise activity and authenticity. Although such approaches are widely used, a detailed understanding of the molecular mechanisms of protein stabilization in low water content amorphous glasses is lacking. Further, whilst deterioration. chemistries are well described in dilute solution, relatively little is known about the extent and mechanisms by which protein integrity is compromised in the glassy state. Here we have investigated the relationship between protein modification and rate thereof, with variation of pH, carbohydrate excipient, temperature and the glass transition temperature using a model protein, lysozyme. Mass spectrometry analysis and peptide mapping confirm that protein modifications do occur in the glassy state in a time-, temperature-, and carbohydrate excipient-dependent manner. There were clear trends between the buffer pH and the primary modification detected (glycation). Most importantly, there were differences in the apparent reactivities of the lysine residues in the glass compared with those previously determined in solution, and therefore, the well-characterized solution reactivity of this reaction cannot be used to predict likely sites of modification in the glassy state. These findings have implications for (i) the selection and combinations of formulation components, particularly with regard to glycation in the glassy state, and (ii) the design of procedures and methodologies for the improvement of protein stability in the glassy state. (c) 2009 American Institute of Chemical Engineers Biotechnol. Prog., 25: 1217-1227, 2009