The fragmentation reactions of a number of a(2) ions ([M-H-CO2](-)) derived from dipeptides have been studied by energy-resolved mass spectrometry, isotopic labeling, and MS3 experiments. The general reaction sequence H2N-CH(R-1)-C(=O)-N(R-2)-CHR3- --> -HN-CH(R-1)-C(=O)-N(R-2)-CH2R3. (1) -HN-CH(R-1)-C(=O)-N(R-2)-CH2R3 --> R3CH2-N(R-2)-C=O- + HN=CHR1 (2) R3CH2-N(R-2)-C=O- --> R3CH2-N(R-2)(-) + CO (3) leading eventually to a deprotonated amine, is shown to occur, a reaction sequence first proposed by Styles and O'Hair (Rapid Commun. Mass Spectrom. 1998, 12, 809) from a study of the a(2) ion derived from glycylglycine. When an amidic hydrogen (R-2 = H) is present, the initial proton-transfer reaction 1 is nonreversible. However, when there is no amidic hydrogen, as in the a(2) ions derived from H-Ala-Pro-OH or H-Gly-Sar-OH, the initial proton-transfer reaction 1 becomes reversible, leading to the interchange of N-bonded and C-bonded hydrogens. Ab initio calculations at the MP2/6=31+G(d) level of the energies and. interconversion pathways of anions derived by deprotonation of glycine N-methylamide show a barrier of 8 kcal mol(-1) for reaction 1, with reaction 2 being 23.8 kcal mol(-1) endothermic. When an amidic hydrogen (R2 = H) is present, the amine-deprotonated species formed in reaction 1 abstracts a proton from the amide nitrogen to form the amide-deprotonated species, the most stable species on the potential energy surface. The system effectively becomes trapped in this low-energy well and exits upon activation by reactions 2 And 3 as observed when glycine N-methylamide is deprotonated directly. When no amidic hydrogen is present; this low-energy state does not exist, and reaction 1 becomes reversible, leading to the interchange of N-bonded and C-bonded hydrogens. In these cases, a significant population of the original a(2) ion is formed, which fragments by the reaction H2N-CH(R-1)-C(=O)-N(R-2)-CHR3---> H2N-CH(R-1)-C=O- + R3CH=NR2 (4).