When cystine is irradiated at pH 1 by 254‐nm u.v. the following yields are observed: 4 cystines → 5.2 cysteines + 2.8NH3. Thus, SH production accounts for only 0.65 of the cystine destruction; further C‐S breakage to give alanine or serine is not efficient. The yields for cystine and glutathione destruction are essentially the same at pH 1. However the presence of the glutamic and glycine residues stabilize the cystine in glutathione so that NH3 is not lost until the peptide bonds are hydrolyzed. Increasing the pH from 1 to 8.6 increases the yield of cystine destruction in glutathione by 50 per cent. The yield of cystine destruction is greater in both compounds when O2 is present during irradiation (e. g. the cysteic acid yield in glutathione is increased by 50 times). The overall production of SH varies by a factor of 2 in the four proteins‐insulin, RNase, trypsin and lysozyme. The present data further support the earlier observation that radiation damage is quite non‐random in RNase: at least two and perhaps three of the four constituent cystines must be disrupted before activity is lost: i.e. the most radiosensitive cystines are not critical for enzymic activity. Similarly, in both trypsin and lysozyme the integrity of the most radiosensitive cystines also does not appear to be critical for the retention of enzymic potential. In insulin, however, all three cystines appear to be crucial for activity and to have approximately equal radiosensitivities. These differences in sensitivity of cystines in different proteins must depend specifically upon energy transfer and/or chemical interactions between the chromophoric groups. If yields are calculated on the basis of those quanta absorbed only in the cystines, values about 5 to 8 times greater than those in the model compounds cystine and oxixized glutathione are obtained. The yields of cystine destruction are much higher in those protiens which contain trypotophan.

Two divergent theories of the mechanisms involved in U.V. inactivation of enzymes have been developed over a period of years. The one proposes that the random destruction of any amino acid residue causes inactivation. The second emphasizes the importance of the disruption of a cluster of specific cystines and hydrogen bonds responsible for the spatial integrity of the active center. Consistent with the latter postulates, previous studies in this series have shown that the number of cystines disrupted in trypsin is correlated with loss in enzymic activity and in ribonuclease the disruption of cystines is very non‐random. In the present experiments only 1–1.1 titrable (with pCMB) SH groups are formed for each cystine disrupted. The available evidence suggests that about 0.1 of the disrupting events produce two SH groups: 0.2 lead to the formation of one ‐SH and one ‐SOnH: and about 0.7 involve the cleavage of a C‐S bond with the appearance of one ‐SH plus an oxide (perhaps ‐C = O) of the C moiety. The disruption of cystine is such that neither serine, alanine, glycine nor H2S is formed in large amounts. The kinetics of SH production emphasize that the various cystines in RNase and trypsin have unequal radiation sensitivities; some must differ by at least a factor of 10. Tentatively it is concluded that u.v. absorbed directly by cystines (ca 20 per cent) causes random, symmetric breakage of S‐S, whereas quanta absorbed in aromatic residues (ca 80 per cent) lead to efficient energy transfer and localization in chromophores adjacent to critical cystines followed by addition of hydrogen to specific C‐S bonds resulting in their cleavage.

Abshct-Studies of purine absorption and emission in seven solvents differing greatly in dielectric constant and hydrogen bonding potential, reveal a variety of solvent effects. For example, the resolution of structure in the absorption spectrum, the position and/or intensity of the X, absorption band, the intensity of fluorescence, the magnitude of the long wavelenth tail, and the position of the XI absorption band are differentially affected-in the order listed-by the solvents tested. Even though it is possible to correlate the extent of decrease in the n-a* tail with increasing solvent dielectric constant, probably alterations in all of these spectroscopic parameters depend most critically upon the ability of the various solvents to form hydrogen bonds with the hydrogen on N9 andfor with the non-bonding electrons on the purine nitrogens: it is tentatively concluded that the probability of hydrogen bonding is directly correlated with the electronegativity of the aza nitrogens (N7 > N3 > Nl).In solvents like isopropanol not all of the non-bonding electrons must be solvated maximally in most purine molecules since there is appreciable fluorescence under conditions where a long wavelength tail is readily observed in the absorption spectrum (alternatively some noabonding electrons may not t e relevant to fluorescence quenching.) Decreases in fluorescence yield are associated with red shifts in the fluorescence maximum, and in the solbents of highest polarity the fluorescence yield is again small indicating that glycerol and water can enhance radiationless tunneling-presumably by altering Franck-Condon configurations and/or improving electronic-vibrational coupling betwen solute and solvent. The quantum yield is uniform throughout the atsorption band for a given solvent, but studies in aqueous buffers varying from pH 1 to 11 show that the fluorescence yield is greater for charged than for neutral molecules. Further, the fluorescence excitation ~eak is red shif.ed in powders. Since phosphorescence is the predominant emission at 77'K and increases in fluorescence can be correlated with the presumed solvation of non-bonding electrons, the singlet excited state of lowest energy in 'unperturbed' purine must be n-n* in nature. The shape of the phosphorescence band and the decay lifetime of -1 sec at 77'K lead to the conclusion that the emitting triplet is a P-H* state. The eight vibrational structures in phosphorescence emission can be readily grouped into two progressions: there is an average separation of about 1300 cm-* between peaks within a given progression, and the two sets are mutually displaced by about 500 cm-l. Individual vibrational peaks are favoured in different solvents and the whole band may be shifted up to 500 cm-l. Even larger shifts are observed in charged purine molecules and in powders (up to 3000 cm-l) and the presumed 0-0 band is not observed.

No abstract