338
The Journal of Physical Chemistry, Vol. 62, No. 3, 1978
M. D. Sevilla and J. B. D'Arcy
An Electron Spin Resonance Investigation of Charge Transfer in Aromatic Peptide 7r-Cation Radicals' Michael D. Sevilla" and James B. D'Arcy Department of Chemistry, Oakland University,Rochester, Michigan 48063 (Received June 24, 1977) Publicationcosts assisted by the U.S. Department of Energy and by the U.S. Army Natick Development Center
Radicals produced by the photoionization of aromatic dipeptides in neutral (8 M NaC104)and basic (8 M NaOD) glasses at 77 K were investigated by ESR spectroscopy. For L-phenylalanyl-L-alanine and L-phenylalanylglycine photoionization in 8 M NaC104glasses results in decarboxylated radicals of the form -CO-ND-CHR. In the cases of L-tyrosyl-L-alanine, L-tyrosylglycine, L-tryptophyl-L-alanine, and L-tryptophylglycine in 8 M NaC104, both decarboxylated radicals and radicals on the aromatic groups are observed. The aromatic group radicals are found to be the phenoxy1 radical in tyrosine containing peptides and the indole T cation in the tryptophan containing peptides. Photoionization of the aromatic peptides in 8 M NaOD results in only aromatic group radicals for phenylalanine and tryptophan containing peptides, whereas both decarboxylated and aromatic group radicals are noted in tyrosine containing peptides. For phenylalanylalanine a study of decarboxylated radical vs. light intensity shows that the radical is produced by a biphotonic process. A concentration study shows the transfer of charge from the photoionized aromatic group to the carboxyl group is intramolecular in nature. Results found for the individual aromatic amino acids and a tripeptide which show only the aromatic group radicals suggest that the transfer of charge is strongly conformationally dependent.
Introduction Techniques employed in this investigation were similar to those employed in our previous work with DNA Studies of the UV photolysis of aromatic amino acids bases15J6 and peptidesa8 Photoionization of aromatic and peptides at low temperatures in aqueous matrices have dipeptides a t concentrations as low as 0.1 mg/mL in 8 M shown that the aromatic ring is photoioni~ed.~-~ The NaC104 glasses at 77 K were performed employing 254-nm photoionization is primarily biphotonic and takes place UV light from a low pressure helical mercury vapor lamp. after absorption of a second photon by the molecule in its Photolysis times as short as a few seconds usually produced lowest excited triplet ~ t a t e . ~These - ~ previous studies of adequate ESR signal intensity. The photoejected electrons the species produced by photoionization have not idenare not stable in the NaC104 glass. Those that are trapped tified the radicals formed in the amino acids on peptides in the glass are photobleached with an incandescent lamp. due to matrix reactions and poor resolution. The C104- acts as electron scavenger and produces 0- as Other studies which have employed photosensitizers in reaction 1. The signal due to 0- can be subtracted from have shown that certain peptides will undergo decarb ~ x y l a t i o n . ~The - ~ mechanism of the decarboxylation is (20,- t e - (310,- t 0 (1) likely energy transfer or charge (hole) transfer. The the spectra leaving only that from the photoionized suggestion t h a t electron a t t a c h m e n t causes molecule, Computer techniques employed in this study decarboxylation5 has been questioned.8 are those detailed in previous investigations.16 Laser flash photolysis studies in aqueous solution of Fremy's salt (peroxylaminedisulfonate) was employed aromatic amino acids and peptides have shown that as a standard for g values and hyperfine splitting5 (AN = phenylalanineg is photoionized principally by biphotonic 13.1 G, g = 2.0056). mechanism whereas tryptophan undergoes photoionization in a single photon process.lOJ1 Tyrosine undergoes both Results and Discussion biphotonic and single photon processes depending on the state of protonation of the phenyl hydroxyl g r o ~ p . ~A ~ ~ ' ~ L-Phenylalanine (Phe) Containing Peptides. In Figure 1A we show the ESR spectrum immediately after phocomparison of the work in solid matrices a t low tempertolysis of 5 mM phe-ala a t 77 K in 8 M NaC104. The atures with solution studies shows that the probability for spectrum consists of a broad low field component and a single photon processes is increased at higher temperatures well-resolved five-line spectrum. The low field component in aqueous solutions. Such single photon ionization is due to 0- which is produced by electron reaction with processes of aromatic compounds in aqueous solutions have Clod- (reaction 1). The 0- spectrum free from other been recently detailed by Getoff.14 radicals is shown in Figure 1B. This species was produced In this work we have investigated the photoionization by the photoionization of K4Fe(CN)6in 8 M NaC1O4 After of a number of peptides containing aromatic amino acids double integration of the spectra and subtraction of the in aqueous glasses. We believe the results found are amount of 1B sufficient to remove the 0- signal from lA, evidence for electron transfer from the carboxyl group to we find the spectrum in Figure 1C. The subtraction the photoionized aromatic ring resulting in decarboxylation showed that 50% ( f 5 % ) of the signal in Figure 1A was reactions. due to 0-. Since the precursor of 0- is the electron we In this paper standard abbreviations are used for the consider this good evidence that the radical shown in amino acid residues in peptides. Figure 1C is a result of a photoionization process. Experimental Section In Figures 2A and 2B we show the ESR spectra found for gly-phe-ala and phe-gly, respectively, after subtraction The peptides used in this study were obtained from of the 0- signal. For gly-phe-ala about 50% of the original Sigma and were the highest grade available. The peptides signal was due to 0-. A greater fraction of 0- appeared employed contained L-amino acid residues. --f
0022-3654/78/2082-0338$01 .OO/O
0 1978 American Chemical Society
Charge Transfer in Aromatic Peptide a-Cation Radicals
The Journal of Physical Chemistry, Vol. 82, No. 3, 1978 339
of 0-. This spectrum is believed to be that of the a cation of the phenyl group@and it is not like that found in Figure 1C. Second, if the a cation were stable in these photoionized peptides the ESR spectra would not depend on the C-terminal residue as they do for phe-ala and phe-gly. Third, the spectrum in Figure 1C shows a 1:4:6:4:1pattern due to four approximately equivalent protons a t 20.8 G, g = 2.003. It clearly suggests a radical of the form RCH-CH3, not the 7 cation. There are two possible structures for this species, I (R = phenyl) or 11. 0 ii
'
ND,+CHCNDCHCH, i
CH,CHCO,'
R I
Figure 1. (A) The first derivative ESR spectrum of radical I (R = phenyl) and 0-produced by the UV photolysis of L-phenylalanyl-L-alanine at 77 K in 8 M NaC104/D20. The markers in the center are spaced 13.1 G apart. The center marker is at g = 2.0056. Spectrum was recorded at 115 K. (8) The ESR spectrum of 0-in 8 M NaC104/D20produced by the photolysis of K,Fe(CN),. (C) The spectrum produced by the subtraction of B from A. The subtraction shows that in A, 50% of the signal is due to 0-and 50% is due to radical I (R = phenyl).
I1
Radical I would result from decarboxylation and radical I1 from a secondary deamination reaction, both presumably after the initial ionization event. Radical I1 can be eliminated as a possibility from the results shown in Figure 2C. This spectrum arises from electron attachment to 2-bromopropionic acid in 8 M NaC104 (reaction 2). The BrCH(CH,)CO,- t e - + Br- t CH,CHCO,-
concentration of the solute was kept high enough so that reaction 2 could effectively compete with reaction 1. The spectrum found for this species extends 112 G, shows well resolved anisotropic components, has an average isotropic splitting of 24 G (4 H), has been fully interpreted in previous work in other matrices,lg and is clearly not that in Figure 1C. We conclude radical I is produced by the photolysis of phe-ala. The lack of resolution of the anisotropic components in Figure 1C is most likely due to broadening caused by the anisotropic nitrogen coupling from the nitrogen atom next to the radical site in radical I. The overall spread of 83 G (4 H at 20.8 G) is significantly less than that found for the spectrum of radical 11. This further distinguishes the two radicals and suggests that the delocalization of the unpaired spin is greater in radical I than 11. The nearly identical results found for the photolysis of gly-phe-ala (85 G width, 4 H at 21.2 G) to those found for phe-ala suggest the similar radical, I11 (R = phenyl). The 0
Figure 2. (A) The ESR spectrum of radical I11 produced by UV photolysis of glycyl-L-phenylalanyl-L-alanine in 8 M NaC104/D20at 77 K recorded at 115 K. Spectra B, C, and D were also recorded at 115 K. (B) The ESR spectrum of radical IV produced by the UV photolysis of Lphenylalanylglycine at 77 K in 8 M NaC104. (C) The ESR spectrum of radical 11, .CH(CH3)C02-, produced by electron attachment to BrCH(CH3)C02-in 8 M NaC104/D20at 77 K. (D) The ESR spectrum of the phenylalanyl a-cation radical produced by photolysis of L-phenylalanine in 8 M NaCI04/D20at 77 K. The signal due to 0-has been subtracted in A, B, and D.
to be found for samples of phe-gly; however, this was due to the fact that the radical produced by the photolysis of phe-gly power saturated much more readily than the radical from phe-ala or g1y~he-ala.l~ Even at the lowest power levels attainable with our present instrument (0.1 mW) saturation was evident. As was stated the presence of 0- in the spectra is evidence that the radicals produced are a result of a photoionization process. The obvious choice for the radical identity which gives rise to Figure 1C might then be the positive ion of the phenyl ring in phe-ala. However several pieces of experimental evidence show otherwise. First, the results of photoionization of L-phenylalanine in 8 M NaC104 at 77 K is shown in Figure 2D after subtraction
(2)
0
li
ii
ND, +CH,CNDCHCNDCHCH, I y-L
R I11
difference in structure between I and I11 would not be detected in glassy ESR spectra. The results found for phe-gly (Figure 2B) shows a triplet of 38.6 G width which is likely due to two interacting protons of 19.3-G splitting each. This suggests a radical of the form of radical IV (R = phenyl). 0 Ii
ND, *CHCNDCH, * I
CH, R IV
The other possible choice .CH2C02-was produced by electron attachment to chloroacetic acid in 8 M NaC104 and shows a very different and more well-defined anisotropic structure which is identical with that previously reported for CH2C02- in 8 M NaOD glasses.20 The 19-G
340
The Journal of Physical Chemistry, Vol. 82, No. 3, 1978
M. D. Sevilla and J. B. D’Arcy
k
V Flgure 3. The ESR spectra of radicals produced by the photolysis of tyrosyl and tryptophyi dipeptides in 8 M NaC104/D20at 77 K after subtraction of 0-.The spectra were recorded at 77 K. (A) ESR spectrum of 55% radical I (R = phenol) and 45% phenoxyl radical from photolysis of L-tyrosyl-L-phenylalanine. (8)ESR spectrum of 40 % radical I V (R = phenol) and 60% phenoxyi radical from photolysis of L-tyrosylglycine. (C) ESR spectrum of predominantly radical I (R = indole) from the photolysis of L-tryptophyi-L-alanine. (D) ESR spectrum of predominantly radical I V (R = indole) from the photolysis of L-tryptophyiglycine. A small amount of tryptophyl A cation is also present in C and D.
splitting due to IV is less than found for .CHzCO2-(22 G). This again suggests an increased delocalization for radicals of type I and IV. Tyrosine ( T y r ) and Tryptophan ( T r p ) Dipeptides. In Figure 3 we show the ESR spectra after photolysis at 77 K of M solutions of tyr-ala (Figure 3A), tyr-gly (Figure 3B), trp-ala (Figure 3C), and trp-gly (Figure 3D) in 8 M NaC104. The background signal due to 0- has been subtracted in these spectra. The alanine containing spectra (3A and 3C) are nearly identical with that found for phe-ala (85 G width). Again, about 50% of the signal was due to 0-. The tyr-gly and trp-gly spectra show the same outer features and the same power saturation behavior as was found for phe-gly. In the cases of tyr-gly, trp-gly, and to a lesser extent tyr-ala and trp-ala we see an increase in the intensity of the central component of the spectra with new hyperfine structure in the tyr containing peptides. The increase in the central component is attributed to radicals on the aromatic group. Evidence for this in tyr-gly and tyr-ala is presented in Figure 4A which shows the tyrosine phenoxyl radical a t 77 K produced by photoionization of L-tyrosine in 8 M NaC104. The 0- signal is subtracted out. The same spectrum as is Figure 4B is found from the photolysis of tyrosine in 8 M NaOD (Figure 4B). The P cation of tyrosine might be expected to be observed at 77 K in a neutral medium at low temperature;21 however it is clear that the photoionization of tyrosine in a basic medium will produce the phenoxyl r a d i ~ a l The . ~ ~similarity ~~ in spectra between Figures 4A and 4B thus suggests that the phenoxyl radical is present in the neutral glass. In room temperature aqueous solutions it has been shown by Dixon and Murphyz3that the substituted phenol cations and their corresponding phenoxyl radicals have very similar hyperfine splittings; however, the g values of the phenoxyl radicals are 2.0044 f 0.0004 while those of the phenol cations are 2.0032 f 0.0004. In addition the pK of the phenol radical cation was found to be -2.0. This pK value and the fact that the measured g value was 2.0045 for both spectra in Figure 4 supports our assignment of
Flgure 4. ESR spectra of the L-tyrosine phenoxyl radicals produced by UV photolysis of L-tyrosine at 77 K in (A) 8 M NaCi04/D20and (B) 8 M NaOD recorded at 77 K.
EXP.
/---
3 1 1 I,
Figure 5. Computer simulation (lower spectrum) of L-tyrosine phenoxyl radical in 8 M NaCIO,/D,O (upper spectrum). Parameters for the simulation are given in the text.
both spectra in this figure to phenoxyl radicals. In the lower spectrum in Figure 5 we show our simulation of the phenoxyl radical spectrum from L-tyrosine in 8 M NaC104. The simulation assumes three protons at 6 G, one proton at 11G, a linearly decreasing line width with field position (6.8-5.2 G), and gaussian lineshapes. The closeness of the fit suggests a correct interpretation. The splittings are those expected from a tyrosine phenoxyl radical with two ortho ring protons at 6 G, one methylene proton a t ca. 6 G, and the other at 11 G. These values are in reasonable agreement with those found for the L-tryrosine phenoxyl radical in aqueous solution if one allows for the fact that the methylene protons are averaged by rotation in solution.22 Computer simulations of the tyr-gly spectrum (Figure 3B) and the tyr-ala spectrum (Figure 3A) show that the phenoxyl radical accounts for 60% of the total intensity in Figure 3B and 45% of the intensity in Figure 3A. We conclude that the photolysis of tyr-ala and tyr-gly produce radicals I and IV (R = phenol), respectively, as well as varying amounts of phenoxyl radical. In the cases of trp-ala and trp-gly the small increase in the central components is attributed to the unreacted ~ , ~ ~ for this tryptophyl indole ring P c a t i o i ~ . ~Evidence comes from the fact that photoionization of tryptophan itself produces a blue colored sample with a singlet ESR spectrum. The blue color is noted in the trp-ala and trp-gly samples and has been associated with the T cation in previous s t ~ d i e s The . ~ ~“deprotonated ~~~ cation” or neutral species is found in our work in basic glasses to have pink (lilac) color.
The Journal of Physical Chemistry, Vol, 82, No. 3, 1978 341
Charge Transfer in Aromatic Peptide nCation Radicals
r
“
t
k!
d\ ‘c “
/
/
1204
801 40
-1
W
, 0 F ,
I@
,
0 . 2
,
,
0 . 4
I‘CARBITRARY
,
,
0 6
,
,
0 8
I-
1
0
UNITS:,
Figure 7, Plot of the relative peak height found for the photolysis of samples of L-phenylalanyl-L-alanine vs. the square of the UV light intensity. Figure 6. ESR spectra produced by the photolysis of tyrosyl containing dipeptides in basic 8 M NaC104/D20at 77 K. (A) spectrum of radical I (R = phenol) and the phenoxyl radical from tyrosylalanine. (B) Spectrum of the phenoxyl radical and a small amount of radical I V (R = phenol) from tyrosylglycine. The fraction of the phenoxyl radical is increased in the basic NaC104 glass over the neutral glass.
The presence of the phenoxyl radical in the spectra of tyr-gly and tyr-ala may be due to the deprotonation of the hydroxyl group of the T cation of the phenol ring. This would be expected to compete with transfer of charge to the carboxyl group. Since the deprotonated T cation would be expected to convert very rapidly to form the phenoxyl radical, the competition would be expected to shift toward phenoxyl radical production when the solution is made basic, (The pK of the OH group is 10 in tyrosine.) To test this hypothesis, experiments were performed in which M NaOD was added to samples of tyr-gly and tyr-ala. Figure 6 shows the results of these experiments. For tyr-ala (Figure 6A) the amount of phenoxyl radical relative to the decarboxylated species was only slightly increased over that found in the neutral medium. For tyr-gly (Figure 6B) the phenoxyl radical was greatly increased. Computer simulations show the phenoxyl radical accounts for 90% of the signal in the basic glass vs. 60% in the neutral solution. Other Peptides i n 8 M NaC104. Two tripeptides, other than gly-phe-gly (Figure 2A), were investigated. They are phe-gly-gly and tyr-gly-gly. For phe-gly-gly we found only the phe T cation. For tyr-gly-gly the signal was predominantly (ca. 80%) the phenoxyl radical. Another signal suggestive of a radical of the form -NDCH2. was also observed. This species could result from secondary deamination or decarboxylation. Several dipeptides with the aromatic group at the C terminal were also investigated (gly-phe, ala-phe). I n these cases results showed radicals of the form PhCHzCHR were produced; however, it is not clear whether these species resulted from decarboxylation or secondary deamination. A r o w t i c Amino Acids and Dipeptides i n 8 M NaOD. Phe-ala, phe-gly, trp-ala, trp-gly, tyr-ala, and tyr-gly were all investigated in 8 M NaOD glasses. Photolysis of the phe containing peptides produced electrons by photoionization but did not result in a spectrum suggesting decarboxylation. Instead the same spectrum as was found for phe in 8 M NaOD was found for these peptides (phe-ala, phe-gly). Analogous results were found for trp, trp-ala, and trp-gly. Only in the cases of tyr-ala and tyr-gly was decarboxylation found in 8 M NaOD. For tyr-gly decarboxylation amounted to 10% of the signal whereas
for tyr-ala the decarboxylation signal amounted to over half the intensity. These results were similar to those found for tyr peptides in basic 8 M NaC104. We conclude that in NaOD glasses the radical remains on the aromatic group in the phe and trp containing dipeptides whereas for tyr containing dipeptides some decarboxylation is found. Proposed Mechanism, Light Intensity, and Concentration Studies. The formation of radicals I, 111, and IV in di- and tripeptides with the aromatic group adjacent to the C-terminal residue is clearly the result of a decarboxylation reaction. The presence of 0- in the spectra suggests the decarboxylation is preceded by a photoionization event. The mechanism we propose to account for the results found in this work is the transfer of the “hole” from the photoionized aromatic ring to the carboxyl group of the adjacent C-terminal residue followed by rapid decarboxylation to produce radicals I, 111, and IV. The photoionization process for phenylalanine containing peptides has been shown to be biphotonic by several workers; however, it would be argued that the decarboxylation results from a single photon process such as triplet energy transfer, and is only incidental to the photoionization process. The fact that the 0- concentration was equal to the concentration of radical I in the phe-ala spectrum and that no other radicals were observed argues against this. In addition we have performed studies of lamp intensity vs. the production of radical I from the photolysis of phe-ala solutions in 8 M NaC104 at 77 K. In order to test for a biphotonic process the lamp intensity was kept low, the solute concentration relatively high (2 X lo-* M), and the photolysis time short. A plot of the concentration of radical I (R = phenyl) vs. the square of the lamp intensity is shown in Figure 7. A least-squares fit of the data to the equation, radical concentration = (constant)(intensity)n,gave n = 2.2 f 0.2. For trp-ala similar experiments resulted in the value n = 2.0 f 0.3. A value of 1.6 for tyr-gly in 6 M NaOH was reported by Santus, Helene, and Ptak.3 We have repeated this experiment on 8 M NaOD and find the same value, 1.6 f 0.3. We conclude that in neutral glasses a biphotonic mechanism for radical production prevails. Thus the results suggest the radicals of the form of radical I originate on a one for one basis from the photoionized species, the x cation of the aromatic ring. From the above it may not be clear whether the transfer of the hole from the 7r cation to the carboxyl group is intramolecular or intermolecular in nature. An inter-
342
The Journal of Physical Chemistry, Vol. 82, No. 3, 1978
molecular mechanism would be expected to manifest itself only at higher concentrations (>lo-’ M) where agglomeration becomes likely in these solutions. However we find that samples with as low as M phe-ala produce excellent ESR intensities. Since these concentrations are well below those at which agglomeration occurs, we conclude the transfer process is intramolecular in nature.
Summary and Conclusions In this work we have found that decarboxylated radicals of the form of I, 111, and IV are produced when di- and tripeptides with an aromatic amino acid residue next to the C terminal are photoionized. The process of radical production is found to be biphotonic and intramolecular in nature. The radical produced is equal in concentration to 0- which originates from the photoejected electron. The mechanism we find most consistent with these results is the intramolecular transfer of an electron from the carboxyl group to the photoionized aromatic ring in the peptide, followed by decarboxylation.26 Since the process is occurring at 77 K in a rigid aqueous glass, the peptides must be in a conformation which places the carboxyl group near the aromatic ring. The fact that the tripeptide phe-gly-gly and the individual aromatic amino acids show no decarboxylation supports the suggestion that the transfer process is strongly conformationally dependent. The incomplete transfer found for tyr and trp peptides in 8 MNaC104 and the complete lack of transfer found for phe and trp peptides in 8 M NaOD is likely due to conformations which lower the probability of hole transfer to the carboxyl group, and/or in the case of tyrosine containing peptides a competing mechanism, i.e., the deprotonation of the P cation of the phenol ring to form the phenoxyl radical. Presumably the stability of the phenoxyl radical would not allow for hole transfer. Finally, the results reported here are expected to be of significance to the radiation chemistry of peptides and proteins. The initial positive charges produced in the peptide would be expected to migrate to the aromatic and carboxyl groups. Once the positive charge is on the aromatic group a number of reactions are possible: deprotonation on tyrosine to form a phenoxyl radical, hydroxyl ion addition to T cations of the aromatic groups,27 and finally in competition with these processes charge
M. D. Sevilla and J. B. D’Arcy
transfer to the carboxyl groups in close proximity resulting in decarboxylation.
References and Notes (1) This research was supported by the US. Department of Energy and by the Food Engineering Laboratory of the U.S. Army Natick Development Center. (2) H. B. Steen, Photochem. Photobid., 9, 479 (1969). (3) (a) R. Santus, C. Helene, and M. Ptak, Photochem. Photobiol., 7, 341 (1968); (b) Y. A. Vladimirov and E. E. Fesenko, ibkf., 8, 209 (1968). (4) (a) J. Moan, Photochem. Photobiol., 25, 591 (1977); (b) Acta Chem. Scand., Sect. A , 31, 281 (1977). (5) A. Meybeck and J. Meybeck, Photochem. Photoblol., 16,359 (1972). (6) R. Poupko, I. Rosenthal, and D. Elad, Photochem. Photobiol., 17, 395 (1973). (7) E. N. Chekvardze, Yu. A. Koslov, and K. M. L’Vov, Stud. Blophys., 35, 189 (1973). (8) M. D. Sevilla and V. L. Brooks, J. Phys. Chem., 77, 2954 (1973). (9) D. V. Bent and E. Hayon, J. Am. Chem. Soc., 97, 2606 (1975). (10) D. V. Bent and E. Hayon, J. Am. Chem. Soc., 97, 2612 (1975). (11) F. Dudley Bryant, R. Santus, and L. I. Grossweiner, J. Phys. Chem.: 79, 2711 (1975). (12) D. V. Bent and E. Hayon, J. Am. Chem. Soc., 97, 2599 (1975). (13) J. Fertelson, E. Hayon, and A. Treinin, J. Am. Chem. Soc., 95, 1025 (1973). (14) N. Getoff in ”Excited States in Organic and Biochemistry,” B. Pullman and N. Goldblum, Ed., Reidel Press, Holland, 1977 (15) M. D. Sevilla, J. Phys. Chem., 80, 1898 (1976). (16) M. D. Sevilla, R. Failor, C. Chrk, R. A. Holroyd, and M. Pettei, J. Phys. Chem., 80, 353 (1976). (17) The tunneling of methyl protons can be an efficient relaxation mechanism. (See L. Kevan and L. Kispert, “Electron Spin Double Resonance Spectroscopy”, Wiley, New York, N.Y., 1976, and M. K. Bowman and L. Kevan, Faraday Discuss., Chem. Soc., in press.) We believe this effect results In the resistanceto power saturation of radicals I and 111. (18) The complete analyses of these complex spectra is in progress. (19) C. Van Paemel, H. Frumin, V. L. Brooks, R. Failor, and M. D. Sevilla, J. Phys. Chem., 79, 839 (1975). (20) M. D. Sevilla, J . Phys. Chem., 74, 2096 (1970). (21) H. C. Box, E. E. Budzinski, and H. G. Freund, J. Chem. Phys., 61, 2222 (1974). (22) H. Taniguchi, H. Hasumi, and H. Hatano, Bull. Chem. SOC.Jpn., 45, 3380 (1972). (23) W. T. Dixon and D. Murphy, J. Chem. Soc., Faraday Trans. 2, 72, 1221 (1976). (24) M. T. Pailthorpe and C. H. Nichols, Photochem. Photobiol., 15, 465 (1972). (25) R. Santus, R. Guermonprez, and M. Ptak, J . Phys. Chem., 74, 550 (1970). (26) A decarboxylation reaction has been observed previously for Nacetylphenylalanine (ref 9). These investigators proposed a mechanism to account for the decarboxylationwhich is similar to that proposed in this work. (27) R. 0. C. Norman, P. M. Storey, and P. R. West, J . Chem. Soc. 6 , 1087 (1970).