Photochemistry and Phofobiology. 1976. Val 23. pp I 6
Pergamon Press
Printed in Great Britain
THE PHOTOSENSITIZED OXIDATION OF TYROSINE DERIVATIVES IN THE PRESENCE OF ALGINATE-I REACTION UNDER HOMOGENEOUS CONDITIONS G. R. SEELYand R. L. HART Charles F. Kcttcring Research Laboratory*, Yellow Springs, Ohio 45387, U.S.A. (Received 9 June 1975; acccpted 23 September 1975) Abstract--The rates of dye-sensitized photooxidation of tyrosine and tyramine to brown products were compared in the presence and the absence of the anionic polysaccharide, alginate. The polyelectrolyte did not affect the reaction when it was sensitized by monochromatic light absorbed mainly by the monomeric form of the dye. In white light, the rate of oxidation sensitized by thionine or phenosafranine was increased in the presence of alginate for tyramine. but not for tyrosine. In the thionine sensitized reaction, the ratio of brown product formation to tyramine consumption increased with decreasing wavelength of monochromatic excitation. These and other phenomcna are believed rclatcd to the formation of complexes between the dyes and some of the oxidation products, and to association between some of the oxidation products and alginate. A mechanism for oxidation of phenols is proposed, based on the addition of Oz ('Ag) across a double bond ortho to the phenolate oxygen. Dycs bind to alginate in monomeric and in aggregated forms; only the monomeric forms of thiazine dyes are photochemically active, but both the monomeric and the aggregated forms of crystal violet are active.
INTRODUCTION
The photosensitized oxidation of certain amino acid residues of enzymes is a widely used technique for probing the structure of active sites. Two phenomena appear in enzyme systems which are not generally encountered in simpler ones. Binding of the sensitizing dye to the enzyme may expose nearby residues to rapid oxidation, and the conformation of the protein may protect other residues effectively from attack (Spikes and MacKnight, 1970). Studies of photosensitized oxidation in systems containing other macromolecules have not been as frequent. The usual experience is that photooxidation is somewhat retarded when the dye is bound to a synthetic polymer (Bellin, 1968). However, binding usually stabilizes the dye against chemical attack and photolysis, and in particular may stabilize a reactive excited state, as is the case for crystal violet (Oster and Bellin. 1957). We were interested in examining the effect of the presence of a high polymer on the rate of a photosensitized reaction, and in particular, whether binding of the substrate to the same polymer to which the sensitizing dye is bound would increase the rate of its reaction over that of substrate not bound. As the type of reaction to be investigated, we chose photosensitized oxidation, the rate of which, when effected by the singlet oxygen mechanism, is limited by the lifetime of 0, ('As) in solution (2 ps in water [Merkel and Kearns, 19721). For a number of reasons we chose to examine the sensitized oxidation of tyrosine (tyr) derivatives in the *Contribution No. 544.
presence of alginate. Alginate is a natural copolymer of mannuronic and guluronic acids, derived from brown algae, which binds many cationic dyes strongly. As sensitizers we used primarily the thiazine dyes, thionine and methylene blue, but also phenosafranine and crystal violet. Tyr derivatives are available with a variety of modifications of the amino acid part which determine their ability to bind to alginate. The oxidation of tyr and other phenols is important biologically and leads to a variety of plant and animal pigments. The photosensitized oxidation of tyr is probably less well understood than that of the other susceptible amino acids. In this paper we compare the oxidation of tyramine and tyr. compounds which should and should not bind readily to alginate, and examine the oxidation of tyramine under a variety of conditions. It is concluded that alginate does not have very much effect on the rate of oxidation of tyramine itself, but docs increase the rate and extent of further reaction of early oxidation products. In the following paper (Part 11) we describe the reaction under heterogeneous conditions, where some distinct differences in rate appear. Although O2 ('Aq) is known to rcact with hindered phenols in organic solvents (Matsuura et ul.. 1969), direct proof of its reaction with tyr in aqueous solution has hitherto been lacking. However, the fact that chlorophyll sensitizes an oxidation apparently similar to ours in basic methanol (Ashkinazi et al., 1967) strongly suggests that tyr, like other amino acids (Nillson et al., 1972), is oxidized by 0, ('Ag). Although the present work is not directly concerned with the mechanism of oxidation, we have confirmed the role of singlet oxygen by demonstrating
2
G. R. SEELYand R. L. HART
a n increase in the rate of oxidation of tyramine in D,O over that in H,O (Nillson et al., 1972). In view of this, it is reasonable to assume that the results we have observed are primarily the work of 0, (‘Ag), and leave the details of a n undoubtedly complicated mechanism for a clarification which we hope to accomplish in the future.
RESULTS AND DISCUSSlON
pH Dependence. In agreement with Weil et al. (1951), Sluyterman (1962), and Bellin and Yankus (1968), we found that oxidation of tyr derivatives goes with appreciable speed only in alkaline solutions, where the phenolate ion is present. For the oxidation of tyr (8 x 10-4M) sensitized by methylene blue a t 660 nm, in well-buffered solutions without alginate, MATERIALS AND METHODS the rate as measured by increase of product absorpSodium alginate was a commercial sample with mannur- tion at 320 nm, was proportional to from onic/guluronic ratio of 2.0 (Seely and Hart, 1974). Solutions were freed of salt by dialysis, and the carboxylate p H 6 to 10. Over the same range, tyr changes from the predominantly phenol form to the phenolate. content was determined by titration. Dyes used in this work were purified by chromatography Concentration dependence. The rate of oxidation of on alumina and recrystallized as chlorides. CHN analyses tyr and tyramine is concentration limited in the agreed with their formulas and molar absorptivities comneighborhood of 10-4M substrate, and approximates pared with the best literature values. Commercial tyr and tyramine HC1 were used as independence of concentration a t 10-3M. Initial received. Tyramine was titrated spectrophotometrically, quantum yields vs initial substrate concentration arc and the results, worked up by the procedure of Martin plotted in Fig. 1, for methylene blue sensitized (1971), were in good agreement with those he had calcu- (660nm) reaction a t p H 8.5 in the presence of lated from work at higher ionic strength. Much of tyramine is in the cationic form, which is attracted to the polymer, alginate. Although the initial quantum yields are similar, it was noted that the yield of tyr consumption up to pH 10. Water used for the solutions was distilled from acid fell more rapidly as the reaction progressed than the dichromate, alkaline permanganate, redistilled, and stored yield of tyramine consumption did. in borosilicate containers. Its conductivity was checked to Buffer concentration. A distinct effect of buffer converify that the ion concentration was low. on the rate of tyramine oxidation could centration Since we were primarily interested in the effects of the presence of polymer on the rate of oxidation, reaction con- not be distinguished. Since buffer salts might compete ditions were chosen to favor their detcction. These condi- with tyramine for alginate binding sites, it was tions are necessarily less than optimal for the reaction. expected that a higher buffer concentration might The substrate was present at concentrations which would be limiting to the reaction rate, usually between and retard the oxidation. In solutions initially set a t p H t o 10-3M tetraethylammonium 1W3M. The pH was set initially between 7 and 10; high 8.5 and 9.5 with pH favors oxidation, but low pH favors binding of sub- borate buffer, small differences in rate were noted, strates such as tyramine to the polymer. The pH falls dur- but because the buffer concentration influenced both ing oxidation as acids are produced. The rate of fall can the rate of p H decrease as the reaction progressed be reduced by buffering the solution, but buffer salts comand the amount of methylene blue sensitizer in monopete with substrate cations for anionic polymer binding sites; as a compromise, we usually buffered the solution meric form (see below), no effect due to buffer alone weakly to the desired initial pH, and let the reaction take could be concluded. its course. Tetraalkylammonium salts of buffering anions Wavelength dependence. Many dyes are strongly and alginate were often used because such cations are less metachromatic when bound t o alginate, and present competitive than others for binding to polysaccharides (Rinaudo et al., 1973). Oxygen was bubbled through the spectra like that of methylene blue in Fig. 2. For reasolutions to maintain saturation. Tetraethylammonium sons which will be discussed in detail elsewhere. we glucuronate was used as a monomeric analog of alginate in some experiments. Reactions were run under “white” (unfilterid 750 W projection lamp) or “monochromatic”(interference filter) irradiation. Light intensities absorbed by the reaction system were measured by difference with a Kettering Radiant Power Meter which had been standardized against an Eppley Thermopile and a National Bureau of Standards lamp. Rates of oxidation were determined by an appropriate Methylene Blue t Alginate analysis of the phenol absorption bands in the UV. The ultimate products of oxidation absorb broadly through the pH 8.5 visible and UV regions, and their increase could best be followed at 320nm, where interference from dye and phenol absorption was slight. On exposure to light, solutions of thiazine dyes such as methylene blue are gradually bleached, in a reaction 0 2 4 6 8 10 which increases with pH but is independent of the presence [tyr] , M of phenols. This reaction is retarded, not enhanced, by the presence of alginate, and further, we have not observed Figure 1. Initial quantum yield of tyrosine and tyramine changes in the viscosity or gel structure of alginates on oxidation vs initial concentration, in the presence of Na prolonged irradiation in the presence of the dyes we have alginate, 6 x 10-4N in carboxylate, 2.6 x M methylused that would betray reaction between the dyes and the ene blue, 2 x 10- ‘ M tetrabutylammonium borate buffer binding polymer. (pH ca. 8.5) and 660 nm irradiation.
The photosensitized oxidation of tyrosines-I
3
the pH was lower during this reaction than it was during the reaction without alginate. 2.6 X Methylene Blue When the oxidation of tyr was compared under A similar conditions, there was little difference in the rate and extent of reaction, and both resembled 0.02 closely the oxidation of tyramine without alginate. 9 The rate of oxidation in white light was greater 0.2 in the presence of alginate, even though much of the 0.0 I thionine was in a photochemically inactive aggregated 0.1 form. It seemed reasonable that if a dye such as phenosafranine were used, which binds to alginate but 500 550 600 650 700 750 has little tendency to aggregate, the rate difference A , nm would be greater. This proved to be so. The increase Figure 2. Wavelength dependence of initial quantum yield in product absorption at 320nm was three times as of oxidation of tyramine, sensitized by methylene blue. great in the presence of alginate as in its absence durComposition of solution: 6 x 10-4M tyramine HCI, ing most of the reaction. 6 x 10-4N Na alginate, 2.6 x IO-'M methylene blue, Reaction in monochromatic light. In sharp contrast 2 x l W 3 M tetrabutylammonium borate huffer, pH 8.5. Upper curve (Ieft ordinate scale): spectrum of methylene with results in white light, the rate or oxidation of blue; lower curve: contribution of monomeric dye to tyramine in monochromatic light, absorbed mainly absorption. Stars: initial quantum yields (right ordinate by bound monomeric dye, is not accelerated by the scale). presence of alginate. In Fig. 5 the rates of increase of product absorption in the near UV, in the presence believe that the spectrum consists of a band of bound of alginate and glucuronate, are compared for sensitidye in monomeric form at 665nm, and a group of zation by methylene blue and by phenosafranine. bands of bound dimeric forms centered about 570 nm. These reactions were started at pH 9.5 to compensate It is naturally of interest whether the bound dimeric for the lower intensity of monochromatic light. forms are more or less active photochemically than With methylene blue, the rate of absorption inthe bound monomer. crease in the presence of alginate is about 2/3 of that Quantum yields of oxidation of tyramine, measured in the presence of glucuronate; this is consistent with by the decrease in the phenol absorption band at the fact that about 63% of the absorbance at 660 nm 276nm, were determined at 660, 600 590, 570 and is of dye in the monomeric form. With phenosa550 nm. Initial quantum yields, plotted in Fig. 2, show franine, where there is little bound dimer, the two a clear proportionality to the fraction of light curves nearly coincide. With thionine, at 600 nm and absorbed by monomeric dye. Bound dimeric dye is initial pH 9.3, rates with glucuronate and alginate photochemically inactive, or nearly so. nearly coincide with the respective curves for methyMonomeric bound thionine is also much more lene blue. active than the dimeric forms, but this situation is Complex associations. We have found that alginate not universal. Crystal violet on alginate shows two increases the rate of oxidation of tyramine, sensitized bands, of monomer at 590nm and of an aggregated by thionine (and phenosafranine) in white light, but species at 520 nm. Both monomer and aggregated dye are active photochemically, and the quantum yield at 520nm is greater than that at 590nm. Unbound dye, which absorbs at 590nm, is much less active. 4 X 10-6 M Thionine The photochemical activity of crystal violet is known to be increased by binding to polymers (Oster and Bellin, 1957). ' Reuction in white light. The rate and the extent P of thionine-sensitized oxidation of tyramine are distinctly increased by the presence of alginate, when the reaction is sensitized by thionine in white light at a relatively low pH. From comparison of reaction without and with alginate (Figs. 3 and 4) it is evident that there is much less photolysis of dye when A , nm alginate is present, though there is then some conversion from dimeric to monomeric form. Loss of phenol bands around 280 nm is more complete when alginate Figure 3. Photosensitized oxidation of tyramine in white is present, and there is a greater increase in product light, without alginate. Composition: 10- 4M tyramine HCI, 4 x 10-'M thionine, 4 x 10-5M Na,CO,; initial absorption in the UV. Without alginate, a broad pH 7.97, final pH 7.07. Spectra taken in 5cm path length product band is evident around 450 nm. The reaction cell; note change of wavelength scale at 330 nm. Numbers was greater in the presence of alginate, even though on spectra are times of irradiation by white light, in min. 6X
M Tyrnmine
6 X
N Alginate
G. R. SEELY and R. L. HART
4
4 X 2 2 X
M Thionine N Alginote
P
240
280
320
400
A , nm
500
600
Figure 4. Photosensitized oxidation of tyramine in white light, in the presence of alginate. Composition as for Fig. 3 but with 2.2 x 10-4N Na alginate. Initial pH 7.56, final, 6.91.
not that of tyr. There was no acceleration with alginate compared to glucuronate in monochromatic (600nm) light. Glucuronate by itself has no effect on these reactions. Weil et al. (1951) noted that in the earlier stages of oxidation of tyr and its derivatives, before brown products appeared, more than 2mol O2 were consumed, and I mol CO, was released. As the same things happened with phenol, they concluded that the aromatic part of the molecule was attacked, rather than the amino acid part. The characteristic aromatic bands in the UV were replaced by broad absorption bands, indicating disruption of the aromatic ring. Weil (1965) found a fairly large temperature coefficient for 0, uptake. If the reaction is allowed to continue, complex changes in the UV suggest that the early products are subject to further reactions, some photosensitized, some not. If a well-oxidized solution of tyramine is passed through a Sephadex G 10-120 column, a series of brown bands emerge, the darker generally, but not always, in front. The spectra of these products, and their origin by oxidative disruption of a phenol, suggest that they are conjugated carbonyl compounds, and that the higher molecular weight ones have been built up from the lower by condensation reactions. We have grounds for believing that the acceleration in the presence of alginate and white light is due to the existence of two sorts of complexes: photochemically active associations between the cationic dyes and anionic oxidation products, and associations between oxidation products and alginate. The former are characterized by broadening and displacement of the spectrum of the dye. The strength of the latter probably depends on the charge on the p-substituent of the phenol. Evidence for these associations consists of direct observation of dissociation of dyeproduct complexes, solubility behavior of oxidation products, and the course of photooxidation in blue and in monochromatic radiation. Identification of complexes of dye with oxidation
products is impossible when alginate is in solution, but becomes possible if the reaction is run without alginate or with alginate in a gelled state. For example, after exhaustive oxidation of p-chlorophenol by thionine in the presence of an alginate gel, the visible spectrum contained broad absorption bands but without much evidence of dye. After acidification and extraction of much yellow-brown material into ether, characteristic bands of dye appeared in the aqueous layer. Similar observations were made with toluidine blue. It would be hard to account for such events unless a complex between the dye and oxidation products existed, in which the spectrum of the dye was broadened almost to the point of unrecognizability. A similar observation was also made with tyramine and thionine, when the reaction was run without alginate but with the dye initially inside a dialysis bag. Although the dye was rapidly and almost completely bleached, acidification of the contents of the bag regenerated a weak dye spectrum. Evidence for association between alginate and oxidation products is less direct. When the reaction is conducted under heterogeneous conditions, with the alginate in a dialysis bag or gelled, it is generally observed that the alginate phase becomes brown, and that the color is hard to extract. For example, after a thionine-sensitized oxidation of tyramine, thc alginate phase was acidified with HCI and extracted with 75% ethanol. Nearly all the remaining dye was removed, but only part of the brown products, a part with bands at shorter wavelengths. Alginate in nature is usually associated with brown phenolic substances, which are removed commercially by bleaching or by ethanol extraction (Smidsrmd et al., 1963).
RADIATION
ABSORBED, EINSTEIN / LITER
Figure 5. Photosensitized oxidation of tyramine in monochromatic light absorbed by bound monomeric form of dye. Increase of product absorption at 320 nm for niethylene blue with alginate).( and with glucuronate (0).and at 326 nm for phenosafranine with alginate (H) and glucuronate (O), in 5 cm path length cell. Composition of solutions: 10-4M tyramine HCI, 3.75 x 10-4N tetraethylammonium alginate or glucuronate, 7.4 x lO-'M methylene blue or 6.1 x 10-fiM phenosafranine; pH adjusted to 9.5 initially with tetraethylammonium hydroxide. Irradiation with 660 nm (methylene blue) or 520 nm (phenosafranine) light.
5
The photosensitized oxidation of tyrosincs- 1
It is not clear whether the brown products are retained by insolubility or entanglement, or whether they are bound to the polyelectrolyte by ionic forces or men by primary chemical bonds. But it is plausible to suppose that the ionic nature of substituents will affect the strength of association and thereby contribute toward explaining the effect of the presence of polymer. Reuction in blue light. The extent of oxidation was often followed by the increase of product absorption at 320nm as well as by the decay of phenol bands further into the UV, because of some uncertainty as to the extent of product absorption in the region of the phenol bands, and the effect on them of pH changes. It was early established that in the case of tyr (X x lW4M, pH 8.38, 660nm irradiation), the two absorption changes were quite accurately linearly correlated for at least 80% of the reaction. With tyramine and other kinds of radiation, however, it is possible t o dissociate the rise at 320nm from the decay of the phenol bands. If broad bands of complexes between thionine and oxidation products are responsible for much of the increase in visible and near UV absorption in white light, then the effect of polymer ought to be evident even at pH > 9 when the reaction is run in blue light, absorbed more by the inert bound dye dimers and by the supposed dye-product complexes than by monomeric dye. This prediction is supported by an experiment using thionine, the results of which are plotted in Fig. 6. The decrease of tyramine and increase of 320nm absorption are both slower with alginate than with glucuronate, because with the former, much of the light is absorbed by photochemically inactive dimers. The ratio of the quantum yield of tyramine loss with glucuronate to that with alginate is nearly 2.6 throughout the reaction. If we define an effective dose for tyramine loss in the preslo4
:1
ence of alginate as (1/2.6) of the absorbed dose, and plot the data against it, the dashed curves of Fig. 6 result. The dashed curve for tyramine loss with alginate now nearly coincides with that for glucuronate, but the curve for 320 nm absorption rises more rapidly than for glucuronate, almost from the beginning. In the presence of alginate, blue light is relatively more effective in sensitizing increased absorption at 320nm than in sensitizing tyramine loss. The same phenomenon is seen in the methylene blue sensitized oxidation of tyramine in monochromatic light (Fig. 7). The shorter the wavelength, the greater the build-up of near UV absorption for the same amount of tyramine oxidized. The divergence becomes more apparent after the reaction has proceeded to an extent of -3o”/,. Mechanism. Although we are not principally concerned with the mechanism of phenol oxidation in
-A [ t y r l X 1 0 4 , M
Figure 7. Methylene blue sensitized photooxidation of tyramine in light of different wavelengths. Composition of solution as in Fig. 2. Comparison of absorbance increase at 320 nm with loss of tyramine.
[ t y r ~
0.4
0.2 0 10
I
0.06
0.04 0.02
lCp0
Figure 6. Photooxidation of tyramine in the presence of glucuronate and alginate, sensitized by thionine in light through Corning 4-94 blue filter. Composition of solutions: 3 x 10-4N tetraethylammonium glucuronate or alginate, 10-4M tyramine HCI, 5 x 10-6M thionine, 2 x 10-4M tetraethylammonium carbonate buffer to establish initial pH 9.28-9.30. Tyramine concentration with glucuronate (0)and alginate ( 0 )calculated from absorbance changes at 239 and 218nm and plotted (left ordinate scale) against total light energy absorbed, J./ I . Product absorbance changes at 320 nm with glucuronate (0) or alginate (B) similarly plotted (right ordinate scale). Dashed lines are data for alginate plotted against effective dose for tyramine destruction, i.e. total dose divided by 2.6.
G. R. SEELYand R. L. HART
6
this paper, a few remarks concerning its probable (Matsuura et al., 1969; Weil, 1965), the known facts course seem to be in order. Acceleration of the rate in the oxidation of tyrosines and other phenols with of tyramine oxidation in DzO supports the assump- an unsubstituted ortho position can be readily tion of a singlet oxygen reaction. Rates of singlet explained if 0, ('Ag) adds across an electron-rich oxygen reactions may be increased by as much as double bond next to the phenolate oxygen, in a mantenfold in D,O over HzO, owing to the longer life- ner like that proposed for eneamines and vinylethers time of 0, ('Aq) in the former solvent. Although the (Bartlett and Schaap, 1970; Foote and Lin, 1968; conditions (5 x 10-4M tyramine, 4.73 x 10-3M Mazur and Foote, 1970). In Fig. 8 the intermediate KHC03, 1.4 x 10-5M methylene blue, pH 10.2, dioxetane (11) opens to 4-R-5-formyl-2,4-penta660nm) were not optimal for expression of a D,O dienoate (111),which adds a second O2 ('Ag) in a reaceffect, the rate of tyramine loss was greater initially tion not sensitive to pH. Decarboxylation of (IV) to by a factor of 2 in DzO, and as tyramine approached give the unsaturated dialdehyde (V) accounts for two depletion, the factor increased to 4-5. These factors O2 taken up and one CO, released. The reactions are no less than what would be expected from the available to (V) are numerous, but one way in which tyramine concentration dependence of quantum yield condensation could occur to build up extended conjuin Fig. I. Further oxidation of products was acceler- gated carbonyl systems is shown. Compounds like ated in DzO, more so perhaps than of tyramine itself, (VI), in enolate form or oxidized to carboxylate, could but the extent of bleaching of the dye was unchanged. be expected to form strong complexes with thionine, Although radical oxidations of phenols are well and at the same time, or alternatively, to bind known, and radical formation has been proposed as strongly to an anionic polyelectrolyte through posia first step in photosensitized oxidation of phenols tively charged groups R.
6R
k
R
m
m
TI
I
I+
-HCOj
CHO CHO r
1
L
J
H20
P
PI Figure 8. The proposed reaction scheme
REFERENCES
Ashkinazi, M. S., I. A. Dolidze and V. A. Yegorova (1967) Biufiz. 12,427--432[Eng. Transl., 488-4931. Bartlett, P. D., and A. P. Schaap (1970) J . Am. Chem. Soc. 92, 3223-3225. Bellin. J. S . (19681 Photochem. Phorohiol. 8. 383-392. Bellin, J. S., and C. A. Yankus (1968) Arch. Biochem. Biophys. 123, 18-28. Foote, C. S., and J. W.-P. Lin (1968) Tetrahedron Letters 3267-3270. Martin, R. B. (1971) J. Phys. Chem. 75, 2657-2661. Matsuura, T., N. Yoshimura, A. Nishinaga and I. Saito (1969) Tetrahedron Letters 1669-1671. Mazur, S., and C . S. Foote (1970) J . Am. Chem. Soc. 92, 3225-3226. Merkel, P. B., and D. R. Kearns (1972) J. Am. Chem. Soc. 94, 1029-1030, 72447253. Nilsson, R., P. B. Merkel and D. R. Kearns (1972) Photochem. Photohiol. 61, 117-124. Oster, G., and J. S . Bellin (1957) J . Am. Chem. Soc. 79, 294298. Rinaudo, M., M. Milas and M. Lafford (1973) J. Chim. Phys. 70, 884-887. Seely, G. R., and R. L. Hart (1974) Macromol. 7, 706-710. Sluyterman, L. A. Ae. (1962) Biochim. Biophys. Acta 60, 557-561. Smidsrgd, O., A. Haug and B. Larsen (1963) Acta Chem. S c a d . 17, 1473-1474. Spikes, J. D., and M. L. MacKnight (1970) In Photochemistry of Macromolecules, (Edited by R. F. Reinisch), pp. 67-83. Plenum Press, New York. Wed, L. (1965) Arch. Biochem. Biophys. 110, 57-68. Weil, L.. W. G. Gordon and A. R. Buchert (1951) Arch. Biochem. Biophys. 33, 90--109.
Pergamon Press
Printed in Great Britain
THE PHOTOSENSITIZED OXIDATION OF TYROSINE DERIVATIVES IN THE PRESENCE OF ALGINATE-I REACTION UNDER HOMOGENEOUS CONDITIONS G. R. SEELYand R. L. HART Charles F. Kcttcring Research Laboratory*, Yellow Springs, Ohio 45387, U.S.A. (Received 9 June 1975; acccpted 23 September 1975) Abstract--The rates of dye-sensitized photooxidation of tyrosine and tyramine to brown products were compared in the presence and the absence of the anionic polysaccharide, alginate. The polyelectrolyte did not affect the reaction when it was sensitized by monochromatic light absorbed mainly by the monomeric form of the dye. In white light, the rate of oxidation sensitized by thionine or phenosafranine was increased in the presence of alginate for tyramine. but not for tyrosine. In the thionine sensitized reaction, the ratio of brown product formation to tyramine consumption increased with decreasing wavelength of monochromatic excitation. These and other phenomcna are believed rclatcd to the formation of complexes between the dyes and some of the oxidation products, and to association between some of the oxidation products and alginate. A mechanism for oxidation of phenols is proposed, based on the addition of Oz ('Ag) across a double bond ortho to the phenolate oxygen. Dycs bind to alginate in monomeric and in aggregated forms; only the monomeric forms of thiazine dyes are photochemically active, but both the monomeric and the aggregated forms of crystal violet are active.
INTRODUCTION
The photosensitized oxidation of certain amino acid residues of enzymes is a widely used technique for probing the structure of active sites. Two phenomena appear in enzyme systems which are not generally encountered in simpler ones. Binding of the sensitizing dye to the enzyme may expose nearby residues to rapid oxidation, and the conformation of the protein may protect other residues effectively from attack (Spikes and MacKnight, 1970). Studies of photosensitized oxidation in systems containing other macromolecules have not been as frequent. The usual experience is that photooxidation is somewhat retarded when the dye is bound to a synthetic polymer (Bellin, 1968). However, binding usually stabilizes the dye against chemical attack and photolysis, and in particular may stabilize a reactive excited state, as is the case for crystal violet (Oster and Bellin. 1957). We were interested in examining the effect of the presence of a high polymer on the rate of a photosensitized reaction, and in particular, whether binding of the substrate to the same polymer to which the sensitizing dye is bound would increase the rate of its reaction over that of substrate not bound. As the type of reaction to be investigated, we chose photosensitized oxidation, the rate of which, when effected by the singlet oxygen mechanism, is limited by the lifetime of 0, ('As) in solution (2 ps in water [Merkel and Kearns, 19721). For a number of reasons we chose to examine the sensitized oxidation of tyrosine (tyr) derivatives in the *Contribution No. 544.
presence of alginate. Alginate is a natural copolymer of mannuronic and guluronic acids, derived from brown algae, which binds many cationic dyes strongly. As sensitizers we used primarily the thiazine dyes, thionine and methylene blue, but also phenosafranine and crystal violet. Tyr derivatives are available with a variety of modifications of the amino acid part which determine their ability to bind to alginate. The oxidation of tyr and other phenols is important biologically and leads to a variety of plant and animal pigments. The photosensitized oxidation of tyr is probably less well understood than that of the other susceptible amino acids. In this paper we compare the oxidation of tyramine and tyr. compounds which should and should not bind readily to alginate, and examine the oxidation of tyramine under a variety of conditions. It is concluded that alginate does not have very much effect on the rate of oxidation of tyramine itself, but docs increase the rate and extent of further reaction of early oxidation products. In the following paper (Part 11) we describe the reaction under heterogeneous conditions, where some distinct differences in rate appear. Although O2 ('Aq) is known to rcact with hindered phenols in organic solvents (Matsuura et ul.. 1969), direct proof of its reaction with tyr in aqueous solution has hitherto been lacking. However, the fact that chlorophyll sensitizes an oxidation apparently similar to ours in basic methanol (Ashkinazi et al., 1967) strongly suggests that tyr, like other amino acids (Nillson et al., 1972), is oxidized by 0, ('Ag). Although the present work is not directly concerned with the mechanism of oxidation, we have confirmed the role of singlet oxygen by demonstrating
2
G. R. SEELYand R. L. HART
a n increase in the rate of oxidation of tyramine in D,O over that in H,O (Nillson et al., 1972). In view of this, it is reasonable to assume that the results we have observed are primarily the work of 0, (‘Ag), and leave the details of a n undoubtedly complicated mechanism for a clarification which we hope to accomplish in the future.
RESULTS AND DISCUSSlON
pH Dependence. In agreement with Weil et al. (1951), Sluyterman (1962), and Bellin and Yankus (1968), we found that oxidation of tyr derivatives goes with appreciable speed only in alkaline solutions, where the phenolate ion is present. For the oxidation of tyr (8 x 10-4M) sensitized by methylene blue a t 660 nm, in well-buffered solutions without alginate, MATERIALS AND METHODS the rate as measured by increase of product absorpSodium alginate was a commercial sample with mannur- tion at 320 nm, was proportional to from onic/guluronic ratio of 2.0 (Seely and Hart, 1974). Solutions were freed of salt by dialysis, and the carboxylate p H 6 to 10. Over the same range, tyr changes from the predominantly phenol form to the phenolate. content was determined by titration. Dyes used in this work were purified by chromatography Concentration dependence. The rate of oxidation of on alumina and recrystallized as chlorides. CHN analyses tyr and tyramine is concentration limited in the agreed with their formulas and molar absorptivities comneighborhood of 10-4M substrate, and approximates pared with the best literature values. Commercial tyr and tyramine HC1 were used as independence of concentration a t 10-3M. Initial received. Tyramine was titrated spectrophotometrically, quantum yields vs initial substrate concentration arc and the results, worked up by the procedure of Martin plotted in Fig. 1, for methylene blue sensitized (1971), were in good agreement with those he had calcu- (660nm) reaction a t p H 8.5 in the presence of lated from work at higher ionic strength. Much of tyramine is in the cationic form, which is attracted to the polymer, alginate. Although the initial quantum yields are similar, it was noted that the yield of tyr consumption up to pH 10. Water used for the solutions was distilled from acid fell more rapidly as the reaction progressed than the dichromate, alkaline permanganate, redistilled, and stored yield of tyramine consumption did. in borosilicate containers. Its conductivity was checked to Buffer concentration. A distinct effect of buffer converify that the ion concentration was low. on the rate of tyramine oxidation could centration Since we were primarily interested in the effects of the presence of polymer on the rate of oxidation, reaction con- not be distinguished. Since buffer salts might compete ditions were chosen to favor their detcction. These condi- with tyramine for alginate binding sites, it was tions are necessarily less than optimal for the reaction. expected that a higher buffer concentration might The substrate was present at concentrations which would be limiting to the reaction rate, usually between and retard the oxidation. In solutions initially set a t p H t o 10-3M tetraethylammonium 1W3M. The pH was set initially between 7 and 10; high 8.5 and 9.5 with pH favors oxidation, but low pH favors binding of sub- borate buffer, small differences in rate were noted, strates such as tyramine to the polymer. The pH falls dur- but because the buffer concentration influenced both ing oxidation as acids are produced. The rate of fall can the rate of p H decrease as the reaction progressed be reduced by buffering the solution, but buffer salts comand the amount of methylene blue sensitizer in monopete with substrate cations for anionic polymer binding sites; as a compromise, we usually buffered the solution meric form (see below), no effect due to buffer alone weakly to the desired initial pH, and let the reaction take could be concluded. its course. Tetraalkylammonium salts of buffering anions Wavelength dependence. Many dyes are strongly and alginate were often used because such cations are less metachromatic when bound t o alginate, and present competitive than others for binding to polysaccharides (Rinaudo et al., 1973). Oxygen was bubbled through the spectra like that of methylene blue in Fig. 2. For reasolutions to maintain saturation. Tetraethylammonium sons which will be discussed in detail elsewhere. we glucuronate was used as a monomeric analog of alginate in some experiments. Reactions were run under “white” (unfilterid 750 W projection lamp) or “monochromatic”(interference filter) irradiation. Light intensities absorbed by the reaction system were measured by difference with a Kettering Radiant Power Meter which had been standardized against an Eppley Thermopile and a National Bureau of Standards lamp. Rates of oxidation were determined by an appropriate Methylene Blue t Alginate analysis of the phenol absorption bands in the UV. The ultimate products of oxidation absorb broadly through the pH 8.5 visible and UV regions, and their increase could best be followed at 320nm, where interference from dye and phenol absorption was slight. On exposure to light, solutions of thiazine dyes such as methylene blue are gradually bleached, in a reaction 0 2 4 6 8 10 which increases with pH but is independent of the presence [tyr] , M of phenols. This reaction is retarded, not enhanced, by the presence of alginate, and further, we have not observed Figure 1. Initial quantum yield of tyrosine and tyramine changes in the viscosity or gel structure of alginates on oxidation vs initial concentration, in the presence of Na prolonged irradiation in the presence of the dyes we have alginate, 6 x 10-4N in carboxylate, 2.6 x M methylused that would betray reaction between the dyes and the ene blue, 2 x 10- ‘ M tetrabutylammonium borate buffer binding polymer. (pH ca. 8.5) and 660 nm irradiation.
The photosensitized oxidation of tyrosines-I
3
the pH was lower during this reaction than it was during the reaction without alginate. 2.6 X Methylene Blue When the oxidation of tyr was compared under A similar conditions, there was little difference in the rate and extent of reaction, and both resembled 0.02 closely the oxidation of tyramine without alginate. 9 The rate of oxidation in white light was greater 0.2 in the presence of alginate, even though much of the 0.0 I thionine was in a photochemically inactive aggregated 0.1 form. It seemed reasonable that if a dye such as phenosafranine were used, which binds to alginate but 500 550 600 650 700 750 has little tendency to aggregate, the rate difference A , nm would be greater. This proved to be so. The increase Figure 2. Wavelength dependence of initial quantum yield in product absorption at 320nm was three times as of oxidation of tyramine, sensitized by methylene blue. great in the presence of alginate as in its absence durComposition of solution: 6 x 10-4M tyramine HCI, ing most of the reaction. 6 x 10-4N Na alginate, 2.6 x IO-'M methylene blue, Reaction in monochromatic light. In sharp contrast 2 x l W 3 M tetrabutylammonium borate huffer, pH 8.5. Upper curve (Ieft ordinate scale): spectrum of methylene with results in white light, the rate or oxidation of blue; lower curve: contribution of monomeric dye to tyramine in monochromatic light, absorbed mainly absorption. Stars: initial quantum yields (right ordinate by bound monomeric dye, is not accelerated by the scale). presence of alginate. In Fig. 5 the rates of increase of product absorption in the near UV, in the presence believe that the spectrum consists of a band of bound of alginate and glucuronate, are compared for sensitidye in monomeric form at 665nm, and a group of zation by methylene blue and by phenosafranine. bands of bound dimeric forms centered about 570 nm. These reactions were started at pH 9.5 to compensate It is naturally of interest whether the bound dimeric for the lower intensity of monochromatic light. forms are more or less active photochemically than With methylene blue, the rate of absorption inthe bound monomer. crease in the presence of alginate is about 2/3 of that Quantum yields of oxidation of tyramine, measured in the presence of glucuronate; this is consistent with by the decrease in the phenol absorption band at the fact that about 63% of the absorbance at 660 nm 276nm, were determined at 660, 600 590, 570 and is of dye in the monomeric form. With phenosa550 nm. Initial quantum yields, plotted in Fig. 2, show franine, where there is little bound dimer, the two a clear proportionality to the fraction of light curves nearly coincide. With thionine, at 600 nm and absorbed by monomeric dye. Bound dimeric dye is initial pH 9.3, rates with glucuronate and alginate photochemically inactive, or nearly so. nearly coincide with the respective curves for methyMonomeric bound thionine is also much more lene blue. active than the dimeric forms, but this situation is Complex associations. We have found that alginate not universal. Crystal violet on alginate shows two increases the rate of oxidation of tyramine, sensitized bands, of monomer at 590nm and of an aggregated by thionine (and phenosafranine) in white light, but species at 520 nm. Both monomer and aggregated dye are active photochemically, and the quantum yield at 520nm is greater than that at 590nm. Unbound dye, which absorbs at 590nm, is much less active. 4 X 10-6 M Thionine The photochemical activity of crystal violet is known to be increased by binding to polymers (Oster and Bellin, 1957). ' Reuction in white light. The rate and the extent P of thionine-sensitized oxidation of tyramine are distinctly increased by the presence of alginate, when the reaction is sensitized by thionine in white light at a relatively low pH. From comparison of reaction without and with alginate (Figs. 3 and 4) it is evident that there is much less photolysis of dye when A , nm alginate is present, though there is then some conversion from dimeric to monomeric form. Loss of phenol bands around 280 nm is more complete when alginate Figure 3. Photosensitized oxidation of tyramine in white is present, and there is a greater increase in product light, without alginate. Composition: 10- 4M tyramine HCI, 4 x 10-'M thionine, 4 x 10-5M Na,CO,; initial absorption in the UV. Without alginate, a broad pH 7.97, final pH 7.07. Spectra taken in 5cm path length product band is evident around 450 nm. The reaction cell; note change of wavelength scale at 330 nm. Numbers was greater in the presence of alginate, even though on spectra are times of irradiation by white light, in min. 6X
M Tyrnmine
6 X
N Alginate
G. R. SEELY and R. L. HART
4
4 X 2 2 X
M Thionine N Alginote
P
240
280
320
400
A , nm
500
600
Figure 4. Photosensitized oxidation of tyramine in white light, in the presence of alginate. Composition as for Fig. 3 but with 2.2 x 10-4N Na alginate. Initial pH 7.56, final, 6.91.
not that of tyr. There was no acceleration with alginate compared to glucuronate in monochromatic (600nm) light. Glucuronate by itself has no effect on these reactions. Weil et al. (1951) noted that in the earlier stages of oxidation of tyr and its derivatives, before brown products appeared, more than 2mol O2 were consumed, and I mol CO, was released. As the same things happened with phenol, they concluded that the aromatic part of the molecule was attacked, rather than the amino acid part. The characteristic aromatic bands in the UV were replaced by broad absorption bands, indicating disruption of the aromatic ring. Weil (1965) found a fairly large temperature coefficient for 0, uptake. If the reaction is allowed to continue, complex changes in the UV suggest that the early products are subject to further reactions, some photosensitized, some not. If a well-oxidized solution of tyramine is passed through a Sephadex G 10-120 column, a series of brown bands emerge, the darker generally, but not always, in front. The spectra of these products, and their origin by oxidative disruption of a phenol, suggest that they are conjugated carbonyl compounds, and that the higher molecular weight ones have been built up from the lower by condensation reactions. We have grounds for believing that the acceleration in the presence of alginate and white light is due to the existence of two sorts of complexes: photochemically active associations between the cationic dyes and anionic oxidation products, and associations between oxidation products and alginate. The former are characterized by broadening and displacement of the spectrum of the dye. The strength of the latter probably depends on the charge on the p-substituent of the phenol. Evidence for these associations consists of direct observation of dissociation of dyeproduct complexes, solubility behavior of oxidation products, and the course of photooxidation in blue and in monochromatic radiation. Identification of complexes of dye with oxidation
products is impossible when alginate is in solution, but becomes possible if the reaction is run without alginate or with alginate in a gelled state. For example, after exhaustive oxidation of p-chlorophenol by thionine in the presence of an alginate gel, the visible spectrum contained broad absorption bands but without much evidence of dye. After acidification and extraction of much yellow-brown material into ether, characteristic bands of dye appeared in the aqueous layer. Similar observations were made with toluidine blue. It would be hard to account for such events unless a complex between the dye and oxidation products existed, in which the spectrum of the dye was broadened almost to the point of unrecognizability. A similar observation was also made with tyramine and thionine, when the reaction was run without alginate but with the dye initially inside a dialysis bag. Although the dye was rapidly and almost completely bleached, acidification of the contents of the bag regenerated a weak dye spectrum. Evidence for association between alginate and oxidation products is less direct. When the reaction is conducted under heterogeneous conditions, with the alginate in a dialysis bag or gelled, it is generally observed that the alginate phase becomes brown, and that the color is hard to extract. For example, after a thionine-sensitized oxidation of tyramine, thc alginate phase was acidified with HCI and extracted with 75% ethanol. Nearly all the remaining dye was removed, but only part of the brown products, a part with bands at shorter wavelengths. Alginate in nature is usually associated with brown phenolic substances, which are removed commercially by bleaching or by ethanol extraction (Smidsrmd et al., 1963).
RADIATION
ABSORBED, EINSTEIN / LITER
Figure 5. Photosensitized oxidation of tyramine in monochromatic light absorbed by bound monomeric form of dye. Increase of product absorption at 320 nm for niethylene blue with alginate).( and with glucuronate (0).and at 326 nm for phenosafranine with alginate (H) and glucuronate (O), in 5 cm path length cell. Composition of solutions: 10-4M tyramine HCI, 3.75 x 10-4N tetraethylammonium alginate or glucuronate, 7.4 x lO-'M methylene blue or 6.1 x 10-fiM phenosafranine; pH adjusted to 9.5 initially with tetraethylammonium hydroxide. Irradiation with 660 nm (methylene blue) or 520 nm (phenosafranine) light.
5
The photosensitized oxidation of tyrosincs- 1
It is not clear whether the brown products are retained by insolubility or entanglement, or whether they are bound to the polyelectrolyte by ionic forces or men by primary chemical bonds. But it is plausible to suppose that the ionic nature of substituents will affect the strength of association and thereby contribute toward explaining the effect of the presence of polymer. Reuction in blue light. The extent of oxidation was often followed by the increase of product absorption at 320nm as well as by the decay of phenol bands further into the UV, because of some uncertainty as to the extent of product absorption in the region of the phenol bands, and the effect on them of pH changes. It was early established that in the case of tyr (X x lW4M, pH 8.38, 660nm irradiation), the two absorption changes were quite accurately linearly correlated for at least 80% of the reaction. With tyramine and other kinds of radiation, however, it is possible t o dissociate the rise at 320nm from the decay of the phenol bands. If broad bands of complexes between thionine and oxidation products are responsible for much of the increase in visible and near UV absorption in white light, then the effect of polymer ought to be evident even at pH > 9 when the reaction is run in blue light, absorbed more by the inert bound dye dimers and by the supposed dye-product complexes than by monomeric dye. This prediction is supported by an experiment using thionine, the results of which are plotted in Fig. 6. The decrease of tyramine and increase of 320nm absorption are both slower with alginate than with glucuronate, because with the former, much of the light is absorbed by photochemically inactive dimers. The ratio of the quantum yield of tyramine loss with glucuronate to that with alginate is nearly 2.6 throughout the reaction. If we define an effective dose for tyramine loss in the preslo4
:1
ence of alginate as (1/2.6) of the absorbed dose, and plot the data against it, the dashed curves of Fig. 6 result. The dashed curve for tyramine loss with alginate now nearly coincides with that for glucuronate, but the curve for 320 nm absorption rises more rapidly than for glucuronate, almost from the beginning. In the presence of alginate, blue light is relatively more effective in sensitizing increased absorption at 320nm than in sensitizing tyramine loss. The same phenomenon is seen in the methylene blue sensitized oxidation of tyramine in monochromatic light (Fig. 7). The shorter the wavelength, the greater the build-up of near UV absorption for the same amount of tyramine oxidized. The divergence becomes more apparent after the reaction has proceeded to an extent of -3o”/,. Mechanism. Although we are not principally concerned with the mechanism of phenol oxidation in
-A [ t y r l X 1 0 4 , M
Figure 7. Methylene blue sensitized photooxidation of tyramine in light of different wavelengths. Composition of solution as in Fig. 2. Comparison of absorbance increase at 320 nm with loss of tyramine.
[ t y r ~
0.4
0.2 0 10
I
0.06
0.04 0.02
lCp0
Figure 6. Photooxidation of tyramine in the presence of glucuronate and alginate, sensitized by thionine in light through Corning 4-94 blue filter. Composition of solutions: 3 x 10-4N tetraethylammonium glucuronate or alginate, 10-4M tyramine HCI, 5 x 10-6M thionine, 2 x 10-4M tetraethylammonium carbonate buffer to establish initial pH 9.28-9.30. Tyramine concentration with glucuronate (0)and alginate ( 0 )calculated from absorbance changes at 239 and 218nm and plotted (left ordinate scale) against total light energy absorbed, J./ I . Product absorbance changes at 320 nm with glucuronate (0) or alginate (B) similarly plotted (right ordinate scale). Dashed lines are data for alginate plotted against effective dose for tyramine destruction, i.e. total dose divided by 2.6.
G. R. SEELYand R. L. HART
6
this paper, a few remarks concerning its probable (Matsuura et al., 1969; Weil, 1965), the known facts course seem to be in order. Acceleration of the rate in the oxidation of tyrosines and other phenols with of tyramine oxidation in DzO supports the assump- an unsubstituted ortho position can be readily tion of a singlet oxygen reaction. Rates of singlet explained if 0, ('Ag) adds across an electron-rich oxygen reactions may be increased by as much as double bond next to the phenolate oxygen, in a mantenfold in D,O over HzO, owing to the longer life- ner like that proposed for eneamines and vinylethers time of 0, ('Aq) in the former solvent. Although the (Bartlett and Schaap, 1970; Foote and Lin, 1968; conditions (5 x 10-4M tyramine, 4.73 x 10-3M Mazur and Foote, 1970). In Fig. 8 the intermediate KHC03, 1.4 x 10-5M methylene blue, pH 10.2, dioxetane (11) opens to 4-R-5-formyl-2,4-penta660nm) were not optimal for expression of a D,O dienoate (111),which adds a second O2 ('Ag) in a reaceffect, the rate of tyramine loss was greater initially tion not sensitive to pH. Decarboxylation of (IV) to by a factor of 2 in DzO, and as tyramine approached give the unsaturated dialdehyde (V) accounts for two depletion, the factor increased to 4-5. These factors O2 taken up and one CO, released. The reactions are no less than what would be expected from the available to (V) are numerous, but one way in which tyramine concentration dependence of quantum yield condensation could occur to build up extended conjuin Fig. I. Further oxidation of products was acceler- gated carbonyl systems is shown. Compounds like ated in DzO, more so perhaps than of tyramine itself, (VI), in enolate form or oxidized to carboxylate, could but the extent of bleaching of the dye was unchanged. be expected to form strong complexes with thionine, Although radical oxidations of phenols are well and at the same time, or alternatively, to bind known, and radical formation has been proposed as strongly to an anionic polyelectrolyte through posia first step in photosensitized oxidation of phenols tively charged groups R.
6R
k
R
m
m
TI
I
I+
-HCOj
CHO CHO r
1
L
J
H20
P
PI Figure 8. The proposed reaction scheme
REFERENCES
Ashkinazi, M. S., I. A. Dolidze and V. A. Yegorova (1967) Biufiz. 12,427--432[Eng. Transl., 488-4931. Bartlett, P. D., and A. P. Schaap (1970) J . Am. Chem. Soc. 92, 3223-3225. Bellin. J. S . (19681 Photochem. Phorohiol. 8. 383-392. Bellin, J. S., and C. A. Yankus (1968) Arch. Biochem. Biophys. 123, 18-28. Foote, C. S., and J. W.-P. Lin (1968) Tetrahedron Letters 3267-3270. Martin, R. B. (1971) J. Phys. Chem. 75, 2657-2661. Matsuura, T., N. Yoshimura, A. Nishinaga and I. Saito (1969) Tetrahedron Letters 1669-1671. Mazur, S., and C . S. Foote (1970) J . Am. Chem. Soc. 92, 3225-3226. Merkel, P. B., and D. R. Kearns (1972) J. Am. Chem. Soc. 94, 1029-1030, 72447253. Nilsson, R., P. B. Merkel and D. R. Kearns (1972) Photochem. Photohiol. 61, 117-124. Oster, G., and J. S . Bellin (1957) J . Am. Chem. Soc. 79, 294298. Rinaudo, M., M. Milas and M. Lafford (1973) J. Chim. Phys. 70, 884-887. Seely, G. R., and R. L. Hart (1974) Macromol. 7, 706-710. Sluyterman, L. A. Ae. (1962) Biochim. Biophys. Acta 60, 557-561. Smidsrgd, O., A. Haug and B. Larsen (1963) Acta Chem. S c a d . 17, 1473-1474. Spikes, J. D., and M. L. MacKnight (1970) In Photochemistry of Macromolecules, (Edited by R. F. Reinisch), pp. 67-83. Plenum Press, New York. Wed, L. (1965) Arch. Biochem. Biophys. 110, 57-68. Weil, L.. W. G. Gordon and A. R. Buchert (1951) Arch. Biochem. Biophys. 33, 90--109.
