FREE RADICAL FORMATION IN RIBOFLAVIN COMPLEXES" BY IRVIN ISENBERG- AND ALBERT SXENT-GYGRGYI INSTIT`[JTE FOR MUSCLE RESEARCH, MARINE BIOLOCHCAL LABORATORY, WOODS HOLE, MASSACHUSETTS Communicated June 30, 1958 If an aqueous solution containing 10 -3 M tryptophan and 1O-3 M riboflavin-5' phosphate is frozen, the resultant sample has a red appearance instead of the yellow shown by riboflavin alone. This red form can also be seen at room temper- ature if more concentrated solutions are used-for example, lo+ M tryptophan 858 BIOCHEMISTRY:ISENBERG AND SZENT-GYORGYI PROC. N. A. 8. and low3 M riboflavin. Proteins that contain tryptophan, as well as certain trypto- phan derivatives, also form red samples with riboflavin. This report will present evidence that this red form is a tryptophan-riboflavin complex in which a riboflavin molecule has taken up one election from the trypto- phan. The riboflavin is then in a semiquinoid form, a free radical. This form of riboflavin was first studied by Kuhn and Wagner-Jaureggl and afterward by Michaelis and his co-workers. 2-4 Since the work of Michaelis, the reduction of riboflavin to its semiquinone form has hardly been studied. Haas reported that when old yellow enzyme was re- duced by hydrosulfite in the presence of TPN, a red color appeared, and he attrib- uted this red color to free radical formation. More recently, BeinerV has studied the spectral properties of the semiquinoid form of riboflavin, and Beinert?, 8 and Ehrenberg and Ludwig9 have further studied the occurrence of free radicals in flavo-protein catalysis. MATERIALS AND METHODS We have found it convenient to study the red riboflavin complex in two ways. For qualitative visual observations, aqueous solutions were frozen in dry ice. In this form, complex formation is greatly enhanced over room-temperature condi- tions, and direct and easy observations may be made as to which samples do con- tain complexes and which samples do not. For quantitative studies the samples were studied at room temperature in phos- phate buffer, pH 6.90, by means of a Beckman DKI recording spectrophotometer. The sample compartment cuvette contained the mixture to be studied-say, ribo- flavin and tryptophan-while the reference cuvette contained riboflavin of a molar- ity equal to the sample. In this way, by balancing out most of the riboflavin ab- sorption, the complex absorption becomes a major component of what is recorded by the instrument. It should be noted that this technique does not lead to an exact balancing-out of the riboflavin. To achieve an exact balance, one would need to use a riboflavin concentration equal to the free riboflavin concentration in the sample cell rather than the total concentration there. It should therefore be recognized that the peak at 500 rnF reported below is not the true peak for ab- sorption by the complex. The true peak will lie at somewhat shorter wave lengths. It also follows that the molar extinction at the peak for the complex absorption will be somewhat high& than that at 500 rnp reported below. RESULTS a) Qualitative Observations.-10-3 M riboflavin-5'-phosphate was frozen in the presence of a number of a.mino acids, tryptophan derivatives, and several trypto- phan-containing proteins. The results are shown in Table 1. A minus sign indi- cates no shift in color from a riboflavin control, while a plus sign means a shift toward the red. b) Quantitative Obseruutions.-Upon balancing low4 M riboflavind'-phosphate with varying concentrations of tryptophan against 10 -4 M riboflavin-5'-phosphate, all in phosphate buffer pH 6.90, a peak at 500 rnp was obtained. Riboflavin-Fi'- phosphate in 10 per cent HCl, reduced to a red form by sodium hydrosulfite, also yielded a peak at 500 rnp when balanced against unreduced riboflavind'-phosphate. VOL. 44, 1958 BIOCHEMISTRY: ISENBERG AND SZENT-GYdRGYI TABLE 1 Partner of 10-a M Riboflavin-5'-Phosphate lOma M tryptophan 10-a M tyrosine ;;I: g g!$$a~ lOma M 5-hydroxytryptamine (seroto?in) l-Benzyl-2-methyl-5-methoxl-N,N-dlmethyl tryptamine hydrochloride l-Benzyl-2-m.ethyl-5-methoxy-tryptamine hydrochloride 1-Benzyl-2,5-dimethyl serotonin 1-Benzyl-2,5dimethyl bufotenine l-Methyl medmain d-Lysergic acid diethylamide tartrate (LSD-25)t 1 Per cent bovine plasma albumin 1 Per cent myosin 0.3 Per cent actomyosin 3 Per cent Casein Appearance at + - 859 -78O C. * The authors would like to thank Dr. D. W. Wooley for giving them samples of these compounds. * The authors thank the Sandor Pharmaceutical Company, Hanover, New Jersey, for a sample of LSD-26. This agrees with the value of 503 rnp reported by Beinert! for the difference maxi- mum of the semiquinoid form in 1 M HCl. Since the red form is stable only in strong acid so1ution,2 the tryptophan must, in some way, stabilize this semi- quinoid form at neutral pH. The most obvious way to do this is by complexing. It will be assumed that a complex forms in a one-to-one fashion, so that Tryptophan + riboflavin ;--) (reduced riboflavin, oxidized tryptophan complex) Let t = total tryptophan concentration in the sample; T = total riboflavin concen- tration in the sample; c = concentration of the complex; k = dissociation constant for the complex; and K = l/k It will be assumed that the extinction of the complex bears a simple Beer's law relationship to the complex concentration. Since we are measuring a difference spectrum and since we have assumed that one molecule of semiquinoid riboflavin forms for every molecule of riboflavin that reacts, we have E = A&l, where E = recorded extinction at 500 rnp; Ae = molar extinction of the complex at 500 rnp minus the molar extinction of oxidized riboflavin at 500 rnp; C = concen- tration of the complex; I = cell path length, in our case 1.000 cm. The tryptophan concentrations used were always at least twenty times that of the riboflavin concentrations, and therefore they necessarily greatly exceeded the com- plex concentrations. Under these conditions t(r - c) -- = C ]c = 1 K or 1 - = 1rK; - K. t Thus a plot of l/t versus l/E should yield a straight line. The intercept on the abscissa yields l/A& while the intercept on the ordinate yields -K. Figure 1 shows some typical data for tryptophan and also typical data for sero- 860 BlOCHEMIS1'RY: ISENBERG AND &ZEN?`-GYdRGYI PROC. N. A. $4. IO 20 J- 30 E 40 FIQ. l.-Plot of inverse concentration of riboflavin-5-phosphate partner, l/t, versus inverse extinction, l/E. the reference cell. The riboflavin-5'-phosphate concentration was 10m4 M in hoth the sample and tonin. It can be seen that straight-line plots do give reasonable representations of the data. A more striking result, however, is that, within experimental error, both plots have the same intercept on the 1rE axis. This is further confirmation that the measured absorption is due to a riboflavin semiquinoid, the semiquinoid having a unique molar extinction coefficient whether in the presence of tryptophan or of serotonin. On the other hand, the two plots have different l/t intercepts. This means that serotonin and tryptophan complex with the riboflavin semiquinone with different strengths, serotonin complexing about seven times as strongly as riboflavin. The data yield AE = 2300 liters per mole cm. ; K for tryptophan equals 60 liters per mole, while K for serotonin equals 400 liters per mole. Since E for oxidized riboflavin at 500 rnp is approximately 2,200, we obtain an approximate value of 4,500 for the molar extinction coefficient of the complex. It is noteworthy that solutions of riboflavin that contain the complex appear to be less fluorescent than pure riboflavin solution. DISCUSSKON There has been a recent upsurge of interest in free radical formation in biological systems, due, in part, to the development and application of the technique of para- magnetic resonance.g* lo It is therefore of some interest that a semiquinoid form of riboflavin can be produced and stabilized at neutral pH by the simple addition of tryptophan or of proteins containing tryptophan or of molecules resembling trypto- phan. Indeed, as shown by serotonin, other molecules may reduce and form com- plexes with riboflavin much more strongly than tryptophan itself. It is clear that some of the work that has been done on free radical formation in flavoproteins may be subject to reinterpretation. For the tacit assumption has VOL. 44, 1958 BIUCHEMISY'RY: ISENBERG AND fXENI'-GYGRGYI 861 always been made that the flavoprotein itself contained no sen~iquinoid form be- fore the addition of a reducer. This now appears unjustified. This realization may also clarify the absorption spectra of certain flavoproteins. Consider the absorption spectrum of old yellow enzyme, as given, for example, by Ehrenberg and Ludwig9 The spectrum has two peaks that appear to be similar to the ab- sorption peaks of riboflavin. In addition, there is a marked shoulder at about 490 rnp. The work reported here suggests that this shoulder results from the ab- sorption of ribo~avi~l-5'-phosphate in a semiquinoid form stabilized by the trypto- phan in the protein. It is possible that complex forn~at'ion may be much stronger in tissue than in vitro. Just as freezing enhances complex formation in a test tube, so it is possible that in closely packed tissues constituents, such as mitochondria, a similar lattice-ordered structure of the water favors association.ll Within the cell the complex formation might also be favored by additional links, as assumed by Nygaard and Theorell"* for the binding of the flavin adenine nucleotide in the old yellow enzyme. So it seems possible that the complex formed by flavins with the proteins, via their tryptophan, has a major biological in~portance. This assumption is supported by the fact that various tissues contain strongly bound flavins in high concentration, The authors are impressed by the very great quantity of strongly bound flavin in the liver. In fact, the brown color of this organ seems to be due to the flavin radical formed in the charge transfer with the protein. The strong brown color of the liver suggests that a considerable part of the structural proteins is present as a Aavin complex, containing the flavin in a free radical condition. It, also follows that the tryptophan or protein, complexing with the ribo~avi~~ and donating an electron to it, is also present as a free radical. The charge transfer between protein and riboflavin might have various biological consequences. So, for instance, if the electron, donated by the protein to the flavin, is drawn from an energy band,13 the resulting hole might make the band conductant, This change would be accentuated if the flavins passed the electron thus accepted to the oxidation system. If the electrons thus lost by the protein were replaced by electrons given off by the dehydrogenated ~netabolites, DPN or TPN, this would mean that the electron transport goes ~~~o~g~ the protein molecule itself and does not take place merely on the surface of the protein, as hitherto believed. In any case, if complex formation between tryptophan and flavin plays a major biological role, then one would expect that substances which form a similar complex with flavins and have a relatively high affinity to it interfere with the normal course of reactions and thus show a definite pharmacological effect. It is int,eresting to note that serotonin has a high affinity for riboflavin, and so has lysergic acid and bufotenine. These substances, which form similar complexes, have been implicated in normal and pathological mental activity. Possibly the described reactions might give an explanation of their mechanism of action on the molecular level. In this connection it is interesting that the other drugs listed in Table 1 have also been implicated in the problems of the activity of the centra1 nervous system. * This research was sponsored by the grant H-2012R of the National Heart Institute, a grant from the Commonwealth Fund, the National Science Foundation, the American Heart Associ- ation for the Aid of Crippled Children, and the United Cerehral Palsy Associations. 1 R. Kuhn, and T. Wagner-Jaureg~, Ber. Chem. Ges., 67,361,1934. 862 BIOCHEMISTRY: TALALAY ET AL. PROC. N. A. 8. p L. Michaelis, M. P. Schubert, and C. V. Smythe, J. Biol. Chem., 116, 587, 1936. 8 L. Michaelis and G. Schwarzenbach, J. Biol. Chem., 123,527, 1938. 4 L. Michaelis, The Enzymes, Vol. II, Part I, ed. J. B. Sumner and K. Myrbtick (New York: (Academic Press, Inc., 1951). 6 E. Haas, BiochemZ., 290,291,1937. * H. Beinert, J. Am. Chem. Sot., 78,5323,1956. 7 H. Beinert, Biochim. et Biophys. Ada, 20,588, 1956. 8 H. Beinert, J. Biol. Chem., 225,465, 1957. 9 A. Ehrenberg and G. D. Ludwig, Science, 127,1177,1958. lo B. Commoner, J. J. Heise, B. B. Lippincott, R. E. Norberg, J. V. Personneau, and J. Town- send, Science, 126,57,1957. 11 A. Saent-Gyorgyi, Bioenergetics (New York: Academic Press, Inc., 1957). If A. P. Nygaard and H. Theorell, Acta Chem. Stand., 9,1587, 1955. 1* A. Szent-Gyorgyi, Science, 93,609,1941.