Electron Transfer in the Flavin Mononucleotide System - Journal of the

Journal of the American Chemical Society. Advanced Search .... One-electron reactions in biochemical systems as studied by pulse radiolysis. II. Ribof...
1 downloads 0 Views 332KB Size
1056

Electron Transfer in the Flavin Mononucleotide System James H. Swinehart Contribution f r o m the Department of Chemistry, University of California, Davis, California. Received October 22,1965 Abstract: A kinetic investigation of partially reduced aqueous flavin mononucleotide solutions at pH 5 is reported. O n t h e basis of the thermodynamic a n d activation parameters the mechanism is formulated in terms of the equilibria FMN

kiz + FMNH2 e (FMN.FMNH2) kzi

(iv)

(FMN.FMNHz)

ka3 e 2FMNH.

(v)

k32

where FMN, FMNH2, ( F M N . FMNH?), a n d FMNH. refer to t h e oxidized, reduced, dimeric, a n d free-radical forms of flavin mononucleotide, respectively. T h e enthalpy a n d entropy for reaction v, K23 = [FMNH.I2/[(FMN* F M N H 2 ) ] , are 9.3 i. 0.3 kcal/mole a n d 15 i 2 eu, respectively. The activation parameters, AH* a n d AS*, for k32 are 5.6 + 0.5 kcal/mole a n d -9 i 2 eu. The activation parameters for kZ3,as calculated from the thermodynamic parameters for K23and activation parameters for k32,are in fair agreement with those experimentally observed. T h e is found t o b e 2.3, which is consistent with t h e transfer of charge taking place in (v) a s a value of k32(HzO)/k32(D20) hydrogen-atom transfer.

In

a previous report a preliminary study of the kinetics of the partially reduced flavin mononucleotide system in the pH range 3.9 to 5.2 using the temperaturejumprelaxation method was described. A complete kinetic investigation of the system in this pH range is reported with special emphasis being placed on a clarification of whether or not the dimeric species, (FMN. FMNHz), is involved directly in equilibria with both the oxidized and reduced forms, FMN and FMNH2, and the free radical, FMNH . . The mechanism by which the transfer of electrons takes place is also investigated to determine whether the mechanism is one of direct electron transfer or of hydrogen atom transfer.

Experimental Section Chemicals. Flavin mononucleotide (riboflavin 5’-phosphate sodium salt) W P S obtained from Mann Research Laboratories, New York, N. Y., and was used without further purification, Especially pure FMN, greater than 99%, was obtained from Dr. L. L. Ingraham, Department of Biochemistry, University of California, Davis, Calif. This sample was obtained by a paper chromatographic separation using a butanol-water-acetic acid mixture (4:5:1). The kinetics proved to be independent of the sample. Sodium dithonite (Merck, reagent), acetic acid (Mallinckrodt, analytical reagent), sodium acetate (Baker and Adamson, (Mallinckrodt, analytical reagent), and hydrochloric acid, 37 reagent), were used without further purification. D10 was purchased from Bio-Rad Laboratories, Richmond, Calif. All solutions were prepared with doubly distilled water. Preparation of Solutions. Solution preparation is dewibed in a previous paper and need not be repeated in detail here.’ Acetateacetic acid was used as the buffer and the final solutions were 0.2 M in total acetate. Apparatus. The pH was measured with a Leeds and Northrup Model 7401 pH meter. Standard buffers were used for calibration. The spectrophotometer used was a Cary Model 14 recording spectrophotometer with a thermostated sample compartment, which was nitrogen flushed. One-centimeter matched cells were used. ‘Temperature control was good to 0.3”. l h e theory of relaxation methods and a technical description of the temperature-jump-relaxation equipment is summarized elsewhere. ’,* The instrument used in this investigation was manufactured by the Messanlagen Studiengesellschaft m.b.H., Goettingen, Germany. (1) J. H. Swinehart, J . Am. Chem. Soc., 87, 904 (1965). (2) M. Eigen and L. deMaeyer in “Technique of Organic Chemistry,”

Vol. VIII, A. Weissberger, Ed., Interscience Publisher, Inc., New York, N. Y., 1963, Part 11, Chapter XVIII.

Measurements. The two absorbancy measurements important for a description of the system are the absorbancy at 840 mM, A84orn,, and the corrected absorbancy at 570 mp, Asiom,. The shoulder of the band which maximizes at 840 mp makes a major contribution to the total absorbancy at 570 m p ~ The corrected absorbancy at 570 mp was calculated by assuming the 840-mp band was symmetrical and subtracting this contribution from the total 570-mp band. The contribution to the absorbancy at 570 mp from the oxidized and reduced forms is small. The error recorded in measurements is the maximum error taken from extreme measurements divided by four. This gives a reasonable estimate of the probable error.3

Results and Discussion The pK values of the various species present in partially reduced flavin mononucleotide solutions are: FMN, -0.14 (in HC1) and 9.8; FMNH., 1.2 (riboflavin) and 7.3;5 and FMNH2, 6.8. At pH 4.7 the possible equilibria are

e (FMN);: e (FMNH*)% FMNH? + F M N H . e (FMNH2.FMNH.) F M N + FMNH;: (FMN.FMNHz) 2FMN 2FMNHz

(i) (ii) (iii)

ki2

(iv)

k2i

k23

(FMN’FMNHz)

2FMNH.

(VI

k32

where the species in equilibria iv and v may be in other combinations (e.g., F M N FMNHz $ 2FMNH., etc.). Calculations using the data of Gibson, Massey, and Atherton6 show that at the greatest FMN and FMNHzconcentrations used less than 10 % of the species are present as the dimers (FMN)z and (FMNHdz. Thus equilibria i and ii are of minor importance. Figure 1 is a plot of Ag,,m+2YS. A,,,,, at 298°K for pH values 2.7 and 5.9. The species contributing to the cor-

+

(3) R. Livingston, “Physic0 Chemical Experiments,” 3rd ed, The Macmillan Co., New York, N. Y., 1948, p 44. (4) Determined spectrophotometrically by the author. (5) There is considerable disagreement on the va!ile of this pK. L. Michaelis and G. Schwarzenbach, J . B i d . Chem., 123, 527 (1938), reported a value of 6.5 for riboflavin. B. Holmstrom, Phorochem. PhorobioL, 3, 97, (1964), reported the value recorded here. Other workers have reported values as high as 8.3 (riboflavin): R. D. Draper and L. L. Ingraham, private communication. (6) Q. H. Gibson, V. Massey, and N. M. Atherton, Biochem. J . , 85 369 (1962).

Journal ojtiie American Chemical Society / 88:5 / March 5, 1966

1057

'T 12

0.08

1

e

0 0

o

0

e 0

0.

e

0

02

0.6

04

08

IO

*570mr

Figure 2. Plot of the relaxation time, I / T vs. 4A570mll at 300, 292, and 282 OK (accuracy of temperature, i1 "C). *840rnr

Figure 1. Plot of absorbancy at 570 mp squared vs. absorbancy at 840 mp at pH values 2.7 and 5.9 (298°K).

rected absorbancy at 570 mp are FMNH. and (FMNHr FMNH .), and at 840 mp the absorbancy is attributed to (FMN.FMNHz).6,7 There is a linear relation between A670mp2 and A840mpin the region in which the [FMN] is greater than or equal to the [FMNH2]. This indicates that equilibrium iii is of minor importance compared to equilibria iv and v. Values of A570mp2/As40mpas a function of temperature are contained in Table I. Table I. Equilibrium Data for Reaction v at p H 4.7

A840mr

17.5 i 1.0 15.0 =t1.0 8 . 5 i 1.0 6.1 i0 . 6 5 . 1 i 0.5 3.7 i0.3 2.9 i0 . 3

314 31 3 299 293 29 1 283 282

where Klz = [FMN .FMNHz]/([FMN][FMNHz]).~If Klp([FMN] [FMNH2]) is much greater than 1, 1 / ~ = kZ3 4k3z[FMNH.]. Theslope of ~ / T V S 4[FMNH.] . yields k32 and the extrapolated intercept is kZ3. The only other reasonable mechanism which would lead to a linear relationship between 1 / and ~ A570mr, [FMNH-1, is one in which the equilibria are

+

+

FMN.FMNHz

102, Ajlomfiz

as discussed in a previous paper, is one in which reaction iv equilibrates rapidly compared to (v).' The form of 1/ T for such a mechanism is

Concn conditions

b C

FMN

eF M N + FMNHz

+ FMNH?

kra

ZFMNH.

(vi) (vii)

ka4

and (vi) equilibrates rapidly compared to (vii). This mechanism leads to a relationship

U

C

U C

The problem becomes a question of whether F M N . FMNHz is a precursor to F M N H . . The results of a [FMN] > [FMNH2]and [FMN] = [FMNH2]. [FMN] > Gibson, et al., indicate that the dimer is a precursor FMNH21. [FMN] = [FMNH?]. to the free radicaL6 The values for the activation parameters of the rate constants obtained lead to the When the In (A570mfi2/&0mfi)is plotted as a function same conclusion. of the reciprocal of the absolute temperature, AH is The values of k32/~570~~, where E570mr is the extinction found to be 9.3 f 0.3 kcal/mole. coefficient of F M N H . at 570 mp obtained from Figure It has been shown in previous work that a linear 2, are (12.8 It 1.0) X lo3 sec-' cm (300"K), (10.0 k relation exists between the relaxation time, I/T, and 1.0) X I O 3 sec-' cm (292OK), and (6.3 f 1.0) X l o 3 sec-' cm (282°K). As determined from Figure 3 &omfi.' Figure 2 shows a plot of 117 vs. 4 X A570mp at three temperatures. It is clear from the accumulated the value of AH* for ksz is 5.6 0.5 kcal/mole. The data that there is not a linear relation between 1 / ~ enthalpy of activation for k23,as calculated from A H and ABIOmror ([FMN] [FMNH2]). Experiments for reaction v, is 14.9 f 0.8 kcal/mole. Fair agreement where the ([FMN] [FMNH2]) was made five times is obtained between this value and AH* determined the concentration normally used, while A570m, was directly using extrapolated kz3 values from Figure 2. The values of k23 are (5.5 f 1.0) X lo3 sec-1 (300"K), kept small, led to no abnormal behavior when I / T was plotted against 4 X (point 1 in Figure 2). (3.5 f 1.0) X lo3 sec-' (292"K), and (1.5 f 1.0) X Thus it can be concluded that the mechanism which lo3 sec-l (282°K). A line as drawn through In relates FMN, FMNHz, FMN.FMNHz, and F M N H . (k23/T) vs. 1 / values ~ in Figure 3 corresponds to 14.9 must lead to a relationship in which l / ~increases kcal. Since the values for k23are obtained from a long linearly with A570mror [FMNH.]. Such a mechanism, extrapolation, the error will be large. b

*

+

+

(7) H. Beinert,J. Am. Chem. Soc., 78, 5323 (1956).

(8) The form for 1 / is ~ incorrect in ref 1 and is corrected here.

Swinehart / Electron Transfer in the Flavin Mononucleotide System

constants k23 and k22 at 298°K is 8 X l e 4 M . A S is calculated to be 15 f 2 eu. The value of K23 is in fair agreement with the value of about l e 4 determined The by Michaelis, et al., for glucoroalloxazene. are at p H values of l e 6Mdetermined by Gibson, et 6.3, which is near the pK of FMNHz, and thus no valid comparison can be made. Table I1 summarizes the results obtained at pH 4.7. Previous work has shown the same mechanism to be operative over a pH range from 2.9 to 5.5. The data in

6.0 8.0

Table 11. Rate and Equilibrium Information at pH 4.7 and 298°K

Valueat 298°K kpl

3.3

3.4

3.5

kt3

3.6

K23

103 x f ( O K - ' )

Figure 3.

kl? ktl

Plots of In (k23/T) and In (k32/T~670ml.l) vs. 1/T.

k12

(6.0 zlc 0.2) X 106 M-1 sec-1 5 X 103sec-l 8 X 10-4 M > l o g M-' sec-' > l o 6 sec-l > i o 3 M-'

AH* or AH, kcal/mole 5.6 & 0.5 14.9 i.0 . 8 0 9 . 3 =!= 0 . 3

AS* or AS. eu -9

j=

2

6 i2a 15 =t2

From a combination of K23 and k32values.

If the extinction coefficient of F M N H . at 570 mp is assumed to be 3050 M-l ~ m - - l ,the ~ value of AS* for kY2is -6 f 2 eu. This value is consistent with the mechanism proposed. [(FMN . FMNH2)] can be calculated from the equilibrium constant Kz3and [FMNH -1, which equals A670mp/3050 M-' cm-l. Values of K12 calculated from the initial concentrations of FMN and FMNHz and the concentrations of (FMN FMNHz) and F M N H . are inconsistent. The values of KI2 decrease with increasing total flavin mononucleotide indicating that the concentrations of (FMN . FMNHz) and FMNH . subtracted from the initial concentration of FMN and FMNH2 are too small. It is quite clear that an extinction coefficient which will satisfy the experimental conditions must be considerably smaller than 3050 M-' sec-I. A wide variety of extinction coefficients for FMNH . are reported in the literature (7009 to 13,600 M-' cm-1).6 An extinction coefficient at the lower end of this range (about 500 M-' cm-l) yields both consistent values of Klz and values greater than I O 3 M-' which are necessary in order to satisfy the condition that Kl.XIFMN] [FMNH2]) is greater than 1. A calculation of K12is very sensitive to the value of c~~~~~ in this region, so the value selected is accurate to cm-'. The value of AS* for kS2calculated ~ 2 0 M-' 0 using e670ml.l= 500 M-l cm-I is -9 =k 2 eu. Since reaction iv is rapid compared to (v) and cannot be observed experiential, kzl > lo6 sec-l. As a result of K12 being greater than lo3, k12 is greater than lo9 M- sec- l. The value of K23 calculated from the ratio of the rate

+

Table I1 are consistent with the mechanism postulated: reactions iv and v. The entropy of activation for a reaction in which the dimerization of uncharged species takes place is generally negative. Such a reaction is represented by the rate constant kS2which has associated with it an entropy of activation of -9 f 2 eu. For the over-all equilibrium v, A S is 15 f 2 eu which is reasonable for the process of this type. The assumption is made that interactions between the solvent and the species involved, (FMN . FMNH2) and F M N H . , are small. This assumption seems reasonable since the molecules are uncharged and flavins are generally very insoluble in water ( M for riboflavin) unless a polar group is added. The question as to the point at which electron transfer takes place (reaction iv or v) and by what mechanism it proceeds (electron or hydrogen atom transfer) is pertinent. From experiments in D 2 0 at 292"K, the ratio of k32(H20)/k32(D20) is found to be 2.3 and = 1.5. The latter result is less kzS(H20)/k2:,(D20) reliable than the former because of the extrapolation necessary to obtain k23. This evidence is consistent with a hydrogen atom transfer which takes place in equilibrium v. Acknowledgment. This investigation was supported by Public Health Service Research Grant G M 11767-02 from the National Institutes of Health. (10) L. Michaelis, M. P. Schubert and C. V. Smythe, J . Biol. Chem., 116, 507 (1936).

(9) B. Holmstrom, Bull. SOC.Chem. Belges, 71, 869 (1962).

Journal of the American Chemical Society

1 88:5

March 5 , 1966

( I f ) L. Michaelis and G. Schwarzenbach, ibid., 123, 527 (1938).