Effect of the Protonation State of the Amino Group on the *OH Radical

---

J. Phys. Chem. 1985,89, 3139-3144

3139

Effect of the Protonation State of the Amino Group on the *OH Radical Induced Decarboxylation of Amino Acids in Aqueous Solution Jorg Monig,* Rita Chapman, and Klaus-Dieter Asmus Hahn- Meitner-Institut fur Kernforschung Berlin, Bereich Strahlenchemie, Berlin 39, Federal Republic of Germany (Received: December 1 1 , 1984)

Decarboxylation of a-amino acids by .OHradicals has been found to be very effective in basic aqueous solutions, Le., under conditions where the amino group is unprotonated and a free electron pair at nitrogen is accessible. The C 0 2 yield measured in y-radiolysis experiments of 13 simple aliphatic a-amino acids accounts for ca. 60-100% of the .OH radicals available. The decarboxylation mechanism is considered to be initiated by an .OH addition to the free electron pair at nitrogen. This .OH adduct or a radical cation resulting thereform is suggested to decarboxylate spontaneously leaving a-amino radicals. These strongly reducing radicals have been identified through their reaction with added halothane, CF,CHCIBr. Competitive processes to the -OH attack at nitrogen are H-atom abstractions from anywhere in the carbon skeleton of the amino acids. The relative reaction rates are shown to depend on the type of the C-H bonds, the degree of C-H bond activation by neighboring groups, and to a significant extent also on structural effects.

Introduction The increased interest in the effects of radiation on living matter shown in the last few decades has stimulated many investigations into the radiolysis of amino acids. They have been subjected to many different kinds of irradiation, including highly energetic X-rays,'-3 y - r a d i a t i ~ n , ~and - ' ~ pulse radiolysis"J2 both in dilute aqueous solution and in the solid state.l3,l4 Product studies following y-radiolysis of aqueous solutions of amino acids have uncovered that hydrated electrons, eaq-, and hydroxyl radicals, .OH, (the two principal primary radicals upon H 2 0 radiolysis) lead to degradation of amino acids.I5*l6 Rate constants for these reactions have been measured by direct and competition metho d ~ . ' ~Pulse - ~ ~ radiolysis was employed to obtain spectra of the transient radicals produced,"J2 while flow experiments using in situ ESR measurements allowed the characterization of the radical species involved.21-25 (1) Stein, G.; Weiss, J. J . Chem. SOC.1949, 3256. (2) Maxwell, C. R.; Peterson, D. C.; Sharpless, N. E. Radiat. Res. 1954, 1, 530. (3) Sharpless, N. E.; Blair, A. E.; Maxwell, C. R. Radiat. Res. 1955, 2, 135. (4) Maxwell, C . R.; Peterson, D. C.; White, W. C. Radiat. Res. 1955, 2, 431. ( 5 ) Weeks, B. M.; Garrison, W. M. Radiat. Res. 1958, 9, 291. (6) Kopldovl, J.; Liebster, J.; Babiclj, A. Int. J. Appl. Radiat. Isot. 1961, 11, 139. (7) Kopoldovl, J.; Liebster, J.; Babickv, A. Int. J. Appl. Radiat. Isor. 1962, 13, 617. (8) Kopldovl, J.; Liebster, J.; Babiclj, A. Int. J. Appl. Radiat. Isor. 1963, 14, 455. (9) Kopoldovl, J.; Liebster, J.; Babickg, A. Int. J. Appl. Radiat. Isor. 1963, 14, 489.

(IO) Kopoldovi, J.; Liebster, J.; Babickv, A. Int. J . Appl. Radiat. Isor. 1963, 14, 493.

(11) Neta, P.; Simic, M.; Hayon, E. J . Phys. Chem. 1970, 74, 1214. (12) MarkoviE, V.; NikoliE, D.; MiEi6, 0. I. Int. J . Radiat. Phys. Chem. 1974, 6, 227. (13) Muto, H.; Iwasaki, M.; Takahasi, Y. J. Chem. Phys. 1977,66, 1943. (14) Samskog, P. 0.;Nilsson, G.; Lund, A.; Gillbro, T. J . Phys. Chem. 1980, 84, 2819. ( 1 5 ) Garrison, W. M. In "Current Topics in Radiation Research"; Ebert, M., Howard, A., Eds.; North-Holland: Amsterdam, 1968; Vol. IV, p 43. (16) Garrison, W. M. Radiat. Res. Reu. 1972, 3, 305. (1 7) Davies, J. V.; Ebert, M.; Swallow, A. J., In 'Pulse Radiolysis"; Ebert, M., Keene, J. P., Swallow, A. J., Baxendale, J. H., Eds.; Academic Press: London, 1965; p 165. (18) Braams, R. In "Pulse Radiolysis"; Ebert, M., Keene, J. P., Swallow, A. J., Baxendale, J. H., Eds., Academic Press: London, 1965; p 171. (19) Scholes, G.; Willson, R. L. Trans. Faraday SOC.1967, 63, 2983. (20) Scholes, G.; Shaw, P.; Willson, R. L.; Ebert, M. In "Pulse Radiolysis"; Ebert, M., Keene, J. P.; Swallow, A. J., Baxendale, J. H., Eds., Academic Press: London, 1965; p 151.

0022-365418512089-3139$01.50/0

It is commonly accepted that reaction of simple aliphatic aamino acids with eaq- leads to deamination via

and in a competitive process to hydrogen atoms according to eoq-

+

/cooR-CH \NH~+

-

H*

,coo-

+

(2)

R-CH \YH2

Hydroxyl radicals, on the other hand, can in principle abstract hydrogen atoms from anywhere in the carbon skeleton. The actual site of hydrogen atom abstraction depends to some extent, however, on the protonation state of the amino group. This effect is indicated by the formation of different radical species at neutral and basic pH, respectively, and has been attributed" to a deactivating influence on the a-C-H bond by the protonated amino function. Hence, it was suggested that the positive charge of +NH3- directs the attacking electrophilic .OH radical to C-H bonds located further away from the a-carbon, Le., most of the .OHradicals react via .OH

+

,COOR-CH

N ' H3'

- F~(-H)-cH

On the other hand with the amino group being unprotonated it was assumed that reaction 4 becomes the main reaction pathway. .OH

+

R-CH

/coo\NH,

-

RC -,

./coo-

+

H20

(4)

NH2

For a number of simple aliphatic amino acids the radiolytic product yields have been measured.I-1° Ammonia, carboxylic acids, and keto acids are the main products. Interestingly, decarboxylation of amino acids was found to be of only minor importance in acid and neutral solution. For example, C 0 2 was formed with as little as G = 0.6 in irradiated oxygen-free 1 mol dm-3 solutions of glycine at p H 6.4, (G refers to the number of molecules formed or destroyed per 100 eV absorbed radiation energy). Kopoldov5 et al. suggested7that the C 0 2 resulted in fact from a secondary process. However, in their investigations into (21) Armstrong, W. A.; Humphreys, W. G. Can. J. Chem. 1967,45,2589.

(22) Taniguchi, H.; Fukui, K.; Ohnishi, S.; Hatano, H.; Hasegawa, H.; Maruyama, T. J. Phys. Chem. 1968, 72, 1926. (23) Smith, P.; Fox, W. M.; McGinty, 0. J.; Stevens, R. D. Can. J . Chem. 1970, 48, 480. (24) Neta, P.; Fessenden, R. W. J. Phys. Chem. 1971, 75, 738. (25) Paul, H.; Fischer, H. Helu. Chim.Acta 1971, 54, 485.

0 1985 American Chemical Society

Monig et al.

3140 The Journal of Physical Chemistry, Vol. 89, No. 14, 1985

the radiolysis of a-amino acids in aqueous solution the total conversion of starting material was as high as 50%. In a recent paper Hiller et a1.26 reported that .OH radical induced oxidation of the sulfur-containing amino acid methionine leads to decarboxylation almost quantitatively. In irradiated N20-saturated solutions of methionine COz is formed with G = 5.4 (i.e., = G(-OH)), irrespective of p H at p H 2 3. It has been shown that the mechanism is essentially based on an oxidation of the amino group by the primarily oxidized sulfur function.26-28 For this decarboxylation to occur it is necessary, however, that the oxidized amino function is located at the same carbon atom as the carboxyl group. On the basis of this consideration any a-amino acid should be prone to decarboxylation as well if direct or indirect oxidation of the amino group is possible. The latter may be expected whenever the amino group is deprotonated. It is therefore somewhat surprising that in the .OH radical induced oxidation of simple a-amino acids the COz formation has been reported to be essentially negligible. However, none of these studies included the p H region, in which the amino group was deprotonated. The -OH radical induced decarboxylation of simple aliphatic amino acids has therefore now been reinvestigated paying particular attention to the basic pH region. Experimental Section The amino acids were obtained from Fluka and were of the purest commercially available, usually analytical grade. They were all used as received. “Millipore”-filtered water, the quality of which corresponded to triply distilled water, was used as a solvent throughout. Reagent grade NaOH and HCIOI were used to adjust the solutions to the appropriate pH. All experiments were carried out at room temperature. The carbon dioxide analysis was performed in basic solution. After irradiation of the solution and prior to analysis an appropriate amount of N a O H was added to adjust the solutions to pH 12 in order to convert COz via C02

+ OH-

HC03-

.

c L

5 ul

1 H 2 min

Figure 1. Ion chromatogramm of an N,O-saturated solution of 5 mol dm-3 a-alanine irradiated at pH IO.

X

t

3.0 m

kU

/

(5)

(pK = 6.37) to an anionic form. The latter can be analyzed by means of high perfo-mance ion chromatography, a technique being capable of separating the carbonate ion from all other anions in solution. A Dionex 2010i ion chromatograph equipped with a 25-cm column HPICE-AS 1 in combination with a conductivity cell was used. Distilled water was used as eluent with a flow rate of 2 cm3 min-’. The very small remaining conductivity of the eluent was suppressed electronically. The elution time of carbonate with a new column was =11 min and the half-width of the peak was 0.9 min. The lower detection limit was found to be about mol dm-3, which incidentally amounts to the “blank” (1-2) X value of carbonate in unirradiated samples. This blank value originated from the N a O H stock solution, which was used for adjusting the pH. The stock was therefore kept under nitrogen. Calibration of the method was obtained by introducing known amounts of C 0 2 gas into unirradiated solutions. The final carbon dioxide concentrations were calculated by applying the ideal gas law. After subtracting the blank value (measured daily) the calibration points were found to fall on one straight line. This finding is conceivable since the detection method involves an absolute measure, Le., conductance of the carbonate ion in a nonconducting eluent. Irradiations were carried out in the field of a 6000-Ci 6oCo-y source. The dose rate was estimated by normal Fricke dosimetry and was about 800 Gy h-’. Details of the glass equipment used have already been published.29 Owing to the high sensitivity of (26) Hiller, K.-0.; Masloch, B.; Gobl, M.; Asmus, K.-D. J . Am. Chem. SOC.1981, 103, 2134.

( 2 7 ) Asmus, K.-D.; Gobl, M.; Hiller, K.-0.; Mahling, S.; Monig, J. J . Chem. SOC.,Perkin Trans. 2, in press. (28) Monig, J.; Gobl, M.; Asmus, K.-D. J . Chem. SOC.,Perkin Trans. 2 , in press. (29) Monig, J.; Krischer, K.; Asmus, K.-D. Chem.-Bioi. Interact. 1983, 45, 43.

100

200

LO0

300

DOSE I

500

Gy

Figure 2. COz yield vs. radiation dose in N20-saturated solutions of 5 X mol dm-3 a-alanine at pH 10.

the ion chromatographic method total doses could be kept low and generally were in the range of 150-600 Gy (1 Gy = 1 J kg-I = 100 rad), which corresponds to a maximum of 7% conversion of starting material. Results Oxidation of simple a-amino acids in basic aqueous solution by .OH radicals yields carbon dioxide. This is evident, for example, mol from irradiated nitrous oxide saturated solutions of 5 X dmW3alanine at pH 10.0. A representative ion chromatogram is shown in Figure 1. The bicarbonate peak ( C 0 2 is converted into bicarbonate at high pH) is well separated and can easily be quantified. The C 0 2 formation depends linearly on the irradiation dose (Figure 2), from the slope of which G ( C 0 2 ) = 5.3 was derived. This value approaches the yield of .OH radicals in N 2 0 saturated solution and indicates that decarboxylation is the almost exclusive reaction pathway in this system. This yield is also much higher than that reported by anyone in the literature. However, as mentioned above, these older studies did not comprise the basic pH region. The pH dependence of an .OH radical induced decarboxylation is exemplified for alanine in Figure 3, where the COz yields (expressed in terms of G) are plotted vs. the pH of the solution. The G values were obtained from similar plots as shown in Figure 2 using at least four points for the linear regression. In neutral solution only very little carbon dioxide is produced, which is in complete accord with all the previous studies. With increasing pH a sharp increase in G(C0,) is observed which reaches a plateau around pH 10, Le., near the pK = 9.87 of the amino group. These

The Journal of Physical Chemistry, Vol. 89, No. 14, 1985 3141

Decarboxylation of Amino Acids

TABLE I: Yields of .OH Radical Induced CO, Formation (Expressed in G) in N,O-Saturated Solutions of 5 X lo-' mol dm-'Substrate at Various pH Values

5-

amino acid

0 -

glycine

6-

H-CH

formula /coo-

PH G(C0,) 6.0 8.0 8.9 10.1

\NH~+

N

0

3-

11.1

0

&-alanine

2-

/COO' CH3-CH

\NH~+

1-

0

~

5

6

7

9

8

1

0

1

1

12

PH Figure 3. pH dependence of the C 0 2 yield (expressed in G) in N 2 0 mol dm-3 cu-alanine. saturated solutions of 5 X

results suggest that the reaction of -OH radicals with alanine yields carbon dioxide only if the amino group is unprotonated, viz.

+

*OH

*OH

+

CH3-CH

CH,-CH

/coo\FH

- - co,

/coo-

2

/coo-

CH3-CHZ-CH

\NH~+

norvaline valine

(CH&CH-CH

/coo\NH~+

no CO,

+

2-aminobutyric acid

(6)

...

norleucine

CH3-(CH&-CH\

/cooNH~+

(7)

The actual shape of the yield vs. p H curve can, of course, not be identified with the thermodynamic pK curve but rather reflects the associated kinetics of reactions 6 and 7 . Other simple aliphatic a-amino acids have also been investigated. In principle, always the same characteristics were found as outlined above. High yields of C 0 2 were only obtained in basic solution as can be seen from the results summarized in Table I. ?-Irradiation of N 2 0 saturated solutions of aliphatic amino acids, which carry the amino and carboxylate substituents not at the same. carbon atom, e.g., y-aminobutyric acid (-OOC(CH2)3-NH2/NH3+) or P-aminoisobutyric acid (-OOC-CH(CH3)-CH2-NH2/NH3+), did not yield any significant amounts of carbon dioxide irrespective of pH, Le., of the protonation state of the amino group. The measured G values are also included in Table I.

/coo-

leucine

(CH3)2CH-C~2-~~

\NH~+

isoleucine

/coo-

CH,-CH,-CH(CH,)-CH

\NH,+

2-aminoisobut yric acid

(CH ) C

,coo-

3 2 \NH~+

serine

CH20H-CH

/coo\NH~+

threonine

CH~-CHOH-CH/

COO-

2.3 3.5 4.0 4.0 4.0

5.9 7.5 8.0 9.0 10.1 11.0 11.9

1.0 1.2 2.3 4.4

5.8 10.0 11.7

0.5 4.2 4.0

6.0 10.0 12.0

0.6 3.5 3.6

6.2 10.3 12.0

0.2 3.6 3.0

6.0 10.0 12.0

0.4 1.6 1.8

10.0 11.9

3.9 4.2

10.1 11.9

3.0 3.2

5.8 10.0 11.6

0.5 5.6 5.6

10.0 12.0

4.5 2.5

10.0

3.8

10.0

4.6

10.0

5.4

6.0 12.0

0.4 0.4

5.8 12.0

0.3 0.3

9.1

0.4

5.2 5.0 5.3

\uti3+

glycine-d

,COOD-CD "ti3'

N,N-dimethylplycine

H-CH

/coo\yHlCH31Z

Reaction Mechanism The results presented so far clearly establish that two prerequisites are necessary for decarboxylation to occur. Firstly, both the amino and carboxylate substituents must be located at the same carbon atom, and secondly, the amino group must be unprotonated, i.e., the free electron pair at nitrogen must be accessible. It is therefore suggested that .OH radical attack occurs at the free electron pair at nitrogen. *OH

+

,cooR-CH \NH,

-

,cooR-CH\

(8)

y e 6H 1

An addition reaction seems most likely because of the known electrophilic character of .OH. The .OH adduct, I, could then fragment via

3-aminobutyric acid 4-aminobutyric acid p-alanine

fragmentation reaction.30 The latter requires the presence of both an electron-donating (-COO-) and an electron-withdrawing group (-OH) in the same molecule. These fragmentation reactions are mainly entropically driven. An analogous although nonradical example has been described in the l i t e r a t ~ r e . ~ ' The - ~ ~ oxidation of a-amino acids by HOC1 or HOBr accordingly yields N-chloroor N-bromo-a-amino acid anions which in aqueous solution fragment spontaneously to give halide ion (Cl- or Br-), imine (HN=CH-R), and C 0 2 . Dissociation of the .OH adduct into the nitrogen-centered radical cation

~

Le., via a process which resembles the well-known Grob-type

~~~

~~

~

~

coo-

I

-

I

R-~H-NH,

CH3-CHlNH3+l-CH2-COG

~

~~~

R-CH-tjH2+ ~

~

~~~

+ ~~

OH~

(10) ~

(30) Grob, C. A. Angew. Chem., Int. Ed. Engl. 1969, 8, 5 3 5 . (31) Langheld, K. Chem. Ber. 1909, 42, 2360. (32) Hand, V. C.; Snuyder, M. P.; Margerum, D. W. J . Am. Chem. SOC. 1983, 105, 4022. (33) Konigsberg, N.; Stevenson, G.; Luck, J. M. J . B i d . Chem. 1960, 235, 1341.

3142

Monig et al.

The Journal of Physical Chemistry, Vol. 89, No. 14. 1985

followed by spontaneous C 0 2 release 0

It

R-CH-RH, may also be envisaged. This type of reaction has been reported to occur in the gas phase34 where electron impact ionization of 2 4 6 8 1 0 1 2 a-amino acids in a mass spectrometer was shown to proceed via number of C - H bonds primary ionization at the nitrogen atom with subsequent loss of Figure 4. Yield of C 0 2at pH 10 vs. the number of C-H bonds present the carboxyl group. Another reaction analogous to reaction 11 in cu-amino acid (0 = glycine; 0 = a-alanine; 0 = 2-aminobutyric acid; has been observed in the solid state. Pulse irradiation of single 0 = norvaline; 0 = norleucine; = 2-aminoisobutyric acid; A = serine; crystals of deutero a-amino acids at 4.2 and 77 K yielded radical A = threonine). cations which on warming decomposed by d e c a r b ~ x y l a t i o n . ' ~ ~ ' ~ Interestingly, amino radical cations have been directly observed in aqueous solutions by pulse radiolysis in a very recent study on 3.c the .OH radical induced oxidation of t r i e t h ~ l a m i n e . ~It~was also reported that flavin-sensitized decarboxylation of (po1y)amino acids 0 proceeds via an intermediate formation of amino radical c a t i ~ n . ~ ~ . ~ ' E The a-amino radicals, R-CH-NH2, formed via (9) or (1 1) have -0 been identified in our system by their reducing properties as will 2.c be discussed in a later section. Evidently, carbon dioxide formation requires .OH radical attack at the free electron pair at nitrogen. The hydroxyl radical, on N the other hand, does not discriminate between different possible 0 0 reaction sites in a molecule to any great extent; Le., reaction 8 1.0 occurs in addition to reactions 3 and 4 in basic solution. Therefore, it is not surprising that with all amino acids the yield of carbon 0 dioxide generally does not account for 100% of the .OH radicals. Decarboxylation is nevertheless the main reaction pathway in basic solution for all a-amino acids investigated. Competing reactions which do not give CO, are considered to 100 200 300 LOO 500 600 700 be H-atom abstractions from the carbon skeleton. The influence DOSE I Gy of reactions 3 and 4 on the carbon dioxide formation is evidenced in Figure 4 which shows the maximum C 0 2 yields (in terms of Figure 5. C 0 2yield vs. radiation dose in N20-saturated solutions of 5 C) obtained at pH 10 as a function of the number of C-H bonds X 10-3mol dm-3 glycine (m) and 5 X mol dm-3 deuterioglycine ( 0 ) at pH 10. present in the molecule. Three different types of a-amino acids are distinguishable. The circles represent a-amino acids of the substituent. Much higher yields of C 0 2 (square in Figure 4) were structure R-CH(NH,)-COO- with R being an alkyl group. An found from this a-amino acid than from the isomeric 2-aminoincreasing number of C-H bonds results in an increasing probbutyric acid, CH3-CH2-CH(NH2)-COO-. The rate constant for ability of H-atom abstraction (3) and (4) at the expense of reaction reaction 4 must therefore be much higher than for H-atom ab8, and is reflected by a decrease in the C 0 2 yield. Glycine seems to be an exception to the rule. But there is a plausible reason for /coo/COOthis as will be discussed later. *OH + R-CH R-C + H2O (4) The triangles represent a-amino acids of the type R'-CH\NH2 \NH2 (NH,)-COO- with R' being an alkyl group carrying a hydroxyl stractions anywhere else in the molecule. But this is e x p t e d since substituent somewhere, e.g., threonine (CH3-CH(0H)-CHthe a-hydrogen is a tertiary one and is also activated by the -COO(NH,)-COO-). Again a longer alkyl chain, Le., more C-H bonds, and -,NH, groups. This conclusion is supported by the results results in a lower C 0 2 yield. But more interestingly, comparison of Anbar et al.38who determined partial rate constants for H-atom with the unsubstituted a-amino acids (represented by the circles) abstraction by -OH at different sites in a molecule. It also explains reveals that for a given number of C-H bonds in the molecule the results from the hydroxyl-substituted a-amino acids (the (corresponding to a vertical line in Figure 4) the hydroxyl-subtriangles in Figure 4) which carry two particularly labile H atoms. stituted a-amino acids yield less carbon dioxide. In other words, Hence, these a-amino acids become relatively less reactive at a-amino acids with additional hydroxyl substituents are more nitrogen and thus evolve less carbon dioxide than the unsubstituted reactive in the side chain (higher rate constant for reactions 3 and substrates. 4), which in turn lowers the probability of reaction 8. All these The same reason applies to the.CO, yield in the glycine system. conclusions imply that the individual rate constants for reaction This amino acid, -OOC-CH2-NH2, carries two activated a-H 8 are equal or almost equal for all amino acids investigated. atoms. Such a configuration increases considerably the probability The importance of relative reaction rates for H-atom abstraction of H-atom abstraction by -OH, which in turn leads to a lower yield is further evidenced by the results obtained from 2-aminoisobutyric of C 0 2 . The measured C 0 2 yield in the glycine system is apacid, (CH,),C(NH,)-COO-, a substrate lacking an H atom at proximately equal to that obtained with 2-aminobutyric acid, the particular carbon which carries both the amino and carboxylate CH,-CH2-CH(NH2)-COO-. From this it is concluded that the rate constant for reaction 4, Le., the abstraction of an a-hydrogen (34) Jones, J. H. Q. Reu. Chem. SOC.1968, 22, 302. atom, is about 5 times higher than the rate constant for abstraction (35) Das, S.; von Sonntag, C. Z . Nuturforsch. E. submitted for publication. of a (3-H or y-H (reaction 3) in these molecules. ( 3 6 ) Traber, R.; Kramer, H. E. A. Hemmerich, P. Biochemistry 1982, 21,

g

. I

. l

7

-

1687. (37) Armstrong, J. S.; Hemmerich, P.; Traber, R. Photochem. Photobiol. 1982, 35, 747.

(38) Anbar, M.; Meyerstein, D.; Neta, P. J . Chem. Soc. B 1966, 742.

The Journal of Physical Chemistry, Vol. 89, No. 14, 1985 3143

Decarboxylation of Amino Acids TABLE II: Yields of Br- Ions Formed in N20-Saturated Solutions of 5 X 1O-j mol dm" Amino Acid and 1 X mol dm3 Halothane Expressed in Terms of C and Normalized to C(C02) in These Systems amino acid pH G(Br-) G(Br-)/G(CO,)

glycine 2-aminobutyric acid 2-aminoisobutyric acid

10.0

4.2

1.05

10.0 10.1

4.4

1.05 1.05

5.9

According to our concept the decarboxylation yield depends on the relative reactivity of .OH toward the amino function and the nature of the carbon-hydrogen bonds of the skeleton. Therefore, .OH oxidation of deuterated glycine, -OOC-CD2-NHZ, should result in a higher COz yield compared with normal glycine, since the rate constant for deuterium abstraction is lower than for hydrogen abstraction. As can be seen from C02yield vs. dose plots in Figure 5 deuterated glycine did indeed evolve significantly more COzat p H 10. This result strongly supports the mechanism outlined above. Identification of a-Amino Radicals Carbon dioxide release is accompanied by the formation of a-amino radicals (reactions 9 or 11) which are very powerful one-electron reductants.3941 It was shown that their reduction potential is similar to those of a-hydroxy radicals." For example, Therefore, a-amino radicals they readily reduce oxygen to 0,-. are expected to be capable of reducing halothane (CF3CHBrC1), for example, a compound which was recently shown to react quantitatively with a-hydroxy radicals.43 (In the absence of suitable scavengers the a-amino radicals undergo disproportionationSz6) .OH radical oxidation of 2-aminobutyric acid at p H 10.0 in mol dm-3 halothane led to the formation the presence of 1 X of bromide ions with G = 4.4. This G value is only marginally higher than G(C0,) in that system. The equivalence found strongly suggests that a-amino radicals are indeed formed and that they are quantitatively scavenged by halothane, according to CH3-CHz-CH-NH2 CF3CHBrCl Br- CF3CHC1 CH3CH2CH=NHz+ (12)

+ +

+

-

Corresponding experiments have been undertaken with other a-amino acids as well. At basic pH values always a very good 1:l correlation between G(Br-) and G(C0,) was obtained (Table 11) which clearly substantiates reactions 9 and 11. Structural Effects Although an increasing number of C-H bonds in the molecule generally decreases the COz yield, structural effects are also important. This is most clearly seen in comparing the results obtained at p H 10 from the isomers leucine, isoleucine, and norleucine (for structures see Table I), all of which contain one H atom at the a-carbon and nine H atoms at more remote positions. Only on this basis one should expect that the yields of CO, are very similar. However, the G(C02) values are 3.9, 3.0, and 1.6 for leucine, isoleucine, and norleucine, respectively. Particularly leucine evolved much more C 0 2 than expected; e.g., G(C0,) was even higher than for valine, a substrate with less C-H bonds, implying that the rate constant for H-atom abstraction from leucine is lower than from valine. This finding must be related to structural reasons. A space filling model of leucine thus reveals that the H atom at the a-carbon is partly shielded by the two bulky methyl groups. As was shown in the previous sections this H atom is more reactive with respect to abstraction by .OH than a phydrogen or y-hydrogen. The shielding of the a-H atom by the (39) Anderson, N. H.; Norman, R. 0. C. J . Chem. SOC.E 1971, 993. (40) Burkey, T. J.; Castelhano, A. L.; Griller, D.; Lossing, F. P. J . Am. Chem. Soc. 1983, 105, 4701. (41) Chandrasekaran, K.; Whitten, D. G. J . Am. Chem. SOC.1980,102, 5119. (42) Hiller, K.-0.; Asmus, K.-D. J. Phys. Chem. 1983, 87, 3682. (43) Mhig, J.; Asmus, K.-D. J. Chem. SOC., Perkin Trans. 2 1984, 2057.

methyl groups results therefore in a net decrease of the reaction probability of .OH with hydrogen atoms. As a consequence, attack of .OH at nitrogen, resulting in a higher CO, yield, becomes more likely. In norleucine no such shielding occurs as is easily seen with a space-filling model. The nonbranched alkyl chain is directed away from the a-carbon leaving its hydrogen freely accessible. Hence, the COz yield from this a-amino acid is fairly low. In isoleucine this shielding effect is less pronounced and G(C02) is found to lie between those obtained from leucine and norleucine. No such shielding effects are evident for valine and its nonbranched isomer norvaline with both substrates yielding the same amount of C 0 2 . As can be seen from the model the two bulky methyl groups of valine are unable to cover the a-hydrogen. Discussion The results presented here give new insight into the mode of reaction of -OH radicals with a-amino acids. In the neutral or acid pH region, Le., where the amino function is fully protonated, they support older findings. Under these conditions abstraction of an H atom is taking place particularly from those positions which are remote from the carbon atom which carries the amino and carboxyl substituents. In the basic pH region, however, which had not been investigated before with respect to end product analysis, the reaction pathway is completely altered. In addition to H-atom abstraction, the .OH radicals can now also attack directly at the nitrogen atom and initiate decarboxylation with simultaneous formation of a strongly reducing a-amino radical. Hydrogen abstraction at the a-position is also found to occur with higher probability is basic solution, but other H atoms still remain susceptible to .OH attack. The latter is clearly established by the decrease of the CO, yield with increasing number of C-H bonds in the molecule. It may be noted that our results are seemingly in contradiction to some published ESR s t u d i e ~ concerned ~ ~ - ~ ~ with transient radicals from amino acids formed at pH 13. In these studies no radicals lacking the carboxylate group were found and it was concluded that CO, evolution from reaction of .OH with amino acids is unimportant over the entire pH range. Since radicals with and without a-carboxyl substituent are clearly distinguishable from each other, the possibility of a misinterpretation of the ESR spectra can be ruled out. However, as we have unambiguously shown now, decarboxylation is the major reaction path in basic solution. Two possibilities are conceivable to account for all the different observations. Firstly, the strongly reducing a-amino radicals could react with HzOz

+ HzOz

R-CH-NH2

-+

*OH

+ OH- + R-CH=N+H2

(13)

thereby leaving for detection only those radicals, which are less readily oxidized.44 Secondly, a-amino radicals could abstract an H atom from the parent compound, viz.

R-CH-NH,

+

R-CH

/coo\NH,

-

R-CHflH,

+

./coo-

R-C\

(14)

NH2

This reaction yields the radical which was directly observed by ESR. Even if the respective rate constants are as low as lo5 mol-] dm3 s-' both reactions would go to completion before ESR detection sets in, since the concentration of both H 2 0 2and amino acid are high (normally 20.1 mol d n ~ - ~ )However, . we are presently not able to distinguish between the two possibilities. Concerning the mechanism of the - O H radical attack at the nitrogen function it has been s ~ g g e s t e d lthat ~ . ~ -OH ~ radicals may abstract a hydrogen atom from the unprotonated amino group.

This would lead to the formation of an aminyl radical, which in fact is just the unprotonated form of the radical cation formulated (44) Koltzenburg, G., personal communication.

J . Phys. Chem. 1985, 89, 3144-3147

3144

in reaction 9. At present this possibility cannot be ruled out as a pathway to CO,. A strong argument against reaction 15 is, however, the finding that .OH oxidation of N,N-dimethylglycine a t pH 10 also yielded considerable amounts of COz.

.OH

+ -0OC-CH,-N(CH,),

-

C02

+ other products

(16)

In N 2 0 saturated solutions of 5 X lo-, mol dm-, dimethylglycine at pH 10 carbon dioxide was formed with G = 5.4. No CO, would have been expected if only aminyl radicals were decomposed by decarboxylation since this substrate carries no hydrogen bound to nitrogen. The two methyl groups increase of course the electron density at nitrogen, making this center more susceptible for addition of the electrophilic .OH radical. This result therefore rather supports th,: mechanism suggested in reactions 8 and 9. The fact

that G(C02) is even higher than that from glycine is most likely due to a shielding of the a-H atoms by the two methyl groups at nitrogen. Acknowledgment. This work has in part been carried out pursuant to a contract with the National Foundation of Cancer Research. It has also been supported by the “Fonds der Chemischen Industrie”. W e thank Dr. G. Koltzenburg for very helpful discussions. Registry No. .OH, 3352-57-6;glycine, 56-40-6; L-a-alanine, 56-41-7; 2-aminobutyric acid, 80-60-4; L-norvaline, 6600-40-4; L-valine, 72- 18-4; L-norleucine, 327-57-1; L-leucine, 61-90-5; L-isoleucine, 73-32-5; 2aminoisobutyric acid, 62-57-7; L-serine, 56-45-1; L-threonine, 72-19-5; glycine-a,a-d,, 4896-75-7; N,N-dimethylglycine, 11 18-68-9; 3-aminobutyric acid, 541-48-0;4-aminobutyric acid, 56-12-2;p-alanine, 107-95-9.

Absolute Rate Coefficient for the Reaction of NO3 with trans-2-Butene A. R. Ravishankara* and R. L. Mauldin I11 Molecular Sciences Branch, Georgia Tech Research Institute, Georgia Institute of Technology, Atlanta, Georgia 30332 (Received: February 1, 1985)

-

The absolute rate coefficient for the reaction, NO3 + trans-2-butene products, has been measured at 298 K with a fast flow apparatus under pseudo-first-order conditions in NO3. NO3 was produced either via F + HNO, reaction or N205thermolysis and detected by either long-path laser absorption at 662 nm or laser-induced fluorescence excited at 662 nm. The pressure in the flow tube was varied from 1 to 4 torr. The obtained value of k2 was (3.78 f 0.41) X cm3 molecule-I s-I, independent of pressure. A brief investigation of the reaction NO3 + isobutene products gave a value of k3 = (3.3 f 0.43) X cm3 molecule-] s-l. Our results are compared with previous less direct measurements of k , and k3. Our value of k2 helps the placing of relative rate constants for NO3 + organics measured in other laboratories on an absolute scale.

-

Introduction The NO, free radical has been known for the past three decades as a reactive intermediate in nitrogen oxide systems.l During the past two decades, it was recognized that NO, is present in the earth’s atmosphere and that NO3could play very important roles in the chemistry of both the lower and the upper atmosphere.2 Therefore, a great deal of effort is being expended to elucidate the details of NO, atmospheric chemistry. NO, has been shown to react with many organic compound^,^^ which, during daylight hours, are removed via reactions with OH and in some instances 0,. In the troposphere, NO, is produced via reactions such as NOz 0, NO3 + O2and N,O, NOz + NO,, and during daylight it is destroyed very rapidly via solar photolysis. However, at night appreciable concentrations of NO, can be present in regions of high NO,r. Any reaction of NO, with organics would, therefore, constitute a nighttime removal mechanism, and would supplement the daytime removal due to OH. Rate coefficients for the reactions of a large number of organic molecules with NO3 have been measured by two different techniques.,” In the first method, which is a competitive technique, the loss rates of organics relative to that for a standard compound are measured. Even though this method yields very useful information regarding relative reactivities of various compounds, the absolute rate coefficients (which are needed for atmospheric lifetime estimations) can be calculated only if the absolute value of the rate coefficient for NO, reaction with a standard compound is known. In the second method, the temporal loss of N 2 0 5(in equilibrium with NO2 and NO,) is measured in the presence and in the absence of the organic compound of interest. From the enhanced loss rate of N 2 0 5in the presence of the organic compound, under the assumption that neither NO, nor N 2 0 5react

+

-

-

*Address correspondence to this author at NOAA/ERL, R/E/AL-2. 325 Broadway, Boulder, CO 80303.

0022-3654/85/2089-3 144$01.50/0

with the organic compound, the absolute rate coefficient for the reaction of N O 3 with the organic compound can be calculated if the equilibrium constant, K,, for reaction 1 is known. UnNO2 + NO, s NlO,; Kl (1) fortunately, there are discrepancies between various determinations of K , (both direct measurements and calculation^).^-^^ Since the measured value of the reaction rate coefficient is directly proportional to K , , any error in K 1 will be present in the rate coefficient as well. For these reasons, we undertook the measurement of the absolute rate coefficient for the reaction NO, trans-2-butene products (2) using a discharge flow apparatus. NO3 was detected via long-path laser absorption a t 662 nm or via laser-induced fluorescence. Reaction 2 was studied under pseudo-first-order conditions in NO3.

+

-

( 1 ) For example see, H. S . Johnston, J . Am. Chem. Soc., 73,4542 (1951).

(2) For example see, P. A. Leighton, “Photochemistry of Air Pollution”, Academic Press, New York, 1961. (3) E. D. Morris, Jr., and H. Niki, J. Phys. Chem., 78, 1337 (1974). (4) S. M. Japar and H. Niki, J . Phys. Chem., 79, 1629 (1975). ( 5 ) R. Atkinston, C. N. Plum, W. P. L. Carter, A. M. Winer, and J. N. Pitts, Jr., J . Phys. Chem., 88, 1210 (1984). (6) (a) R. Atkinson, J. N. Pitts, Jr., and S. M. Aschmann, J . Phys. Chem., 88, 1584 (1984); (b) R. Atkinson, C. N. Plum, W. P. L. Carter, C. N. Plum, A. M. Winer, and J. N. Pitts, Jr., J . Phys. Chem., 88, 2361 (1984); (c) R. Atkinson, W. P. L. Carter, C. N. Plum, A. M. Winer, and J. N. Pitts, Jr., I n t . J . Chem. Kinef.,16, 887 (1984); (d) R. Atkinson, S. M. Aschmann, A. M. Winer, and J. N. Pitts, Jr., Enuiron. Sci. Technol. 18, 370 (1984). ( 7 ) E. C. Tuazon, E. Sanhueza, R. Atkinson, W. P. L. Carter, A. M. Winer, and J. N. Pitts, Jr., J . Phys. Chem., 88, 3095 (1984). (8) C. A . Smith, A. R. Ravtshankara, and P. H . Wine, J . Phys. Chem., 89, 1423 (1985). (9) K. C. Kircher, J. J . Margitan, and S. P. Sander, J . Phys. Chem., 88, 4370 (1984). (10) M. W. Malko and J. Troe, I n t . J . Chem. Kinet., 14, 399 (1982). ( 1 I ) D. Perner, A. Schmeltekopf, R. H . Winkler, H. S. Johnston, J . G. Calvert, C. A. Canrell, and W. R. Stockwell, J . Geophys. Res., 90, 3807 (1985).

0 1985 American Chemical Society