HYDROLYSIS O F 4(5)-HYDROXYMETHYLIMIDAZOLYL ACETATE
Oct. 5, 1964
[CONTRIBUTION FROM
THE
4109
DEPARTMENT OF CHEMISTRY, CORNELL UNIVERSITY, ITHACA, N. Y.]
Hydrolytic Reactions of 4(5)-Hydroxymethylimidazolyl Acetate, 1-Methyl-5-HydroxymethylimidazolylAcetate, 4(5)-Chloromethylimidazole,
and 1-Methyl-5-chloromethylimidazole BY THOMAS C. BRUICE'A N D J, L. H E R Z ~ RECEIVEDMAY4, 1964 In a previous study from this laboratory (Bruice and Fife), it was shown t h a t 4(5)-hydroxymethylimidazolyl acetate ( I ) undergoes facile OH@ catalyzed (koa) or general-base catalyzed (acetate, pyridine, imidazole) hydrolysis. From the magnitude of the rate constants, and the establishment t h a t I undergoes facile alkyloxygen scission on reaction with nucleophiles, it was proposed t h a t the mechanism(s) of hydrolysis involved proton abstraction and elimination of acetate ion with the formation of a metastable diazofulvene intermediate (11). Reanalysis of the original data has shown no contributions of an s N 1 nature t o the solvolytic process. T o determine if t h e hydroxide ion catalyzed hydrolysis proceeds through pre-equilibrium formation of an anion intermediate or via proton abstraction concerted with elimination of acetate ion, the deuterium solvent isotope effect mas determined (80'). The ratio of kobsda/kobsdD was found, on the basis of the conductometrically determined pK2 values for imidazole in H20 and DzO, t o be more in accord with an anionic interxnediatetrue specific base-catalyzed elimination. The previous supposition of an elimination mechanism has found support in the observation t h a t I-methyl-5-hydroxymethylimidazolylacetate (1111, unlike the ester I . exhibits The aqueous hydrolysis tendencies t o only acyl-oxygen scission on reaction with nucleophiles (MeO@,"*OH). of 111 (78", p H 5.49 t o 9.20) was found t o occur with contributions from a spontaneous and an OH@ catalyzed rate, whose deuterium solvent isotope effects were determined t o be kH/kD = 1.4 and 0.58, respectively, in accord with intramolecular nucleophilic and BAC*catalyzed hydrolysis, respectively. Both reactions were found t o be quite facile ( k o H = 40 to 220 times t h a t for the alkaline hydrolysis of p - and m-nitrobenzyl acetates, respectively). The aqueous solvolysis of 4( 5)-chloromethylimidazole(1:) yields only 4(5)-hydroxymethylimidazole. T h e rate of solvolysis cannot be either an intramolecular Sx2 elimination of C l e (the rate is much toc fast when compared t o the hydrolytic rates for nitrogen mustards) or a n s N 2 attack of OH@(koa would approach a diffusioncontrolled process). The reaction, therefore, must be an elimination of C15 from the neutral base species. The ratio of the rate constants for the neutral solvolysis of Y t o the similar constant for 1-methyl-5-chloromethylimidazole ( X ) is 1 : 1.2, supporting an intermediate carbonium ion rather than a neutral diazofulvene intermediate. Further substantiation of a n Ssl mechanism for \.' is provided by the finding of a deuterium solvent isotope effect of kH/kD = 1.25, a value previously determined for the solvolysis of a t-butyl chloride. A common ion effect was observed in the solvolysis of V, further substantiating carbonium ion formation. The solvolytic rate constant for \.' is about t h a t for chloromethyl ethyl ether.
Introduction Bruice and Fife3 studied the alkaline hydrolysis of 4 (5)-hydroxymethylimidazolyl acetate (I) and found that the reaction' proceeded by alkyl-oxygen scission. "Normal" ester hydrolysis, except for some special cases (e.g., as the acid-catalyzed hydrolysis of t-alkyl esters4 and the alkaline hydrolysis of some benzhydryl systems5), proceeds by acyl-oxygen scission. The
+
BH
f
CHSCO2'
ester I was found to be subject to general-base-catalyzed hydrolysis by acetate ion, imidazole, and pyridine. A diazofulvene intermediate (11) was postulated. It was not determined, however, whether hydroxide ion itself acts as a general base (1) or as a specific base (2). If (1) and/or ( 2 ) is the mechanism by which I hydrolyzes, substitution of the ionizable hydrogen by a methyl group should alter the mechanism of hydrolysis. It was also of interest to determine whether (1) or (2) is general for all derivatives of 4(5)-X-methyl-imidazoles regardless of the nature of X, or whether i t is the mode of reaction for the hydrolysis of only derivatives where X- is a poor leaving group such as acetate ion.
0
Results and Discussion6
II
The kinetics of the hydrolysis of 1-methyl-5-hydroxymethylimidazolyl acetate (111) in water a t 78' and Ia
( 6 ) Glossary of symbols used in this paper. k(,hsc] = observed pseudofirst-order rate at a n y constant p H . aH = hydrogen ion activity a s measured b y t h e glass electrode; K, = dissociation constant of a monobasic acid; Ki and K2 = first and second dissociation constants for a diprotic acid, ~ . ,g imidazole, [ I r n ~ ]= [ I m H @ ] f [ I m ] [ I m 3 1 , [ST]= stoichiometric
+
(1) Career Investigator, National Institutes of Health, ( 2 ) Predoctoral fellow, National Institutes of Health. P a r t of t h e work t o be submitted by J. L. Herz in partial fulfillment for t h e P h . D . Degree, Cornell University. (3) T. C . Bruice and T.H. Fife, J . A m Chem. Soc., 83, 1124 (1961). (4) C f , for example, C . A Bunton a n d 1. I.. Wood, J . Chem. Soc., 1.522 (1955); V R. Stimson and E . J. Watson, ibrd., 2848 (19541, W. von E. Doering and H . H . Zeiss, J A m . Chem. Soc., 7 5 , 4733 (1953). ( 5 ) ( a ) M . P. Balfe, A . Evans, J. Kenyon, and K . N . N a n d i , J . Chem. Soc., 803 (1946); (b) M . P Balfe, E A W. D o w n e r , A. A. E v a n s , J. Kenyon. R. Poplett, C E. Searle, and A . L Tarnoky, i b i d . , 797 (1946).
__ ImH
@
Im
[ST] = c I S , l , [ S i ] = I S H @ ] , [SI, i [SSI,etc.; KLv = autoprotolysis constant for H 2 0 ; K, = n a [ O H @ ] , KD,O = autoprotolysis constant for IhO, superscripts H and D refer t o t h e various constants in H z 0 and D 2 0 , respectively concentration of any substrate S;
THOMAS C. BRUICEAND J. L HERZ
41 10
Vol. SG
3
0
0
1
-;
1
/"
150
-
3
I
I
/
pH or pD
Fig. 1 -Log kubad-pH (and p D ) profile for the hydrolysis of I-methyl-5-hydroxymethylitriidazolyl acetate (111) ( i 8 " , solvent water. g = 1.0 M ) . p = 1.0 JI (with KCI) were followed a t constant pH on a pH-stat over the pH range 5.49 to 9.20. In Fig. 1 is plotted the logarithm of the observed pseudo-first-
order rate constant ( k o b s d ) us. pH. In the pH range studied, the reaction was found to be extremely facile. The value of the pseudo-first-order rate constant k&sd a t any pH is provided by 3. The values of k ' o ~ ,OH,
T/ I
and ko are the calculated second-order rate constants for attack of OH" on the conjugate acid and free base forms of I11 and the spontaneous rate of solvolysis of 111, respectively. The value of KIappis the kinetically apparent acid dissociation constant of the conjugate acid of I11 and was determined from an Eadie plot7 of the kinetic data. The points of Fig. 1 are experimental and the curve has been drawn employing the values of ko = 3.34 X lo-* min.-' and KOH = 3.99 X 102 1. mole-' min.-'. The value of k o ~ when , compared to the same constant for I,3determined under the same conditions (8.95 X 1041. mole-' min.-I), indicates that a drastic decrease in the hydrolytic rate accompanies substitution of the active hydrogen of I by a methyl group to give 111. The hydrolytic reaction for I11 occurring a t low pH values (Z.e.$ koH or k,) may pertain t o : (a) a normal BAc2 hydrolysis of the conjugate acid of I11 as in (4) ;
6
0
4
K,
20
16
24
28
32
36
/aH xIO?
(b) an intramolecular general-base-catalyzed hydrolysis as in ( 5 ) , (c) an intramolecular nucleophilic catalysis as in ( 6 ) , or conceivably a carbonium ion mechanism as in (7). Imidazole has been found to act as a general-base catalyst in the hydrolysis of various esters and as an intramolecular nucleophilic catalyst for the hydrolysis of esters of 4(5)-(hydroxyethy1)imidazole 8b The probability of a spontaneous elimination of acetate ion from I11 to yield a carbonium ion (7) as an explanation for the results of the pH-rate profile a t low pH is readily discounted If I11 undergoes carbonium ion formation during hydrolysis then compound I should do the same In fact, compound I might be anticipated to afford a carbonium ion a t the same rate as I I I . g If formation of a carbonium ion makes a contribution to the over-all rate for the hydrolysis of I , then the observed rate constant should be equal to the sum of the rate constants for anionic elimination and carbonium ion formation ( k + ) Under the conditions kabsd = k + m r b o n r u r n ion
K i>> a H
+
kc,lirnina+ion
(8)
so t h a t 8 simplifies to 9
kohsd
=
k+
+
(krK2,'aH)
(9)
A plot of k & d vs. K,/aH should give a straight line of slope k,K2/K, and of intercept on the k o b s d axis of k + . In Fig. 2' is plotted k o b s d vs. K,laH using the data of Bruice and Fife.' Inspection of the plot of Fig. 3 reveals t h a t the intercept of the line on the k o b s d axis is a t zero. From this one may infer t h a t there is no contribution from a carbonium ion mechanism in the hydrolysis of 111.
I
(6)
12
8
Fig 2-Plot of kobsd 18s K W / U Hfor the hydrolysis of 4(5)= hydroxymethylirnidazolyl acetate ( I ) ( X Osolvent , water, 10 M )
employed
CH3
1
50t /o
CH3
(7)
(7') T . C. Bruice and J 31 S t u r t e v a n t , J A m C h r m Soc , 81, 2860 (1959). T. C. Bruice, i b i d , 81, ,5444 (1959).
(8) (a) T. C . Bruice, T H. Fife, J J B r u n o , and P Benkovic. ibid , 8 4 , 3012 (1962); ( b ) U K Pandit and T C Bruice, tbid , 82, .3:386 (1960) (9) l - ~ ~ e t h y l - 5 - c h l ~ , r o m e t h y l i m i d a z o(lX e ) a n d 1(5)-chlorornethylimidazole ( V ) , which both f o r m carbonium ions, hydrolyze at t h e s a m e ratethis s t u d y
Oct. 5 , 1964
HYDROLYSIS O F 4(3)-HYDROXYMETHYLIMIDAZOLYL ACETATE
T o ascertain if 111is susceptible to alkyl-oxygen cleavage on attack by base the half succinate ester of 1methyl-5-hydroxymethylimidazole(IV) was subjected to alkaline methanolysis and I11 was allowed to react with hydroxylamine in aqueous solution. l-Methyl-5hydroxymethylimidazolyl hydrogen succinate (IV) was solvolized in methanol containing 2 equivalents of methoxide ion, 1 equivalent being used to react with the carboxyl group of the ester to form the sodium salt. If there is alkyl-oxygen cleavage, disodium succinate would be formed (10). If, on the other hand, the reaction proceeds via acyl-oxygen scission (1l a ,b) the sodium salt of monomethyl succinate would be isolable. ~
-'i n
.-
E
Y
cn
a YO
CT 0
0.6I
I
1
I
4111 I
1
I
c
0.2t -0.2
- -0.6
0
pH or pD. Fig. ?,.-Log k o b a d - p H (and p D ) profile for the hydrolysis of 4(5)-chloromethylimidazole (30°, solvent water, fi = 1.0 M ) .
I CHs
08)
ti
CH3
+
CH3
e
ti
amine under these same conditions, only a 20% yield of acethydroxamic acid is obtained.' The reaction of I11 with hydroxylamine, therefore, proceeds predominantly by acyl-oxygen scission. Mechanisms (4), ( 5 ) ) and (6) remain as possible explanations for the "spontaneous" hydrolysis of 111. Mechanism ( 5 ) entails a concerted proton abstraction from water, accompanying its attack on the carbonyl carbon in the rate-determining step and a kinetic solvent isotope effect (kH/kD) of a t least 2 would be expected.8 Nucleophilic catalysis by imidazole (6), however, is known to proceed without a kinetic solvent isotope effectlo and for a B ~ c 2mechanism (4) the isotope effect should be considerably smaller than 1.0. Included in Fig. 1 is a plot of log kobsd vs. p D for the hydrolysis of 111 in DzO. The straight line portion of the curve, having a slope of 1.0, is shifted over to the left from the line in HzO because the autoprotolysis constant for DzO is about 1/6 of t h a t for HzO. The observed pseudo-first-order rate coefficient obeys the expression
1
The experiment was performed in the following manner : 2 equivalents of sodium methoxide was added to a solution of ester in absolute methanol. The solution was sealed from the air and stirred in a nitrogen a t mosphere for 3 hr. whereupon a large excess of anhydrous ether was added. The salt was filtered and dried, and an infrared spectrum was taken. This was superimposable on the spectrum of an authentic sample of the sodium salt of monomethyl succinate. The isolation of monomethyl succinate indicates acyloxygen scission for specific base methanolysis of VI, and one would expect the same behavior for 111. When I11 was treated with hydroxylamine in aqueous solution near neutrality a 92.5yGyield of acethydroxamic acid was obtained. When I, which was shown to undergo alkyl-oxygen scission, is treated with hydroxyl-
where koD = 2.93 X min.-' and koD = 6,87 X l o 2 1. mole-' min.-'. A glass electrode correctionll to the pH meter reading was made to obtain pD, and the concentration of ODe was calculated assuming that K w / K ~ isn the ~ same at 78' as it is at 30'. The determined kinetic solvent isotope effect koH/koD = 3.34 X 10-2/2.93 X = 1.1, is in keeping with (6). In addition, a B ~ c mechanism 2 involving the conjugate acid of I11 as in (4) may be firmly ruled out. Thus, it is known that the electrostatic facilitation of a trimethylammonium ion substituted p in ethyl acetate provides a rate acceleration of but cu. tenfold." If (4) were in operation, the second-order rate constant would be -lo4 greater than OH for 111, a factor 103 greater than for an electrostatic effect 6 t o the ester bond and 102 greater than an electrostatic effect o to the ester bond. l 3 T h a t intramolecular attack is apparently (10) Cf.,for example, B. M . Anderson, E. H. Cordes, and W P. Jenks, J . Bioi. C h e m . , 2S6, 455 (1961). (11) T.H.Fife and T. C. Bruice, J. P h y s . C h e m . , 6 0 , 1079 (1961). (12) G . Aksnes and J. E . Prue, J . Chem. Soc., 103 (19.59) (13) R. P. Bell and F. J. Lindars, ibid , 4601 (1964)
4112
THOMAS C. BRCICEA N D J. L. HERZ
Vol.
R j
favored over a general-base-catalyzed or Bac2 mecha3 0 ° , K,IKD?o = 6.6,” SO t h a t (I) would predict nism is surprising on the basis of the anticipated steric kobsdH..’kobsdl-) 2 1 3 . The value of the observed rate strain in forming the bicyclic system and from the constant, if mechanism (2) is the hydrolytic pathway for anticipated repulsion of substituents (from StuartI at pH 9.65 (where K 1 >> a”), is (15). Briegleb models). The ratio of the rate constants for k o b s d = krK2?aH (15) the hydrolysis of 111 to the hydrolysis of its homolog, 4(5)-hydroxyethylimidazolyl acetate, l 4 is reasonable on SinceaH = a D the basis of the ratios of the rates of formation of fiveand six-membered rings. ’s, I6 kohsdH _ _k,HK2H ~ (16) The deuterium solvent kinetic isotope effect for the k o b s d D - krDKzD second-order hydroxide ion dependent rate for the T o evaluate the magnitude of the observed isotope hydrolysis of 111, koH/kOD = 3.99 X 102/6.87 X 102 = effect to be expected from (2), one must estimate 0.58, is in keeping with a “normal” B ~ c 2ester hyk r H i k r D , and a value for k z H / k z 1 ’ must be obtained. drolysis. Because OD3 is a better nucleophile than OHe, an isotope effect of 0.5-0.7 is usually o b ~ e r v e d . ’ ~ ~No ~ measurements have been made to determine the isotope effect on the second dissociation constant of The value of koH for I11 is from 40 to 220 times greater imidazoles Ke. Since I readily decomposes in water, than the value of koH for the alkaline hydrolysis of mespecially so a t high pH’s, the p K 2 of imidazole itself or p-nitrobenzyl acetates, and approximately equal was measured. For a limited series of imidazole bases, to k o for ~ the hydrolysis of dichloroethyl acetate8 (as T.4:3,?”proit has been found that pK2 = 0.94 pK1 extrapolated estimating AH* to be 12 kcal. mole--’ for pK2 of 13.(121 for I . The spectrophotometrividing a the latter reaction). Thus, assuming an approxically dgtermined thermodynamic pK2 for imidazole is mately constant steric effect, the imidazole ring has 14.52,’2 SO that KzH8i. ( 2 1 ) Using t h e h e a t of ionization for I used by Bruice a n d F i f e , 2a value of 6 0 was calculated io!- ~ K a tI 30’ ( 2 2 ) H n‘alha a n d R M‘. Isensee, .I. 0 i . y C h ~ m . 21, . 70%(1%56) (2.’1) P Ballingel a n d F A L o n g , J . 47n. Chiv2 S n c , 81,10.70 I I M d Y ) $ 2 1 ) K P Bell, “ l h e Proton in C h e m i s t r y , ” Coinell University Press, I t h a c n , S 1-, 19.59, p . 188 ( 2 5 ) N C I.i, P T a n g , and R h l a t h u r , J P h y s (‘hem , 66, 1074 (1961). (26) l i I. P y m a n , .I. C h r m . Sor , 9 9 , 668 (1011)
HYDROLYSIS OF ~(~)-HYDROXYMETHYLIMIDAZOYL ACETATE
Oct. 3 , 1964
of V was added dropwise to a saturated aqueous solution of K C N a t O o , a good yield of 4(5)-cyanomethylimidazole was isolated along with a small amount of 1.2-di-(4(5)-imidazolyl)propionitrile (18). Turner?' found that V rapidly hydrolyzed in aqueous solution a t room temperature to form 4(5)-hydroxymethylHNyCH&l+KCN @-N H
CHSCHZOH
4113
'OH
Y
no reaction (17)
V
V
p,CHAHCN L N H
imidazole. Although no kinetic measurements were made, he found t h a t 0.1 g. of V was completely hydrolyzed after 13 min. One can easily hydrogenate V to 4(5)-methylimidazole, and the reaction of V with amines is quite exothermic. The kinetics of the hydrolysis of V were studied a t 30' and the ionic strength was kept a t 1.0 by the use of KC1. The solvolysis was so facile, however, t h a t the reaction kinetics could not be followed (by using the pH-stat) a t pH values greater than 3.75. A plot of kobsd, the pseudo-first-order rate constant for the hydrolysis of V, vs. p H is shown in Fig. 3 . The reaction exhibits apphrent first-order hydroxide ion dependence (as indicated by the slope of 1.0 for the pH-log kobsd plot) so t h a t the kinetically equivalent mechanisms of type (1) or ( 2 ) may be ruled out. Thus, if a mechanism similar to (2) were operative, the rate equation for a specific-base-catalyzed hydrolysis, as in the case of I, would have the form of (19).
VI
Assuming a low steady-state concentration for t h e carbonium ion (VI), one may derive the expression
where [Cle] is the concentration of chloride ion in solution. If the chloride ion concentration is kept constant, and the pseudo-first-order rate constants are measured in a p H range where U H >> KI, ( 2 3 ) reduces to (24). It is of interest to note that VI, the carbonium
ion generated in the hydrolysis of the conjugate base of V, is the conjugate acid of the diazofulvene intermediate 11. Structure VI is analogous to the intermediate proposed for the rearrangement of phenyl-3-methylindolylammonium salts2*( 2 5 ) .
A plot of log kobsd vs. pH should, if the reaction followed ( 2 0 ) , have a slope of 2.0. Since the line of Fig. 3 has a slope of 1.0, specific base catalysis need not be considered as a possible mechanism. Since the kinetic measurement was made under conditions where U H >> K 1the apparent first-order dependence of kobsd on OHe is ascribed to the ionization of V to its conjugate base and the spontxneous solvolysis of this species. By analogy to mechanism ( l ) , one might expect water to act as a general base, in an E'2 type elimination (21). Since the imidazole nucleus possesses aromatic character one might alternatively anticipate the hydrolysis of the conjugate base of V to be similar to the hydrolysis of benzyl chlorides (ie., via a carbonium ion, a mechanism also consistent with the observed kinetics (22)).
(29) (a) B . Cohen, E. R Van Artsdalen, and J . Harris, J . A m . Chem S o c . , 70, 281 (1948); ( b ) P. D. Bartlett, S . D. Ross, and C . G S w a i n , i b i d . ,
(27) R. A . Turner, C . F. Huchner, and C. R. Scholz, J . A m . Chem. Soc., 7 1 , 2801 (1949).
69, 2971 (1947); (c) P. i). Bartlett, J. W. David, S . D. Ross, and C . G. Swain, i b i d . , 69, 2977 (1947).
One might also think of V as the conjugate acid of a nitrogen mustard, a class of compounds which are very hydrolytiwlly labile. Mechanism (26) is similar to the mechanism by which mustards h y d r o l y ~ e . ~ ' h l (28) J. H . Thesing and H . Meyer, B e r . , 87, 1084 (1954).
THOMAS C. BRUICEA N D J. L. HERZ
4114
though intermediate VI1 would appear to be very highly strained, one cannot a priori rule out such structures as unstable. The diazirines30 have a ring system which is just as strained. Much work has been
done on the kinetics of the hydrolysis of tertiary nitrogen mustards.z9 T h e compounds cyclize in a relatively fast step, followed by a much slower hydrolysis (28). For example, the rate of cyclization for bisdiethyl-bchloroethylamine (VIII) in HzO is 1.1 min.-' a t 30'.
Ix In 1 : 3 acetone-water, kl and kz for the hydrolysis of ethylbis-P-chloroethylamine are 0.24 and 0.0017 min.-', respectively. Since the hydrolysis of V was followed by the rate of formation of HC1, if mechanism (26) is operative, the measured rate constant is t h a t for the formation of the cyclic intermediate VII. The magnitude of the rate constant for the hydrolysis of V, when compared to the rate constants for the cyclization of mustards, is much too great to be explained in terms of (26). The value of kl for the conjugate base of V calculated on the basis of (26) is E 2 X l o 3 min.-l, while the rate of cyclization of VIII, which forms a much less strained ring intermediate ( I X ) , is 1.1 min.-'. Cyclization, therefore, would appear to be a very unlikely pathway for the hydrolysis. T h e products from the hydrolysis of V were analyzed by paper chromatography, and the results showed unequivocally t h a t V hydrolyzes to give pure 4(5)-hydroxymethylimidazole. Secondary nitrogen mustards usually produce some dimeric materials under hydrolytic conditionszeb so that product analysis also points to the improbability of (26). Direct displacement of chloride ion by hydroxide (29) in an S N reaction ~ on V is also consistent with the kinetics. The rate constant, k o H , calculated on the basis of (29), is 2.8 X 1 0 1 0 1. mole-' min.-', a number exceeding t h a t for a diffusion-controlled process.
VOl. 86
One is left, therefore, with a choice between (21) and ( 2 2 ) as the mechanism for the hydrolysis of t h e conjugate base of V. If mechanism (21) is operative, substituting the 1-position of V should change both the mechanism and the magnitude of the hydrolytic rate constant. If, on the other hand, a carbonium ion is formed (22), the rate of hydrolysis for the N-H and the N-substituted compounds should be of approximately the same order of magnitude. l-iMethyl-3chloromethylimidazole hydrochloride ( X ) was synthesized, and its kinetics of hydrolysis were measured
x a t 30' and pH 3.35. T h e value of the observed pseudofirst-order rate coefficient determined for two runs was 6.01 f 0.61 X lo-' min.-'. At pH 3.33, V has an observed rate constant for solvolysis of 9.70 X 10-' min.-'.
One can reasonably assume that the ratio of KIV to KIX should be the same as the ratio of K 1 for imidazole to K 1 for N-methylimidazole-approximately 2.0.3 1 Thus krx/krv = 1.2, or within experimental error, as predicted by mechanism ( 2 2 ) )the rate constants are the same. The fact t h a t V and X hydrolyze a t the same rate is additional evidence that V does not hydrolyze via (26), since one would expect the rate of intramolecular attack for a tertiary compound (X) to be different from t h a t of a secondary compound. One can also use a DzO solvent isotope effect as a diagnostic test to distinguish between (21) and ( 2 2 ) . In Fig. 3 there is plotted the log of the observed pseudofirst-order rate constants for the hydrolysis of V in DzO vs. pD. As can be seen from (24), a plot of kobsd vs. K,/uH should give a straight line of slope krK1/K,. Let us call this slope X. By solving for k, and dividing we get
krH - XHK,KID _
krD
-
XDK~,~K~H
(31)
A plot of kobsd vs. K,/aa in HzO and in D20may be found in Fig. 3. The values of XH and XD are 3.1 X 1O'O and 6.1 X l o l o , respectively. Unfortunately, K1 for V cannot be determined; however, if one allows the assumption t h a t the electronic effect on the pK1 of V of a side chain chloride group is the same h s that of a p nitrobenzoyl group, then the pK1 of V would be the same as t h a t for 4(5)-hydroxymethylimidazolyl 9nitrobenzoate (XI).
XI
pK1 of X I was found to be 5.61. Using this value, a LaMer type plotz5gives a value for K I D I K l Hof 0.38. For the first dissociation constant, ( K l ) for imidazoles, (30) Cf.,f o r example, "Advances in Heterocyclic Chemistry," Vol. 2 , Academic Press, Inc., K e a . York, N. Y., 1963, p. 125.
(31) K. H o f m a n n , "Imidazole and its Derivatives. P a r t I," Interscience Publishers, Inc., N e w York, N.Y., 1953, p 16.
Oct. 5, 1964
HYDROLYSIS OF ~(~)-HYDROXYMETHYLIMIDAZOLYL ACETATE
a Lahler type plot is a p p l i ~ a b l e . ~Substituting ~ these numbers into eq. 31 gives a value of krH/krD = 1.25. Fortuitously, the deuterium solvent kinetic isotope effect for the hydrolysis of t-butyl chloride, which solvolyzes via a carbonium, is 1.25.32 The rate of hydrolysis of V, if the reaction does indeed proceed via a carbonium ion, should be sensitive to the concentration of chloride ion in solution (see 23) and one should be able to observe a "mass law" effect.
-I
41-
0
4115
0.2
0.4
0.6
0.8
I.o
CCI'I M. Fig. 4.-Plot of the reciprocal of the determined pseudo-firstorder rate constants determined a t constant p H v s . the reciprocal of the chloride ion concentration (30°, solvent water, p = 1.0 M with LiC104).
Experimental where k o b s d X is the observed pseudo-first-order rate constant a t any chloride ion concentration (x) and kobsdo is the pseudo-first-order rate constant when [C13] = 0. A plot, therefore, of l / k o b s d o os. [CIS] should give a straight h e of slope k-l,'(kobsdok-z) and an intercept of 1/kobsdo. Such a plot is found in Fig. 4. The hydrolysis was carried out a t various concentrations of LiCl with LiC1O4 added to keep the ionic strength a t 1.0. Both the salt bridge and the calomel electrode contained saturated solutions of LiC1. From the slope of Fig. 4, which is 1.85, and from a knowledge of kobsdo, one calculates t h a t k-l/kz = 3.86. Since k l = k o b s d o a H / K 1 , an assumed pK1 of 5.6 for V leads to k , 2 X l o 3 min.-'. One might inquire as to how a rate constant of 2 X l o 3 min.-' for the hydrolysis of V compares to the rate of solvolysis of other compounds which undergo carbonium ion formation. The a-halo ethers solvolyze in water to form carbonium ions a t a rate which probably exceeds the corresponding rate for any other group of compounds reported to date. Perhaps the most hydrolytically labile a-halo ether is chloromethyl ethyl ether ( X I I ) 3 3 which has a first-order hydrolytic rate constant of 0.14 min.-' in 2% water-dioxane. Attempts were made in our laboratories to measure the rateof hydrolysis of chloromethyl methyl ether in pure water, but the reaction was over before mixing was completed. B i j h ~ n ealso ~ ~ reported the rate of hydrolysis of chloromethyl phenyl ether ( X I I I ) , and its variation with water concentration in dioxane- water solvent mixtures. If one takes Bohme's data for the hydrolysis of X I I I , and plots log kobsd ZU. the mole fraction of water in the solvent, a smooth line is obtained. At a mole fraction of 0.35, which was the highest concentration of water employed, the line already begins to flatten out owing to selective solvation. The data could therefore be extrapolated to 1 0 0 ~ ifG one assumes the continuing form of the curve to give an answer correct within an order of magnitude. The data for the hydrolysis of the ethyl ether X I 1 (in 2 7 , aqueous dioxane) was fitted to the function describing the solvent sensitivity for the phenyl ether X I I I . An extrapolated rate constant of -4 X 102min.-' is obtained for the hydrolysis of X I 1 in 1007G HzO. Both XI1 and XI11 should have approximately the same solvent sensitivity, since their transition states are probably very similar. I t would appear, therefore, that X hydrolyzes to form a carbonium ion a t least as fast (if not faster) than any other compound whose solvolysis has been studied. (32) P. 121. Laughton and R. E Robertson, C n n . J . Chem , 34, 1714 (1956). (33) H . Bohme a n d A . Conies, B e r . , 89, 719 (19.56).
4( 5)-Hydroxymethylimidazole hydrochloride,3 4 4( 5)-chloromethylimidazole hydrochloride ( V),36 4( 5)-hydroxymethylimi-
dazoyl acetate ( I \ , 3and 1-methyl-5-chloromethylimidazolehydrochloride ( X ) 3 5were synthesized by literature procedures. Freshly recrystallized p-nitrophenyl acetate was the same as from previous studies.16 Melting points were determined on a Nalge electric melting point block, and are uncorrected. l-Methyl-5hydroxymethylimidazole was synthesized by the method of Jones36; however, a crucial step in the methanolysis of sarcosinonitrile was omitted in the experimental procedure, so that the full preparation of the methyl ester of sarcosine is reported herein. The remainder of the preparation proceeds smoothly as described. A mixture of 135 g . (2.0 moles) of methylamine hydrochloride, 270 ml. of 3776 formaldehyde, and 100 ml. of water were stirred in a round-bottom flask a t 0" for 3 hr. X cold concentrated solution of 130 g. (2.0 moles) of potassium cyanide was added in portions over a period of 2 hr., the temperature not being allowed t o rise above 8'. Solid COZ was added from time t o time to ensure a COI atmosphere. The mixture was stirred for 1 additional hour, and was then extracted with two 500-ml. portions of ether. After drying the ethereal solution over calcium oxide, the ether was removed by flash evaporation. The crude sarcosinonitrile was dissolved in 2.5 1. of d r y methanol previously saturated in the cold with HC1. The solution was allowed to stand overnight a t room temperature and was then heated under reflux for 4.5 hr. The ammonium chloride which had precipitated was removed by filtration, and the filtrate was flash evaporated t o a volume of 800 ml. The solution was once again filtered through a pad of Filter-cell, and the filtrate was evaporated t o dryness. The residue was.dissolved in 100 ml. of water, and the water was removed under vacuum. Formylation was then carried out as described.36 1-Methyl-5-hydroxymethylimidazolylAcetate (111) Hydrochloride.-To 0.50 g. of 1-methyl-5-hydroxymethylimidazole(0.0045 mole) was added 6.0 ml. of acetyl chloride. The mixture was stirred magnetically under reflux for 2 hr., after which time an additional 2.5 tnl. of acetyl chloride was added and refluxing was continued for 1 more hr. The material was cooled and poured into about 350 ml. of anhydrous ether. The precipitate was filtered and washed with ether until the odor of acetyl chloride could no longer be detected. The ester hydrochloride was dissolved in the minimum amount of absolute alcohol, anhydrous ether was carefully layered over the solution, and the mixture was refrigerated overnight. The resulting crystals were filtered and the procedure was repeated until a constant melting point was obtained. The final product was dried in vacuo over P z O ~ ; yield 0.48 g. ( 5 5 % ) of 111, m.p. 130-131'. Anal. Calcd. for CiH1,S20zCl: S , 14.70; C1, 18.60. Found: S , 14.42; C1, 18.32. 1-Methyl-5-hydroxymethylimidazolyl Succinate (IV).1-Methyl-5-hydroxymethyliniidazole ( 1.O g., 0.009 mole) was added t o a suspension of 1.0 g. (0.01 mole) of succinic anhydride in 10 ml. of dry xylene. The mixture was stirred magnetically under reflux for 5 h r . , the reaction mixture was cooled, and the product collected by filtration. The precipitate was repeatedly washed with boiling xylene to remove the unreacted anhydride; 1.78 g. (94% ) of 1-1nethyl-5-hydroxymethyliniidazolyl succinate -.
(34) J R . Totler and W. J. D a r b y , "Organic Syntheses," Coll. Vol 111, John Wiley a n d Sons, Inc., N e w Y o r k , X . Y . , 193.5, p 460. ( 3 5 ) R . G. Jones and K C. McLaughlin, J , A m . Chem Soc , 71, 2-144 (1949). (36) R . G Jones, ibid , 71,644 (1949).
4116
THOMAS C. BRUICEAND J. L. HERZ
was obtained, m.p. 149-155' (dec.) The compound was recrystallized twice from ethanol-acetonitrile and dried in vacuo over P20a. I n this manner 0.78 g . of analytically pure material was obtained, m . p . 169-171" (decomposed compound was p u t into melting block a t 140'). Anal. Calcd. for C ~ H U O ~ NC,~ :50.94; H , 5.70; K, 13.20. Found: C , 50.90; H , 5.68; N , 13.02. Sodium Monomethyl Succinate .-Monomethyl succinate was prepared from the alcohol and succinic anhydride"; 0.58g. of the ester was dissolved in 30 ml. of methanol in which was suspended 1.16 g. of a n ion exchange resin (Bio-rex 70, Cal-Biochem, 100200 mesh, sodium form). iZfter 20 min., the suspension was filtered and a large excess of anhydrous ether was added t o the filtrate. The precipitated salt was filtered and washed with ether; 0.49 g. of the sodium salt was obtained (717,). 4( 5)-Hydroxymethylimidazolyl p-Nitrobenzoate (XI).-To 0.2 g. (0.002 mole) of 4(5)-hydroxymethylirnidazole suspended in 35 ml. of hot ethyl acetate was added a solution of 0.55 g . (0.003 mole) of freshly recrystallized p-nitrobenzoyl chloride in 5 ml. of ethyl acetate. T h e solution was refluxed for 7 5 min. T h e resulting suspension was filtered, and the ester hydrochloride was washed with benzene and then with ether; 0.48 g . (947,) of a compound melting a t 164-166' was obtained; after repeated recrystallization from absolute alcohol and alcohol-acetonitrile mixtures, the melting point was 186-187.5'. A n a l . Calcd. for C I ~ H I O O ~ S ~ C ,I :46.57; H , 3.55; K, 14.81; C1, 12.50. Found: C, 46.21; H , 3.72; S , 15.09; CI, 12.38. Reaction of IV with Sodium Methoxide in Methanol.-To a solution of 0.102 g. (0.00189 mole) of sodium methoxide (Matheson Coleman and Bell) in 10 ml. of anhydrous methanol was added 0.20 g. (0.000943 mole) of 1-methyl-5-hydroxymethylimidazolyl succinate (IV). The solution was stirred in a nitrogen atmosphere a t room temperature for 3 h r . ilnhyrous ether was added to the solution until no more precipitate was seen t o form. The precipitate was filtered, washed with anhydrous ether, and dried over P20sunder vacuum oven a t 50". The yield was 0.0682 g. (47%) of a compound whose infrared spectrum was identical with t h a t of an authentic sample of sodium monomethyl succinate. Reaction of I11 with Hydroxylamine .-To determine the extent of reaction of I11 with hydroxylamine, the reaction of p-nitrophenyl acetate ( p - S P A ) with hydroxylamine was used as a reference. It was assumed t h a t $ - S P A reacted t o provide a quantitative yield of acethydroxamic acid. The standard solutions have been described previously38and a reaction time of 45 min. a t 65' was employed. The extent of reaction t o form acethydroxamic acid as determined from a Beer's line employing p-SPA as standard was 92.5%. Product Analysis for the Hydrolysis of V.-A small amount (ca. 0.01 9 , ) of 4(5)-chloromethylinlidazole ( a s the hydrochloride V ) was dissolved in -5 ml. of water, and let stand a t room temperature for 2 hr. I-arying amounts of the solution were spotted on Whatman N o . 1 paper and paper chromatographed for 12 hr. in both 3 : 1 propanol-1 *V acetic acid and 3 : 1 propanol-0.2 iv XH3. The chromatograms were developed with Pauli coupling reagent as described by Ames and Mitchell.39 Only one spot was found for any of the chromatograms. This spot had the same R , value a s an authentic sample of 4(5)-hydroxymethylimidazole employed a s a control. Some chromatograms were developed with ninhydrin spray t o detect any amines which might result from ring cleavage, but no other product b u t 4(5)-hydroxymethylimidazole could be found. Kinetics.-All the rate constants reported herein were determined on a Radiometer TTT l b pH-stat assembly a s previously d e ~ c r i b e d . ~Constant temperature, fO.l O , in the microtitration cell was obtained with a Haake circulating bath. The sample solution (ca. 2 ml.) was titrated with -0.01 N KOH ( P = 1.00 with KCI) delivered through an Alga trubore syringe. T h e pseudo-first-order rate constants were calculated by the method of Guggenheim,@ and some data was spot checked by plotting log ( a / ( a - x ) against time. (37) W . A , Bone, J. J. Sudborough, a n d C. H. G. Sprankling, J. Chem. Soc., 86, 535 (1904). (38) T . C . Bruice a n d F. H. h l a r q u a r t , J . A m . Chem. Soc., 84, 365 (1962). (39) B . N . Ames a n d H. K. IIitchell, i b r d . , 74, 262 (1952). (40) E. A. Guggenheim, Phii. M a g . , a , 538 (1926).
Vol. 86
pK2 of Imidazole.-The pK2 of imidazole in H 2 0 and 99.5y0 D10 was measured conductometrically by the method of Ballinger and Long.23 When imidazole is dissolved in dilute S a O H , the following equilibrium is established
e
H
The resistance (corrected for viscosity), R,, of an imidazole solution in dilute NaOH is related t o the resistance of a solution of the same concentration of NaOH in the absence of imidazole, Ro, in the following way
R, _ Ro
AcNaOH
+ (1 - ~ X , O +H ~a ~ , ~ m(34)s ~~
XcXa@
where A,Sa@, X,OHS, and Acheare the ionic mobilities of the ions a t their respective concentrations. The equivalent conductance of the alkaline solvent a t concentration C is A,NaOH. Both in H20 and D 2 0 , the ionic mobility of the imidazole anion, Ache, was determined graphically.2o The values are 28.8 and 23.6 mhos, respectively. The ratio of the ionic mobilities in HzO and DzO, as might be expected from a consideration of Stoke's law, was the inverse ratio of the viscosities of the solvents. Solving (34) for cy and substituting into (33), K may be calculated. Thus, from a knowledge of K,,.,K? may be obtained. Because of its high pK2, concentrated solutions of imidazole were employed so that it was necessary to make viscosity corrections. The corrected resistance R, = R i ( v ~ / q i )where , R Lis the measured resistance and v ~ / v i is the ratio of the viscosity of the alkaline solvent t o t h a t of the imidazole solution. Conductivity measurements were made on the apparatus described by Ballinger and Long.23 The cell constant for the conductivity cell was 1.86155 cm .-I. Viscosities were determined using a Cannon-Fenske type viscometer, and the data obtained were reproducible t o better than 0.2Yc. The temperature was kept a t 25.0 f 0.01" for all measurements. Imidazole (Eastman Kodak White label) was recrystallized twice from a mixture of acetone-petroleum ether (60-90°), and was dried in uaruo over a mixture of P2OS and paraffin wax. Standard stock solutions of S a O H and S a O D in H20 and D,O were made up. The imidazole was weighed into a dry 50-nil. volumetric flask and 10.0 ml. of the stock S a O H solution was added. The imidazole was brought into solution and H20 (or D 2 0 ) was added to the mark. Only CO?-free deionized HnO was used, and the D 2 0 employed was boiled t o exclude Con. Conductivity measurements were made on these solutions directly; however, the solutions were filtered free of any dust particles before viscosities were determined (Table I ) . The average value of K 2for imidazole is 2.48 f 0.01 X in H20 and 5.15 i 0.07 X 10-I6in D 2 0 . TABLE I Imidazole, Kz M R , ohms SO/? Re,, x 10-16 Solvent: 0.03008 M hTaOH in H20 0 262.61 1,0000 262.6 1.006 332.24 0,9084 301.8 2.47 288.5 2.48 0.6583 308.45 9353 277 2 2.48 9656 ,3555 287,04 2.47 ,9777 271.2 277.41 ,2090 Solvent: 0.01013 J4 KaOD in ca. 99.5% D20 0 1198.6 1 0000 1199 0 9065 1397 3.99 0,9739 1540.9 1358 5 24 1462 6 ,9282 ,7283 5.22 ,9691 1308 1308.0 ,2996
Acknowledgment.-This work was supported by grants from the National Institutes of Health and the National Science Foundation.
