J . Org. Chem., Vol. 43, No. 9, 1978
Model Studies of Thiamin Catalysis
1713
Model Studies of Thiamin Catalysis. Comparison of the Effects of Heteroatoms at Annular Positions on Side-Chain Kinetic Acidity John A. Zoltewicz* and John K. O'Halloran Department of Chemistry, Uniuersity of Florida, Gainesuille, Florida 3261 I Received September 8, 1977 Creneral-base-catalyzed deprotonation of the 2-methyl groups of 2,3-dimethylbenzothiazolium (II), 1,2,3-tri(IV) ions was studied by means of hydrogen-deutemethylbenzimidazolium (III),and 1,2,3,-trimethylimidazolium riuin isotope exchange. At 7 5 "C and 1.0 M ionic strength 11,111,and IV show the following relative reactivity order toward deuterioxide ion: 3.0 X IO5; 3.4 X lo2; 1,respectively. Toward pyridine I1 is 9.5 X lo3 times more reactive than 111. The Br#nsted fl value for a series of bases, excluding water and lyate ion, reacting with I1 is 0.63. A more limited series of bases gives the value 0.8 for 111. A comparison is made between the reactivities of sp2- and sp3-hyhridized carbon centers in deprotonation reactions involving thiazolium ions.
Thiamin (I) or vitamin B-1 when catalyzing carbon-carbon bond forming reactions is converted to an ylide and subsequently to an enamine intermediate, both of which act as nucleophiles toward carbonyl ele~trophiles.l-~ Many studies have been carried out to obtain an understanding of the various steps in the multistep catalyzed conversions. For example, the ability of several annular heteroatoms to facilitate ylide formation5 in deprotonation reactions, eq I, has been explored. HC€H?CH-
3% -7
CH,
c1-
CHAR
I, R = 2-methyl-4-amino-5-pyrimidinyl
CH
CH
>
Results show that reactivity increases in the order X=NCHs, S, and Os6By contrast, little is known about the influence of similar heteroatoms on the kinetic acidity of an alkyl side chain which gives an enamine rather than an ylide on proton loss. In this article we examine the effects of annular heteroatoms on the kinetic acidities of ions 11-V containing an acidic methyl group. Equation 2 shows the enamine which forms
CH
CH >
1
11, x = s 111, X = NCH,
VI,
x=s
v,x=o
CH \
I
tN,&CH.
+T
CHI
n. when fused-ring heteroaromatic ions 11, 111, and V undergo deprotonation. The accompanying paper explores the influence of methyl and hydroxy substituents bonded to the 2methyl group of I1 on the kinetic acidity of this p o ~ i t i o n . ~ 0022-3263/78/1943-1713$01.00/0
Results No evidence exists to indicate that the 4-methyl and 5-phydroxyethyl substituents of the thiazolium ion ring of I provide other than expected, small, electronic effects on deprotonation reactions at position 2 under biological conditions. For this reason these substituents were not incorporated into our models. In order to simplify our models a methyl group was added to N-3 in place of the pyrimidinylmethyl substituent. Because of their ready availability and because they contain the essential structural elements, compounds 11-V were selected as model substrates. The conjugate base of 11, 3-methyl-2-methylenebenzothiazolene (VI), which is formed as an intermediate in our reactions has been isolated.* All of the studies employ NMR to determine rates of carbon deprotonation. Reactions were carried out using DzO as the reaction medium, leading to replacement of H by D in the side chains. Except where indicated, the reaction temperature is 75.0 "C. 2,3-DimethylbenzothiazoliumIon (11). Deprotonation of the C-methyl group of this substrate was found to take place readily in 0.1 M DC1. Catalysis was also observed with the following buffers, listed in order of increasing basicity: 3chloropyridine, phthalazine, formic acid, acetic acid, pyridine, and 2,6-lutidine (Table I). Buffer ratios and pD were varied in order to determine whether water, buffer base, and deuterioxide ion contributed to the general-base-catalyzed deprotonation reactions. Pseudo-first-order rate constants were analyzed using eq 3
where kD,o, k g , and koD are second-order rate constantsfor water, buffer, and deuterioxide ion bases, respectively, and K, and KwDare the dissociation constants of buffer and the ion product of D20, respectively. Significant catalysis by deuterioxide ion was observed only with the most basic buffer, 2,6-lutidine. Thus, the value of hoD was established using the data derived with this buffer. Application of the term k o ~ [ o D - ]to the data obtained with other buffers indicates that deuterioxide ion measurably influences reactivity only in two other kinetic runs, those a t pD 4.09 involving acetic acid and pyridine buffers. The kinetic contribution of this base represents about 13% of the total catalysis in both cases. With these two kinetic runs and also the lutidine studies as exceptions, deprotonation in the other buffered media was catalyzed only by water and by buffer base. Since the koD term was obtained with only one buffer, another method was employed as a check. This involved the use of a pH-stat in place of a buffer to adjust the pD of the reaction medium. From runs a t pD 5.25 and 5.45 this approach gave C 1978 American Chemical Society
1714 J . Org. Chem., Vol. 43, No. 9, 1978
Zoltewicz and O'Halloran
Table I. Kinetic Results for Hydrogen-Deuterium Exchange a t the 2-Methyl Group of 2,3-Dimethylbenzthiazolium Ion (11) in Buffered D20 at 75.0 "C and 1.0 M Ionic Strength Buffer
PKa
PDU
DC1 3-Chloropyridine
__
__
3.07
Phthalazine'
3.79
2.49 2.65 2.73 3.20 3.37 3.22 3.33 3.33 3.35 3.75 4.09 3.98 4.09 5.10 5.20 5.43 5.51
Formic acid
4.03
Acetic acid
5.12
Pyridine
5.27
2,6-Lutidine
6.44
DC1f.g Acetic acidfsh
__
. _
5.25
Total buffer, M 0.1 1.20 x 10-1 4.80 x 2.40 x 1.01 x 10-1 2.12 x 10-2 4.40 x 10-2 3.33 x 10-2 1.10 x 10-2 2.20 x 10-2 4.20 x 10-2 8.33 x 10-3 4.20 X 8.44 x 10-3 1.20 x 10-1 6.05 X 6.05 x 10-3 2.00 x 10-3 0.1 4.54 x 10-2 8.48 x 10-3
4.27
4.27
*
Obsd
Calcd
lo5 k + , s-1
lo5 k 4 , s-1
2.65 f 0.23c 13.15 8.00 5.23 34.9 11.45 12.1 11.5 5.17 8.48 23.7 9.60 22.1
8.33 99.3 60.8 38.7 40.6 0.010 0.270 0.0544
12.3 7.80 5.61 34.9 11.7 12.2
11.6 5.61 8.82 20.4 11.0 21.8 8.42 96.5 68.8 35.8 35.7
__
0.260' 0.0567'
k B , M-ls-1
(4.90 x lo-' d j 3.84 X 1.56 X 10-2 1.62 x 10-2
1.00 x 10-1 8.90 x 10-2 1.57 x 10-1 (B) 3.08 x 104 (OD-) (1.81 x 10-9 d ) 5.81 x 10-4 1
Measured at 75.0 "C. Using equation 3. Average of four determinations. d k+/[D2O]. e 2,3-Diazanaphthalene. i At 25.0 "C and 0.5 M ionic strength. g Vi* = 22.4 kcal/mol; AS* = -4.12 eu. VI* = 20.7 kcal/mol; AS* = -4.12 eu. Neglects any contribution from OD-.
/;i
I/" I.. 6
3
/
../ 4
/
-4
j
I
-0
L +.4-6 0
-2
L
1
,
0
30
'2
,
14
:
PKa
Figure 1. Brpinsted plot for hydrogen-deuterium exchange at the %methyl group of I1 in D20 at 75 "C and 1.0 M ionic strength. Buffers include: 1,water; 2,3-chloropyridine;3, phthalazine; 4, formic acid; 5 , acetic acid; 6, pyridine;7,2,6-lutidine;and 8, deuterioxide ion. No statistical corrections have been applied. The least-squaresline does not include point; 1 , 7 , and 8.
a value of 3.80 0.16 X lo4 M-l s-l which is only 23% larger than that derived from the lutidine studies. The quality of our results may be assessed by comparing the observed pseudo-first-order rate constants with those calculated with the aid of eq 3 and the second-order constants listed in Table I. In carrying out the computations the buffer-derived value of koD was employed. The largest differences are found in the results obtained with acetic acid and lutidine buffers where the results with the poorest agreement differ
by about 14%. Differences between observed and calculated values are substantially less in all other cases. The spread in second-order rate constants between the least (DzO) and most (OD-) reactive bases is a factor of 6.3 X lolo. The variation in reactivity for buffer bases unrelated to solvent is a factor of 41. A Brdnsted plot may be constructed using the results in Table I. Without making any statistical corrections for the number of basic sites in a buffer a single plot (Figure 1) is obtained, excluding points for water, deuterioxide ion, and 2,6-lutidine. The slope, 6 , is 0.63 and the intercept is -4.305. The correlation coefficient is satisfactory, being 0.991. On the basis of this plot, 2,g-lutidine is 3.7 times less reactive than expected from a consideration of its pKa value, no doubt reflecting a modest steric hindrance to general base c a t a l y ~ i s . ~ Water and deuterioxide ion deviate; observed reactivity in both cases is about 8 times less than that calculated. Negative deviations for these are not uncommon. The reduced reactivities of the two bases with carbon acids is said to be a consequence of a lack of hydrogen bonding between reactants in the ground state.lOJ1 When rate and equilibrium constants are statistically corrected for the two basic centers found in phthalazine and in the carboxylate anions, two correlation lines now are generated, one for carboxylic acids and one for pyridines. The line for carboxylate ion bases is displaced above that for the pyridines. Phthalazine now deviates from the pyridine correlation line in the sense that its reactivity is about 4 times greater than that predicted by its basicity. Phthalazine is known to show an enhanced reactivity ( a effect) toward esters,'* but it is likely that the present deviation is not significant. Thorough studies of a-effect nucleophiles in carbon deprotonation reactions have failed to uncover enhanced reactivities.13J4 Although it is customary to make statistical corrections in constructing Br$nsted plots,15 good correlations using uncorrected data are known for carbon deprotonation.13 Some kinetic information was also obtained a t 25.0 O C so that the reactivity of I1 could be compared with that of the C-methyl group of its acyclic relative VII, N,N-dimethyl-
J. Org. Chem., Vol. 43, No. 9, 1978
Model Studies of Thiamin Catalysis
1715
Table 11. Kinetic Results for Hydrogen-Deuterium Exchange at the 2-Methyl Groups of 1,2,3-trimethylbenzimidazoliumIon (111) and of 1,2,3-TrimethylimidazoliurnIon (IV) in Buffered D20 a t 75.0 "C and 1.0 M Ionic Strength Compd I11
IV
Buffer DC1 Pyridine
PK, -5.27
Phosphate
7.10
Glycine
9.17
Carbonate
9.93
PD"
__ 5.60 5.76 6.61 6.98 7.12
Buffer, M 0.1
7.50
0.303 0.0303 0.140 0.0200 0.280 0.300
7.83
0.0550
7.87
0.110 0.220 0.120 0.350 0.275
8.17 10.00 10.25 10.86
Obsd k,, s - ~
Calcd