The term ñ stability of a compound is associat- ed with the determination of the rate of its changes during the time of a storage under defined condi- tions. The rate of degradation of the substance depends on the external factors (temperature, light, pH of the medium) and on the chemical properties of the substance determined by its chemical structure.
Earlier we investigated hydrolytic stability of 5-ethyl-5-phenyl-2-thioxo-4,6(1H,3H,5H)-pyrim- idinedione (thiophenobarbital, I) and its N-mono- (II) and N,Nídimethyl (III) derivatives in solution (1-3). Now, we were interested how this stability is influenced by an addition of β-cyclodextrin (β-CD) ñ a macrocyclic compound build of seven D-glu- copyranose units connected with (1→4) glycosidic bonds. The UV spectrophotometry was used for determination of rate constants of hydrolysis of compounds I-III without and with different β- cyclodextrin concentrations under different pH and temperature conditions. Assuming the complex for- mation between the investigated thiobarbiturates and β-cyclodextrin, suggested by NMR spectra, an attempt was made to elucidate the structure of these complexes by molecular modeling using semiempir- ical AM1 parametrization (4-8).
MATERIALS AND METHODS
2-Thiophenobarbital (I) and its mono N- methyl derivative (II) were synthesized by conden-
sation of diethyl ethylphenylmalonate with thiourea and N-methylthiourea, respectively (9), whereas N,Ní-dimethyl derivative (III) was obtained by methylation of I with Lawessonís reagent (10).
Pure β-cyclodextrin (Cyclolab Ltd., Budapest, Hungary), D2O (99,9%) and CDCl3 (99,8%) (Deutero GmbH, Germany), standard buffer solu- tions and other chemicals from POCH (Gliwice, Poland) were used.
Phosphate buffer acc. to Michaelis, as well as borate buffer and sodium hydroxide solutions acc. to Bates and Bower (11) were used as reaction media for pH ranges 5-8, 8-11, and 11-12, respectively.
Kinetic measurements
Standard solutions of the investigated com- pounds (c = 10-3M) were prepared in ethanol. 1 mL of the standard solution was diluted to 10 mL with the appropriate buffer containing 0.0, 0.5, 1.0, and 1.5 % of β-CD and the UV spectra of the solutions were recorded (apparatus UV-VIS CECIL BioAquarius 7250, England). Kinetic runs were carried out at constant temperature at different pH values measured with microcomputer pH-meter P- 315M (Elmetron, Poland) and combined electrode ESAg-201W calibrated against standard buffer solutions of pH 4, 7 and 10. The absorbance values at the maximal wavelength of the hydrolyzed solu- tions were measured at appropriate time intervals.
The pH ranges and temperatures for the runs were:
STABILITY OF 2-THIOPHENOBARBITAL AND ITS N-METHYL DERIVATIVES IN THE PRESENCE OF β-CYCLODEXTRIN
GRZEGORZ ØUCHOWSKI, MONIKA TARSA, ANNA STASIEWICZ-URBAN and JACEK BOJARSKI*
Department of Organic Chemistry, Medical College, Jagiellonian University, Medyczna 9, 30-688 KrakÛw, Poland
Abstract: The rates of hydrolysis of thiophenobarbital and its N-mono- and N,Ní-dimethyl-derivatives were determined under different conditions of pH and temperature using UV spectroscopy. They were compared with those obtained in the presence of different concentrations of β-cyclodextrin. It was found that the com- pounds degrade with different rates and β-cyclodextrin retards the hydrolysis. The formation of complexes between the investigated compounds and β-cyclodextrin was proved by 13C NMR and ROESY spectra and molecular modeling. The inclusion with the phenyl substituent into the β-cyclodextrin cavity is preferred.
Keywords:thiophenobarbital, N-methyl derivatives, hydrolysis, β-cyclodextrin, complexes
335
* Corresponding author: e-mail: mfbojars@cyf-kr.edu.pl
Table 1. Calculated values of hydrolysis rate constants (k) for the investigated compounds
0.0% β-CD 0.5% β-CD 1.0% β-CD 1.5% β-CD
k × 10-5 k × 10-5 k × 10-5 k × 10-5
pH [s-1] log k
[s-1] log k
[s-1] log k
[s-1] log k Compound I
5.98 5.065 -4.295 3.693 -4.433 3.647 -4.438 3.306 -4.481
6.59 6.772 -4.169 6.281 -4.202 5.407 -4.267 4.727 -4.326
6.91 8.827 -4.054 8.575 -4.067 8.069 -4.093 7.642 -4.117
7.51 10.818 -3.966 10.073 -3.997 9.450 -4.025 8.007 -4.097
8.51 11.581 -3.936 10.458 -3.981 9.962 -4.002 8.751 -4.058
8.9 13.099 -3.883 11.233 -3.950 10.636 -3.973 10.527 -3.978
9.82 26.103 -3.583 21.859 -3.660 21.204 -3.674 18.824 -3.725
10.84 31.203 -3.506 31.132 -3.507 28.138 -3.551 27.312 -3.564
Compound II
6.00 0.148 -5.873 0.054 -6.265 0.041 -6.598 0.024 -6.735
6.44 0.264 -5.579 0.068 -6.190 0.057 -6.252 0.051 -6.340
7.07 1.476 -4.831 0.319 -5.496 0.232 -5.635 0.218 -5.661
7.54 3.277 -4.492 0.650 -5.187 0.516 -5.291 0.456 -5.342
7.85 5.738 -4.242 1.637 -4.786 1.164 -4.935 1.095 -4.961
8.72 6.647 -4.178 3.467 -4.461 2.546 -4.594 2.155 -4.667
9.15 7.133 -4.147 4.693 -4.329 3.619 -4.441 3.119 -4.506
9.98 7.036 -4.155 5.809 -4.242 4.583 -4.344 4.026 -4.399
10.48 7.477 -4.126 5.890 -4.230 4.653 -4.332 4.297 -4.367
11.05 12.986 -3.887 9.716 -4.013 7.686 -4.114 7.130 -4.147
11.57 17.171 -3.767 12.033 -3.921 9.745 -4.013 8.457 -4.073
12.07 35.056 -3.455 18.413 -3.735 16.443 -3.784 14.060 -3.852
Compound III
5.98 5.345 -4.272 3.292 -4.483 2.852 -4.545 1.584 -4.800
6.93 6.311 -4.199 4.573 -4.340 2.885 -4.540 2.875 -4.541
7.52 9.572 -4.019 7.879 -4.104 4.675 -4.330 4.473 -4.3494
8.07 19.488 -3.710 13.201 -3.879 7.804 -4.108 7.766 -4.110
8.43 79.891 -3.098 45.350 -3.343 25.669 -3.591 17.039 -3.769
8.8 174.873 -2.757 99.517 -3.002 59.477 -3.226 34.931 -3.457
9.98 285.141 -2.545 276.172 -2.559 274.745 -2.561 176.263 -2.754
10.84 516.013 -2.287 402.742 -2.395 343.829 -2.464 299.589 -2.524
5.98 ñ 10.84 at 70OC, 6.00 ñ 12.07 at 25OC and 5.98 ñ 10.84 at 50OC for I, II, and III, respectively (Table 1).
The activation energies were calculated acc. to Arrhenius equation from the rate constants at 70, 60, 50 and 30OC for I, 40, 35, 30, 25 and 20OC for II and 65, 60, 55 and 50OC for III (Table 2 and 3). Deter- mination of the constants of complex formation (Kc) and dissociation (kc) was done in the buffer solution of pH 9.98 at 35OC for all the investigated com- pounds.
NMR spectroscopy
All spectra were recorded on Varian Mercury- VX spectrometer operating at 300.08 MHz. Chemical shifts are referenced to solvent lock signal (D2O or CDCl3). Samples were prepared by the following pro- cedure: for β-CD ñ 0.01 g, for the investigated com- pounds ñ 0.002 g and for the mixture of β-CD and barbiturates 0.01 g of β-CD and 0.002 g of the inves- tigated compound were dissolved in 0.7 mL of D2O.
1H NMR and 13C NMR spectra were recorded for all samples. 1H NMR spectra were acquired with 16
scans, 13C NMR spectra were acquired with 16000 scans. 2D ROESY spectra for mixtures of the investi- gated thiobarbiturates and β-CD were measured with 0.4 s mixing time, 128 repetitions for each of 256 t1
increments. All spectra were recorded at 28OC.
Molecular modeling
Optimizations of geometry of the investigat- ed compounds and β-CD yielded the energies of
free compounds (Ebarband E‚-CD). Then, to the inner space of the β-CD toroid the investigated com- pound was introduced with pyrimidine or phenyl ring as far as possible (with increments of 0.1 Å) and the optimization of geometry was repeated. In the next step the ring was rotated inside the β-CD cavity by 15O. This procedure gave 440 starting structures for which the geometry was optimized acc. to (12, 13). The obtained energies of the com-
Table 2. Calculated rate constants of hydrolysis reaction at various temperatures for the investigated compounds
0.0% β-CD 0.5% β-CD 1.0% β-CD 1.5% β-CD
Temp. 1/T k×10-4 k×10-4 k×10-4 k×10-4
[K] [K-1] [s-1] log k [s-1] log k [s-1] log k [s-1] log k Compound I
303 0.00325 0.2470 -4.607 0.2183 -4.661 0.1986 -4.702 0.1959 -4.708 323 0.00310 1.0139 -3.994 0.7898 -4.103 0.6926 -4.160 0.5288 -4.277 333 0.00300 1.8994 -3.721 1.8143 -3.741 1.8130 -3.742 1.6510 -3.782 343 0.00292 3.1203 -3.506 3.1132 -3.507 2.8138 -3.551 2.7312 -3.563
Compound II
293 0.00341 0.5273 -4.2782 0.39589 -4.402 0.31537 -4.501 0.28329 -4.548 298 0.00336 0.7915 -4.1017 0.62538 -4.204 0.50784 -4.294 0.41895 -4.378 303 0.00330 1.3606 -3.8663 1.03570 -3.985 0.87817 -4.056 0.79699 -4.099 308 0.00325 1.9718 -3.7053 1.52292 -3.817 1.26619 -3.898 1.16238 -3.935 313 0.00319 2.8840 -3.5403 2.26053 -3.646 1.92625 -3.715 1.74523 -3.758
Compound III
323 0.00310 10.042 -2.998 4.933 -3.307 4.442 -3.352 2.645 -3.578
328 0.00305 14.014 -2.854 10.819 -2.966 7.128 -3.147 5.521 -3.258
333 0.00300 19.213 -2.716 18.369 -2.736 15.922 -2.798 10.599 -2.975 338 0.00296 28.511 -2.545 27.617 -2.559 27.475 -2.561 17.626 -2.754
Table 3. Activation energies (Ea), and rate constants (k) and half-life times (t50% ) of hydrolyses calculated for temp. 35OC and pH = 9.98
0.0% β-CD* 0.5% β-CD 1.0% β-CD 1.5% β-CD
Compound k×10-4 t50% Ea k×10-4 t50% Ea k×10-4 t50% Ea k×10-4 t50% Ea
[s-1] [h] [kJ/mol] [s-1] [h] [kJ/mol] [s-1] [h] [kJ/mol] [s-1] [h] [kJ/mol]
I 0.273 7.06 64.06 0.226 8.51 67.77 0.205 9.40 68.60 0.185 10.42 68.66 II 1.959 0.98 65.78 1.521 1.27 66.75 1.280 1.50 69.18 1.154 1.67 71.06 III 3.177 0.61 62.57 0.825 2.33 103.62 0.533 3.61 113.78 0.342 5.64 115.25
* published earlier [1]
Table 4. Calculated hydrolysis reaction rate constants for pure and complexed compounds, k0/kcfraction and complexation constants (temp.
35oC, pH=9.98)
Hydrolysis reaction rate constant Complexation
free complexed compound reaction constant
compound (calculated)
Compound k0×10-5 [s-1] kc×10-5[s-1] k0/kc Kc[M-1]
I 2.727 1.235 2.208 100.243
II 19.585 5.085 3.852 96.629
III 31.768 0.531 59.850 672.238
(Figure 1). This proves that the Lambert-Beerís Law holds and the UV spectrophotometry may be used for kinetic measurements.
The changes of logarithms of absorbance dur- ing the reaction course for the solutions of the inves- tigated compounds at different pH values were lin- ear (Figure 2) thus proving that the processes were of the pseudo-first order. The rate constants for the runs (Table 1) were calculated acc. to the expres- sion:
log(A-A∞) = log(A0-A∞)- k⋅t 2.303
where k is the rate constant in [1/s]; t is time in sec- onds; A0is the absorbance of the solution at time t = 0; A is the absorbance of the solution at time t and A∞is the final absorbance at time t∞.
Activation energies (Ea) and rate constants and half-times for hydrolyses of I-III calculated for 35OC and pH = 9.98 are presented in Table 3.
The observed rate constants (kobs) for hydrolyses of I-III in the presence of β-CD depend on the rate
Figure 1. Plot of the absorbance as a function of the concentration of compound I
Figure 2. Plot of log(A-A∞) = f(t) dependence in buffers with var- ious pH (cβ-CD=1.0%) for compound I
Table 5. 13C-NMR chemical shift change for β-CD after addition of the investigated compounds
Chemical shift change [ppm]*
Atom (C) Compound I Compound II Compound III
1 0.197 0.070 0.116
2 0.060 0.074 0.101
3 0.034 0.071 0.074
4 0.091 0.101 0.143
5 0.019 0.063 0.059
6 -0.057 0.048 0.051
* ∆δ = δcompñ δβ-CD; positive values indicate downfield shift.
Table 6. Energy gain for complex formation (calculated by AM1 method)
Pyrimidine ring Phenyl ring insertion insertion Compound [kcal/mol] [kcal/mol]
I (unionized) 3.214 2.924
I (anion) 25.256 21.520
II (unionized)* 3.653 2.626
II (anion)* 24.651 19.401
III (unionized) 2.907 2.208
* published earlier [12]
plexes (Ecomp) served for estimation of their struc- tures. The calculations were run both for the unionized molecules (for I-III) and for monoan- ions (for I and II) using semiempirical parame- trization AM1 (4-8) and programme GAMESS version 20 JUN 2002 (R2) (14) from the State University of Iowa.
RESULTS AND DISCUSSION
Linear relationship was found by regression between the UV absorbance and the concentration of the investigated thiobarbiturates within the inves- tigated concentration range (2.0 × 10-5ñ 1 × 10-4M)
constants for hydrolysis for free (ko) and complexed (kc) molecules and the complexation constant Kc) as presented in the scheme:
degradation products and are given by the expression (15):
kobs= k0+kcKc
1+Kc[CD]
From this it follows that
1 = 1 - 1 k0- kobs Kc(ko- kc)[CD] (k0-kc)
and the rate constants for complex degradation and complex formation may be calculated by Lineweaver-Burk method (15,16).
Figure 3 represents the example of the rela- tionship
1 = f
(
1)
k0ñ kobs [CD]
and the respective values (for 35oC and pH = 9.98) are presented in Table 4.
The formation of complexes of the investigated compounds with β-CD was confirmed by comparison of their 13C NMR spectra with that of free β-CD. The shifts of respective signals are listed in Table 5. An additional proof of complexation was obtained from the ROESY spectra where correlation signals between the protons from β-CD cavity and the protons of the investigated compounds were observed (Figure 4).
Figure 5. Structures of the most persistent complexes of neutral form of compound I with β-CD. A ñ view from the top, B ñ view from the side
Figure 3. Plot of 1
= f
(
1)
function for compound I k0ñ kobs [CD]Figure 4. ROESY spectrum of compound I and β-CD complex (important ROE signals are indicated by arrow)
Molecular modeling
In the case of complexation the following rela- tionship should hold:
Ecomp< Eβ-CD+ Ebarb
The complex with the lowest energy was cho- sen as the most probable. The energetic gain values for penetrations of β-CD cavity with phenyl or pyrimidine rings of the investigated compounds are presented in Table 6 and the example of the most stable complex of compound and β-CD may be seen in Figure 5.
CONCLUSIONS
The hydrolytic degradation of all the investi- gated compounds is catalyzed by OH- ions, as proved by increasing rate constants of the process concomitant with an increase of the pH values of the reaction medium (Table 1). The process demon- strates the pseudo-first order kinetics as evidenced by linear relationship between log (A-A∞) and time (Figure 2).
The overall rate of hydrolysis is the highest for compound II and the lowest for compound I at all the investigated concentrations of β-CD. However, for the hydrolysis of complexes alone the order of decreased rate is III < I < II. Thus, β-CD inhibition of hydrolysis is the highest for compound III, i.e.
these complexes are the most stable. The methyl substitutions at the nitrogen atoms of the parent compound I favors the hydrolytic degradation of the respective molecules.
Both 13C NMR and ROESY spectra confirm the formation of complexes between β-CD and the investigated compounds and the most favorable pen- etration of β-CD cavity is by the phenyl ring of thio- phenobarbital and their N-mono- and N,Ní-dimethyl derivatives.
Acknowledgment
The grant no. G-22-8 from the Interdisciplinary Centre for Mathematical and Computer Modelling of Warsaw University is gratefully acknowledged.
REFERENCES
1. Tarsa M., Bojarski J.: Bull. Pol. Acad. Sci., Chemistry 45, 63 (1997).
2. Tarsa M., Øuchowski G., Bojarski J.: Acta Pol.
Pharm.-Drug. Res. 60, 247 (2003).
3. Tarsa M., Øuchowski G., Bojarski J.: Acta Pol.
Pharm.-Drug Res. 60, 253 (2003).
4. Dewar M., Thiel W.: J. Am. Chem. Soc. 99, 4499 (1977).
5. Dewar M.J.S., McKee M.L., Rzepa H.S.: J. Am.
Chem. Soc. 100, 3607 (1978).
6. Davis L.P.: J. Comput. Chem. 2, 433 (1981).
7. Dewar M.J.S., Zoebisch G., Healy E.F.: J. Am.
Chem. Soc. 107, 3902 (1985).
8. Dewar M.J.S., Reynolds C.H.: J. Comput.
Chem. 7, 140, (1986)
9. Crossley F.S., Miller E., Hartung W.M., Moore M.L.: J. Org. Chem. 5, 238 (1940).
10. Stasiewicz-Urban A., Kubaszek M., Øylewski M., Ceg≥a M., Bojarski J.: Polish J. Chem. 78, 2105 (2004).
11. Physicochemical adviser (Poradnik fizyko- chemiczny), WNT, Warszawa, 1974.
12. Øuchowski G., Zborowski K., Bojarski J.:
Chem. Anal. 45, 767 (2000).
13. Zborowski K., Øuchowski G.: Chirality 14, 632 (2002).
14. Schmidt M.W., Baldridge K.K., Boatz J.A., Elbert S.T., Gordon M.S., Jensen J.H., Koseki S., Matsunaga N., Nguyen K.A., Su S.J., Windus T.L., Dupuis M., Montgomery J.A.: J.
Comput. Chem. 14, 1347 (1993).
15. Loftsson T.: Drug Des. Deliv. 1, 22 (1995).
16. Lineweaver H., Burk D.: J. Am. Chem. Soc., 56, 658 (1934).
Received: 9.09.2005
