Complexes of N-methylxanthines with carboxylic acids DENYSCOOKAND ZEPHYRR. REGNIER

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/17/17 For personal use only.

Research Laboratories, Food and Drug Directorate, Department of National Health and Welfnre, Ottawa, Carzada Received May 10, 1968 Complexes of caffeine and theobromine, with acetic and salicylic acid, have been studied by means of infrared spectroscopy. In the complexes with aromatic acids a fairly strong hydrogen bond exists between the carboxylic OH group and the nitrogen atom N-9 in the imidazole ring. The acetic acid complexes involve no strong intermolecular forces and are ~robablvweak lattice com~lexes. The complex& with aromatic acids are discussed in ;elation to theories of lipid solubility and gastrointestinal absorption. Canadian Journal of Chemistry, 46, 3055 (1968)

Introduction The formation of complexes of molecules of the xanthine series, caffeine, theobromine, and theophylline with carboxylic acids has been inferred from a variety of measurements (1-4). Althoughit has been suggested that the complexes with aromatic carboxylic acids are a result of interactions between the z electrons of such rings (4), their physical structure is still in doubt. A study of the infrared (i.r.) spectra of such complexes seemed a possible way of solving this structural problem. This method has already been used to show that in salts of xanthines with strong acids protonation occurs at N-9 in the imidazole ring for caffeine (5), theobromine (6), and theophylline (7), while the spectra of the salts of the last two chemicals with strong bases has enabled the easy identification of their anions (7). This preliminary work has considerably facilitated the interpretation of the i.r. spectra of the complexes. Results and Discussion The principal absorption bands in the spectra of the free bases and the complexes with aromatic acids are listed in Table I.

'

Caffeine Complexes Caffeine - Acetic Acid Complex Several diagnostically useful differences are observed in the spectra of caffeine and those of its salts (6). An increase in carbonyl frequency, v,, and in v, for the C,-H bond, are both 'A complete listing of the bands in the spectra has been Ned with the Depository of Unpublished Data. Photocopies may be obtained free of charge from The Depository of Unpublished Data, National Science Library, National Research Council of Canada, Ottawa, Canada.

attributed to the effect of the positive charge. In addition new bands are present in the salt spectra due to stretching, in-plane bending, and out-of-plane bending of the protonated N:-H group. For the acetic acid complex of caffeine the i.r. spectrum shows nearly all the caffeine bands to be at the same frequency as in the free base, with additional bands due to acetic acid. Of these the strongest is a t 1720 cm-'. The monomer has its carbonyl frequency a t w 1790 cm-' (8), and the dimer at 1715 cm-' (i.r., liquid) and 1675 cm-' (Raman, liquid) (9). The crytsalline solid which is composed of infinite chains of hydrogen-bonded monomers (10) has a doublet at 1659/1645cm-I (i.r., solid, 0 "C) (1 1). The acetic acid in the caffeine - acetic acid complex must therefore be in the dimeric form and has little or no interaction with the caffeine molecules. This conclusion is in accord with the physical properties of the complex: it loses acetic acid readily on exposure to air reverting to caffeine. (The hydrate of caffeine also is efflorescent.) Other bands in the spectrum are at frequencies similar to those in the liquid acetic acid dimer (9) with minor shifts in some bands connected with the O H group vibrations as a result of a slightly different molecular environment. A slight weakening of the acetic acid dimer hydrogen bond strength in the complex may be inferred from the slightly higher value of voH and lower value of yo, in the complex compared to the free base. The alternative structure of a free caffeine molecule fitting into a lattice composed of infinite of hydrogen-bonded units can be dismissed as there is no new band

CANADIAN JOURNAL O F CHEMISTRY.

VOL. 46,

1968

TABLE I Principal absorption bands in the infrared spectra of the free bases and their complexes with aromatic acids

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/17/17 For personal use only.

Caffeine Base

Salicylic acid complex

Theobromine Salicylic acid complex

Base

Salicylic acid

NOTE:Assignments on the left are for the free bases and the complexes. Those on the right are for salicylic acid.

in the i.r. spectrum of the complex at -1650 cm-' (1 1). CafSeine - SalicyIic Acid Complex The complex is equimolar. In it the salicylate portion is not in the form of its anion since its i.r. spectrum would show strong bands due to vibrations of the COO- group a t -1590 and 1390 cm-'. The acid must therefore be unionized, but it is not in the same form in the complex as is found in the pure crystalline acid hydrogen-bonded centrosymmetrical dimer (12) because the prominent bands due to COOH stretching in the acid are absent from the spectrum of the complex. This information coupled with the elevation of v,, and v,, in the caffeine part of the complex is conclusive evidence of the interaction at N-9 of a carboxylic acid COOH group, which gives similar effects but of smaller magnitude than a protonating acid. Structure 1 would be consistent with these results. Salicylic acid deuterated at the COOH and 0

R

R

H

\o% \

/

1, R1 = R3 = R7 = CH3 2, R1 = H, R3 = R7 = CH3

O H was prepared and the caffeine complex of this also was made. The s ~ e c t r u mof the salicylic acid-d, : caffeine complex did not contain the distinctive bands centered near 2418122331 2072 cm-' of salicylic acid-d,. Bands associated with the COOH and OH groups at 2440, 1310, 1238, 851, 809, and 786 cm-' in the caffeinesalicylic acid complex disappear on deuteration. None of the bands associated with the caffeine molecules alter significantly on deuteration. The spectrum of salicylic acid in dilute solution in CCl, shows two carbonyl bands at 1662 and 1694 cm-I, the latter being due to the monomer. Differences of 30-60 cm-' have been noted in vc0 between monomeric and dimeric forms in benzoic acid (13), acrylic acid (14), and acetic acid (8). In the case of a solution of fatty acid in pyridineICC1, the difference in carbonyl frequency between solution dimer and complexed monomer (COOH . . . N) was about 20 cm- (1 5). The band centered a t 1666 cm-' in the caffeine : salicylic acid complex is therefore probably made u p of contributions from carbonyl group vibrations of both molecules. Another possible structure was found to be untenable after a consideration of Dreiding models. This was a planar model with the phenolic OH group rotated 180" away from its intramolecular hydrogen bond with the COOH group, and forming a 0-H .. .N-9 bond (about 2.7 A long) with the caffeine molecule. Short H . . . H contacts of 1.6 A between C,-H (caffeine) and C,-H (salicylic acid), and 2.5 A between the phenolic OH and N3CH3 group

'

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/17/17 For personal use only.

COOK AND REGNIER: COMPLEXES O F N-METHYLXANTHINES

make this model entirely unrealistic. Rotation of the OH group out of the salicylic acid molecular plane would relieve some of these short contacts but seems unlikely in view of the great stabilization by internal hydrogen bonding. Moreover, the carboxylic acid group in this conformation would be free to form the dimer but this has been shown above to be an unlikely structural factor. Theobromine Complexes Theobromine - Acetic Acid Complex The contour of the multiple NH stretching band near 3050 cm-l in the complex is very similar to that in the free base, but the peaks are at slightly higher frequency. The stretching vibration of the CH group in the imidazole ring is also at higher frequency. The same hydrogen bonding that occurs in the free base therefore probably persists in the complex. It has been suggested that this takes the form of a centrosymmetrical dimer involving Nl-H . . . 0=C6 hydrogen bonds (6). The slightly higher frequencies in the complex imply a small lengthening of the hydrogen bond, also indicated by the decrease in the out-of-plane NH deformation ,,y from 860 to 837 cm-' in the complex. The carbonyl frequency of the acetic acid portion can be easily identified at 1720 cm-l, at the same position as in the caffeine complex. The dimeric form is indicated again with little interaction between the two components of the complex. Theobromine - Salicylic Acid Complex The spectrum of the theobromine - salicylic acid complex does not show the strong OH stretching bands of the free acid in the 3000 cmregion, neither does it contain bands due to the salicylate anion. There is, however, a group of bands very similar in position and in shape to those in free theobromine, and again this suggests that the hydrogen bonding that occurs in the free base also occurs in the complex. A broad band containing much structure centered near 2465 cm-l is assigned to the OH stretching vibration of the COOH group in the salicylic acid and may also contain a contribution from the internally hydrogen bonded OH group attached to the ring. The same type of structure (2) as for the caffeine complex (1) is proposed, and many of the bands due to the acid portion are at similar frequencies in both complexes. The absence of any decreased theobromine

3057

carbonyl frequency rules out interaction at a carbonyl group. Carbonyl frequencies of the base are raised slightly, as in the caffeine complex. Deuteration studies showed similar changes taking place as in the caffeine - salicylic acid complex, with bands a t 2500, 2465, 1305, 1230, 810, and 765 cm-l disappearing. In addition, the Nl-H hydrogen stretching and bending vibrations at 3050, 1482, 1140, and 860 cm-I due to the theobromine molecule are shifted on deuteration. Caffeine is known to alter the gastric absorption rate of salicylic acid in the rat when the two are administered together (16, 17). The precise significance of the results is not known, however, since the two research groups used a different protocol. The summaries of their work appear to present diametrically opposed views. It is possible to rationalize these apparently opposite views by constructing the graph in Fig. 1 showing % salicylic acid not in stomach as a function of time. With salicylic acid alone group A (17) find an increasing amount absorbed which can be extrapolated to reach the 1 h values of group B (16) thus giving a broad maximum to the curve, which is essentially an absorption or blood level curve. With the salicylic acidlcaffeine mixture, a higher level is reached earlier, followed by a quicker decline, which again extrapolates to the lower levels of group B after 1 h. These blood level characteristics, a lower flatter maximum for the slowly absorbed material, and a sharper higher peak for the faster absorption are quite consistent with our knowledge of the absorption process. Within the experimental errors and standard deviations the data of both groups fit this scheme. The lipid solubility/percent ionized concept (18) has been highly successful in explaining many facets of drug absorption. In the present context the complex of salicylic acid and caffeine by the formation of the 0-H . . . N hydrogen bond effectively removes these polar groups from the environment, and is therefore much more lipid soluble, which is one of the main criteria for more rapid and complete absorption, other things being equal. These experimental absorption figures (16, 17) are also confirmed by a very early pharmacokinetic study (19) available to us only in abstract (20) which concluded in effect that when salicylic acid was administered with caffeine to

3058

CANADIAN JOURNAL O F

CHEMISTRY. VOL. 46,

1968

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/17/17 For personal use only.

100

-

./.-.--.\ Knoll

ef ol

0

SA

+

Levy

et ol

\

CAFF.

\

\

\ \ \

-

--- \

-\-

A--

\}

SA

SA

+

EtOH

+ EtOH + CAFF.

0

I I

7

I

TIME (MINI FIG.1. Disappearance of drug from rat stomach asa function of time. 0,ref. 16; A, unstirred; 0 ,static stirrer; 0, rotating stirrer, ref. 17.

rabbits, it gave (i) a maximum blood level earlier, (ii) a lower total quantity in blood, and (iii) a greater urinary excretion rate than when salicylic acid was administered alone. Allowing for the less advanced technical methodologv of that day, one could only take exception to conelusion (ii). species difference may account for this. Experimental

Houses Limited; theophylline from K and K Laboratories Inc.; and salicylic acid from Anachemia Chemicals Limited.

Materials

Equivalent quantities of each component were dissolved in absolute ethanol. On evaporation of the solvent, the complex remained as a white solid, m.p. 138-138.5'.

--

Caffeine (anhydrous) was obtained from the Aldrich Chemical Company; theobromine from the British Drug

caffeine - ~~~~i~

complex

Caffeine was recrvstallized from a small auantitv of glacial acetic acid in-the form of long needle~,*m.~. same as caffeine. The crystals lost acetic acid completely after exposure to air for 16 h. Caffeine - Salicylic Acid Cornplex

~

COOK AND REGNIER: COMPLEXES

The complex could also be prepared by thoroughly grinding together very small equimolar quantities of each component in an agate mortar.

Can. J. Chem. Downloaded from www.nrcresearchpress.com by MICHIGAN STATE UNIV on 01/17/17 For personal use only.

Theobromine - Acetic Acid Complex Theobromine was recrystallized from acetic acid giving a white solid, m.p. similar to that of theobromine. It lost acetic acid on exposure to air. Theobro~nine- Salicylic Acid Complex A mixture of aqueous solutions of equimolar amounts of the two components on evaporation of the solvent gave the complex as white needles, m.p. 208" (decomp.). The stoichiometry of the complexes was checked by titrimetric methods, giving results in all cases close to equimolar values. The spectra were recorded on a Perkin-Elmer 621 grating spectrometer, and were all of Nujol-fluorolube mulls of the solid materials. 1. T. HIGUCHIand D. ZUCK. J. Am. Pharm. Assoc. Sci. Ed. 41,lO (1951), and later papers. 2. R. LABES. Arch. Exp. Pathol. Pharmakol. 158, 42 (1 971-1) ,----,. 3. G. PIROLI. Boll. Chim. Farm. 103.345 (1964). and A. N. MARTIN. J. Pharm. Sci. 4. R. S. SCHNAARE

OF

N-METHYLXANTHINES

3059

5. D. COOKand Z. R. REGNIER.Can. J. Chem. 45, 2895 (1967). 6. D. COOKgnd Z. R. REGNIER. Can. J. Chem. 45, 2899 (1967). 7. D. COOKand Z. R. REGNIER.Unpublished results. 8. M. HAURIEand A. NOVAK. J. Chim. Phys. 62. 137 (1965). and A. NOVAK. J. Chim. Phys. 62, 146 9. M. HAURIE (1965). 10. R. E.'JONES and D. H. TEMPLETON.Acta Cryst. 11, 484 .- . (1,- 958). --11. M. HA"& and A. NOVAK. Spectrochim. Acta, 21, 1217 (1965). 12. W. COCHRAN.Acta Cryst. 6,260 (1953). and K. H. WEBB. 13. G. ALLEN.J. G . WATKINSON. ~~ectrochim Acta. . 22. 807 (1966). 14. W. R. FEAIRHELLER, JR. and J. E. KATON. Spectrochim. Acta, 23A, 2225 (1967). 15. R. S. ROY. Spectrochim. Acta, 22,1877 (1966). 16. G. LEVYand R. H. REUNING.J. Pharrn. Sci. 53, and G . LEVY. J. 1471 (1964). R. H. REUNING harm. ~ c i 56, : 843 (1967). 17. K. R. KNOLL,R. P. PARKE,and H. A. SWARTZ. J. Pharm. Pharmacol. 18,540 (1966). 18. P. A. SHORE,B. B. BRODIE,and C. A. M. HOGBEN. J. Pharm. Exp. Therap. 119,361 (1957). 19. C. W. MYONG. Nippon Yakurigaku Zasshi, 25,102 (1938), Breviaria 8-9. 20. Chem. Abstr. 32, 38263 (1938). ~

\

-

~

~