Chemistry of 4-hydroxycoumarin

4-Hydroxycoumarin, a highly reactive compound, is reviewed with re•pect to its synthesis, tautomerism, reactions and struc ture-activity relationship of its derivatives for anticoagulant property.


Introduction
4-Hydroxycoumarins not only occur in nature but also possess very good physiological activities. Link and coworkers found that cattle feeding on spoiled sweet clover hay suffered from a condition characterized by sharp increase of blood clotting time. The pathogenic haemorrhagic principle of sweet clover hay was found by them to be 3,3'-methylenebis-( 4-hydroxycoumarin) (2) popularly called Dicoumarol. It was synthesized by reacting 4-hydroxycoumarin (l) with formaldehyde and found to be a good anticoagulant of blood. Since this discovery and also because of its high reactivity, the chemistry of 4-hydroxycoumarin has assumed great importance.
Pauley and Lockeman synthesized 1 by intramolecular Claisen condensation of methyl acetyl salicylate (13) by adding metallic sodium to the molten ester. They reported yield of 55% of 1, which was not repeatable. Link and coworkers modified the conditions for the reaction by keeping the temperature at 240-250° and obtained the yield of22% of pure l. Sonn and Bauer, and Schoder synthesized 4hydroxycoumarin derivatives having hydroxyl substituents in the benzene ring by the application of Hoesch synthesis. **Deceased
Mentzer and coworkers carried out the thermal condensation of phenols with ethyl monosubstituted malonates by heating the reaction mixture at 200-280° for 2-4 days and obtained 3-substituted-4-hydroxycoumarins in poor yields. It was found by Trivedi that if the above reaction is carried out in high boiling inert solvent, viz. diphenyl ether at its boiling point, the reaction time is reduced to 4 h in the case of reactive phenols and 8 h in the case of less reactive phenols. The yield is greatly improved and the products are cleaner and do not require vacuum sublimation as was found necessary earlier. When resorcinol ( 14a) was condensed with ethyl methylmalonate (18) in refluxingdiphenyl ether it gave 3-methyl-4,7-dihydroxycoumarin (19) as obtained by Mentzer and Vercier. When resorcinol (l4a) was condensed with ethyl benzyl malonate (20), it furnished 3-benzyl-4,5dihydroxycoumarin (21) and not 3-benzyl-4, 7-dihydroxycoumarin (22) obtained by Mentzer and Vallet. Thus the course of the reaction was changed by carrying out the reaction in refluxing diphenyl ether.
Boyd and Robertson reported that o-hydroxyacetophenones and their OJ-substituted derivatives 27 readily condensed with diethyl carbonate in the presence of sodium to give 4-hydroxycoumarin derivatives 28 in good yields. This is a very convenient method for synthesizing 4-hydroxycoumarin derivatives and has been explored by a large number of workers 1 • 4 · 7 -20 . Ziegler and Junek prepared 4hydroxycoumarins in excellent yields by cyclization of diary\ ma\onates (29) in the presenc~ of anhydrous aluminium chloride at \80°. Garden et al. synthesized 6-methoxy-4hydroxycoumarin (30) by condensing hydroquinone dimethyl ether (31) with malonyl chloride (32) in the presence of anhydrous aluminium chloride using carbon disulfide as solvent. Rao and Sundaramurthy 21 condensed phenyl acetate (33) with malonyl chloride (32) in the presence of anhydrous aluminium chloride using nitrobenzene as the solvent and obtained 3-acetyl-4-hydroxycoumarin (34). Ziegler and Gelfert carried out the condensation of phenol (35) with diethyl malonate (36) in the presence of phosphorus oxychloride to obtain 1. Shah, Bose and Shah prepared 4hydroxycoumarin derivatives in good yields by the condensation of phenols with malonic acid in the presence ofa mixture of zinc chloride and phosphorus oxychloride at 65°. Shah el a/. 22 31 also reported a similar sulfur assisted carbonylation of o-hydroxyacetophenone derivatives in the presence of a mixture ofDBU or triethylamine and the temperature being kept at 80° and the time reduced to 4 h and obtained 4hydroxycoumarin derivatives in good to excellent yields.
. Spectroscopic evidence has also been utilized to show the existence in one or other forms of 1. UV spectrum is not !le!r!ul in this case but theIR spectru1:n could easily distin-guish between structures I and 53. Thus 4-methoxycoumarin (55) showed carbonyl band at 1710 cm-1 (CHCI 3 ) whereas 2-methoxychromone (54) showed the same at 1635 cm-1 (CHCI 3 ) 37 . IR spectra of 4-hydroxycoumarin (I) in solid state showed weak bands at 3000, 2720 and 2500 cm-1 (OH) and a strong carbonyl band at 1704 cm-1 , however, the carbonyl band shifted to 1730 cm-1 in dioxane solution indicating the presence of intramolecular hydrogen bond- ing. Porter el a/. 38 · 39 studied the IR spectra of 4-hydroxycoumarins by isotopic replacement of carbonyl carbon with 13 C and also by replacement of hydrogen atoms at 3 and 4 positions by deuterium and identified the C=O stretching frequency. Thus carbonyl frequency of I at 1700 cm-1 (s, br) shifted to 1658 cm-1 (m, sh) in the case of I 2-13 c:

N·NH-Ph
while it shifted to 1690 cm-1 (s, br) in the case of I 3,4-d 2 and identified the primary carbonyl band. In the case of 3substituted-4-hydroxycoumarin and 3-substituted-4-alkoxycoumarin, the replacement of carbonyl carbon by 13 C, the carbonyl stretching frequency was identified as the highest frequency strongly absorbing in 1550-1750 cm-1 region. The carbonyl band varied from 1664 cm-1 in inter-or intramolecularly hydrogen bonded derivative to 17 I 8 cm-1 • No evidence for existance of2-hydroxychromone tautomer was found except in the case of anhydrous 4-hydroxycoumarin in solid state as it showed two bands at 1700 (apyronyl group) and 1670 cm-1 (y-pyronyl group). Cussans and Huckcrby 40 studied De NMR spectra of I and established that it has benzo~a-pyronc structure amlnot 53 as its chemical shift of C-2 is almost near C-2 of coumarin. C-2 of I is at 8 162  X-Ray studies of 4-hydroxycoumarin monohydrate reveal that it crystal! izes with four molecules of I and four molecules of water in an orthorhombic unit cell. Each water molecule is hydrogen bonded to carbonyl oxygen of two adjacent 4-hydroxycoumarin molecules with inter-oxygen distance 2.59 and 2.73 A. The oxygen of each water molecules is in turn hydrogen bonded to enolic hydroxyl of a third 4-hydroxycoumarin molecule with an inter-oxygen distance of 2.8 A 41 . X-Ray studies of all other 4-hydroxycoumarin derivatives confirm that they always crystallize as 4hydroxycoumarin and not as 2-hydroxychromone 42 . In the mass spectral studies of 4-hydroxycoumarin, the fragmentation occurs completely in a different manner. CO expulsion from the molecular ion is virtually suppressed, while the base peak at mlz 120 corresponds to the loss of ketene (C 2 H 2 0) fragment from the molecular ion. It was proposed that the fission of the heterocyclic ring occurs by retero-Diels-Alder reaction, a mode which has been rationalized by assuming the participation of tautomeric chromandione molecular ion. The ion at mlz 121 arises from H transfer reaction in the course of retero-Diels-Aidcr cleavage. An intense fragment at m/: 92 arises by the loss of CO from [RDA]+ ion arised from4-hydroxycoumarin. This observation suppo11s the enol structure which is obtained from deuterium labeling experiment. The mass spectral studies of I, its acetyl derivative and deuterium labeled derivatives on the above lines clearly establishes the enolic structure of 143 The controversy of the tautomerism of 2,3,4-trioxochroman-3-arylhydrazone (56) 44 and 3-arylazo-4-hydroxycoumarin (59) 45 -" 6 is finally settled in the favor of structure 56 by IR and 1 H NMR studies 47 and also confirmed by 1 H NMR of 15 N-phenylhydrazone derivative of 2,3.4-trioxochroman48.

Miscellaneous reactions of 4-hydroxycoumarin :
Compound 1 on bromination in chloroform gave 3bromo-4-hydroxycoumarin ( 196a ). Nitration of I with nitric acid in acetic acid gave 3-nitro-4-hydroxycoumarin ( l96b) which on hydrogenation gave 3-amino-4-hydroxycoumarin (196c ). Sulfonation of l with fuming sulfuric acid gave 4-hydroxycoumarin-3-sulfonic acid (196d). 1 when reacted with acetylchloride in the presence of pyridine and piperidine gave 3-acety 1-4-hydroxycoumarin ( 196e ). I when condensed with aliphatic acid in the presence of phosphorus oxychloride gave 3-acyl-4-hydroxycoumarin (196t), while whh aromatic acid chlorides, it gave the corresponding esters 197. Treatment of l with acetic anhydride gave 4acetoxycoumarin ( 197a) below 155°and 3-acetyl-4-hydroxycoumarin (196e) above 180°. 197a on heating above 180° rearranges to 196e 119 . 1 when reacted with acetic anhydride in the presence of pyridine gave 4-acetoxy-3-(N-acetyl-1' ,2'-dihydro-2' -pyridyl)coumarin (198) 120 . 4-Benzoyloxycoumarin (197b) on Fries rearrangement in the presence of anhydrous aluminium chloride gave 3-benzoyl-4-hydroxycoumarin (196g). When the rearrangement to 197b was carried out in refluxing trifluoroacetic acid for 15 h, an abnormal product 5-benzoyl-4-hydroxycoumarin ( 199) was obtained 121   When 1 was refluxed with phenyl isocyanate in dimethyl sulfoxide in the presence of triethylamine, 3phenylcarbamoyl-4-hydroxycoumarin (253) was obtained. The use of pyridine as solvent yielded urethanes (254) 155 Junek carried out the addition reaction of I with tetracyanoethylene (255) and obtained the adduct 256. Reid et a/. 156 -157 carried out the photochemical addition of I with cyclohexene in methanol by irradiation ofthe reaction mixture through a silica filter with a medium pressure mercury lamp and obtained 257 which readily formed acetate 258 by treatment with acetic anhydride and boron trifluoride etherate. Pyrolysis of258 gave 259 (Scheme 39). Haywood and Reid 5 8 carried out the intramolecular photoaddition of 4-(but-3-enyloxy)coumarin (260) as above and obtained 26I. The product 262 was obtained when 4-(pent-4-enyloxy)coumarin was used (Scheme 40). Since the discovery of dicoumarol (2), which is the only bis type of 4-hydroxycoumarin used in the medicine today, many attempts have been to establish the structure-activity relationship of 4-hydroxycoumarin derivatives. The correlation of chemical structure with anticoagulant activity of 4-hydroxycoumarin derivatives has been studied by Link et al., Mentzer et al. and Chmielewska and Cieslak. These studies have clearly pointed out that minimum structural requirement is 4-hydroxycoumarin unit with a substituent in position-3 bearing a suitably placed carbonyl group and a phenyl ring separated from the 4-hydroxycoumarin by one carbon atom. For maximum activity, a bis arrangement as in dicoumarol (2) was considered to be necessary. Seshadri et a/. 165 have synthesized many bridge substituted dicoumarols and found them to be much less active than dicoumarol. Thus any alteration in the structure of dicoumarol results in decrease in activity. Compounds with an alkyl or aryl group in 3-position showed diminished activity. 3-Benzyl-4-hydroxycoumarin (270a) has slight anticoagulant activity but the acetonyl derivative warfarin (132) is a powerful anticoagulant and rodenticide which is marketed. Warfarin and phenprocoumon (270b) have an asymmetric carbon atom in them. Both (+)Rand (-)S enantiomers of warfarin have been tested for anticoagulant activity in rats and the (-)S isomer is 5-8 times potent than (+)R. Commercially available warfarin for use in man, however, is a racemic mixture. Other compounds which are related to warfarin and possess anticoagulant activity are coumarchlor (270c), sintron (270d). Chmielewska and Cieslak analysed the structural requirement for anticoagulant activity from the viewpoint of their vitamin K antagonism. They postulated that the active form of vitamin K can be represented by formulae 271 and 272. On the other hand, an anticoagulant which is antivitamin K should have the structure 273 and 274 which is cyclic hemiacetal obtained from appropriate 3-substituted-4-hydroxycoumarin. For such acetal formation, the carbon chain in position-3 should carry a carbonyl group in position 2' or 3'. Link et al. synthesized cyclocoumarol, a cyclic ketal and found that it possesses greater activity than dicoumarol (Scheme 42).