Insoluble versus soluble polymer-assisted synthesis. A first approach for the preparation of a Biginelli compound

A comparison is established between the experimental conditions associated with the multistep synthesis of ethyl 1,2,3,4-tetrahydro-6-methyl-2-oxo-4-phenylpyrimidine-5-carboxylate from Merrifield’s resin and soluble polymers, derived from poly(ethylene glycol), poly(styrene-co - allyl alcohol), or dendritic aliphatic polyols.


Introduction
Since the birth of combinatorial chemistry there has been an exploding interest in polymerassisted synthesis as the method generally yields final products in a high state of purity.Insoluble resins are often used because they are easy to handle.However long reaction times and excess of reagents are required to complete the reactions.Those drawbacks can readily be avoided by employing soluble polymeric auxiliaries. 2Nevertheless, that late approach remains underestimated probably because it is too often associated with the erroneous idea of the unpleasant manipulation of gummy products.
In a recent paper, 3 we demonstrated that poly(styrene-co-allyl alcohol) is a suitable soluble support for the preparation of various nitrogen and oxygen heterocycles.We now wish to extend that work by disclosing our preliminary results on the solid-phase and liquid-phase synthesis of a member of the Biginelli compounds which are known to exhibit a broad range of biological effects. 4More particularly we focused our attention on a comparison between the experimental conditions associated with the use of each kind of supports.

Results and Discussion
Target compound Alkyl 1,2,3,4-tetrahydro-2-oxopyrimidine-5-carboxylates have been described by P. Biginelli 5 more than a century ago.The original synthesis (Scheme 1) is a 3-component reaction involving urea, a ketoester, and an aldehyde under acidic conditions. 6Recently it has been observed that the resulting heterocycles can also be obtained through a solid-phase strategy (Scheme 1) involving condensation of a polymer bound thiouronium salt with an aralkylidene ketoester, followed by hydrolytic cleavage. 7 Scheme 1. Preparation of Biginelli Compounds.

Choice of the supports
The most common insoluble 4-(chloromethyl)phenyl-containing polymer is the commercially available Merrifield's resin.In order to establish the proposed comparison we reasoned that 4-(chloromethyl)phenyl moieties could be introduced on soluble matrices by reacting commercially available hydroxylated soluble polymers with 4-(chloromethyl)benzoic acid.For that purpose we directed our attention to the following three classes of supports: poly(ethylene glycol) (PEG), poly(styrene-co-allyl alcohol) (PSCAA), and dendritic aliphatic polyols of the Boltorn™ family. 8he chosen Merrifield's resin had a Cl content of 2.4 meq/g.In the PEG series we selected three polymers having average molecular weights (average M n ) of 1500, 4600, and 10000 respectively and three poly(ethylene glycol) methyl ethers having average molecular weights of 750, 2000, and 5000 respectively.PSCAA had an average molecular weight of 2200.The average molecular weights of the dendritic polyols were 2100, 3500, and 5100 respectively.The corresponding hydroxyl indexes (meq/g) are collected in the Table .Among the reported values let us point out the impressive functionality of the Boltorn™ polymers, which bear 8.7-9.1 meq of hydroxyl groups per gram of material.
Esterification (Scheme 2) was performed in the presence of dicyclohexylcarbodiimide (DCC) and a catalytic amount of 4-(dimethylamino)pyridine (DMAP).Dichloromethane was chosen as the solvent except for the dendrimers, which are more soluble in dioxane.In a first set of experiments we decided to perform the reactions at room temperature for 24 hours.After filtration of dicyclohexylurea (DCU), filtrates were concentrated under reduced pressure to afford the expected polymers as indicated by IR (carbonyl band at 1718 cm -1 ) and NMR ( 1 H NMR: δ CH 2 Cl at 4.8 ppm; 13 C NMR: δ C=O at 165 ppm).Due to the high value of the initial hydroxyl content of the dendrimers, the increase of weight of the polymers was also indicative of the success of the esterification.However let us emphasize that that increase of weight is associated with a dramatic decrease of their loading as it falls from 9 meq/g to 4 meq/g.In order to optimize experimental conditions, we determined that the theoretical amount of DCU precipitated out of the media after 4 hours of contact between the reagents.No remarkable structure dependence was detected.Coupling with thiourea From Merrifield's resin we observed that the quantitative addition (no more spectral change, no more increase of weight) of thiourea could be achieved by refluxing a mixture of the resin and thiourea (5 equivalents) in dimethyl formamide (DMF) for 24 hours (Scheme 3). 7,9 early, use of soluble polymers enables to decrease reaction times and does not require an excess of thiourea.All of them readily reacted in boiling ethanol and the thiouronium bound polymers were isolated after cooling and evaporation of the solvent.For the PEG and PEGME derivatives, an optimized contact period of 1 hour was determined by testing the media for the presence of thiourea with bismuth chloride. 10PSCAA-based polymer and dendrimers were entirely derivatized within 4 hours.

Formation of the dihydropyrimidine skeleton
In order to build the heterocyclic moiety, the thiouronium bound polymers were treated with ethyl 2-benzylidene-3-oxobutanoate in the presence of a base (NaHCO 3 ) (Scheme 4). 7,11 om the insoluble polystyrene derivative the experiment was conducted with an excess of reagents (4 equivalents based on the announced initial content of Cl of the resin) in DMF at 100 0 C. The progress of the reaction was followed by IR (appearance of a carbonyl band at 1685 cm - 1 ).No more change of the spectrum of the resin was observed after 6 hours.From the soluble polymers the reactions were performed in boiling ethanol in the presence of one equivalent (calculated on the initial content of hydroxy groups) of the arylidene component and two equivalents of NaHCO 3 .By monitoring the reactions by GC we determined that a maximum amount (75 -85 %) of ester was consumed within 4 hours independently of the nature of the polymer.Here again we did not determine any straightforward relationship between yields and the structure of the supports.

Cleavage from the support
The final pyrimidine derivative was obtained by cleavage of the heterocycle from the support under the conditions described by Kappe: use of a refluxing mixture (volumes are given for 1 g of the starting resin) of dioxane (10 mL), ethanol (10 mL), acetic acid (2 mL), and water (2 mL) (Scheme 5). 7Under those conditions, the expected heterocycle is soluble in the mixture of solvents and can be isolated after concentration of the reaction medium.
From the insoluble resin we obtained the pyrimidine in an overall yield of 65 % (based on the Cl content).An optimized refluxing period of 10 hours was determined, as experiments involving a prolonged heating did not improved the yield.
From the PEG-and PEGME-based polymers, the reaction media were heated under reflux for 4 hours (optimized).After concentration of the solution and treatment with water the heterocycle was isolated in overall yields (based on the OH content) ranging from 55 to 70 %.Reasonably we could not attribute the observed differences to structural changes of the supports.Under similar experimental conditions, the cleaved PSCAA-based polymer precipitated and was filtered.The filtrate was concentrated and the residue was treated with water to afford the pyrimidine in 60 % overall yield (based on the OH content).Boltorn™-based polymers exhibited a similar behavior.

Conclusions
This work compares the efficiency of an insoluble resin and a series of soluble polymers for the preparation of ethyl 1,2,3,4-tetrahydro-6-methyl-2-oxo-4-phenylpyrimidine-5-carboxylate.As expected use of soluble supports enables to reduce reaction times and does not require an excess of reagents.More interestingly from both types of polymers yields are comparable and in each family they do not seem to be dependent on the molecular weight of the starting polymers, thus enabling to take benefit of the macromolecular properties of each one.This work also demonstrates that protocols involving soluble polymers can readily be established and that such supports can reasonably compete with their insoluble counterparts which are often more expensive.

Experimental Section
General Procedures . 1 H and 13 C NMR spectra were obtained using a Bruker AMX-300 spectrometer (300 MHz for 1 H and 75 MHz for 13 C at 7 T); chemical shifts (δ) are given in ppm using TMS as internal reference.IR spectra were recorded on a Perkin-Elmer FTIR 1760K spectrometer.Solvents are commercially available (Aldrich Co, Acros Organics) and were used without further purification.PEGs and PEGMEs were dried by azeotropic distillation before use.
For the soluble polymers, spectroscopic data collect the most striking information recorded in the case of the functionalization of a PEGME only, similar characteristics were observed from the other soluble auxiliaries.
The authors are grateful to Bo Pettersson (Perstorp Speciality Chemicals AB) for a gift of various samples of Boltorn™.

Figure 1 .
Figure 1.Structure of the target compound.

Scheme 5 .
Scheme 5. Formation of the target compound by from the support.

Table 1 .
Characteristics of the polymers involved in the study