Generation of 500-Member Library of 10-Alkyl-2-R1,3-R2-4,10-DihydrobenzoDihydrobenzo[4,5]imidazo[1,2-a]pyrimidin-4-ones

Representative benzimidazopyrimidinones were previously reported to be intercalating antitumor agents. In this work, we used 2-substituted 4,10-dihydrobenzo Dihydrobenzo[4,5]imidazo[1,2-a]pyriminin-4-ones for their diversification by regioselective alkylation. Under the conditions established, the alkylation gave 10-alkyl derivatives which permitted the parallel generation of a 500-member library of the title compounds.

Condensed benzimidazoles represent a suitable platform for the construction of such planar molecules applying the strategy of diversity oriented synthesis. We were interested in the generation of a library of diverse compounds based on the dihydrobenzo [4,5]imidazo [1,2-a]pyrimidin-4-one core ( Figure 1). Figure 1. Dihydrobenzo [4,5]imidazo [1,2-a]pyrimidin-4-one core. We aimed to generate a 500-member library of dihydrobenzo [4,5]imidazo [1,2-a]pyrimidin-4-ones by introducing the initial diversity points in the positions C-2 and C-3 and then diversifying the obtained compounds by regioselective alkylation reactions using alkylating agents of different types in parallel format. In view of the scaffold structure ( Figure 1) there are at least three products of monoalkylation that can be expected: the N1-and N10-substituted derivatives and the product of O-alkylation. Thus, uniform conditions for regioselective alkylation of dihydrobenzo [4,5]imidazo [1,2a]pyrimidin-4-ones are required.

Results and Discussion
Several synthetic approaches to the dihydrobenzo [4,5]imidazo[1,2-a]pyrimidin-4-one scaffold 1 have been found in the literature [17][18][19][20][21][22]. All of them utilize the reactivity of 2-aminobenzimidazole 2 with 1,3-dielectrophiles of several types. However, because of the commercial ability of different β-ketoesters 3, we have applied them [23] for the synthesis of starting compounds 1a-p. In contrast to the original protocol [23] where authors heated a neat mixture of 2-aminobenzimidazole 2 and some liquid β-ketoesters 3, we applied DMF as a solvent to make the procedure general for different starting materials including solid esters 3e-h ( Table 1). Application of other solvents (EtOH, AcOH, dioxane) at reflux resulted in lower yields.
In our initial experiments on the alkylation we used the representative 1a in the reaction with 4-methylbenzyl chloride 4a under different reaction conditions. Strong alkali media (NaH or KOH in DMF, DMSO or dioxane) gave mixtures of alkylation products and therefore were not acceptable. Milder reaction conditions (NaHCO 3 or K 2 CO 3 in acetone, NEt 3 in DMF) led to low yields or to products contaminated with the starting material 4a. However, application of K 2 CO 3 (3 equiv) in DMF at 90 °C for 2h resulted in the formation of only one isomer 5aa with 81% yield following a simple aqueous workup (Scheme 1). Application of such conditions is very convenient for parallel synthesis because the reaction media does not reflux and the reaction can be carried out in a simple sealed vessel. These conditions were suitable for application of a representative of N-substituted 2-chloroacetamides 4o ( Table 2). The product 5ao was obtained in the same manner with 77% yield (Scheme 1).   The information on alkylation of the dihydrobenzo [4,5]imidazo[1,2-a]pyrimidin-4-ones 1 turned out to be absent in the literature, thus it was not possible to predict with certainty which isomer out of three mentioned above was formed. The NOE experiment, irradiation of the product 5aa with the resonance frequency of its CH 2 protons at 5.58 ppm has demonstrated a close location of the methylene group and the upfield CH proton of the benzimidazole ring (doublet at 7.63 ppm) which agrees with either for O-alkylated or N10-alkylated product. To undoubtedly determine the structure of the product 5aa we have carried out a counter-synthesis (Scheme 1) of this compound starting from 2-amino-1-(4-methylbenzyl)benzimidazole 6 described previously [24]. The products obtained in these two alternative ways were identical. This means that the upfield doublet (at 7.63 ppm) of the benzimidazole ring belongs to the 9-C-H proton but not to the 6-C-H. Consequently, the latter gives its downfield doublet at 8.48 ppm. This alternative synthetic pathway leading to the compound 5aa gives lower yield (54% isolated yield). Also in this protocol the main diversity point (the alkylation agent) was introduced not at the last synthetic step, which was less convenient for the library generation. Alkylation of the starting derivative 1a using a representative of the N-substituted 2-chloroacetamides 4o led to one regioisomer 5ao. The discussed signal assignment for compound 5aa gives opportunity to determine structure of the product 5ao, the positive NOE between the protons of the CH 2 group (5.29 ppm) and the upfield benzimidazole 9-C-H proton (doublet at 7.71 ppm) fully confirms the N10-alkylation.
The specified conditions allowed us to perform regioselective alkylation of the 2-(het)arylderivatives 1a-g with all the applied chemotypes of alkylation agents (representatives of benzyl clorides, 2-chloroacetic acid derivatives and representatives of alkylbromides or iodides, see Table 2 for selected alkylation agents 4a-u). However, a series of pilot experiments with the alkylation of 2-methyl-dihydrobenzo [4,5]imidazo[1,2-a]pyrimidin-4-one 1i by representatives of the N-substituted 2-chloroacetamides 4l-u led to isolation of a mixture of two isomeric products. At the same time, the application of other alkylation agents was successful in this case. The same result was observed for the 2-tert-butyl derivative 1m. Also, in the case of the starting compounds 1o,p, we could not achieve the regioselectivity with alkylators of all chemotypes. Thus, we had to exclude the 2-alkyl derivatives 1i-n from the starting set used in the alkylation with the N-substituted 2-chloroacetamides 4l-u, and we did not use the derivatives 1o,p for the library generation.

Scheme 2.
Generation of a 500-member library of 10-alkyl-2-R 1 ,3-R 2 -4,10dihydrobenzo [4,5]imidazo[1,2-a]pyrimidin-4-ones 5.  The library was generated in parallel format (Scheme 2, Figure 2) and the structure and purity of every congener were checked by 1 H-NMR. In about 85% cases the products were isolated with very good average yield (about 70%) and met the purity requirements (maximal total level of impurities less then 10% as determined by 1 H-NMR). In those cases (about 10%) where the impurity level was higher than 10% the product purity was improved by heating in EtOH to remove soluble remains of the starting materials. In about 5% of cases the products did not satisfy the purity requirements either after the simple workup or after the additional purification. As a result, the desired 500-member library contained different compound chemotypes whose distribution is illustrated in Figure 3. As one can see, the starting 2-(het)arylderivatives 1a-g contributed much more significantly to the total number of compounds because of their selectivity in the reaction with the diverse N-substituted 2-chloroacetamides 4l-u.