A novel and practical synthesis of iclaprim

A novel and facile synthesis of iclaprim was reported. Started from Trimethoprim (TMP), the amino-protection and Friedel-Crafts acetylation with acetic anhydride were simultaneously completed in CH2Cl2 with SnCl4 as catalyst. The Knoevenagel condensation of 2,4-diamino-5-(2-acetyl-3-hydroxy-4,5-dimethoxybenzyl)pyrimidine with cyclopropyl carboxaldehyde followed by the intramolecular Michael addition in the buffer system (pyrrolidine and acetic acid) installed the key framework (chromanone 13). The dehydration was catalyzed by H2SO4 so that the formation of 5-cyclopropyl-2,3-dimethoxy-4,5,6,6a,7,12-hexahydronaphtho[1,8-bc]pyrimido[5,4-f]aze pin-9-amine, an impurity of iclaprim reported at the first time, could be minimized. In the end, iclaprim was obtained in a total yield of 21%.


Introduction
Dihydrofolate reductase (DHFR) is an enzyme essential for bacterial survival that is also an excellent target for antibacterial drug development [1]. TMP is an inhibitor of DHFR used as an antibacterial agent in clinic for many years [2]. Iclaprim (as a racemate), a derivative of TMP was developed by Motif Bio plc. The clinical trial demonstrated that iclaprim was effective for treatment of acute bacterial skin and skin structure infections [3]. Since iclaprim is active against multi-drug resistant Gram-positive pathogens such as methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Staphylococcus aureus [4,5], the synthesis has attracted much attention.
The structure of iclaprim is composed of two heterocycle rings: 2H-chromene and pyrimidine. Two kinds of synthetic strategy existed in the related patents [6,7,8], depending on which heterocycle was built at first.

Results and discussion
In route 3, the amino groups were protected with the steric hindrance groups, such as pivaloyl or isobutyryl. However, our preliminary experiment found that the following Friedel-Crafts acetylation product was contaminated by the N-acetyl byproducts, i.e., one or two hindrance groups were partially replaced by acetyl group in the Friedel-Crafts reaction. To resolve the problem, we attempted the same reagent both for the amino-protection and for the Friedel-Crafts reaction. So, as the first step in our new route to iclaprim, TMP was reacted with 5 equiv of acetic anhydride in toluene by refluxing to provide 2,4-diacetamido-5-(3,4,5-trimethoxybenzyl)pyrimidine (9) in a 86% yield. 6 To screen the reaction conditions of the Friedel-Crafts acetylation of compound 9 to form 2,4-diacetamido-5-(2-acetyl-3,4,5-trimethoxybenzyl)pyrimidine (10), two acetic agents (acetic anhydride and acetic chloride) with different lewis acid (AlCl3, SnCl4, and TiCl4) in CH2Cl2 or CH2ClCH2Cl were evaluated. To our surprise, in these cases, AlCl3 didn't work at all. Finally, the best result (yield 55%) was obtained by reaction of compound 9 with SnCl4 (2 eq) and acetic anhydride (5 eq) in dichloroethane at refluxing.
After TMP was successfully transformed into compound 10 by two steps, the one-pot synthesis was attempted. Initially, the conversion of TMP to compound 10 was performed by 5 equiv of acetic anhydride and 2 equiv of SnCl4 in dichloroethane under refluxing in a reasonable yield (46%, Table 1, entry 1), that was equivalent to the yields from two steps. However, there was a limitation from the standpoint of largescale preparation. When TMP was charged on a 100-gram scale, a severe emulsification occurred during phase separation in the work-up, probably due to the poor solubility of compound 10 in dichloroethane. Therefore, some solvents that are suitable for the Friedel-Crafts acetylation were tested by solubility experiment. From Tab 2，it was seen that Compound 10 is much more soluble in chloroform than dichloroethane. So, the reaction was carried out in chloroform and give an exciting yield (93%, Table 1, entry 2).
The modification on the amount of SnCl4 and acetic anhydride ( Table 1, entry 2-5) found that the best result (95%, entry 4) was achieved when acetic anhydride decreased to 4.5eq.
Because of the concern of toxicity of chloroform in pharmaceutical industry, the alternative (dichloromethane) was used, and the reaction (entry 6) gave the almost same 7 result although the amount of solvent was increased. a: volume (ml)/weight (g) of compound 8 For the demethylation, the preliminary experiments started with the amino-protected compound 10 and lewis acid (AlCl3 and BBr3) and various solvents (acetonitrile, toluene, dichloroethane, and dichloromethane) were tested, the best result was obtained with BBr3 in dichloromethane to provide 2,4-diacetamido -5-(2-acetyl-3-hydroxy-4,5-dimethoxybenzyl)pyrimidine in a 21% yield. The reason of a such low yield was due to the formation of some by-products (partial de-protection in 8 2,4-diacetamido), that made the purification complicated. Therefore, our attention turned to demethylation of 2,4-diamino-5-(2-acetyl-3,4,5-trimethoxybenzyl)pyrimidine (11).
The key step in the synthesis of iclaprim is the cyclization of compound 12 with cyclopropyl carboxaldehyde to form The reaction should involve the Knoevenagel condensation of compound 12 with cyclopropyl carboxaldehyde followed by the intramolecular Michael addition by the ortho-phenolic group. In general, the Knoevenagel condensation was catalyzed by organic base such as the secondary amines. So, the reaction was carried out in the presence of pyrrolidine, and it gave a too complicated mixture to purify (Table 3, entry 1), in which beside compound 13, some impurities formed by the intermolecular addition of the Knoevenagel product with pyrrolidine were found by the 1 HNMR data. Therefore, diisopropylamine, the sterically hindered base was employed, but no improvement was observed (entry 2). And then, the buffer system (pyrrolidine1.5eq/acetic acid 1eq) was attempted, a 44% yield of the compound 13 was achieved (entry 3). Addition of 1.5 equiv 9 acetic acid to the reaction led to a 15% raise in the reaction yield (entry 4). The other modification such as increasing the amount of acetic acid or the aldehyde didn't produce the better result (entries [5][6][7]. Surprisingly, in the tests with piperidine, a very similar base with pyrrolidine (entry 8), the reaction did not take place (most of the starting material 12 was recovered). was prepared by reduction with 0.5 equiv of NaBH4 in 80% yield.
The final step was the dehydration. In the beginning, compound 13 was heated under refluxing in methanol under the persence of TsOH . H2O (Table 4,

Conclusions
In summary, the novel and practical synthesis of iclaprim was accomplished.
Beginning with TMP, the very cheap material, our new synthetic routes only included six 12 reaction steps with an overall yield of 21%, which was almost four times as much as those existing processes. In addition, all purification only involved recrystallization, without column chromatography. This process offers the distinctive advantages over the reported routes to synthesize iclaprim.

Experimental Section
Solvents and reagents from vendors were used as received unless otherwise indicated. Melting points were determined on a capillary melting point apparatus and were uncorrected. NMR spectra were taken on a Bruker Avance III instrument operated at 400 or 600 MHz for 1 H-NMR and 100 MHz for 13  After cooling to rt, the reaction solution was poured into 1500 ml ice water. The organic phase was separated and the aqueous layer was extracted with dichloromethane (3 × 300ml). Combine the organic phase and wash with saturated aqueous Na2CO3. The organic layers were dried over anhydrous NaSO4. Evaporation of the dichloromethane to 13 afford product 10 (277.0 g, 96.1%). It was directly used for the next step without purification. Mp: 204-206℃. 1   1.8 L of CH2Cl2, cooled to ca -10 °C, boron tribromide (410ml, 0.41mol, 17% in dichloromethane, ca. 1mol/L) was added dropwise to the reaction mixture. And then, the 14 mixture was stirred at room temperature for 5 h. The reaction mixture was cooled to 0-5 °C and quenched with 400ml methanol, stirred at rt for 1 h, concentrated in vacuo, water (0.9 L) was added, the mixture was stirred for 8 h at 0-5°C. The crude product 12 was isolated by filtration, and crystallized from isopropanol to give 56.2 g of product 12 as a white solid (Yield: 65.0%). Mp. 217℃. 1