Optimization of a Key Step of Synthesis of L-Acosamine and L-Daunosamine Derivatives

An optimization of synthesis of 3-N-(allyloxycarbonyl)-4-O-acetyl-3-amino-1,5-anhydro-2,3,6-trideoxy-L-arabino-hex-1-enitol (D-2), a key intermediate in the six-step synthesis of daunosamine glycoside, was performed at laboratory scale. The D-2 compound was obtained according to the following scheme: where R = allyl, benzyl. While taking into account a remarkable sensitivity of the D-2 yield (varied from a few percent to ~50%) on the reaction conditions, a series of experiments was performed in order to determine the most suitable reaction conditions both to increase the D-2 yield and, simultaneously, to reduce the amount of the unconsumed D-1 substrate. Six reaction parameters (temperature of the reaction, time, D-1 concentration, type of sodium carbonate, temperature of hydrolysis, and amount of isocyanate denoted as x1-x6 variables) were selected for optimization and two reaction responses (D-2 yield and D-1 content, denoted as y1 and y2, respectively) were monitored. The optimization was performed in two steps. In the course of the first step (8 experiments), it was realized that two of five variables studied, i.e., x1 and x2 selected from x1-x5, can be variables of major importance controlling the D-2 synthesis. In the course of a subsequent second step of optimization (next 15 experiments), a set of six variables (including the amount of chlorosulfonyl isocyanate, x6) was studied. The exploration of the reaction response surface (y1, y2) was supported by mathematical modeling with the use of the multiple linear regression and the partial least squares methods. As a result, we localized those regions of reaction parameters that lead to high, max ~60% (determined by high performance liquid chromatography (HPLC)), yield of D-2 and contain less than 5% of D-1. We also determined regions of low yield that can facilitate control of reaction parameters in case of their perturbations. The present optimized D-2 yield at laboratory scale corresponds well with the 53-56% yield of D-2 obtained at the pilot-plant scale.