ИСТИНА

Aspartoacylase (ASPA) belongs to carboxypeptidase family and has a Zn-atom in its active site [1]. ASPA catalyzes the hydrolysis of N-acetyl-aspartate (NAA) namely breaks up the peptide bond producing aspartic and acetic acids which further proceeds to myelin synthesis [2]. NAA was revealed as a major amino acid derivative in mammalian brain half a century ago and was shown to be present at high concentrations in most regions of mammalian brain. During postnatal development NAA is found in both neurons and oligodendrocytes but becomes located in neurons at maturity. It was shown that concentration of NAA corresponds to mitochondrial function, which provides an opportunity to measure the quantity of NAA in human brain in vivo. Thus, reduction of NAA concentrations was found in several psychiatric disorders such multiple sclerosis, Alzheimer’s disease. Abnormally high level of NAA is associated with Canavan disease (CD) [3]. A defect in human aspA gene that encodes for aspartoacylase has been identified as a cause of CD – a fatal neurodegenerative disorder for which there is currently no treatment. CD effects the CNS characterized by dysmyelination and spongiform degeneration of white matter in children. The first computational modeling of ASPA-catalyzed deacylation mechanism was reported by C. Zhang by means of QM/MM based SCC-DFTB model. Since the final structure of EP complex in the mechanism proposed could be considered insufficient as Glu178 remains protonated whereas resulting aspartate amino group is uncharged, we decided to complete it. The authors believe that it would be highly desirable to carry out QM/MM study in order to elucidate all the catalytic effects of ASPA [4]. Our primary goal was to calculate an inclusive potential energy profile for the chemical stage of ASPA-catalyzed NAA hydrolysis and to estimate the activation barrier for a first stage of its mutant analogue E285A. For an accurate comparison of calculated energy values with experimentally obtained kcat and Km [3] a complete catalytic cycle was modeled which involved computation of substrate-binding free energy and free energy of products’ release. The TOF (turnover frequency) of human brain ASPA (kcat) was evaluated in terms of Energetic span model for a complete Michaelis-Menten catalytic cycle [5]. Our minor goal was to elucidate the experimental relation between ASPA-activity and NAA-concentration, which revealed putative existence of side binding site and corresponding allosteric regulation of human brain ASPA [4]. Hybrid QM/MM method was implied for computational modeling of ASPA hydrolysis reaction. A QM part was treated with DFT/PBE0/6-31G**, and MM-part was characterized with AMBER99 force field. For prior minimum energy profile (MEP), series of relaxed scans were performed with elastic constant set to 5 a.e. Selected stationary points - ES, I1, I2, I3, I4, I5 and EP - were confirmed in series of unconstrained QM/MM optimizations. The optimization was deemed converged if absolute difference in total energies between consecutive optimization cycles fell below 10-6 a.u. For saddle point optimizations Hessian matrixes were computed for putative transition states structures with subsequent optimizations following the imaginary frequencies. After saddle point optimization the imaginary frequencies were recalculated in support of transition state structure selected for each stage of the profile. When the saddle points were located, we have verified that the descent forward and backward correctly led to the respective minimum energy structures. The free energies of noncovalent stage of catalytic cycle were computed with molecular dynamics Umbrella sampling method.