ИСТИНА

Motivation and Aim: Escherichia coli is one of the best investigated organism, which is widely used both in microbiology and for biotechnological production of different biologically active compounds. The special stage in microorganism metabolism investigations started with implementation of 13C metabolic flux analysis (13C-MFA). This approach allows quantitative estimation of actual carbon metabolic fluxes distribution in vivo. For example, using 13C-MFA the role of two transhydrogenases encoded by genes pntAB and udhA was experimentally determined and characterized [1]. However, there are still some unsolved questions, even about central metabolism of E. coli. In particular, there are two opposite points of view about flux distribution in 6-phosphogluconolactonase deficient (Δpgl) mutant of E. coli. For example, applying the flux balance analysis (FBA) to the well known E. coli strain BL21, which has pgl deficient genotype, investigators assume that oxidative branch of pentose phosphate pathway (oxPPP) is inactive in this strain [2]; this assumption influences on prediction of biotechnological potential of this strain based on genome-scale model. This assumption is based on experimental data of accumulation of product of glucose-6-phosphate dehydrogenase reaction (G6PDH reaction), 6- phosphogluconolactone [3], which is a substrate of 6-phosphogluconolactonase reactions (PGL reaction). In contrast, 13C-MFA of BL21 [4] and pgl deficient MG1655 strain mutant [5], carried out also in our laboratory, has demonstrated only insignificant decrease of flux through oxPPP; the spontaneous hydrolysis of 6-phosphogluconolacton is proposed as the simplest explanation of this phenomenon. In this work we hypothesized that flux through oxPP pathway is actually reduced by near two times, but not to zero, in pgl mutant. Methods and Algorithms: For 13C-MFA of E. coli strains grown on minimal medium with 20 %/80 % mixture of [1-13C]- and [U-13C]-glucose as sole carbon source. Flux calculation and statistical analysis of calculated carbon flux distribution have been provided the OpenFLUX2 software [6] on the basis of labeling data of proteinogenic amino acids and monosaccharides from cellular polymers. Results: Comparative 13C-MFA of two E. coli strains MG1655 and MG1655 Δpgl has shown that the flux solution with the best accordance with experimental data (minimal value of sum of squared residuals, SSR) corresponds to carbon flux through the oxPPP of 18 % of input glucose in case of Δpgl mutant in contrast with 25 % in parent strain. The minimal value of SSR is commonly used but conditional criterion. Formally, any flux distribution providing SSR value which passes χ2-criterion can be recognized as acceptable [7]. However flux calculation under assumption that pgl inactivation blocks oxPPP led to only statistically unacceptable solution (even minimal SSR value do not pass χ2-criterion). We determined that flux through PGL reaction responded to χ2-criterion on minimal 10 % level. It means that the 2.5-fold decrease in flux through the oxPPP in pgl deficient mutant can reflect the real situation.