SUMMARYTen DEAE (2-(diethy1amino)ethyl) dextran samples were investigated by means of static and dynamic light scattering, viscometry and size-exclusion chromatography (SEC) in combination with on-line small-angle laser light scattering (LALLS) and viscometry (VISC). In dilute solution the behavior of DEAE-dextran was compared with that of unsubstituted dextran and the molecular weight (M) dependences of the radius of gyration R, , hydrodynamic radius R , , intrinsic viscosity [ q ] , second virial coefficient A , and zaverage diffusion coefficient were determined. The relationships for DEAE-dextran dissolved in a 0,8 molar sodium nitrate solution were nearly the same as for dextran dissolved in water with 0,05 wt.-Yo sodium azide and gave the same exponents. The molecular weight dependence of the intrinsic viscosity cannot be described by a Kuhn-Mark-Houwink relationship with a constant exponent. The slope in the plot of log [q] versus IogMdecreases with increasing molecular weight which indicates the occurrence of branching. By means of SEC/LALLS/VISC measurements the molecular weight distributions were determined.The distributions were calculated (1) directly from the light scattering signal, (2) from a calibration line obtained by light scattering data of a DEAE-dextran sample with a broad distribution and (3) from the intrinsic viscosity distribution obtained by the on-line viscosity/refractive index detector in combination with the [q]-M relationship. In order to get the correct molecular-weight dependence of the intrinsic viscosity it is necessary to determine the molecular weight distribution directly by LALLS (technique 1) and to combine this with the appropriate intrinsic viscosity data from the viscometer. Only the third technique, which is an extension of technique 1, gave satisfactory results over the whole molecular weight region observed.

When the transketolase‐deficient and D‐ribose‐producing Bacillussubtilis strain ATCC 21951 was grown in a glucose (200 g l−1)‐based medium (Kintaka et al. 1986), only 11 g l−1D‐ribose was synthesized, in addition to acetic acid (12 g l−1) and acetoin plus 2,3‐butanediol (24 g l−1), within 1 week of fermentation. After optimizing the process conditions at 2 l fermentor scale (simplified medium composition, pH 5·0 or 6·0, highly oxidative (1000 rev min−1, 3 vvm)), 40 g l−1D‐ribose was obtained from 200 g l−1D‐glucose, in addition to 14·5 g l−1 acetoin, during 1 week of fermentation. By partially substituting D‐glucose with D‐gluconic acid (100 g l−1D‐glucose plus 50 g l−1D‐gluconic acid) under highly oxidative (1000 rev min−1, 3 vvm) and pH‐controlled (pH 6·5) conditions, D‐ribose productivity increased (45 g l−1) and acetoin formation (7·5 g l−1) dropped, as did the fermentation time (3·5 d). The mixed carbon substrate procedure here developed provides an excellent alternative to the less efficient glucose‐based processes described so far.

WhenBacillus subtilis strain ATCC 21951, a transketolase-deficient D-ribose-producing mutant, was grown on Dglucose plus a second substrate which is metabolized via the oxidative pentose phosphate cycle (D-gluconic acid, D-xylose, L-arabinose or D-xylitoI), D-glucose did not catabolite repress metabolism of the second carbon source. The D-ribose yield obtained with the simultaneously converted carbon substrates, significantly exceeded that when only D-glucose was used. In addition, the concentration of glycolytic by-products and the fermentation time significantly decreased. Based on these findings, a fermentation process was developed with B. subtilis strain ATCC 21951 in which D-glucose (100 g L 1) and D-gluconic acid (50 g L -1) were converted into 45 g L -1 of D-ribose and 7.5 g L -1 of acetoin. A second process, based on D-glucose and D-xylose (100 g L -1 each), yielded 60 g L -1 of e-ribose and 4 g L -1 of acetoin plus 2,3-butanediol. Both mixed carbon source fermentations provide excellent alternatives to the less efficient D-glucose-based processes used so far.

No abstract

No abstract