Modeling solid-to-solid biocatalysis: integration of six consecutive steps.

A quantitative model for the conversion of a solid-substrate salt to a solid-product salt in a batch bioreactor seeded with product crystals is presented. The overall process consists of six serial steps (with dissolution and crystallization each in themselves complex multistep processes): solid-salt dissolution, salt dissociation into an ionic substrate and a counter-ion, bioconversion accompanied by biocatalyst inactivation, complexation of the ionic product with the counter-ion, and salt crystal growth. In the model, the consecutive steps are integrated, including biocatalyst inactivation and assuming that salt dissociation and complexation of ions are at equilibrium. Model parameters were determined previously in separate independent experiments. To validate the model, either dissolved or solid Ca-maleate was converted to solid Ca-D-malate by permeabilized Pseudomonas pseudoalcaligenes in a batch bioreactor seeded with Ca-D-malate crystals. The model very well predicted the concentrations of all components in the liquid phase (Ca-maleate, Ca(2+), maleate(2-), D-malate(2-), and Ca-D-malate) and the amounts of the solid phases (Ca-maleate. H(2)O and Ca-D-malate. 3H(2)O), especially when high initial amounts of Ca-maleate. H(2)O and Ca-D-malate. 3H(2)O were present.

Growth of Ca-D-malate crystals in a bioreactor.

To develop a bioreactor for solid-to-solid conversions, the conversion of solid Ca-maleate to solid Ca-D-malate by permeabilized Pseudomonas pseudoalcaligenes was studied. In a bioreactor seeded with product (Ca-D-malate) crystals, growth of Ca-D-malate crystals is the last step in the solid-to-solid conversion and is described here. Crystal growth is described as a transport process followed by surface processes. In contrast to the linear rate law obeyed by the transport process, the surface processes of a crystal-growth process can also obey a parabolic or exponential rate law. Growth of Ca-D-malate crystals from a supersaturated aqueous solution was found to be surface-controlled and obeyed an exponential rate law. Based on this rate law, a kinetic model was developed which describes the decrease in supersaturation due to Ca-D-malate crystal growth as a function of the constituent ions, Ca(2+) and D-malate(2-). The kinetic parameters depended on temperature, but, as expected (surface-controlled), they were hardly affected by the stirring speed.

D-malate production by permeabilized Pseudomonas pseudoalcaligenes; optimization of conversion and biocatalyst productivity.

For the development of a continuous process for the production of solid D-malate from a Ca-maleate suspension by permeabilized Pseudomonas pseudoalcaligenes, it is important to understand the effect of appropriate process parameters on the stability and activity of the biocatalyst. Previously, we quantified the effect of product (D-malate2 -) concentration on both the first-order biocatalyst inactivation rate and on the biocatalytic conversion rate. The effects of the remaining process parameters (ionic strength, and substrate and Ca2 + concentration) on biocatalyst activity are reported here. At (common) ionic strengths below 2 M, biocatalyst activity was unaffected. At high substrate concentrations, inhibition occurred. Ca2+ concentration did not affect biocatalyst activity. The kinetic parameters (both for conversion and inactivation) were determined as a function of temperature by fitting the complete kinetic model, featuring substrate inhibition, competitive product inhibition and first-order irreversible biocatalyst inactivation, at different temperatures simultaneously through three extended data sets of substrate concentration versus time. Temperature affected both the conversion and inactivation parameters. The final model was used to calculate the substrate and biocatalyst costs per mmol of product in a continuous system with biocatalyst replenishment and biocatalyst recycling. Despite the effect of temperature on each kinetic parameter separately, the overall effect of temperature on the costs was found to be negligible (between 293 and 308 K). Within pertinent ranges, the sum of the substrate and biocatalyst costs per mmol of product was calculated to decrease with the influent substrate concentration and the residence time. The sum of the costs showed a minimum as a function of the influent biocatalyst concentration.

IgG and hybridoma partitioning in aqueous two-phase systems containing a dye-ligand.

The effect of the important ATPS- and buffer parameters on IgG and hybridoma partitioning in ATPSs containing a PEG-dye-ligand was studied. Objective was to establish selection criteria for effective ligands for extractive fermentations with animal cells in ATPSs. In the presence of 1% PEG-dye-ligand the binding of IgG to the PEG-ligand was affected severely by the Na-chloride concentration. The tie-line length and pH affected IgG partitioning to a lesser extent. The desired partitioning of IgG into the top phase, was only obtained when, in addition to the 10 mmol/kg K-phosphate buffer, no Na-chloride was present. In an ATPS culture medium, with +/- 35 mmol/kg Na-bicarbonate and 60 mmol/kg Na-chloride, increasing the PEG-dye-ligand concentration up to 100% did increase the partition coefficient, but was not effective in concentrating the IgG in the top phase of ATPS culture medium at a pH of 7.8. Furthermore, addition of the PEG-dye-ligand to ATPS culture medium changed the hybridoma cell partitioning from the bottom phase to the interface.

Hybridoma and CHO cell partitioning in aqueous two-phase systems.

The partitioning of mouse/mouse hybridoma cell line BIF6A7, mouse/rat hybridoma PFU-83, and CHO DUKX B11-derived cell line BIC-2 in aqueous two-phase systems (ATPSs) of poly(ethylene glycol) (PEG) and dextran was studied. The partitioning of BIF6A7 was investigated systematically by using a statistical experimental design. The aims were to identify the key factors governing cell partitioning and to select ATPSs with suitable cell partitioning for extractive bioconversions with animal cells. The influence of five factors, i.e., the poly(ethylene glycol) molecular weight (PEG MW), dextran molecular weight (Dx MW), tie-line length (TLL), pH, and the ratio of potassium phosphate to potassium chloride, defined as the fraction KPi/(KPi + KCl), on BIF6A7 cell partitioning was characterized by using a full factorial experimental design. The cell partitioning ranged from complete partitioning into the interface to an almost complete partitioning to the lower phase. In all cases less than 1% of the cells partitioned to the top phase. The potassium phosphate fraction had the largest effect on cell partitioning. Low potassium phosphate fractions increased the proportion of cells in the lower phase. To a lesser extent the other factors also played a role in the cell partitioning. The best partitioning for the BIF6A7 cell line was obtained in ATPSs with PEG MW = 35,000, Dx MW = 40,000, TLL = 0.10 g/g, pH 7.4, and KPi/(KPi + KCl) = 0.1, where 93% of the cells were present in the lower phase. The previously reported partitioning of BIF6A7 cells in ATPS culture medium, corresponded well with the current findings. The partitioning of mouse/rat hybridoma cell line PFU-83 and CHO cell line BIC-2 was studied in an ATPS culture medium with PEG 35,000, dextran 40,000, TLL = 0.12 g/g, and hybridoma culture medium. Both cell lines partitioned almost completely into the lower phase. Moreover, the PFU-83 cell line was able to grow in the ATPS hybridoma culture medium. This ATPS hybridoma culture medium therefore seems to be suitable for extractive bioconversions with a wide range of hybridoma cell lines, provided that their product can be partitioned into the upper PEG-rich phase.

Separation of hybridoma cells from their IgG product using aqueous two-phase systems.

The partitioning of IgG in aqueous two-phase systems (ATPSs) of PEG and Dextran was studied systematically using a statistical experimental design. Aim was to improve the separation of hybridoma cells and their IgG product, by identifying the key variables governing IgG partitioning, and by comparing the IgG partitioning data with the hybridoma cell partitioning data obtained in previous work. The influence of five factors, i.e. the poly(ethylene glycol) molecular weight (PEG Mw), dextran molecular weight (Dx Mw), tie-line length (TLL), pH and potassium phosphate fraction (KPi/(KPi+KCl)), on IgG partitioning was characterized using a full-factorial experimental design. In all of the ATPS's the IgG partitioned predominantly into the lower phase. The partition coefficient varied between 0.78 (Variable settings: PEG Mw = 6000, Dx Mw = 500000, TLL = 0.10 g g-1, KPi/(KPi+KCl) = 1.0 and pH = 7.4) and 0.0002 (Variable settings: PEG Mw = 35000, Dx Mw = 40000, TLL = 0.20 g g-1 KPi/(KPi+KCl) = 1.0 and pH 6.6). The tie-line length, the dextran molecular weight and the PEG molecular weight had the most pronounced effect on IgG partitioning. Matching the partitioning data of the IgG product with previously obtained data of the hybridoma cell partitioning, showed that within the experimental design no ATPS could be found giving a good separation of the hybridoma cells and their IgG product. There are, however, ATPS's available in which the cells partition to, and grow in the lower dextran-rich phase. To achieve a good separation of the hybridoma cells and their IgG product in these ATPSs, the IgG product has to be specifically extracted into the PEG-rich top phase. For this purpose the use affinity ligands coupled to PEG may offer a solution. Therefore, a number of commercially available dye-resins was screened for their ability to bind the BIF6A7 IgG antibody. The mimetic green 1 A6XL dye-resin was found to bind BIF6A7 IgG. The dye-ligand coupled to PEG was used to manipulate the IgG partitioning in an ATPS. In the presence of the PEG-ligand, the IgG partitioned almost completely to the top phase. The IgG-partition coefficient increased three orders of magnitude, resulting in a 25-fold higher IgG concentration in the top phase than in the bottom phase.