Premium
Simulated oxygen and glucose gradients as a prerequisite for predicting industrial scale performance a priori
Author(s) -
Kuschel Maike,
Takors Ralf
Publication year - 2020
Publication title -
biotechnology and bioengineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.136
H-Index - 189
eISSN - 1097-0290
pISSN - 0006-3592
DOI - 10.1002/bit.27457
Subject(s) - process engineering , a priori and a posteriori , mixing (physics) , scale (ratio) , scale up , corynebacterium glutamicum , biochemical engineering , computer science , impeller , mass transfer , biological system , computational fluid dynamics , environmental science , mechanical engineering , chemistry , engineering , aerospace engineering , chromatography , physics , philosophy , epistemology , biochemistry , classical mechanics , quantum mechanics , biology , gene
Abstract Transferring bioprocesses from lab to industrial scale without loss of performance is key for the successful implementation of novel production approaches. Because mixing and mass transfer is usually hampered in large scale, cells experience heterogeneities eventually causing deteriorated yields, that is, reduced titers, productivities, and sugar‐to‐product conversions. Accordingly, reliable and easy‐to‐implement tools for a priori prediction of large‐scale performance based on dry and wet‐lab tests are heavily needed. This study makes use of computational fluid dynamic simulations of a multiphase multi‐impeller stirred tank in pilot scale. So‐called lifelines , records of 120,000 Corynebacterium glutamicum cells experiencing fluctuating environmental conditions, were identified and used to properly design wet‐lab scale‐down (SD) devices. Physical parameters such as power input, gas hold up, k L a , and mixing time showed good agreement with experimental measurements. Analyzing the late fed‐batch cultivation revealed that the complex double gradient of glucose and oxygen can be translated into a wet‐lab SD setup with only few compartments. Most remarkably, the comparison of different mesh sizes outlined that even the coarsest approach with a mesh density of 1.12 × 10 5# / m 3was sufficient to properly predict physical and biological readouts. Accordingly, the approach offers the potential for the thorough analysis of realistic industrial case scenarios.