Automatic inducer addition and harvesting of recombinant Escherichia coli cultures based on indirect on‐line estimation of biomass concentration and specific growth rate

Abstract

This article describes a novel bioreactor configuration for production optimization of recombinant proteins in Escherichia coli. Inducer addition and harvesting are controlled on-line based on indirect estimation of biomass concentration and specific growth rate from addition of NaOH to maintain constant pH. When either a predetermined biomass concentration is reached or the cultures have obtained, a constant specific growth rate inducer is introduced automatically. The induction period is ended by automatic harvesting of the cultures either at a predetermined biomass concentration or when substrate (in this study glucose) is depleted, detected as an increase of pH, or dissolved oxygen tension. During harvesting, metabolic activities are quenched within 3 min by cooling of the cell suspension. The system has been used to optimize expression of glutathione S-transferase (GST) fusion protein of the ligand binding domain of mouse peroxisome proliferator-activated receptor, GST-PPARalpha LBD. Total yield of GST-PPARalpha LBD was independent of the time of inducer addition as long as the length of induction period corresponded to at least 0.25 cell divisions while the yield of soluble GST-PPARalpha LBD, the only active form, increased with the length of induction period. Highest yields were obtained when the inducer was added at low cell concentration as soon as constant specific growth rate was detected, resulting in induction periods corresponding to 3.4 +/- 0.4 cell divisions. The specific growth rate remained almost constant for one cell division after inducer addition, whereafter it decreased. No decrease of specific growth rate was observed when inducer was added in the lag-phase, and no soluble protein was produced. These results suggest that solely soluble GST-PPARalpha LBD acts as a growth inhibitor and that GST-PPARalpha LBD is expressed predominantly as inclusion bodies immediately after inducer addition whereas the proportion expressed as soluble protein is increased after 1 h of induction. Compared to the procedures, which are generally used for protein expression in the laboratory, this system is less labor intensive, it automatically provides recording of biomass concentration and specific growth rate, and it allows direct comparisons between expression of different proteins and performance of different constructs since the induction period is linked to growth.

Evaluation of software sensors for on-line estimation of culture conditions in an Escherichia coli cultivation expressing a recombinant protein

“…Our observations are in contrast to previous results, for example those of Eriksen et al (2001), who reported good agreement for biomass concentration estimation after addition of inducer. These authors, like others (Lee et al, 1988;Jensen and Carlsen, 1990), remark that recombinant protein expression generally affects the physiology and specific growth rate in E. coli but that the effect is greatly dependent on the particular protein expressed.…”

“…The strong promoter system used for the GFP expression in our study may have resulted in a different culture response. As previously observed (Eriksen et al, 2001;Sundström and Enfors, 2008), in the initial culture phase the software sensor estimation is affected by the supply of ammonia salts and amino acids in the specific media used. Since the medium used in the present study included constituents such as yeast extract, ammonium chloride, and sulfate, as well as a balance of CO 2 /HCO 3 , the initial deviation we obtained could to some extent be explained by this (cf.…”

“…The titration method has been successfully applied in a variety of microbial cultures (Christiansen Lykkegaard and Eriksen, 2002;Hoffmann et al, 2000;Eriksen et al, 2001;Jenzsch et al, 2006a;Sundström and Enfors, 2008). Ammonia has proved a particularly useful titrant in E. coli cultivations operated as batch or fed-batch, due to its stoichiometric uptake by the cells.…”

“…Greek letters NH 3 biomass specific growth rate from titrant (g g −1 h −1 ) feed density of feed (g L −1 ) feed accumulated feed (g) Ä NIR factor for NIR signal to biomass concentration (g L −1 /%) NIR biomass (NIR) specific growth rate (h −1 ) NH 3 titrant pump runtime (h) mechanistic model derived from first principles, typically on the basis of basic mass balance and kinetic equations (Axelsson et al, 1988;Eriksen et al, 2001;Cornelissen et al, 2003;Sundström and Enfors, 2008). Estimation methods from control theory, such as extended Kalman observers (Jenzsch et al, 2006a,b) or asymptotic observers, have proved particularly useful (Veloso et al, 2009).…”

Software sensors for fermentation processes

“…The approach to use base titration data to estimate biomass and specific growth rate has been evaluated and described in the literature for E. coli [12][13][14][15], yeasts [16][17][18][19] and other organisms [20]. The sensor is useful for instance in processes where induction at a certain cell density or growth rate is wanted.…”