Alternative production process strategies in E. coli improving protein quality and downstream yields

Process Biochemistry 44 (2009) 1039–1045 Contents lists available at ScienceDirect Process Biochemistry journal homepage: www.elsevier.com/locate/procbio Alternative production process strategies in E. coli improving protein quality and downstream yields Jordi Ruiz, Jaume Pinsach, Gregorio Álvaro, Glòria González, Carles de Mas, David Resina, Josep López-Santı́n * Departament d’Enginyeria Quı´mica, Escola Tècnica Superior d’Enginyeria, Unitat de Biocatàlisi Aplicada associada al IIQA (CSIC), Universitat Autònoma de Barcelona, Edifici Q, 08193-Bellaterra, Spain A R T I C L E I N F O A B S T R A C T Article history: Received 30 May 2008 Received in revised form 14 May 2009 Accepted 19 May 2009 Process strategies for production of recombinant rhamnulose 1-phosphate aldolase (RhuA) in Escherichia coli were found to have an important impact on downstream processing when preparing the enzyme for its use as immobilized biocatalyst. First, a continuous inducer feed was implemented in substrate limited fed-batch cultures to overexpress RhuA with a hexa histidine-tag (6xHis-tag) at its N-terminus. The final specific RhuA level was 180 mg g 1 DCW, but the final specific enzyme activity (1.7 AU mg 1 RhuA) was considerably lower than expected. Only 55% of immobilization yield was achieved when immobilized metal affinity chromatography (IMAC) was used to purify and immobilize RhuA from cellular lysate in a single step. Western blot analyses showed that only 20% of overexpressed RhuA kept the whole 6xHistag at the end of the culture due to partial proteolysis. Two different growth strategies improved protein quality and immobilization yield: Keywords: E. coli Protein quality One step purification-immobilization Substrate in excess fed-batch cultures Continuous inducer dosage Aldolase (i) Temperature reduction to 28 8C in substrate limited operation decreased proteolysis and allowed higher specific activities, 210 mg g 1 DCW. The enzyme activity increased to 4 AU mg 1 RhuA and purification-immobilization yield to 93%. (ii) A novel fed-batch operational procedure, working at high glucose concentration was implemented. High aldolase levels, 233 mg g 1 DCW, were reached at the end of the culture. The final enzyme activity was also higher than 4 AU mg 1 RhuA, and 95% of immobilization yield was achieved. For both cases, Western blot analyses showed that 80–100% of overexpressed RhuA kept the whole 6xHis-tag at the end of the culture, confirming that recombinant protein quality had been improved. ß 2009 Elsevier Ltd. All rights reserved. 1. Introduction Nowadays, biotechnology is seen as an important tool for the sustainable industrial development. This requires the development of new enzymes, processes, products and applications [1]. Enzymes are used in a wide range of applications and industries [2], and they are a promising alternative/complement to chemical synthesis for producing chiral synthons that are used as building blocks to obtain highly active pharmaceuticals. Aldolases catalyze C–C bond formation with defined stereochemistry yielding enantiomerically pure products, even when the starting materials are non-chiral substrates [3]. Rhamnulose 1-phosphate aldolase (RhuA) catalyses the reversible cleavage of L-rhamnulose 1-phosphate to dihydroxyacetone phosphate * Corresponding author. Tel.: +34 935811806; fax: +34 935812013. E-mail address: josep.lopez@uab.cat (J. López-Santı́n). 1359-5113/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.procbio.2009.05.007 (DHAP) and L-lactaldehyde in vivo. The potential of RhuA as biocatalyst in asymmetric synthesis is based on the fact that the reaction can be reversed in vitro, and it accepts a wide range of both natural and non-natural substrates for the stereoselective synthesis of carbohydrates and other derivatives of pharmaceutical interest [4–7]. The production of recombinant proteins in Escherichia coli has been the subject of numerous studies. Several genetic as well as cultivation strategies have been developed to increase recombinant protein yields [8–10]. However, limitations still exist to produce them at high levels in a biological active form, and proteolysis has been reported to impact recombinant protein production processes [11–13]. Little work has been done on the control of proteolysis in high cell density fed-batch cultures [14] and although affinity tags have been reported to have a positive impact on yield, solubility and even folding and reduction of proteolysis [15], partial proteolysis has been identified as the cause for reduced IMAC recovery yields [16]. 1040 J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 The rhaD gene has been cloned and high volumetric productivities were achieved in high cell density cultures using E. coli M15 [pREP4] pQErham [17]. Nevertheless, IPTG pulse induction clearly overloaded host cell metabolism, reduced activities per gram of biomass were obtained and plasmid loss was observed [18]. To ensure plasmid maintenance, a glycine auxotrophy-based plasmid maintenance system (E. coli M15DglyA [pREP4] pQEabrham) was constructed [19]. Fed-batch operation with continuous IPTG feed was implemented to avoid the metabolic burden on host cells produced by pulse induction and to increase the specific activity yields [20]. The recombinant aldolase was expressed as a fusion protein to a hexa-hisitidine tag. Thus, one-step immobilization and purification of clear lysate on affinity chromatography (IMAC) support [21] could be used after fermentation broth processing. This technology has not only the advantage of reduction in enzyme loss and time processing during the purification stages, but also allows the obtention of an immobilized enzymatic derivative suitable for application in synthetic processes. When having a look at production and downstream processing as a whole, cost breakdown showed that downstream processing can hamper the economical viability of some recombinant protein production process [22]. Thus, integrated production and downstream process strategies should be developed. In this work, the linkage between recombinant 6xHis-tagged RhuA production strategy and downstream process by one-step purification and immobilization will be investigated in order to improve the yield and quality of the final product. 2. Materials and methods 2.1. Bacterial strain and plasmids The K-12 derived strain E. coli M15DglyA [pREP4] harbouring the vector pQEabrham was used for rhamnulose 1-phosphate aldolase overexpression. This is a system based on glycine auxotrophy to ensure plasmid stability which avoids the need of antibiotic supplementation [19]. Transcription of the rhaD gene is under the control of the strong T5 promoter in a low copy number plasmid derived from the pQE-40 (Qiagen). Frozen stock aliquots containing glycerol prepared from exponential phase cultures grown in Luria-Bertani media (LB) were stored at 80 8C. 2.2. Media composition LB medium, with a composition of 10 g L 1 peptone, 5 g L 1 yeast extract and 10 g L 1 NaCl, was used for the preinoculum preparation. A defined mineral medium, utilising glucose as the sole carbon source, was used for inocula and for bioreactor experiments. The medium for shake flask cultures was composed of 5 g L 1 glucose, 11.9 g L 1 K2HPO4, 2.4 g L 1 KH2PO4, 1.8 g L 1 NaCl, 3 g L 1 (NH4)2SO4, 0.11 g L 1 MgSO4
7H2O, 0.01 g L 1 FeCl3, 0.03 g L 1 thiamine and 0.72 mL L 1 of trace elements solution. The batch phase of bioreactor cultivations was composed of 20 g L 1 glucose, 11.9 g L 1 K2HPO4, 2.4 g L 1 KH2PO4, 1.8 g L 1 NaCl, 3 g L 1 (NH4)2SO4, 0.45 g L 1 MgSO4
7H2O, 0.02 g L 1 FeCl3, 0.1 g L 1 thiamine and 2.86 mL L 1 of trace elements solution. The feed medium for high-cell-density fermentations consisted of 487 g L 1 glucose, 9.7 g L 1 MgSO4
7H2O, 0.5 g L 1 FeCl3, 0.34 g L 1 thiamine, 64 mL L 1 trace elements solution and 0.5 mL L 1 of antifoam (Sigma). The trace elements solution composition was defined elsewhere [23], but the amount of Zn was doubled because 0.103 mg Zn g 1 DCW were required for the recombinant protein to be fully active [18]. Phosphates were not included in the feeding solution in order to avoid co-precipitation with magnesium salts. Instead, a concentrated phosphate solution containing 500 g L 1 K2HPO4 and 100 g L 1 KH2PO4 was pulsed during the fed-batch phase to avoid their depletion when necessary (calculated yield YXP = 18 g DCW g 1 P). 2.3. Cultivation conditions Preinoculum cultures were grown from glycerol stocks in a 100 mL shake flask containing 15 mL LB media and incubated overnight at 37 8C in a rotary shaker at 250 rpm. To prepare inoculum, 5 mL of preinoculum cultures were transferred aseptically to a 0.5 L shake flask containing 100 mL of defined medium which was incubated at 37 8C for 5 h at 250 rpm. For bioreactor experiments, 80 mL of inoculum culture were transferred to the bioreactor containing 720 mL of defined medium. All growths were carried out using a Biostat1 B bioreactor (Sartorius) equipped with a 2 L fermentation vessel. The pH was maintained at 7.00  0.05 by adding 25% NH4OH solution to the reactor. The temperature was kept at 37 8C or 28 8C depending on the experiment. The pO2 value was maintained at 50% of air saturation by adapting the stirrer speed between 450 and 1120 rpm and supplying air (enriched with pure oxygen when necessary) at a space velocity of 2 vvm. The end of the batch phase was identified by a reduction in the oxygen consumption rate and an increase in pH. A simple mathematical model based on mass balances and substrate consumption kinetics was used in an open-loop mode to control the specific growth rate at a constant value by an exponential feed medium addition [24]. In substrate inhibited cultures, a concentrated glucose solution was pulsed to reach 60 g L 1 of glucose and the exponential feeding strategy was continued in order to maintain an almost constant concentration inside the bioreactor and the same specific growth rate. Induction started 1 h later. In all cases, IPTG was continuously fed into the bioreactor to tune the transcription rate below values which burdened the host cell metabolism. The continuous inducer feed was designed to progressively increase the IPTG concentration maintaining a relation between 0.4 and 0.6 mmol IPTG g 1 DCW [20]. 2.4. Broth processing The fermentation broth was centrifuged at 10,000 rpm for 20 min at 4 8C using a Beckman J2-21 M/E centrifuge. Harvested cells were resuspended in lysis buffer: 43 mM Na2HPO4, 7 mM NaH2PO4, 20 mM Imidazol, 300 mM NaCl (pH 8) at a ratio of 1 mL buffer: 0.3 g harvested cells. Resuspended cells were lysed by one-shot high pressure disruption (Constant Systems LTD One Shot) at 2.57 kbar and at constant temperature of 4 8C. The crude cell lysate was centrifuged at 14,000 rpm for 35 min at 4 8C and cell debris was rejected. Enzymatic activity of the supernatant was measured and sodium azide was added to keep a concentration of 0.02% (w/w) to avoid biological degradation of the clear lysate. 2.5. RhuA purification-immobilization 10 mL aliquots of clear lysate at the appropriate activity concentration were employed for one-step purification immobilization on Co-IDA support (Chelating Sepharose FF Amersham Biosciencies-GE Healthcare with Co2+ chelated on it). One millilitre sample was used as reference and it was kept under mild horizontal agitation at 4 8C. Its activity was measured both at the beginning and end of the immobilization process. The second sample (9 mL) was added to 1 mL of Co-IDA support, and the residual activity of the suspension and supernatant was measured until the adsorption–desorption equilibrium was reached. 2.6. Analytical methods 2.6.1. Monitoring bacterial growth Growth was followed by optical density measurements at 600 nm (OD600). The samples were diluted in deionised water until the measurement was within the linear range of the spectrophotometer. The dry cell weight (DCW) was measured by centrifugation of aliquots of the broth. The pellets were washed twice with deionised water and dried at 110 8C until constant weight. For other determinations, 1 mL of culture was centrifuged. The supernatant was then used for glucose, organic acids, ammonium and phosphate measurements. Glucose and organic acids were analyzed by HPLC (Hewlett Packard 1050) on an Aminex HPX-87H (Biorad) column at 25 8C with IR detector (HP 1047) using 15 mM H2SO4 (pH 3.0) as eluent at a flow rate of 0.6 mL min 1. Ammonium and phosphates were determined by colorimetric kit assays (Dr. Lange) following the supplier instructions. To quantify the product concentration during cultures, broth samples were withdrawn and centrifuged. The pellet was resuspended in 100 mM Tris
HCl (pH 7.5) to a final OD600 = 3 for enzyme determination. Cell suspensions were placed in ice and sonicated with four 15 s pulses with 2 min intervals in ice between each pulse using a Vibracell1 model VC50 (Sonics & Materials). Cellular debris was removed by centrifugation and the clear supernatant was collected for product analysis. 2.6.2. Product quantification and characterization Total protein content was determined using a Coomassie1 Protein Assay Reagent Kit (Pierce). To determine the percentage of RhuA amongst the rest of intracellular soluble proteins, 12% polyacrylamide SDS-PAGE gels were performed in a Miniprotean1 II instrument (Bio-Rad) according to the manufacturer’s instructions and quantified by Kodak Digital Science1 densitometry software. Determination of RhuA activity was carried out as described previously [17]. One unit of RhuA activity was defined as the amount of enzyme required to convert 1 mmol of rhamnulose 1-phosphate in DHAP per minute at 25 8C under the assay conditions. 2.6.3. Western blotting analyses Western blots were performed after SDS-PAGE proteins were transferred to a nitrocellulose membrane. 6xHis-tag proteins were detected with mouse anti-6xHIS antibodies (Roche); anti-mouse IgG alkaline phosphatase conjugate (Sigma) was used as a secondary antibody. Detection was performed with the Alkaline Phosphatase Conjugate Substrate Kit (BioRad). J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 1041 Fig. 1. Biomass, substrate, specific activity and specific RhuA content profiles along time in a substrate-limited fed-batch culture at 37 8C. The vertical dotted line shows when continuous IPTG supply was started. 3. Results and discussion 3.1. Production of RhuA in substrate-limited fed-batch cultures. Single step purification and immobilization Fig. 1 shows the time profiles of a glucose-limited fed-batch process for RhuA production performed at 37 8C. The culture was operated by using an exponential substrate feed, and a continuous inducer feed was switched on at early stages of the culture to progressively increase the inducer concentration up to 13 mM [20]. Although the continuous inducer feed kept the specific production rate at low values to avoid the metabolic burden effect (data not shown), cell growth stopped 24 h after the IPTG feed had been started, when the biomass concentration was approximately 30 gDCW L 1 (see Fig. 1). The contents of recombinant RhuA at that point was close to 180 mg RhuA g 1 DCW (corresponding to 27% of the total soluble protein). Fig. 1 also shows that the specific activity (activity units per gram of dry cell weight) increased similarly to the specific RhuA contents (milligram of recombinant protein per gram dry cell weight) and the biomass concentration, reaching 500 AU g 1 DCW, but suddenly decreased to 310 AU g 1 DCW when cell growth stopped, in spite of the steady values of specific RhuA contents. After processing the fermentation broth as described in the analytical methods section, the intracellular lysate was submitted to stirred tank adsorption on Co-IDA resin in order to purify and immobilize 6xHis-tag RhuA in a single step [21]. A proper evaluation of the recovery yield required the previous determination of the enzymatic load which allowed the measure of the immobilized biocatalyst activity in absence of internal diffusion limitations. Several immobilization runs starting from a clear lysate with and average aldolase activity of 21.6 AU mL 1 were carried out in a range between 1 and 50 AU mL 1 support. The relation between enzymatic load offered and retained activity in the support was evaluated and is presented in Fig. 2. The retained immobilized activity was calculated by subtracting the suspension activity to the activity remaining in the supernatant, while the actual measured activity was determined directly over the immobilized derivative. As can be seen, the measured enzymatic activity of the solid biocatalyst was considerably lower than the theoretical one when 10 AU mL 1 support were exceeded. It has to be noticed that the offered enzymatic loads between 1.5 and 20 AU mL 1 support corresponded to a recombinant protein amount (mg RhuA per mL of support) lower than the maximum capacity of the resin (recommended values of 12 mg mL 1 support to avoid saturation). The enzymatic load at which the catalytic process started to be controlled by mass transfer limitations instead of the reaction rate was then established between 5 and 10 AU mL 1 support. Concerning to total immobilization yields, moderate values between 55 and 65% were achieved in all cases. These yields were calculated as the activity difference between reference (blank retaining 100% activity under immobilization conditions) and supernatant in order to account for the entire enzyme linked to the support via the histidine tag (both active and deactivated). These values were lower than expected compared to the yields obtained for a related aldolase (FucA) of around 95% using Co-IDA supports [21]. Even though the amount of RhuA did not overtake the maximum acceptable load, the yields achieved for each immobilization seemed to be independent from the amount of RhuA offered. These results suggested that part of RhuA had lost the histidine tag and was not able to adsorb on the IMAC support. Fig. 2. Influence of the enzymatic load on Co-IDA immobilization of RhuA from 37 8C substrate-limited fermentation broth. Measured RhuA activity in the immobilized derivative (5) and retained immobilized activity (*) vs. offered activity in the immobilization process at 4 8C and pH 8. Dashed line indicates a theoretical 100% retained activity. 1042 J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 recombinant RhuA synthesis combined with the starvation experienced by cells due to substrate limitation [11,13,25]. To overcome these problems, we focused on alternative growth strategies to minimise the deleterious effects on protein quality, i.e. activity units per milligram of pure RhuA and integrity of the hexahistidine tag: (a) Lowering process temperature, in order to improve protein folding and reduce the activity of proteases. (b) Substrate-in-excess fed-batch cultures to minimise the starvation-induced stress. 3.2. Substrate-limited fed-batch cultures at 28 8C Fig. 3. Specific RhuA activity and specific 6xHis-tag contents in recombinant RhuA along the induction phase in a substrate-limited fed-batch culture at 37 8C. The vertical dotted line shows when continuous IPTG supply was started. Thus, Western blot analyses were performed to check for the 6xHis tag in RhuA along the induction phase in the substratelimited fed-batch culture. In Fig. 3, the Western blot data are presented as fraction of initial specific intensity (calculated for the first induced sample) along time. As can be seen, the specific content of 6xHis-tag in RhuA decreased along the production phase and, at the end of the culture, only 20% of the initial 6xHis-tag mg RhuA 1 was detected. This fact might be attributed to a partial proteolysis affecting the recombinant protein tag. The results are compatible with the obtained immobilization yields of 55% as the Western blot is specific for 6 histidine residues per tag. A fraction of the proteolyzed protein could keep enough histidine residues to be attached to the column and account for the remaining 35% yield. In fact, this possible partial proteolysis did not only affect the affinity tag, but also the biological activity of the protein. The specific activity of the recombinant enzyme (also shown in Fig. 3) was never above 3.4 AU mg 1 RhuA, and as low as 1.7 AU mg 1 RhuA at the end of the culture. Since 6.2 AU mg 1 RhuA had been reported as the specific activity for pure RhuA [17], a significant decrease of the specific enzyme activity was evidenced. Proteolysis could probably be switched on because of the stress caused by the The reduction of the production process temperature could be a promising strategy to improve RhuA quality because it has been reported that the conformational quality and functionality of protein increase in parallel to reduced culture growth temperature [26,27]. Moreover, protein degradation rate will decrease with decreased temperature in accordance to the Arrhenius’ Law and protein misfolding often occurs at high temperatures due to either heat damage or due to an over increase of protein synthesis rate [14]. In consequence, an additional culture was performed reducing the fed-batch phase temperature from 37 8C to 28 8C, and employing the same induction strategy as in the previous experiment. The results are presented in Fig. 4. In this case, the specific RhuA levels increased along the whole induction phase reaching 210 mg RhuA g 1 DCW (around 40% of the total intracellular protein) at the end of the culture. As can be seen, the profiles of enzyme activity and RhuA amount are more parallel than at 37 8C, reaching 850 AU g 1 DCW. The fermentation broth was submitted to the same downstream processing than the 37 8C one and one-step purification/ inmobilization of the intracellular lysate was performed by offering 3 AU mL 1 support (without overtaking the maximum acceptable protein load). The obtained yield was around 93%, allowing an almost complete recovery of the produced aldolase in active immobilized form. Western blot results agreed with the immobilization data. The specific content of 6xHis-tag in RhuA was nearly constant along the induction phase, indicating that all the obtained active aldolase contained enough histidine molecules to be linked to the Co-IDA resin. Fig. 4. Biomass, substrate, specific activity and specific RhuA content profiles along time in a substrate-limited fed-batch culture at 28 8C. The vertical dotted line shows when continuous IPTG supply was started. J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 Fig. 5. (a) Initial specific growth rates vs. initial glucose concentrations in shake flask cultures. (b) Specific acetate production rate (qAc, g acetate g specific growth rate in shake flask cultures. The average value of the activity units per mg of RhuA was around 4 AU mg 1 RhuA, clearly higher than the average at 37 8C (2.6 AU mg 1 RhuA). The above data are confirming the suitability of temperature reduction to improve recombinant protein quality and significantly increasing the downstream yields. 3.3. Substrate in excess fed-batch cultures An alternative operational strategy to minimise the deleterious effects of substrate limitation on recombinant RhuA was implemented in fed-batch cultures by using an excess of glucose. In the standard glucose-limited high cell density cultures, IPTG addition triggers induction in cells which are already short of resources and stressed due to limited carbon source [28], compromising cell growth and yielding lower specific protein levels than in the case of unlimited cultures [20]. Thus, it would be desirable to induce high cell density cultures in which cells had enough resources available, i.e. an excess of glucose. Under glucose in excess, carbon catabolite repression takes place, reducing the AMP and CRP levels [29,30]. Thus, IPTG transport into the cell is expected to be mainly by diffusion because permease (lacY gene products) expression should be repressed [31,32]. For the same reason, the expression of the recombinant aldolase should take place at a lower level than in absence of glucose [33]. If IPTG transport and protein expression rates could be compatible with growth in the presence of glucose in excess, a novel strategy for recombinant protein production would be possible. Although it implies an increase in the fermentation costs, 1043 1 DCW h 1 ) vs. initial any possible increase in productivity (even moderate) could justify its use, as recombinant proteins are usually high-added value products. Preliminary shake flask cultures were performed to characterize growth in defined medium. Ten non-induced shake flask cultures were grown at initial glucose concentrations ranging from 0 to 90 g L 1. Biomass, substrate and acetate concentrations were measured along time in all cultures. Large differences were observed in the case of cultures grown at high initial glucose concentrations. When plotting the initial specific growth rates (m0) vs. initial glucose concentrations for each culture, a substrate inhibition growth pattern was obtained (Fig. 5a). High specific acetate production rates (qAc, g acetate g glucose 1 h 1) were measured at specific growth rates above 0.50 h 1. Acetate accumulation was switched off at a specific growth rate below 0.32 h 1, and below 0.20 h 1 it was not detected (Fig. 5b). This critical specific growth rate was similar to the one determined in substrate-limited fed-batch cultures [20]. Then, for this particular strain, inhibiting acetate concentrations could be avoided growing at specific growth rates lower than 0.20–0.25 h 1 by maintaining the glucose concentration above 60 g L 1 in the bioreactor. Therefore, this strategy could be potentially used to obtain high cell density cultures. A glucose in excess fed-batch experiment was performed at 37 8C (Fig. 6). The growth started as usual, using a predefined exponential feed addition profile. As described in Section 2, growth was inhibited by pulsing a concentrated glucose solution to reach 65 g glucose L 1 1 h before IPTG feed was switched on. After the glucose pulse, the specific growth rate was kept at low enough Fig. 6. Biomass, substrate, specific activity and specific RhuA content profiles along time in a substrate-inhibited fed-batch culture. Glucose concentration was increased to 65 g L 1 at 19 h, by adding 100 mL of a concentrated glucose solution. The vertical dotted line shows when continuous IPTG supply was started. 1044 J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 values to avoid acetate accumulation and IPTG concentration increased from 10 to 25 mM during the 5.5 h induction phase. The IPTG range was determined from previous experiments under substrate in excess (data not shown), which showed that 10 mM IPTG was necessary to trigger RhuA expression and the bacterial growth was unaffected below 25 mM. At the end of the culture, biomass concentration was above 40 g DCW L 1. The biomass build-up stopped when the specific RhuA levels reached 233 mg g 1 DCW (corresponding to 39% of the total soluble protein). Fig. 6 also shows that the specific activity increased similarly to the specific content of RhuA reaching 1130 AU g 1 DCW at the end of the culture. In contrast to the case of substrate limitation, no specific activity decrease was observed when cell growth stopped. After processing the fermentation broth, a clear lysate with a measured aldolase activity of 33 AU mL 1 was obtained. The single step purification-immobilization was carried out by offering 5 AU mL 1 and the immobilization yield achieved in this case was 95%. As in the 28 8C operation, almost all the activity was able to adsorb on the IMAC support. Western blot analyses confirmed that the loss of 6xHis tag in RhuA had been minimised. Proteolysis effects over the recombinant protein decreased, and the specific content of 6xHis-tag in RhuA was nearly constant along the production phase and 80% of the overexpressed RhuA kept the whole 6xHis-tag at the end of the induction phase. These results are compatible with the 95% immobilization which include the attachment of 6xHis tagged protein and also partially hydrolyzed protein still containing enough histidine residues to adsorb onto the IMAC support. The negative effects of proteolysis over RhuA biological activity were also diminished as the average specific activity of the recombinant enzyme was around 4 AU mg 1 RhuA. Thus, RhuA quality has been significantly improved reaching the same levels than the operation at low temperature. 4. Conclusions The standard substrate-limited fed-batch strategy was found to have a negative impact on recombinant 6xHis-tagged RhuA production and downstream. Starvation-induced proteolysis affected both the enzyme biological activity as well as its recovery yield. To overcome these problems, two alternative operational procedures have been proposed to improve aldolase quality and the yield of the immobilization and purification step. The operation at low temperature (28 8C) allows increasing the specific enzyme activity when compared to the case of substrate limited fed-batch cultures. In addition, the recombinant enzyme could be almost quantitatively recovered in a single step. On the other hand, an alternative growth strategy at nonlimiting substrate levels has been implemented in high cell density cultures. The culture can be operated at high cellular levels and submitted to continuous inducer supply for RhuA overexpression. When compared to substrate-limited cultures, higher inducer concentrations were required, although still much lower than the standard pulse concentrations used in the range 500 mM to 1 mM [33]. The required minimum value for aldolase expression moves from 4 to 10 mM and the IPTG amount affecting cell growth from 12 to 25 mM. It has to be taken into account that inducer uptake is envisaged to be different depending on the case. In substrate in excess cultures, the inducer was not uptaken as effectively as in the case of glucose limitation, and probably a lower fraction of the IPTG added was transported into the cell. Carbon catabolite repression strongly influenced inducer uptake when glucose levels were high, and IPTG diffusion was probably the limiting step for RhuA overexpression. In contrast, at limiting glucose levels, IPTG was also transported by permeases, which enhance its transport into the cytoplasm [31,34]. The successful operation at high glucose levels allowed obtaining similar protein quality improvements than at lower temperatures and a quantitative recovery of the obtained enzyme in active immobilized form. These findings could justify the higher glucose costs associated with this operation mode. Acknowledgements This work was supported by the Spanish Programme on R&D, Project numbers CTQ2005-01706 and CTQ2008-00578, and by DURSI 2005SGR 00698 Generalitat de Catalunya. The Department of Chemical Engineering of the UAB constitutes the Biochemical Engineering Unit of the Reference Network in Biotechnology of the Generalitat de Catalunya (XRB). J.P. and J.R. acknowledge the Spanish MEC for predoctoral grants. References [1] Lorenz P, Eck J. Metagenomics and industrial applications. Nature Reviews Microbiology 2005;3:510–6. [2] Kirk O, Borchert TV, Fuglsang CC. Industrial enzyme applications. Current Opinion in Biotechnology 2002;13:345–51. [3] Gijsen HJM, Qiao L, Fitz W, Wong CH. Recent advances in the chemoenzymatic synthesis of carbohydrates and carbohydrate mimetics. Chemical Reviews 1996;96:443–73. [4] Fessner WD, Sinerius G, Schneider A, Dreyer M, Schulz GE, Badia J, et al. Enzymes in organic-synthesis. 1. Diastereoselective enzymatic aldol additions—L-rhamnulose and L-fuculose 1-phosphate aldolases from Escherichia coli. Angewandte Chemie-International Edition in English 1991;30: 555–8. [5] Liu KKC, Kajimoto T, Chen LR, Zhong ZY, Ichikawa Y, Wong CH. Use of dihydroxyacetone phosphate dependent aldolases in the synthesis of deoxyazasugars. Journal of Organic Chemistry 1991;56:6280–9. [6] MorisVaras F, Qian XH, Wong CH. Enzymatic/chemical synthesis and biological evaluation of seven-membered iminocyclitols. Journal of the American Chemical Society 1996;118:7647–52. [7] Samland AK, Sprenger GA. Microbial aldolases as C–C bonding enzymesunknown treasures and new developments. Applied Microbiology and Biotechnology 2006;71:253–64. [8] Makrides SC. Strategies for achieving high-level expression of genes in Escherichia coli. Microbiological Reviews 1996;60:512. [9] Baneyx F. Recombinant protein expression in Escherichia coli. Current Opinion in Biotechnology 1999;10:411–21. [10] Shiloach J, Fass R. Growing E. coli to high cell density—a historical perspective on method development. Biotechnology Advances 2005;23:345–57. [11] Ramirez DM, Bentley WE. Characterization of stress and protein turnover from protein overexpression in fed-batch E. coli cultures. Journal of Biotechnology 1999;71:39–58. [12] Rozkov A, Schweder T, Veide A, Enfors SO. Dynamics of proteolysis and its influence on the accumulation of intracellular recombinant proteins. Enzyme and Microbial Technology 2000;27:743–8. [13] Jordan GL, Harcum SW. Characterization of up-regulated proteases in an industrial recombinant Escherichia coli fermentation. Journal of Industrial Microbiology & Biotechnology 2002;28:74–80. [14] Rozkov A, Enfors SO. Analysis and control of proteolysis of recombinant proteins in Escherichia coli. Advances in Biochemical Engineering and Biotechnology 2004:89:163–95. [15] Waugh DS. Making the most of affinity tags. Trends in Biotechnology 2005;23:316–20. [16] Bentley WE, Madurawe RD, Gill RT, Shiloach M, Chase TE, Pulliam-Holoman TR, et al. Generation of a histidine-tagged antibotulinum toxin antibody fragment in E. coli: effects of post-induction temperature on yield and IMAC binding-affinity. Journal of Industrial Microbiology & Biotechnology 1998; 21:275–82. [17] Vidal L, Durany O, Suau T, Ferrer P, Benaiges MD, Caminal G. High-level production of recombinant His-tagged rhamnulose 1-phosphate aldolase in Escherichia coli. Journal of Chemical Technology and Biotechnology 2003; 78:1171–9. [18] Vidal L, Ferrer P, Alvaro G, Benaiges MD, Caminal G. Influence of induction and operation mode on recombinant rhamnulose 1-phosphate aldolase production by Escherichia coli using the T5 promoter. Journal of Biotechnology 2005;118:75–87. [19] Vidal L, Pinsach J, Streidner G, Caminal G, Ferrer P. Development of an antibiotic-free plasmid selection system based on glycine auxotrophy for recombinant protein production in E. coli. Journal of Biotechnology 2008; 134:127–36. [20] Pinsach J, de Mas C, López-Santı́n J. Induction strategies in fed-batch cultures for recombinant protein production in Escherichia coli: application to rham- J. Ruiz et al. / Process Biochemistry 44 (2009) 1039–1045 [21] [22] [23] [24] [25] [26] [27] nulose 1-phosphate aldolase. Biochemical Engineering Journal 2008;41: 181–7. Ardao I, Benaiges MD, Caminal G, Álvaro G. One step purification–immobilization of fuculose-1-phosphate aldolase, a class II DHAP dependent aldolase, by using metal-chelate supports. Enzyme and Microbial Technology 2006;39:22–7. Datar RV, Cartwright T, Rosen CG. Process economics of animal-cell and bacterial fermentations—a case-study analysis of tissue plasminogen-activator. Bio-Technology 1993;11:349–57. Durany O, de Mas C, Lopez-Santin J. Fed-batch production of recombinant fuculose-1-phosphate aldolase in E. coli. Process Biochemistry 2005;40:707– 16. Pinsach J, de Mas C, Lopez-Santin J. A simple feedback control of Escherichia coli growth for recombinant aldolase production in fed-batch mode. Biochemical Engineering Journal 2006;29:235–42. Gottesman S, Maurizi MR. Cell biology: enhanced: surviving starvation. Science 2001;293:614–5. González-Montalbán N, Garcı́a-Fruitós E, Villaverde A. Recombinant protein solubility-does more mean better? Nature Biotechnology 2007;25:718–20. Gottesman S, Maurizi MR. Regulation by proteolysis—energy-dependent proteases and their targets. Microbiological Reviews 1992;56:592–621. 1045 [28] Andersson L, Strandberg L, Enfors S-O. Cell segregationand lysis have profound effects on the growth of Escherichia coli in high cell density fed-batch cultures. Biotechnology Progress 1996;12:190–5. [29] Hogema BM, Arents JC, Bader R, Eijkemans K, Inada T, Aiba H, et al. Inducer exclusion by glucose-6-phosphate in Escherichia coli. Molecular Biology 1998;28:755–65. [30] Ishizuka H, Hanamura A, Inada T, Aiba H. Mechanism of the down-regulation of cAMP receptor protein caused by glucose is an important determinant for catabolite repression in Escherichia coli: role of autoregulation of the crp gene. EMBO Journal 1994;13:3077–82. [31] Sanden AM, Bostrom M, Markland K, Larsson G. Solubility and proteolysis of the Zb-MaIE and Zb-MaIE31 proteins during overproduction in Escherichia coli. Biotechnology and Bioengineering 2005;90:239–47. [32] Jensen PR, Hammer K. Artificial promoters for metabolic optimization. Biotechnology and Bioengineering 1998;58:191–5. [33] Donovan RS, Robinson CW, Glick BR. Optimizing inducer and culture conditions for expression of foreign proteins under the control of the lac promoter. Journal of Industrial Microbiology 1996;16:145–54. [34] Jensen PR, Westerhoff HV, Michelsen O. The use of lac-type promoters in control analysis. European Journal of Biochemistry 1993;211:181–91.