(DE3): ammonia effects - Springer Link

Appl Microbiol Biotechnol (2009) 82:249–259 DOI 10.1007/s00253-008-1756-z

BIOTECHNOLOGICAL PRODUCTS AND PROCESS ENGINEERING

Impact of oxygen supply on rtPA expression in Escherichia coli BL21 (DE3): ammonia effects Hengwei Wang & Fengqing Wang & Dongzhi Wei

Received: 9 June 2008 / Revised: 15 October 2008 / Accepted: 18 October 2008 / Published online: 18 November 2008 # Springer-Verlag 2008

Abstract In shake flasks, good oxygen supply tended to decrease rtPA expression in media containing only yeast extract and tryptone, while oxygen limitation would increase rtPA synthesis in the same medium. Our data showed that though the drop of rtPA expression in the 20-ml cultures of LBG or 2YTG was accompanied with a severe acetate accumulation, it was actually caused by low ammonia. The rtPA expression level could be significantly improved by increasing culture ammonium ion up to 500 mM. The effects of exogenous high ammonia on cell growth and rtPA expression were further examined in shake flasks and a 4-l fermentor. The high ammonia had no significant impact on cell growth and oxygen respiratory activity but significantly depressed the activities of glutamine synthetase/ glutamate synthase and glutamate dehydrogenase, suggesting that ammonium ion as a nitrogen source improved the protein expression by mediating ammonia-assimilating enzymes. We thus propose in our work that E. coli cells, which were grown to a certain density to produce rtPA, would undergo nitrogen starvation under the low ammonia conditions even when the organic nitrogen sources remained abundant. The scale-up of rtPA production from shake flasks to fermentors could be readily achieved in the media containing rich ammonium ion. Keywords Escherichia coli . Ammonia . Scale-up . Fed-batch cultivation . Recombinant protein expression . rtPA H. Wang : F. Wang : D. Wei (*) Newworld Institute of Biotechnology, State Key Laboratory of Bioreactor Engineering, East China University of Science & Technology, P.O. Box 311, 130 Meilong Road, Shanghai 200237, People’s Republic of China e-mail: [email protected]

Introduction Scale-up processes of recombinant protein production from shake flasks to fermentors generally suffer from the same trouble of lowering the protein expression level. The problem could be caused by a multitude of factors, and presently, there is no common method to deal with them (Lee 1996; Makrides 1996). A great difference between shake flasks and fermentors is the supply and limitation of oxygen (Losen et al. 2004; Qoronfleh 1999; Ryan et al. 1989). However, oxygen supply has been reported to have variable effects on heterogeneous gene expression in recombinant Escherichia coli cells and would increase or decrease the expression of heterogeneous proteins, and up to now, there are no conclusive comments on the topic. Qoronfleh (1999) found that the difference in the effect of oxygen supply on the accumulation of p24Gag protein observed in OD650 5 and OD650 10 inductions was quite dramatic and difficult to explain, for the obvious differences between the conditions were the relatively minor difference of cell density and the time lag of 45 min for the latter. Ryan et al. (1989) found that in bioreactor experiments, though operations at high airflow rates were beneficial for cell mass formation, the specific β-lactamase activity would decrease markedly, and these operations were not necessarily the best for the intracellular protein accumulation. Losen et al. (2004) studied the effects of oxygen limitation and medium composition on E. coli fermentation in shake flask cultures, but they could not explain why the specific activity of benzoylformate decarboxylase (BFD) expressed in E. coli SG13009 cultured in TB medium at 400 rpm and filling volume 10 ml was much lower than that at 150 rpm and filling volume 25 ml, though the former conditions indicated a better oxygen supply. In addition, though ammonia sometimes is used as a base to adjust pH in

250

fermentors, little attention has been given to the physiological responses resulting from the difference of ammoniaassimilating conditions between shake flask and fermentor cultivations. It is well documented that recombinant protein expression by E. coli cells in shake flasks varies with shaker speed or culture volume. In this study, the recombinant E. coli BL21 (DE3)(pET) expressing human tissue plasminogen activator derivative K2P (rtPA) (Saito et al. 1994; Zhao et al. 2003) was taken as an example to clarify the correlation between oxygen supply and the recombinant protein expression. We found that the effects of oxygen supply on heterogeneous gene expression in E. coli BL21 (DE3) cells may be correlated with culture ammonia. The profiles of rtPA expression in E. coli were further investigated in lowand high-ammonia cultivations in shake flasks and a 4-l fermentor. The characterization of the cultivations would help to understand better the mechanism of the recombinant protein expression decrease in scale-up and gain some insight into factors influencing microbial behaviors in shake flask and fermentor cultivations, which would permit good duplication of results between them.

Materials and methods Bacterial strains and plasmids E. coli BL21 (DE3) was used in all experiments. The strain was transformed with the plasmid pET-22b (+) carrying the human tissue plasminogen activator (tPA) gene derivative K2P described previously (Saito et al. 1994; Zhao et al. 2003). Flask cultivations Luria–Bertani (LB) liquid medium containing 5 g/l yeast extract (Oxoid, UK), 10 g/l tryptone (Oxoid, UK), 10 g/l NaC1, and 100 mg/l ampicillin (Amp) was used for the starter cultures and LB-Amp agar plates were used to propagate colonies of transformed competent cells. For E. coli cultivation and recombinant protein expression, buffered LB-glycerol medium (LBG), 2× YT-glycerol medium (2YTG), and high-ammonia YT-glycerol medium (YTGA) thereof were prepared as shown in Table 1. Cultures were grown at 37°C, 200 rpm on a rotary shaker with an orbital diameter of 30 mm (Hualida Laboratory Equipment, China) in all experiments. The medium pH in shake flasks was controlled at 7.1±0.3 by adding 20– 100 μl solution of 25% NaOH every 2 h in situ. One hundred milligrams per liter of ampicillin was added at inoculation, and 1 mM IPTG was added at induction. It should be noted that the ammonia contained in the organic

Appl Microbiol Biotechnol (2009) 82:249–259 Table 1 Media for recombinant protein expression Components

Yeast extract (g/l) Tryptone (g/l) (NH4)2SO4 (mM) Glycerol (g/l) Na2HPO4·12H2O (g/l) KH2PO4 (g/l)

Media LBG

2YTG

YTGA

5 10 0 10 9 1

10 16 0 10 9 1

2.5 5 50 10 9 1

In the buffered media, Na2HPO4·12H2O and KH2PO4 were used to give a concentration of 32 mM phosphate and the pH was adjusted to about 7.2. The stock of glycerol (500 g/l) was autoclaved separately LBG buffered LB-glycerol medium, 2YTG buffered 2× YT-glycerol medium, YTGA buffered YT-glycerol medium with 100 mM ammonium ion.

nitrogen sources was about 0.47±0.02 mM for 10 g/l yeast extract and 1.38±0.05 mM for 10 g/l tryptone in their aqueous solutions. Fed-batch cultivations Fifty milliliters liquid LB-Amp was used for 8–10 h starter culture as described above. The fed-batch cultivations were carried out in a 4-l bioreactor Biostat B (B. Braun Biotech, Germany) with an initial culture volume of 3.7 l. The highammonia medium used for fed-batch cultivation contained 10 g/l yeast extract, 10 g/l tryptone, 33 g/l (NH4)2SO4 (500 mM ammonium ion), and the low-ammonia medium contained only 20 g/l yeast extract and 20 g/l tryptone as nitrogen sources. In addition, both media contained 4.5 g/l Na2HPO4·12H2O and 0.5 g/l KH2PO4. Culture glycerol was monitored offline and controlled at 10±5 g/l by using a glycerol stock of 500 g/l. The temperature was maintained at 37°C, and the pH was controlled at 7.1±0.1 using a 25% NaOH solution. The dissolved oxygen (DO) was monitored using a polarographic oxygen electrode and the inlet airflow rate was kept at 2.0 l/min per liter broth. In both cases 100 mg/l ampicillin was added at inoculation, and 1 mM IPTG was added to induce rtPA expression at about OD600 11. Determination of cultivation parameters Cell growth was monitored by measurement of the optical density (OD) at 600 nm. Samples were diluted with deionized water to keep OD readings in the linear range between 0.005 and 0.450, and 1.0 OD unit at 600 nm corresponded to 0.385 g dry cell weight (DCW) per liter culture of E. coli cells. In addition, DCWs were determined from 10-ml culture collected by centrifugation. Cells were collected by centrifugation for 5 min at 4,000×g,

Appl Microbiol Biotechnol (2009) 82:249–259

resuspended in deionized water, centrifuged again, and dried at 115°C to weight constancy in about 24 h. The specific growth rate μ was calculated according to the formula X2 =X1·eμΔt. Glycerol concentrations were determined based on the enzymatic method of Fossati and Prencipe (1982). Ammonium ion concentrations were determined based on the phenol-hypochlorite reaction reported by Weatherburn (1967). Acetate was determined by high-performance liquid chromatography. The initial kLa values were estimated with a polarographic DO probe (63.2% response time=12 s) by the dynamic gassing-in and gassing-out method (Boogerd et al. 1990; Tribe et al. 1995). The specific oxygen uptake rate (SOUR) was determined from a linear decrease of DO in cell suspension in a closed flask attached to a polarographic oxygen probe. Five to 20 ml of culture was added to a 120-ml flask, which contained about 100 ml air-saturated medium (5 g/l yeast extract, 5 g/l glucose, 50 mM KH2PO4, pH 7.0); then, the residual space in the flask was quickly filled with the air-saturated medium and closed with a rubber plug and the DO probe. The culture was then kept in suspension by magnetic stirring at about 3,000 rpm, and the DO decrease was recorded. SOUR was calculated from the DO slope and DCW. All the experiments were carried out at 25°C. The oxygen concentration in the air-saturated medium was about 0.20 mM, and SOUR was expressed in mmol O2/(min·g DCW). Protein concentrations of cell-free extracts were determined by the Lowry method (Lowry et al. 1951). Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was carried out as described by Laemmli (1970) using a separating gel of 12.5% acrylamide and approximately 15 μg cellular protein for each cell sample. rtPA was expressed in E. coli as inclusion body, the expression levels were estimated by Coomassie blue staining SDS-PAGE analysis (Zhao et al. 2003), and the yields were determined from rtPA expression levels and total cell protein per liter broth. rtPA fibrinolytic activity was estimated by the fibrin plate method, and its amidolytic activity was evaluated spectrophotometrically by the method of Verheijen after renaturation and purification (Margareta et al. 1991; Verheijen et al. 1982; Zhao et al. 2003). Enzyme assays Cells were harvested by centrifugation at 10,000×g for 2 min, re-suspended in cold potassium phosphate buffer (20 mM, pH 7.5), centrifuged again, and frozen at −20°C for sonication and enzyme assays in 48 h. The thawed cell pellets were sonicated on ice in 5 ml cold potassium phosphate buffer (20 mM, pH 7.5). Five hundred microliters sample from each of the cell extracts was applied to protein determination, and the left was centrifuged at 4°C,

251

10,000× g for 10 min to clear the extract for enzyme assays. The glutamine synthetase activity was measured by the γglutamyltransferase assay (Woolfolk et al. 1966) based on the formation of glutamyl hydroxamate. Nicotinamide adenine dinucleotide phosphate (NADP)-dependent glutamate synthase activity was assayed based on the method of Miller and Stadtman (1972) with some modifications. Standard assay mixtures contained 100 mM Tris–HCl (pH 8.0), 0.5 mM ethylenediaminetetraacetic acid (EDTA), 5 mM L-glutamine, 1 mM α-ketoglutarate, and 0.15 mM nicotinamide adenine dinucleotide phosphate (reduced form) (NADPH). NADP-dependent glutamate dehydrogenase activity was determined based on the method of Maurizi and Rasulova (2002). The assay mixture contained 100 mM Tris–HCl (pH 9.0), 0.5 mM EDTA, 50 mM ammonium chloride, 1 mM α-ketoglutarate, and 0.15 mM NADPH. The millimolar extinction coefficient of NADPH

b

Fig. 1 Expression of rtPA in different media under different oxygen supply levels. a Comparison of kLa values in a 250-ml shake flask for different culture volumes at 200 rpm on a rotary shaker with an orbital diameter of 30 mm. b SDS-polyacrylamide gel electrophoresis analysis of rtPA expression in 250-ml shake-flask cultures under different conditions of oxygen supply and media. Lane 0 E. coli BL21 (DE3) as a control, absent of the recombinant K2P plasmid; lanes 1, 2 and 3 rtPA expression in the 20-ml cultures of LBG, 2YTG, and YTGA, respectively. Lanes 4, 5 and 6 rtPA expression in the 100-ml cultures of LBG, 2YTG, and YTGA, respectively. The expression of rtPA was carried out as described in “Materials and methods,” and the protein synthesis continued for 8 h. The arrow shows the rtPA protein band of 39 kDa

252


used was 6.22 at 340 nm in all determinations (Miller and Stadtman 1972). Enzyme activities were averaged over three determinations with the corresponding standard deviation and presented in micromoles per minute per milligram protein.

the 20-ml cultures, the increase of yeast extract and tryptone in the 2YTG medium could not improve the protein expression level.

Results

In this experiment, cell responses to the different oxygen supply levels were examined on cell growth, glycerol consumption, acetate, and ammonium ion production of E. coli BL21 (DE3) in shake flask cultures (Fig. 2). Figure 2a shows that there existed a significant difference in the growth behaviors of the 20 and 100-ml cultures, since in the latter cultivation mode, the cells grew into stationary phase in 4 h due to oxygen limitation. During recombinant protein synthesis, the rates of glycerol consumption (mM h−1) and acetate production (mM h−1) were 2.8±0.2 and 1.2±0.1 for the 100-ml cultures, and for the 20-ml cultures, 12.0±0.5 and 5.2±1.2, respectively (Fig. 2b,c). The final concentration of acetate in the 20-ml cultures of LBG and 2YTG reached 38±4 and 52±4 mM, respectively, and in the 100-ml cultures, reached about 19 mM (Fig. 2b). A marked and sustaining increase of extracellular ammonium ion in the 100-ml cultures of LBG and 2YTG was detected over the whole cultivation, even in the stationary phase (Fig. 2d). Although the profiles of ammonium accumulation

Expression of rtPA under different oxygen supply levels in different media The volumetric oxygen transfer coefficient kLa, as an index of the rate of oxygen diffusion through the gas–liquid surface, indicates the system capacity of oxygen supply and may be affected by the agitation and aeration of the system. In shake flasks, different oxygen supply levels could be achieved by changing culture volume or shaker speed, and the values of the initial kLa of 20- and 100-ml cultures in a 250-ml flask were estimated at 110±15 h−1 and 28±5 h−1 at 200 rpm, 37°C (Fig. 1a). Figure 1b shows the effect of different oxygen supply levels and different media on rtPA expression. In comparison with the 20-ml cultures, the rtPA expression level was improved approximately twice when cells were cultivated under the lower oxygen supply conditions in the 100-ml cultures of LBG or 2YTG. For

Fig. 2 Effects of oxygen supply on E. coli fermentations in shake flask cultures. Profiles of cell density OD600 (a), glycerol consumption (b), acetate production (c), and ammonium ion accumulation (d) of E. coli BL21 (DE3)/pET22b(+)-rtPA in 250-ml shake flask cultures are shown. The open symbol represents the 20-ml cultures, and the filled symbol represents the 100-ml cultures. The upright line marks the induction start

Effects of oxygen supply on E. coli fermentations in shake flask cultures


in the 20-ml cultures before the late exponential phase were similar to that in the low-oxygen supply cultures, a decrease of ammonium followed thereafter, due to the higher demand for cell growth from 5 h into fermentation. As shown in Fig. 2, the low levels of growth rate (Fig. 2a), glycerol consumption (Fig. 2b), acetate production (Fig. 2c), and high levels of ammonium ion accumulation (Fig. 2d) were accompanied with the enhanced recombinant protein expression in the 100-ml cultures of LBG and 2YTG (Fig. 1b). Figure 3a shows that the acetate accumulation (48± 4 mM) in the 20-mL YTGA cultures, which was presumably believed to decrease the protein expression, was similar

Fig. 3 The high-ammonia cultivations of E. coli in YTGA medium. Cell growth (filled square), glycerol consumption (filled circle), acetate excretion (filled triangle), ammonia consumption (filled inverted triangle), and SOUR (open square) of E. coli BL21 (DE3)/pET22b(+)-rtPA grown in the 20-mL (a) and 100-mL (b) YTGA cultures at 37°C are shown. In each case, the initial concentration of ammonium ion was about 100 mM. The upright line marks the induction start

253

to the situations shown in the 20-ml LBG or 2YTG cultures (Fig. 2c). Acetate in the 100-mL YTGA cultures was about 16 mM at the end of cultivation (Fig. 3b). The residual ammonium ion was about 87 mM in the 20-mL cultures and 91 mM in the 100-mL cultures (Fig. 3). The relationship between rtPA expression and accumulation of acetate or ammonia in the shake flask cultures is summarized in Table 2. The accumulation of up to 48 mM acetate in YTGA cultures was not accompanied with a marked drop of rtPA expression, suggesting that the accumulation of high acetate in the 20-ml LBG or 2YTG cultures might not be responsible for the drop of rtPA expression.

254


Table 2 Effects of endogenous acetate and ammonia accumulation in shake flask cultures on rtPA expression Cultures

LBG, 20 mL 2YTG, 20 mL YTGA, 20 mL LBG, 100 mL 2YTG, 100 mL YTGA, 100 mL

a

Cultivation results Acetate (mM)

NH4+ (mM)

rtPA expression

38±4 52±4 48±4 18±3 20±3 16±2

1.1±0.1 3.1±0.2 87±4 9.0±0.4 13.5±0.3 91±4

− − + + + +

b

The final concentrations of culture acetate and ammonia were averaged over two measuring results. The rtPA low (−) and high (+) expression levels are also shown by SDS-PAGE in Fig. 1b.

Effects of exogenous ammonium ion and acetate on E. coli cultivations and rtPA expression in shake flask cultures To further clarify the role of acetate and ammonia in regulating rtPA expression in shake flask cultivations, cell responses to different concentrations of exogenous acetate and ammonium ion were examined on cell growth and rtPA production. Supplementation of the 20-ml LBG cultures with different concentrations of ammonium ion in the form of (NH4)2SO4 apparently increased the rtPA expression level under the conditions of 0 or 200 mM exogenous acetate (Fig. 4a), and surprisingly, even up to 200 mM acetate sodium had no obvious inhibitive effects on rtPA expression in the 100-ml LBG cultures or in the 20-ml high-ammonia LBG cultures (Fig. 4b). The addition of 50 or 250 mM sodium sulfate into cultures, which can lead to a similar osmotic stress on E. coli cells caused by the addition of ammonium sulfate of the same concentration, did not increase rtPA expression (Fig. 4c), suggesting that the improvement of rtPA expression by addition of ammonium sulfate could not be due to osmotic pressure or sulfate anion. As shown in Fig. 5, only up to 500 mM ammonium ion in the 20-ml cultures could cause an obvious inhibition on the cell specific growth rates of the first 3 h (Fig. 5a), and 50–200 mM acetate in the 20ml LBG cultures containing 500 mM ammonium ion would seriously inhibit the cell specific growth rates and the final OD600 (Fig. 5b). Our experiments showed that, for the 20-ml cultures of LBG or 2YTG (Fig. 2), the rtPA expression drop was not due to the high concentrations of acetate excreted but the ammonia depletion in broth. For the 100-ml cultures (Fig. 2d), the endogenous ammonia from cell catabolism of organic nitrogen sources (McFall and Newman 1996) continuously accumulated in broth, especially when μ was near 0 h−1 due to the limitation of poor oxygen supply. While the level of ammonia in the 20-ml cultures could not be maintained during recombinant protein synthesis, for E. coli cells could reuse the secreted ammonia when the good oxygen supply permitted a high cell density.

c

Fig. 4 Effects of exogenous ammonium and acetate on rtPA expression by E. coli BL21 (DE3). The cells were cultivated for 3 h and induced for 8 h as described in “Materials and methods.” a Effects of exogenous ammonium ion on rtPA expression in E. coli BL21 (DE3). The 20-ml LBG cultures containing 0 or 200 mM acetate sodium were supplied with (NH4)2SO4 to give a final ammonia concentration of 0, 100, and 500 mM, respectively. b Effects of exogenous acetate sodium on rtPA expression in E. coli BL21 (DE3). The 100-ml LBG cultures or the 20-ml LBG cultures containing 500 mM ammonia were supplied with acetate sodium to give a final concentration of 50, 100, and 200 mM, respectively. c Effects of Na2SO4 on rtPA expression by E. coli BL21 (DE3). The 20-ml LBG cultures were supplied with 50 or 250 mM Na2SO4 to produce an osmotic stress equivalent to the one caused by (NH4)2SO4 of the same concentration. M Protein marker, C0 rtPA expression in LBG cultures without sulfates. The arrow indicates the rtPA protein band

Comparison of low- and high-ammonia fed-batch cultivations in a 4-l fermentor The profiles of recombinant protein expression were further compared in fed-batch cultivations, in the low-ammonia culture containing abundant yeast extract and tryptone as organic nitrogen sources and in the high-ammonia culture with less organic nitrogen sources (see “Materials and methods”). Under our fed-batch conditions, glycerol con-


255

Fig. 5 Effects of exogenous ammonium ion and acetate on cell growth of E. coli BL21 (DE3). The cells were cultivated for 3 h and induced for 8 h as described in “Materials and methods.” a The 20-ml LBG cultures in 250-ml shake flasks were supplied with (NH4)2SO4 to give a final ammonia concentration of 0, 50, 100, and 500 mM,

respectively. b The 20-ml LBG cultures containing 250 mM (NH4)2SO4 were supplied with acetate sodium of 50, 100, and 200 mM, respectively. For each culture, cell growth at OD600 (open square, open circle) and the specific growth rates (filled square, filled circle) during the first 3 h of cultivation are shown

centration was controlled at 10±5 g/l, and the accumulation of acetate continued over the recombinant protein synthesis at a nearly constant rate of 24 mM h−1 and reached about 240 mM at the end of cultivation (Fig. 6a,b), which was five to six times as much as that of the 20-ml cultures in shake flasks. While our data showed that, compared with shake flask cultivations, the high acetate stress in the fedbatch cultures had no significant destructive impacts on the profiles of rtPA expression (Fig. 1b vs. Fig. 7b). The SOUR values were a reflection of the cell oxygen utilization capacity and thus of the cell metabolism intensity. The cell capacity of metabolism and ability of survival may be impaired by carbon or nitrogen starvation (Atkinson et al. 2002; Bockman et al. 1986; Groat et al. 1986), cumulative oxidative damage (Imlay 2003), or acetate stress (Axe and Bailey 1995; Kleman and Strohl 1994). The drop of cell SOUR (Figs. 3 and 6) after induction is suggested to result from the metabolic load of recombinant protein synthesis to cellular capacities (Neubauer et al. 2003). Over the fedbatch cultivations, similar profiles of cell growth, acetate production, and SOUR (drop form 9.1 ± 0.2 to 4.3 ± 0.3 mmol O2/(min·g DCW)) were observed in Fig. 6a,b during rtPA expression, while in the low-ammonia cultivation, the culture ammonia reached the peak of about 10 mM at induction start and dropped to below 2.0 mM in 4 h thereafter (Fig. 6a). In the high-ammonia cultivation, the culture ammonia also decreased significantly (from 472 to 360 mM), not mainly due to cell assimilation but the dilution by addition of glycerol solution and sampling (Fig. 6b). The final OD600 reached 30.2 in the low-ammonia cultivation and 27.5 in the high-ammonia cultivation (Fig. 6), while the rtPA expression level of the latter was about three times as much as that of the former (Fig. 7).

Nitrogen-assimilating enzymatic activities in low- and highammonia cultures Ammonia, as the preferred source of nitrogen for bacterial growth, could be mainly assimilated into glutamate and glutamine to serve as the key nitrogen donors for biosynthetic reactions by glutamine synthetase/glutamate synthase (GS/ GOGAT) or glutamate dehydrogenase (GDH) pathway (Leigh and Dodsworth 2007; Merrick and Edwards 1995; Reitzer 2003). In E. coli cells, GS assimilates ammonia by converting glutamate to glutamine and GOGAT, then transfers the amido group of glutamine to α-ketoglutarate, making two glutamate. The GS/GOGAT pathway is adenosine triphosphate consuming and is used by E. coli in energy-rich environments. GDH makes glutamate from αketoglutarate and ammonia and is employed in energylimited environments (Leigh and Dodsworth 2007; Reitzer 2003). Their enzymatic activities in the fed-batch cultivations were shown in Fig. 8. For the low-ammonia cultivation (Fig. 6a), the activities of GS/GOGAT and GDH increased constantly when the cell density was above OD600 11 and the ammonia concentration dropped continuously from 9.1 to near 0 mM. While under the high-ammonia cultivation conditions (Fig. 6b), their activities stayed constant at a low level after induction at the same cell density, for the culture ammonia was kept above 360 mM.

Discussion To achieve a high yield of recombinant protein, it is a common practice to focus on improving the high cell density cultivation techniques; as a result, optimization of

256


a

b

Fig. 7 Time-course profiles of rtPA expression in fed-batch cultivations. SDS-PAGE analysis of cellular proteins on induction of 0, 1, 2, 4, 6, and 8 h from the low-ammonia (a) or the high-ammonia cultivation (b) are shown. The arrow indicates the rtPA protein band

Fig. 6 Time-course profiles of fed-batch cultivations in a 4-l Biostat B bioreactor. Cell growth (filled square), acetate excretion (filled triangle), ammonia (filled inverted triangle), and SOUR (open square) of E. coli BL21 (DE3)/pET22b(+)-rtPA grown in the low-ammonia cultivation (a) or the high-ammonia cultivation (b) are shown. Before induction, the agitation speed (em dash) was increased by 200 rpm each time to keep a DO level above 30%; after induction, the agitation speed stayed at 1,000 rpm and both culture DO levels remained at about 15%

media composition, batch and fed-batch fermentation techniques have been developed (Lee 1996), while the techniques too often result in a drop of recombinant protein expression in scale-up from shake flasks to fermentors. The drop of heterogeneous protein expression could be caused by increases in fermentors of plasmid instability (Grabherr et al. 2002), acetate accumulation (Jensen and Carlsen 1990), or probably of oxidative damage (Lu et al. 2003), etc. Our work indicated that, in the scale-up from shake

flasks to fermentors, growing recombinant E. coli cells in complex media to a high cell density, the concentration of culture ammonia could be correlated with recombinant protein expression. In support of this explanation, the experiments were conducted in LBG, 2YTG, and YTGA medium in 250-ml shake flasks. The former two media contained high concentrations of yeast extract and tryptone and no exogenous ammonia, while the YTGA medium contained low concentrations of yeast extract and tryptone and a high concentration of ammonia. For the LBG or 2YTG medium containing only yeast extract and tryptone, good oxygen supply conditions in the 20-ml cultures allowed a high cell density above OD600 4.0, which would deplete the secreted ammonium ion in culture due to bacterial assimilation, and at the same time, resulted in a high acetate concentration of 38–52 mM (Fig. 2). Under the poor oxygen supply conditions in the 100-ml cultures, the growth of E. coli cells was limited, the culture ammonium could be accumulated, and the final concentrations of acetate were less than 20 mM (Fig. 2). The rtPA expression level in the 100-ml


257

Fig. 8 Profiles of nitrogen-assimilating enzymatic activities in fed-batch cultivations. The activities of GS (a), GOGAT (b), and GDH (c) are shown. The open symbols represent the low-ammonia cultivation, and the filled symbols represent the high-ammonia cultivation

cultures of LBG or 2YTG was about three times of that in the 20-ml cultures (Fig. 1b). For the YTGA medium, rtPA expression was markedly improved in the 20-ml cultures compared with that in LBG or 2YTG medium of the same culture volume (Fig. 1b), though the oxygen supply conditions similarly led to a high cell density and acetate concentration (Fig. 3). The data showed that, though the improvement of rtPA expression seemed to be a direct result of the limitation of oxygen supply and decreases of cell density by increasing the culture volume in shake flasks (Fig. 1), the constant changes in medium were ignored to some extent. With the increase of oxygen supply, cell density would increase and some substance in culture, which is helpful to enhance the heterogeneous protein expression, would be depleted. Perhaps, this could provide a reasonable explanation for the drop of BFD specific activity in TB medium at 400 rpm and filling volume of 10 ml (Losen et al. 2004). Based on our data, we suggest that this essential substance is ammonium ion. It should be pointed out that though the high ammonium supply (100 mM or 500 mM) was shown to have positive effects on improving rtPA expression in our study, the benefits of a supplement of exogenous1 mM) is available (Merrick and Edwards 1995; Reitzer 2003). The ammonia assimilating enzymes have long been studied in bacteria nitrogen regulation (Leigh and Dodsworth 2007; Reitzer 2003), while up until now, there are few documents on their

258

relation to recombinant protein expression in E. coli cells. In our experiments, the enzymes of ammonia assimilation were assayed in crude extracts from cells grown at high (>360 mM) or low (