Biotechnol. Prog. 1995, 1I, 475-470
475
Metabolic Engineering of Escherichia coli To Enhance Recombinant Protein Production through Acetate Reduction Aristos k Aristidou3 Ka-Yiu Sari,*'+ and George N. Bennett* Departments of Chemical Engineering and Biochemistry and Cell Biology, Institute of Biosciences and Bioengineering, Rice University, P.O. Box 1892, Houston, Texas 77251-1892
Genetic and metabolic engineering provide powerful and effective tools for the systematic manipulation and fine tuning of cellular metabolic activities. In this study, successful application of such techniques to enhance recombinant protein production by reducing acetate accumulation in Escherichia coli is presented. The alsS gene from Bacillus subtilis encoding the enzyme acetolactate synthase was introduced into E. coli cells using a multicopy plasmid. This newly introduced heterologous enzyme modifies the glycolytic fluxes by redirecting excess pyruvate away from acetate to acetolactate. Acetolactate is then converted to a nonacidic and less harmful byproduct acetoin, which appears in the broth. Furthermore, comparative fermentation studies show that the reduction in acetate accumulation leads to a significant improvement of recombinant protein production. The expression of a model recombinant CadAlPgalactosidase fusion protein, under the control of a strong pH-regulated promoter, was found to increase by about 60% for the specific protein activity (to a level of 30% of total cellular protein) and 50% in terms of the volumetric activity in a batch fermenter. In fed-batch cultivation, the engineered strain achieved a volumetric recombinant protein yield of 1.6 million units/mL (about 1.1 g/L of P-galactosidase), which represented a 220% enhancement over the control strain. In the meantime, acetate excretion was maintained below 20 mM compared with 80 mM for the control, and the final cell density was improved by 35%.
Introduction One of the long-sought goals in recombinant protein production processes is to achieve a high cloned gene expression level and high cell density. Unfortunately, under these demanding conditions, the amount of acetate accumulated in the reactor increases precipitously. Escherichia coli belongs to the group of microbes known as mixed acid producers that are capable of fermenting pyruvate to a number of potentially harmful acidic byproducts. During aerobic or anaerobic growth in the presence of excess carbon source, as in the case of high cell density cultures, certain physiological events take place that lead to the accumulation of pyruvic acid that needs t o be dissipated (Doelle et al., 1981; Meyer et al., 1984). Furthermore, oxygen transfer limitations that can arise very rapidly in dense cultures will initiate a second cascade of physiological events leading again to pyruvate accumulation (Smith and Neidhardt, 1983). For both cases the cell will attempt to balance carbon fluxes that result mainly with acetate as the major byproduct (Holms, 1986). Acetate, is a lipophilic agent that is harmful to cell growth (Adams, 1988; Luli and Strohl, 1990). Moreover, experimental results in our laboratory (Chou et al., 1995) agree well with common observation that recombinant gene expression is greatly reduced for acetate accumulation above 15-25 mM (Bauer et al., 1990; Jensen and Carlsen, 1990; Shimuzu et al., 1988). The goal of this work was to improve recombinant protein production under high cell density and high expression level. The approach adopted here was to genetically engineer E. coli cells in an effort to systematically alter their cellular metabolic activities. This +
Department of Chemical Engineering. Department of Biochemistry and Cell Biology.
technique is believed to be more robust and easier to implement than the traditional methods that attempt to provide an ideal growth environment in order to minimize acetogenesis. Metabolic engineering was achieved by cloning the gene for the Bacillus subtilis acetolactate synthase (ALS)in E. coli, an enzyme capable of catalyzing the conversion of pyruvate to nonacidic and less harmful species. The choice of this gene was based on the low acetate production pattern of organisms that naturally contain this enzyme (Holtzclaw and Chapman, 1975). The ALS gene was successfully expressed in E. coli using a multicopy plasmid, resulting in a strain that retains desired growth characteristics and yet maintains low acetate accumulation (Aristidou et al., 1994). The byproduct, acetoin, was shown to be 50 times (in mole basis) less toxic to cells than acetate (Aristidou et al., 1994). This strain was further found to have superior properties in terms of recombinant protein production.
Experimental Protocol Bacterial Strains and Plasmids. E. coli GJTOOl was used throughout this work (Tolentino et al., 1992). Plasmid p M 1 5 bearing the alsS gene from B. subtilis was constructed in our laboratory (Aritidou, 1994) using pACYC184 (Chang and Cohen, 1978) as the cloning vector. Plasmid pSM552-545C- is a cadAllacZ fusion vector used for recombinant protein studies (Meng and Bennett, 1992; Chou et al., 1995). For recombinant protein studies, a two-plasmid system was employed (Figure 1). Cells were transformed with either plasmid pAAA215 (ALS)or plasmid pACYC184 (control) in conjunction with the recombinant protein vector pSM552545c-.
8756-7938/95/3011-0475$09.00/0 0 1995 American Chemical Society and American Institute of Chemical Engineers
Biotechnol. Prog., 1995,Vol. 11, No. 4
476 ~
1.OeC6
,
30000 25000 6
E odiGJTOOl
E mli GrroOl
ai
o.oe+o
0
5 30
A: pSM552-54SC
IO
I5
20
25
30
35
~""""""'""""""""""""
40
i
40
r
0
t o
120
C:pACYC184
&Galactosidase
&Galactosidase
Figure 1. Two plasmid systems: both hosts carry plasmid pSM552-545C- for P-galactosidase production; the second plasmid is either pAAA215 (ALS)or pACYC184 (control).
Cultivation Media. For batch cultivation, experiments were performed in a 2.5-L BioFlo 111 benchtop fermentor (New Brunswick Scientific) with a working volume of 1.5 L. The pH was measured by a glass electrode (Phoenix) and adjusted to the desired setpoint by acid or alkaline additions. For experiments involving the pH-regulated promoter, the pH was initially maintained a t a value of 7.5 (to minimize the baseline expression of this promoter), and at the point of induction it was lowered to a value of 6.0 (ODs00 GZ 2). Dissolved oxygen (DO) was monitored using a polarographic oxygen electrode (Phoenix) and was sustained above 50%saturation by 0 2 enrichment. The agitation speed was maintained at 500 rpm. Fed-batch cultivations were conducted in the same fermentor. The starting conditions were similar to those of the batch process, with the exception that 1.25 L of SB was used as the start up medium and the agitation speed was maintained at 750 rpm. Following protein induction as in the batch case, the fed-batch mode was initiated at an ODs00 = 15. The feed media (750 ml) contained (per liter) 230 g of yeast extract, 50 g of glucose, 5 g of (NH4)2S04, and 1 g of MgC12. The feed solution was delivered by a controllable peristaltic pump operating at a predetermined exponential feed rate, F, given by F = Foekt,where F, is the initial feed rate, t is the time elapsed since starting the feed, and k is a time constant with a value equal to the desired specific growth rate, p (h-l). Assays. Cell growth was followed by measuring offline the optical density a t 600 nm using a spectrophotometer (Spectronic 1001, Bausch & Lomb). Collected samples were first chilled in an ice bath and then centrifuged at SOOOg and 4 "C for 10 min (Sorvall SS-34 rotor). The supernatant was then separated from the pellet and stored chilled for further analyses. The pellet was resuspended in cold phosphate buffer (10 mM P04'-, 10 mM Mg2+,1 mM EDTA, pH = 7.5) and stored on ice. Cell extracts used for protein assays were obtained by a 6 min sonication of the above cell suspension with an ultrasonic cell disrupter (Heat Systems Ultrasonics). /?-Galactosidase assays were performed at 28 "C in triplicate using o-nitrophenol-/?-D-galactopyranoside (ONPG) as the substrate (Miller, 1972). Specific activity is expressed in Miller Units (MU): 1MU = [1000&&/ OD~oo*mL.minl.Volumetric activity is obtained from the product of Miller Units and the culture optical density, and is reported in units/mL. The residual glucose present in the fermentation broth was determined enzymatically (Hexokinase kit, Sigma). Extracellular metabolites, such as acetate, ethanol, and acetoin, were analyzed with a
5
IO
I5
20
25
30
35
40
Cell Easily (OD6m)
Figure 2. Typical fermentation profiles for recombinant protein production (A); acetate and acetoin excretion and glucose consumption (B).Comparison of results from two experiments: one where the ALS is expressed (filled symbols) and a control run (open symbols). Solid diamonds represent acetoin for the control strain. Cells were grown in a batch culture at 27 "C using SB media supplemented with 2% glucose.
Varian 3000 gas chromatograph equipped with a 6 ft glass column (1.4 in. thick x 2 mm i.d.1 packed with 100/ 120 mesh Chromosorb 101 coated with 0.5% free fatty acid phthalate, FFAP (AllTech), and a flame ionization detector (FID). Nitrogen was used as the carrier gas a t a flowrate of 30 mL/min. The sample was acidified (20 pL of 50% HZSOdmL sample) before injection to assure complete protonation of the organic acids.
Results Acetate Reduction and Its Effect on Heterologous Gene Expression. The ability of the ALS-containing system to overexpress a model recombinant protein was evaluated. The ZacZ gene encoding the protein @-galactosidase was cloned on a ColEl-based plasmid compatible with pAAA215. Its expression is regulated by a powerful pH-inducible promoter that has been developed recently (Aristidou et al., 199; Chou et al., 1995; Tolentino et al., 1992). Experiments were performed in a well-controlled fermenter under optimal conditions for the particular expression system. The expression of @-galactosidasewas initiated by changing the reactor pH from a value of 7.5, where the induction level is minimal, to a value of 6.0 that has been determined to be the optimum induction pH (Aristidou et al., 1991; Chou et al., 1995). Both plasmids were stably maintained throughout the experiments as monitored by standard plating technique (Maniatis et al., 1982). Batch Cultures. The presence of ALS can prevent acetate accumulation even at a high cell density as shown in Figure 2 for a typical fermentation run. The acetate concentration in the control system increases monotonically and reaches a maximum level of about 35 mM. The highest acetate level for the ALS system, however, remains less than 10 mM throughout the experiment. While keeping acetate accumulation in check, the ALS strain produces a less harmful product, acetoin, instead. The acetoin concentration increases at about the same rate as acetate in the control system. Excreted acetoin reached a final concentration of 36 mM, which is well
Biotechnol. Frog., 1995, Vol. 11, No. 4
477
Table 1. Summary of Results from Batch Fermentation Experiments: Effect of ALS on Growth, Acetate Accumulation, and Recombinant Protein @-Galactosidase)Production in E. coli GJTO01” glucose strain ALS control ALS control
(g/L)
OD600
10 10
25.1 27.2 39.6 37.1
20 20
0
acetate acetoin P-Gal P-Gal (mM) (mM) (MU) (units/mL) 10.2 22.7 23854 508900 8.4 nd 12294 357280 17.4 30.7 25 162 805650 15466 573973 35.6 nd
Recombinant protein yields are reported as either volumetric (units/mL) or specific (MU) activities. All values represent maximum attained quantities. Batch cultivation a t 27 “C in SB 10 or 20 gL.glucose. (I
+
below its inhibitory level (San et al., 1994). The final optical density for the ALS system is about 15% higher. More importantly, the ALS system has higher expression levels of the recombinant protein @-galactosidasewith a specific activity near 25 000 MU, which was about 60% higher than the control. The volumetric productivity was about 50% higher. The effectiveness of the ALS system at higher cell densities was evaluated using superbroth supplemented 10 or 20 g/L of glucose. Whereas comparable biomass yields were obtained at each glucose concentration, acetate levels were clearly reduced by the presence of the ALS,especially at the higher glucose concentration (Table 1). In the meantime, significant enhancements in P-galactosidase production, in both specific and volumetric activities, were observed for the two media. Fed-Batch Cultivation. A series of fed-batch experiments were conducted to investigate the performance of the ALS system in a high-cell density environment. The strategy adopted in this study was to operate the fedbatch at a sufficiently high growth rate, with the ALS system in the meantime acting as a negative feedback mechanism to keep acetate levels below their critical value. Figure 3 compares the performance of the ALS strain with that of the control strain using a feeding profile F ( m m ) = 6e0,25t.In both cases, the feed was started at an OD of 10, which coincided with a sharp increase in the rate of glucose consumption. By using the exponential feed profile, it was possible to maintain the growth rate of the ALS system between 0.15-0.2 hr-l, which eventually reached a final OD of 86 (CDW about 30 g/L). The control strain followed a similar growth profile up to an OD of 50, after which point the growth rate declined substantially reaching a final OD of only 63 (panel A). Acetate excretion, which was maintained below 20 mM for the ALS strain, increased sharply for the control especially during the feeding phase, reaching a final value of 80 mM (panel B). During the same time period, the ALS strain accumulated 35 mM of acetoin. More importantly, a substantial enhancement in recombinant protein production was attained by the ALS strain achieved (Figure 3, panel C). For the mid-growth phase, the specific @-galactosidasecontent was maintained at around 25 000 MU for the ALS strain compared with 15 000 MU for the control. Specific productivities were reduced by about 7000 MU for both stains for the last part of the fermentation, with the ALS strain maintaining the lead. Final volumetric protein yields reached 1.6 and 0.5 million units/mL for the ALS strain and the control, respectively. In other words, the engineered ALS strain achieved a 220% improvement in volumetric recombinant protein production, while reaching a final cell density that was 35% higher than that of the control. The expression levels attained by the ALS
0 0
4
8
12
16
20
24
3.0ct4
2 . 0 ~ ~ A6
2.5CH4
1.5ct6
2.OCH4
3 9
=
I.Oet6 $
-2
1sct4 1.0c+4 J.OSi.3
5.0ct5
o.oCH0
o.oCH0
0
4
8
12
16
20
9
Q
24
Time after induction fir)
Figure 3. Fed-Batch cultures of GJTOOl:pSM552-545C- harboring either the alsS gene (closed symbols) or the control plasmid (open symbols). Time profiles for cell growth and glucose consumption (A), acetate and acetoin excretion (B), and P-galactosidase production (C).
strain correspond to more than 1 g/L of recombinant protein, a level that coincides with top yields reported for process employing optimized fed-batch cultivations.
Discussion The results reported here demonstrate two important accomplishments through the use of genetic and metabolic engineering of an E. coli strain. First, the accumulation of acetate was minimized by introducing the ALS enzyme into the metabolic stm‘cture of E. coli. The newly added ALS directs excess pyruvate away from the common acetate formation pathway. The enzyme competes effectively with other pyruvate utilizing enzymatic systems to form acetolactate and carbon dioxide as a major metabolite. The minimal effects of the ALS system on cell growth rate and biomass yield indicate that the ALS competes favorably enough to suppress acetate formation yet not to such an extent that is harmful to the cell. Second, the presence of the ALS significantly improves the overproduction of a recombinant protein. This finding, which indirectly supports the hypothesis of deteriorating productivity due to acetate accumulation, may have important applications in large-scale production of heterologous proteins. Fed-batch fermentations are very commonly used for achieving high concentrations of recombinant proteins (Shimuzu et al., 1988; Yee and Blanch, 1992). High cell densities are usually achieved by regulating the addition of the limiting substrate, which is usually glucose. One of the challenges in maximizing the performance of the fed-batch culture is to maintain an optimal feeding profile. Underfeeding will inevitably limit the growth rate, and moreover it will diminish the recombinant protein yield (Ramirez and Bentley, 1993; Yee and Blanch, 1992). On the other hand, overfeeding will lead to substrate accumulation, which in the case of glucose will lead to catabolic repression and subsequent excretion
Biotechnol. Prog., 1995, Vol. 11, No. 4
470
of inhibitory acidic byproducts (Robbins and Taylor, 1989; Shimuzu et al., 1988). Experimental results indicate that the genetically engineered strain is less sensitive to the presence of high glucose concentrations, as illustrated by its capability to maintain a reduced acetate level even in a nutrient-rich environment. Since maintaining an “optimal” nutrient feeding profile is a major technical challenge in most fedbatch cultivations, results from the current study suggest that the ALS system can be used to alleviate the problems of overfeeding and thus widen the operating window for high cell density cultivation processes.
Acknowledgment This material is based in part upon work supported by the Texas Advanced Technology Program under Grant No. 003604-011 and 003604-035 and the National Science Foundation (BES-9315797). Literature Cited Adams, M. R. Growth inhibition of food-borne pathogens by lactic and acetic acids and their mixtures. Intern. J. Food Sci. 1988,23,287-292. Aristidou, A. A.; Meng, S.-Y.; Bennett, G. N.; San, K.-Y. Characterization of a pH-inducible promoter system; National AIChE meeting, Los Angeles, CA, 1991. Aristidou, A. A.; San, K.-Y.; Bennett, G. N. Modification of Central Metabolic Pathway in Escherichia coli to Reduce Acetate Accumulation by Heterologous Expression of the Bacillus subtilis Acetolactate Synthase Gene. Biotechnol. Bioeng. 1994,44,944-951. Bauer, K.;Bassat, A. B.; Dawson, M.; De La Puente, V. T.; Neway, J. 0. Improved expression of human interleukin-2 in high-cell-density fermentor cultures of Escherichia coli K-12 by a phosphotransacetylase mutant. Appl. Enuiron. Microbiol. 1990,56,1296-1302. Chang, A.C. Y.; Cohen, S. N. Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the p15A cryptic miniplasmid. J . Bacteriol. 1978,134,11411156. Chou, C.-H.; Aristidou, A. A.; Meng, S.-Y.; Bennett, G. N.; San, K.-Y.; Characterization of a pH-Inducible Promoter System for High-Level Expression of Recombinant Proteins in Escherichia coli. Biotechnol. Bioeng. 1995,in press. Doelle, H. W.; Ewings, K. N.; Hollywood, N. W. Regulation of glucose metabolism in bacterial systems. Adu. Biochem. Eng. / Biotechnol. 1981,23,1-36. Holms, W.H. The central metabolic pathways of Escherichia coli: relationship between flux and control at a branchpoint,
efficiency of conversion to biomass, and excretion of acetate. Curr. Top. Cell. Reg. 1986,28,69-105. Holtzclaw, W. D.; Chapman, L. F. Degrative acetolactate synthase of Bacillus subtilis: Purification and properties. J . Bacteriol. 1975,121,917-922. Jensen, E. B.; Carlsen, S. Production of recombinant human growth hormone in Escherichia coli: Expression of different precursors and physiological effects of glucose, acetate, and salts. Biotechnol. Bioeng. 1990,36,1-11. Luli, G. W.; Strohl, W. R. Comparison o f growth, acetate production and acetate inhibition of Escherichia coli strains in batch and fed-batch fermentations. Appl. Environ. Microbiol. 1990,56,1004-1011. Maniatis, T.; Fritsch, E. F.; Sambrook, J. Molecular Cloning; Cold Spring Harbor Laboratory: Cold Spring Harbor, 1982. Meng, S.-Y.; Bennett, G. N. Regulation of the Escherichia coli cud operon: Location of a site required for pH induction. J . Bacteriol. 1992,174,2670-2678. Meyer, H.-P.; Leist, C.; Fiechter, A. Acetate formation in continuous culture ofEscherichia coli K12 D1 on defined and complex media. J . Biotechnol. 1984,I , 335-358. Miller, J. H. Experiments in Molecular Genetics; Cold Spring Harbor Laboratory: Cold Spring Harbor, 1972. Ramirez, D. M.; Bentley, W. E. Enhancement o f recombinant protein synthesis and stability via coordinated amino acid addition. Biotechnol. Bioeng. 1993,41,557-565. Robbins, J. W.; Taylor, K. B. Optimization of Escherichia coli growth by controlled addition of glucose. Biotechnol. Bioeng. 1989,34,1289-1294. San, K.-Y.; Bennett, G. N.; Aristidou, A. A.; Chou, C.-H. Strategies in high level expression of recombinant protein in Escherichia coli. Ann. N . Y. Acad. Sci. 1994 Shimuzu, N.; Fukuzono, S.; Fujomori, K.; Nishimura, N.; Odawara, Y. Fed-batch cultures of recombinant Escherichia coli with inhibitory substance concentration monitoring. J . Ferm. Technol. 1988,66,187-191. Smith, M. W.; Neidhardt, F. C. Proteins induced by anaerobosis in Escherichia coli. J . Bacteriol. 1983,154, 336-343. Tolentino, G. J.; Meng, S.-Y.; Bennett, G. N.; San, K.-Y. A pHregulated promoter for the expression of recombinant proteins in Escherichia coli. Biotechnol. Lett. 1992,14, 157-162. Yee, L.; Blanch, H. W. Recombinant protein expression in high cell density fed-batch cultures of Escherichia coli. Biol Technol. 1992,IO, 1550-1556. Accepted May 26, 1995.@ BP950030J @
Abstract published in Advance ACS Abstracts, July 1, 1995.
