Biotechnol. Bo$.lQQ4, 10, 160-164
160
Effect of Glucose on the Expression Parameters of Recombinant Protein in Escherichia coli during Batch Growth in Complex Medium Xiaoli Li and Kenneth B. Taylor* Department of Biochemistry, The Fermentation Facility, University of Alabama at Birmingham, Birmingham, Alabama 35294
The expression parameters were determined at different phases of batch growth of Escherichia coli JM101/pYEJ001 in complex medium and a t different conditions of glucose addition. Thus the plasmid content, the RNA content, the RNA synthesis rate, the specific recombinant mRNA content, the specific recombinant mRNA synthesis rate, the recombinant chloramphenicol acetyltransferase content, and the overall protein synthesis rate were determined during growth with no glucose, with initial glucose, with glucose feeding during stationary phase, and with initial glucose plus glucose feeding during stationary phase. The results show that the specific rate of total RNA synthesis was enhanced in the presence of glucose a t both exponential and stationary phases, while recombinant mRNA synthesis was enhanced only a t stationary phase by glucose feeding. However, the steady-state level of the recombinant mRNA was not changed under the same conditions. In addition, although the overall protein synthesis rate a t exponential phase was enhanced in the presence of glucose, the specific recombinant protein level was unaffected. The specific synthesis rate of recombinant mRNA varied inversely with the plasmid content during exponential and stationary phases. Furthermore, changes in the specific activity of the recombinant protein were not correlated with either the changes in the specific synthesis rate of the recombinant mRNA or the overall protein synthesis rate. Therefore, the specific activity of the recombinant protein is not universally limited by its gene transcription rate or the overall protein synthesis capacity.
Introduction Knowledge about intermediate-state variables is important in the formulation of hypotheses and structured models for the control and optimization of recombinant gene expression in Escherichia coli. In addition, studies of the effects of the environmentaland growth conditions on the state variables associated with the expression of recombinant protein (r-protein) are needed for the formulation of empirical models. Furthermore, identification of the important molecular mechanisms will provide insight for the formulation of hypotheses about other similar systems. The level of an r-protein can be regulated by various factors, including the gene dosage, the transcript concentration, the translation rate, and the product stability. Lee et al. (1985) simulated the correlation between the plasmid content and the r-protein level and found that the level of r-protein and the plasmid copy number were negatively correlated at high copy number (greater than 60). Furthermore, in studies on the effects of plasmid copy number on the expression parameters, both Peretti and Bailey (1989) and Wood and Peretti (1990) found that the r-protein level did not correlate positively with the plasmid content at high copy number (greater than 60) and that the synthesis of the recombinant mRNA did not limit the r-protein level. They concluded that the translational machinery limited r-protein production in E. coli. The plasmid content is negatively correlated with growth rate supported by different media in batch culture (Seo and Bailey, 1985). Seo and Bailey (1986) also observed an inverse correlation between the plasmid content and the 87567938/94/3010-0180$04.50/0
growth rate in chemostat culture. Also, Stueber and Bujard (1982)reported that the plasmid content increased significantlyat stationary phase in batch culture. Cashel and Budd (1987) have demonstrated that the RNA synthesis rate of E. coli is correlated closelywith the specific growth rate. Furthermore, Hansen et al. (1975) have observed that both the growth rate and the synthesis rate of RNA decreased dramatically when glucose uptake was blocked during steady-state growth. The most common configuration for the production of recombinant protein is in some form of batch growth in complex medium. The incorporation of glucose in the medium, either initially or by feeding,commonly enhances growth and productivity (Robbinsand Taylor, 1989;Li et al., 1990). Thus,a number of investigatorshave reported the results of studies of one or two parameters as a function of growth phase or conditions. However, very few have reported studies of most or all of the important expression parameters simultaneously in the strain under the same seta of conditions. We report here the results of the determination of five host parameters and five expression parameters simultaneouslyas a function of growth phase and the effect of glucose addition on each of these parameters. Specifically, the biomass level, the glucose concentration, the plasmid content, the specific synthesis rate of total RNA, the steady-state level of total RNA, the specific synthesis rate of chloramphenicol acetyltransferase (CAT) mRNA, the steady-state level of CAT mRNA, the overall protein synthesis rate, the CAT specific activity, and the CAT stability were measured.
0 1994 American Chemical Society and Amerlcan Institute of Chemlcai Engineers
B k ~ t e d " .Rog,, 1994, Vol. 10, No. 2
Materials and Methods Materials. All radioactive compounds ([3H]uracil, PHIthymidine, P2P1dCTP, and P5S1methionine) were obtained from Amersham. Yeast RNA, salmon testes DNA, and Micrococcus luteus DNA were obtained from Sigma Chemical Co. All other chemicals and reagents were of analytical grade. Microorganism and Plasmid. E. coli strain JMlOl (thi,A(luc-proAB), [P, traD36,prolLB, lads, AM151)was transformedwith plasmid pYEJOO1. The plasmid carries a CAT (chloramphenicol acetyltransferase) gene that is controlled by two lac operators and a tac promoter. The tacpromoter is a hybrid of the trp promoter and the lacW5 promoter, and it is not subject to glucose-inducedcatabolite repression (De Boer et al., 1983). In a strain with the lads gene, the tac promoter is inducible by isopropyl 8-Dthiogalactopyranoside (IPTG) (Kapralek et al., 1991). Bacterial Cultivation. The basal medium for all experimentswas Luria broth (LB) (tryptone, 10g/L; yeast extract, 5 g/L; and NaC1,5 g/L) supplemented with 5 g/L yeast extract. The medium also contained 0.1 mM IPTG. The medium was supplied with glucose in three ways: (i) medium with initial glucose supplementation (0.2 % , sterilized separately), which was depleted at the middle exponential phase; (ii) no initial glucose, but fed (Robbins and Taylor, 1989) with glucose from the late exponential phase (ABM)= 7); (iii) initial glucose (0.2%), which was replenished by feeding from the late exponential phase (A650 = 4). All fermentations were performed in a 16-L fermentor equipped with dissolved oxygen (DO) and pH control (SF116, New Brunswick Scientific). The inoculum (1% , v/v) was an overnight culture in LB with 250 mg/L chloramphenicol grown in a shaker at 37 "C. The pH of the fermentor culture was controlled at 7.2 with NaOH (10N) and either Hap04 (40% ) or glucose (240g/L) feeding aspart of the feeding schedule (Robbins and Taylor, 1989). The DO level was controlled at 50% saturation in all experiments. The antifoaming agent was MAZU DF6OP. Analytical Methods. Cell dry weight was determined as in previous work (Li et al., 1989,1992). The plasmid DNA was separated by agarose gel electrophoresis. The plasmid content was determined by scanning the plasmid band on a negative of a photograph of the agarose electrophoresis gel (Li et al., 1992). The plasmid stability was determined by replica plating, and the CAT activity was assayed spectrophotometrically (Li et al., 1992). Glucose was measured with a Sigma Glucose [HKl20 kit. Fructose in 3 mL of medium was converted to glucose with glucose isomerase (100 units/mL) at 37 "C for 1 h, and lactose in 3 mL of medium was converted to glucose and galactose with /3-galactosidase (100 units/mL) at 37 "C for 1h. The pH of the medium was adjusted to 7.0 with 0.01 M HEPES. The concentration of glucose from either fructose or lactose was then determinedas described above. Pure fructose and lactose were used as standards. Synthesis Rate of DNA. The method of DNA pulse labeling was adopted from that of Kellenberger et al. (1962). A sample of the culture was pulse-labeled during exponential phase (Asso= 1.0) and during stationary phase, the latter of which was diluted with culture supernatant to the same density as that at the exponential phase. PHIThymidine (2.7 pCi) was added to 1mL of culture, and 2 min later (37 "C) the cells were poured over 0.3 mL of chilled 17% trichloroacetic acid (TCA). The sample was collected on membrane filters (0.22 pm), washed with distilled water, dried, and counted in a liquid scintillation counter.
161
Synthesis Rates of Total RNA and the Steady-State Levelof Total RNA. RNA LabelingandPreparation. The RNA was pulse-labeled with [SHIuracil (Peretti et al., 1989) in a culture sample at exponential phase (Am = 1.0) and in one at stationary phase, diluted so that Am = 1.0. [3HlUracil(100pCi) was added to 10 mL of cell suspension, and 2 min later (37 "C) the cells were poured over 20 mL of crushed frozen stop buffer (20 mM TrisHC1 (pH 7.3),5 mM MgC!l2,20 mM NaN3, and 400 pg of chloramphenicol/mL)(Rose et al., 1970). The cells were centrifuged and resuspended in 1 mL of medium C (40 mM NHrC1,40 mM Na#04, and 50 mM NaC1). In order to reduce RNase, the glassware and plasticware for the isolation of RNA were treated as described by Maniatis et al. (1982a,b),and the solutionswere treatedwith diethyl pyrocarbonate as described by Wood and Peretti (1990). The cells were lysed after the addition of 1mL of lysing mixture (0.1 M NaCl, 0.5 % SDS, and 0.01 M EDTA) and incubation at 95 "C for 20 s. The RNA was extracted three times with phenol saturated with ATE buffer (Tris, 0.01 M; azide, 0.01 M and EDTA, 0.001 M) and precipitated with ethanol. The RNA was dried in a vacuum freeze-dryer and dissolved in 0.5 mL of ATE buffer containing 5 mM MgSOd. DNase I was added to a final concentration of 5 units/mL, and the mixture was left at room temperature for 30 min. After the addition of EDTA to a final concentration of 5 mM, the RNA was again extracted, precipitated, and dried as described above.
Total RNA Synthesis Rate and the Steady-State Level of Total RNA. The extracted RNA was dissolved in 2X SSC (0.15 M saline and 0.015 M sodium citrate),and the steady-state level of total RNA (mg/mg of dry cell) was measured by the determination of Am. An absorbance of 1.0 at 260 nm was equivalent to 50 pg of RNA/mL (Dennisand Nomura, 1975). A portion (5pL) of the sample was counted with 5 mL of ScintiverseI1liquid scintillation cocktail (Fisher Scientific)in a liquid scintillation counter for the specific synthesis rate of total RNA (cpm/mg of dry cell/min). Specific Synthesis Rate of CAT mRNA. The synthesis rate of CAT mRNA was determined by hybridization of the [3Hluracil-labeledRNA described above,with excess CAT DNA probe immobilized with a slot-blot apparatus (Peretti et al., 1990). CAT DNA Probe Preparation. Plasmid pYEJOOl DNA was isolated from JM101/pYEJ001 cells by the alkaline lysis procedure and purified by CsCl equilibrium centrifugation (Maniatiset al., 1982). The isolated plasmid was cleaved with Hind111to yield the CATgene fragment, which was isolated by electrophoresis on a 1% agarose gel containing 0.5 pg/mL ethidium bromide. The slice of agarose gel containing the CAT gene fragment was cut out, and the DNA was recovered by direct extraction (Maniatis et al., 1982). The DNA was dissolved in TE buffer and the concentration was determined spectrophotometrically (Dennis and Nomura, 1975). The DNA probe was denatured in 0.3 M NaOH at 100 "C for 10min, chilled, and diluted with an equal volume of cold 2 M NH4OAC to 16 pg/mL. DNA probe (2.5 pglslot) was immobilized by filtration through a nitrocellulose filter (0.2 pm) in a microfiltration apparatus (Bio-Rad). The filter was air-dried and baked at 80 "C for 2 h in a vacuum oven. Hybridization. The areas with adsorbed probe DNA were cut from the filter and prehybridized at 42 "C for 1 h in a tube containing 0.3 mL of hybridization buffer [yeast RNA, 1 mg/mL; salmon testes DNA, 0.1 mg/mL; SDS, 0.1% (v/v); 5X Denhardt's solution (ficoll, 1 g/L; poly-
B & t ~ r w / Rag.., . 1994, Vol. 10, No. 2
102
vinylpyrrolidone, 1 g/L; BSA, 1 g/L); formamide, 50% (v/v);poly(ethyleneglycol),10% (v/v);and 5X SSPE buffer (0.9 M NaC1, 50 mM NaHnPOr, and 5 mM EDTA (pH 7.4))], The RNA, prepared as described above, was heatdenatured at 80 "C for 20 min, and samples were added to the tubes to initiate the hybridization. For each determination, four RNA samplescorresponding to 0.024, 0.048,0.096,and 0.192 mg of biomass were hybridized to ensure excess probe and linear hybridization capacity of the probe. Pulse-labeled RNA was also incubated with blank filter and hybridized against Micrococcus luteus DNA probe (2.5 pg/probe) to determine the background level. All hybridizations were incubated for 60-70 h at 46 "C, values previously determined to be optimal. After hybridization, the filters were washed three times with wash 1(SSPE and 0.1% SDS, 15 min at 60 "C) and once with wash 2 (0.1X SSPE and 0.1% SDS). The filter was further washed twice with 2X SSC and digestedwith RNase A (5 pg/mL in 2X SSC) for 1 h at 37 "C to remove nonspecific binding of the labeled RNA. The filters were again rinsed with wash 1and dissolved in 1mL of ethylene glycol monoethyl ether for 15min. The Samples were then counted for radioactivity in the liquid scintillation counter. Steady-StateLevel of CAT mRNA. The steady-state level of CAT mRNA was measured by hybridization of radioactively labeled coding sequences of the CAT gene to cellular RNA fixed on a nitrocellulose filter (Woodand Peretti, 1990). DNA Probe Labeling. CAT gene (25ng), prepared as described above, was oligolabeled (50pCiof [(r-32PldCTP, 3000 Ci/mmol) with the protocol of the Random Primers DNA labeling system (Gibco BRL). A specific activity of 4.3 X 108cpm/pg of CAT DNA was obtained. The labeled probe, diluted with unlabeled probe to 2.8 X 1Oe cpmlpg, was denatured by heating for 15 min at 100 "C and immediately cooled on ice. Hybridization. The cellular RNA, prepared as described above, was denaturated at 80 "C for 10 min and fixed on the filter (1 pg/slot) of the microfiltration apparatus as described above for DNA. Each RNA slot was cut from the filter paper and put in a tube with 0.3 mL of the hybridization buffer to prehybridize for 1h at 42 OC. The labeled DNA probe was added to each tube containing the filter-bound RNA and incubated for 18 h at 42 "C. The filter was then washed three times with wash 1 (SSPE and 0.1% SDS) at 25 "C for 15 min each and once with wash 2 (0.1X SSPE and 0.1% SDS) at 64 "C for 15 min. The filters were then dried and counted in a liquid scintillation counter. To ensure an excess of the DNA probe, three probe concentrations (53,106, and 212 ng) were used for each RNA measurement. The results showed that a probe concentration above 106 ng was in excess. The probe was also incubated with a blank filter, a filter with adsorbed yeast RNA, and a filter with adsorbed pYW0Ol plasmid (denatured) as two negative controls and one positive control, respectively. Overall Protein Synthesis Rate. A sample of 2 mL of culture at exponential phase (ABM)= 1)or stationary phase (diluted to A ~ M=)1with culture supernatant) was labeled with 0.02 pCi of [3%]methionine for 1min at 37 "C, and then unlabeled methionine was added to a final concentration of 0.5 mM for 1min (Kuriki, 1987). The reaction was stopped by the addition of 0.1 vol of 100% TCA, and the samplewas filtered on a 0.22-pm filter. After the precipitates were washed with hot 5 % TCA, the filter was counted in a liquid scintillation counter. CAT Degradation Rate. The culture sample at exponential or stationary phase was transferred from the
10.w
I 1
o/.-o
1'5 4
0 '
1.2
0
I 0
-
l2O0
i
-
1.w 0.9
0
P
v
0.8
(D
E
I 4
0.10
-
0
m 0.3
0.01
0.0 0
2
8
4
E
0
10
Time (hr)
Figure 1. Host parameters and expression parameters during the c o m e of culture in the b a d medium.
fermentor to a flask in a shaker at 37 "C. Streptomycin was added into the sample to a final concentration of 250 pg/mL/Am. Sampleswere withdrawn from the flask every 30 min from 0 to 180 min, and CAT specific activity was measured. CAT specific activity in culture without streptomycin was measured as a control.
Results Base-Line Parameters without Glucose Addition. The basal medium contained only negligible amounts of glucose (16 mg/L), fructose (6.8 mg/L),and lactose (none). Consistent with the high cellular growth rate (Figure l), the total RNA content, the specific synthesis rate of total RNA, and the specific synthesis rate of total protein were higher at exponential phase than at late exponentialphase or stationary phase. In addition, the specific synthesis rate of CAT mRNA (cpm/min/mg of dry cell) at exponential phase was about 8-10 times higher than the rates at stationary phase (Table l),and the steady-state level of CAT mRNA was 2-3 times higher at exponential phase thanat stationaryphase (Table 1). In contrast,the plasmid content (mg/g of dry biomass) in the stationary-phase culture was 30 times higher than that in exponential phase (Figure 1). In addition, the maximum CAT specific activity was obtained at stationary phase (Figure 1). The specific activity of CAT was stable at both exponential and stationary phases (over 95% of the activity was retained) over a 3-h period after protein synthesis was blocked (data not shown). Effects of Glucose. When glucose was added to the medium prior to inoculation, ita concentration decreased to zero in late exponential phase. In those experiments in which it was instituted, glucose feeding started in late exponential phase and continued into stationary phase. In the latter experiments, the glucose concentration remained at a low (ca. 0.1 g/L) but constant level. With every parameter measured,initial glucose exerted ita effect during exponential phase, and glucose feed exerted its effect in the late exponential and stationary phases. Glucose was associated with an increase in the specific synthesis rate of total RNA (Table 1). Initial glucose produced an increase during exponential phase, and glucose feeding produced an increase during stationary phase. In the latter experiments without initial glucose, the rate of RNA synthesis during stationary phase became greater than that during exponential phase. Glucose feeding resulted in an increase in the specific synthesis rate of CAT mRNA at stationary phase, but initial glucose did not produce a corresponding increase during exponentialphase (Table 1).Thus, in experimenta
Bbtechnol. Prog.., 1994,Vol. 10,No. 2
16s
Table 1. Effects of Glucose on the Synthesis Rate of Cellular RNA, the Synthesis Rate of b o m b i n a n t CAT "A, the Cellular RNA Content, the CAT mRNA Level, and the Total Protein Synthesis Rate at Different Growth Phases. determination no initial glucose 0.2% initial glucose 0.2% initial glucose feed glucose feed only exp stat exp stat exp stat exp stat growth phase 0.29 4.68 1.98 2.83 4.78 2.1 0.2 4.57 total RNA synth rate (105cpm/mg/min),*15% 2.14 0.23 2.47 1.79 1.89 3.45 CAT mRNA synth rate (loa cpm/mg/mh), *12% 2.51 0.3 0.15 0.05 0.18 0.09 0.16 0.05 RNA content (mg/g of dry cell), &3% 0.17 0.07 2.7 0.79 3.3 0.79 2.98 1.1 CAT mRNA level (105 cpm/mg cell), *lo% 3.0 1.4 12.5 2.2 13.5 3.5 5.6 3.6 total protein synth rate (104cpm/mg/min),*6% 6.3 3.7
+
0 exp, exponential phase; stat, stationary phase. The standard error of each determination, in percent, was estimated from the resulta of multiple determinations under a single set of conditions (0.2% initial glucose).
with glucose feeding but no initial glucose, the rate of CAT mRNA synthesis at stationary phase became higher than that at exponentialphase. In contrast, initial glucose produced an increase in the rate of total protein synthesis during exponential phase, but glucose feeding did not produce a correspondingincreaseduring stationary phase. This result further accentuates the difference in this rate from the exponential phase to the stationary phase. Glucose added to the culture at stationary phase produced the same parameter changes as glucose feeding. Furthermore, the effects of glucose on the synthesis rate of total RNA in the host strain, JM101, were the same as those in the recombinant strain. However, glycerol fed in the same manner produced no significant changes in the parameter values. Variation of the IPTG concentration (0.01-0.4 mM) affected neither the specificsynthesis rate of CAT mRNA nor the specific activity of CAT. None of the other parameter values are significantly different in the presence of glucose from their values in its absence.
Discussion The effects of glucose, particularly those at stationary phase, are apparently specific to glucose, since they are not observed in the presence of glycerol, an alternative carbon source. The glucose effect of increasing the rate of RNA synthesis, total RNA in both exponential and stationary phases, and mRNA in stationary phase was unexpected. However, the fact that the steady-state level of RNA remained much the same in the present results shows that the degradation rate also increased. Peretti and Bailey (1987) and Wood and Peretti (1990,1991)have reported that the synthesis rate and degradation rate of recombinant mRNA both increased simultaneouslyupon an increase in plasmid copy number or in the inducer (IPTG)concentration. The ribosomal RNA synthesis rate, which represents more than 50 % of the total transcriptional activity in E. coli at high growth rates (Bremer and Dennis, 1987), is modulated by amino acid concentration (stringent control) and growth rate (Gausing,1977;Bremer and Dennis, 1987; Cashel and Rudd, 1987). However, since neither the amino acid level nor the growth rate correlates with the presence of glucose, the increase in RNA synthesis must be a fundamental effect of glucose. Hansen et al. (1975) reported that, when glucose uptake was blocked with a-methyl glucoside (a-MG) during steady-state growth, both growth rate and the RNA synthesis rate were reduced drastically in relaxed and stringent strains of E. coli. That the decrease in the RNA synthesis rate was not correlated with the size of the ppGpp (guanosine5'-diphosphate 3'diphosphate) pool suggests that the decrease in RNA synthesis was not due to the stringent response. However, it is unclear whether the decrease in the synthesis rate of
RNA was due to the reduced growth rate or due to the glucose uptake itself. Although both the plasmid content and the specific CAT activity are lower during exponential phase than during stationary phase, both the CAT mRNA content and synthesis rate are higher at exponential phase than at stationary phase. The relatively high CAT mRNA level and synthesis rate coincident with a low plasmid content during the exponential phase indicate that the transcription rate increases until the plasmid content limits any further increase. This hypothesis explains why the transcription rate of CAT mRNA is not increased in the presence of glucose in exponential phase. It also explains why the productivity of recombinant protein is sometimes increased when the biomass growth rate is controlled at a modest level. A modest growth rate results in a higher plasmid content, which would alleviate the limitation in the rate of expressionof recombinant protein. Wood and Peretti (1990) reported that, when the plasmid copy number increased by a factor of 10, via copy number mutation, the synthesis rate of 8-lactamase mRNA in recombinant E. coli in chemostat cultivations increased 2.5 times. However, there was a sharp increase in the mRNA synthesis rate (75 times) upon a further 3.3X copy number increase. They suggested that this might be due to a derepression phenomenon at high copy number. The results reported here at consistent with those associated with an increase in the plasmid content (JMlOl/pYEJ001 has a calculated copy number of 200 per cell at stationary phase). The hypothesis that the protein synthesis machinery limits recombinant gene expression at exponential phase is not consistent with the increase in protein synthesis at exponential phase caused by glucose with no change in the content of recombinant protein. In late exponential phase and stationary phase, the higher plasmid content coincident with the lower mRNA synthesis rate and RNA content is consistent with the hypothesis that the transcription machinery or the translation machinery, but not the plasmid content, limits the rate of recombinant gene expression. However, the fact that glucose increases the transcription rate dramatically without affectingthe level of recombinant protein eliminates the hypothesis that transcription per se is limiting. Therefore, at stationary phase either the level of mRNA, by virtue of its degradation rate, or the translation machinery itself limits the expression of recombinant protein in the strain investigated. That the overallprotein synthesis rate at stationary phase was not enhanced in the presence of glucose feeding, probably because of amino acid(s) limitation, supports the hypothesis that the translation rate is a significant limiting factor. In general, the protein degradation rate during periods of balanced growth is negligible (Bremer and Dennis, 1987) and increases during starvation for carbon or other required nutrients (Miller, 1987). The
164
Bbtechd. Prog., 1994, Vd. 10, No. 2
Li, X.; Robbins, J. W., Jr.; Taylor, K. B. Effect of the levels of dissolved oxygen on the expression of recombinant proteim in four Escherichia coli strains. J. Ind. Microbiol. 1992,9, 1-10. Maniatis, T.; Fritach, E. F.; Sambrook, J. Large-scale isolation Literature Cited of plasmid DNA. In Molecular Cloning: Cold Spring Harbor Bentley, W. E.; Compala, D. S. A novel structured kinetic Laboratory Press: Cold Spring Harbor, NY,1982a;pp 87-96. modelling approach for the analysis of plasmids instability in Maniatis, T.; Fritach, E. F.; Sambrook, J. Gel electrophoresis. In recombinant bacterial cultures. Biotechnol. Bioeng. 1989,33, Molecular Cloning; Cold Spring Harbor Laboratory Press: 49-61. Cold Spring Harbor, NY, 1982b; pp 164-167. Bremer, H.; Dennis, P. P. Modulation of chemical composition Miller, C. G. Protein degradation and proteolytic modification. and other parameters of the cell by growth rate. In Escherichia In Escherichia coli and Salmonella typhimurium, Cellular coli and Salmonella typhimurium, Cellular and Molecular and MolecularBiology;Neidhardt, F. C., et al., Eds.;American Biology; Neidhardt, F. C., et al., Eds.; American Society for Society for Microbiology: Washington, D.C., 1987;pp 680Microbiology: Washington, D.C., 1987;pp 1527-1542. 691. Cashel, M.; Rudd, K. E. The stringent response. In Escherichia Peretti, S.W.; Bailey, J. E. Simulations of hostplasmid interaccoli and Salmonella typhimurium, Cellular and Molecular tions in Escherichia coli: copy number, promoter strength, Biology; Neidhardt, F. C., et al., Eds.; American Society for and ribosomebinding site strength effects on metabolicactivity Microbiology: Washington, D.C., 1987;pp 1410-1438. and plasmid gene expression. Biotechnol. Bioeng. 1987,29, De Boer, H. A,; Comstock, L. J.; Wasser, M. The tac promoter: 316-328. A functional hybrid derived from the trp and lac promoters. Peretti, S.W.; Bailey, J. E.; Lee, J. J. Transcription from plasmid Proc. Natl. Acad. Sci. U.S.A. 1983,80,21-25. genes, macromolecularstability, and cell-specificproductivity Dennis, P. P.; Bremer, H. Macromolecular composition during in Escherichia coli carrying copy number mutant plasmid. steadystate growthof Escherichia coliB/r. J. Bacteriol. 1974, Biotechnol. Bioeng. 1989,34,902-908. 119,270-281. Robbins, J. W., Jr.; Taylor, K. B. Optimization of Escherichia Dennis, P. P.;Nomura, M. Regulation of the expression of coli growth by controlled addition of glucose. Biotechnol. ribosomal protein genes in Escherichia coli. J.Mol. Biol. 1975, Bioeng. 1989,34,128+1294. 97,61. Rose, J. K.; Mosteller, R. D.; Yanofsky, C. Tryptophanmesaenger Gaming, K. Regulation of ribosome production in Escherichia ribonucleic acid elongation rates and steady-state levels of coli: synthesis and stability of ribosomal RNA and of ribosomal tryptophan operon enzymes under various growth conditione. protein messenger RNA at different growth rates. J. Mol. J. Mol. Biol. 1970,51,541-550. Biol. 1977,115,335-354. Ryan, W.; Parulekar, S. J. Recombinant protein excretion in Hansen, M. T.; Pato, M. L.; Molin, S.; Fill, N. P.; Meyenburg, Escherichia coli JMlOl[pUCS]: Effects of plasmid content, K. V. Simple downshift and resulting lack of correlation ethylenediaminetetraacetate, and phenethyl alcohol on cell between ppGpp pool size and ribonucleic acid accumulation. membrane permeability. Biotechnol. Bioeng. 1991,37,430J. Bacteriol. 1975,122, 585-591. 444. Iijima, S.;Kai, K.; Mizutani, S.;Kobayashi,T. Effects of nutrients Ryan, W.; Parulekar, S. J.; Stark, B. C. Expression of ,9-lactamase on a-amylase production and plasmid stability in fed-batch by recombinant Escherichia coli strains containing plasmids culture of recombinant bacteria. J . Chem. Biotechnol. 1986, of different size: effects of pH, phosphate, and dissolved 36,539-546. oxygen. Biotechnol. Bioeng. 1989,34,309-319. Kapralek, F.; Jecmen, P.; Sedlacek, J.; Fabry, M.; Zadrazil, S. Seo, J.-H.; Bailey, J. E. Effects of recombinant plasmid content Fermentation conditions for high-level expression of the caton growth properties and cloned gene product formation in promoter-controlled calf prochymosin cDNA in Escherichia Escherichia coli. Biotechnol. Bioeng. 1985, 27, 1668-1674. coli HB101. Biotechnol. Bioeng. 1991,37,71-79. Seo, J.-H.; Bailey, J. E. Continuous cultivation of recombinant Kellenberger, E.; Lark, K. G.; Bolle, A. Amino acid dependent Escherichia coli: existence of an optimum dilution rate for control of DNA synthesis in bacteria and vegetative phage. maximum plasmid and gene product concentration. BioProc. Natl. Acad. Sci. U.S.A. 1962,48,1860-1867. technol. Bioeng. 1986,28,1590-1694. Kuriki, Y. Response to temperature shifts of expression of the Stueber, D.; Bujard, H. Transcription from efficient promoters amp gene on pBR322 in Escherichia coli K12. J. Bacteriol. can interfere with plasmid replication and diminish expremion 1987,169,2294-2297. of plasmid specified genes. EMBO J. 1982,1,13~1404. Lancaster,M. J.;Sharp,R. J.;Court, J.R.;McEntee,I.D.;Melton, Wood, T. K.; Peretti, S. W.Depression of protein synthetic R. G.; Sherwood, R. Production of cloned carboxypeptidase capacity due to cloned-gene expression in E. coli. Biotechnol. G2 by Escherichia coli: genetic and environmental considBioeng. 1990,36,865-878. erations. Biotechnol. Lett. 1989, 10,699-704. Wood,T. K.; Peretti, S. W.Constructionof aspecialized-riboeome Lee, S. B.; Bailey, J. E. Analysis of growth rate effect on vector for cloned-gene expression in E, coli. Biotechnol. productivity of recombinantEscherichia coli populations using Bioeng. 1991,38,891-906. molecular mechanism models. Biotechnol. Bioeng. 1985,26, 66-73. Accepted November 23, 1993.. Li, X.;Robbins, J. W., Jr.; Taylor, K. B. The production of betagalactosidase in Escherichia coli in yeast extract enriched e Abstract published in Advance ACS Abstracts, February 15, 1994. medium. J. Ind. Microbiol. 1990,5,85-94. specific activity of CAT, however, was stable at both the exponential and stationary phases after protein synthesis was blocked.
