`Volume 16 No.9 (Sept.J994) p.89J-896
`Received as revised 7th July
`
`A STUDY OF THE EFFECT OF SPECIFIC GROWTH RATE AND
`ACETATE ON RECOMBINANT PROTEIN PRODUCTION OF
`ESCHERICHIA COLI JM107
`
`Claire Tumerl *;Malcolm E. Gregory2 and Michael K. Tumerl
`
`1 Advanced Centre for Biochemical Engineering, University College London, Torrington Place,
`London WCIE 7JE; and 2Centre for Process Systems Engineering, Imperial College of
`Science, Technology and Medicine, London SW7.
`
`ABSTRACT
`The growth of Escherichia coli in fed-batch and continuous culture is examined and results
`show that a-amylase production is strongly dependent on specific growth rate (dilution rate) of
`the culture and production is greatest at an intermediate rate. Using continuous culture, it has
`also been found that the presence of acetate above a certain concentration reduces both ~max•
`and the production of recombinant protein.
`
`INTRODUCTION
`Culture pH, temperature and dissolved oxygen are among the variables which affect growth
`and recombinant protein production in Escherichia coli; each is routinely measured and
`controlled to optimum values. Now that monitoring techniques are more advanced, other
`variables such as specific growth rate, and the concentrations of nutrient sugars and excreted
`metabolites can also be controlled. However, before implementing any control strategy, it is
`important to know what effects these variables might have on a fermentation system.
`
`E. coli cells excrete acetate when they grow rapidly or when the medium contains an excess of
`sugar. Holms (1986) suggested that the production of acetate enables the cells to grow faster,
`by partially uncoupling glycolysis and the TCA cycle. Whatever the reason, acetate is an
`undesirable fermentation product. It disrupts the proton motive force, and inhibits the growth
`of the cells and their ability to produce recombinant protein (Sun et al. 1993; Luli et al. 1990;
`Yang 1992; Pan et al. 1987; George et a/.1992). Furthermore, the effect of acetate on the
`growth of recombinant cells is more marked than on that of host cells (Koh et al. 1992). It is
`therefore important to devise a suitable control strategy, by limiting carbon source feeding, to
`keep acetate concentrations within acceptable limits.
`
`Although the rate of growth and sugar concentration in the medium affect acetate production,
`they also have an effect on recombinant protein production which may be independent of the
`acetate concentration. Thus all the factors need to be examined together. A number of
`researchers have worked on the effect of specific growth rate (dilution rate) in continuous
`culture on the production of recombinant protein (Fu et al. 1993; Nancib et al. 1992; Brown et
`al. 1985; Park et al. 1990) and their results show that product yields were generally higher at
`
`891
`
`BEQ 1038
`Page 1
`
`
`
`low dilution rates. In contrast, little has been reported on the control of specific growth rate in
`fed-batch culture.
`
`This paper examines the role of specific growth rate and acetate concentration on the production
`of recombinant protein in E. coli fermentations, and demonstrates that the effect of these two
`variables is independent. The effect of acetate concentration on the maximum specific growth
`rate of the organism is also discussed.
`
`MATERIALS AND METHODS
`Organism:
`The organism used in this study was a recombinant Escherichia coli JMI07 carrying the high copy number
`plasmid pQR126 (Bahri and Ward, 1991). The latter contains the genes for a-amylase, which is secreted into
`the periplasm, and for resistance to kanamycin A, which was used for plasmid selection. The lac promoter
`which controls the a-amylase gene is subject to catabolite repression by glucose. For this reason, we used
`galactose as the carbon source in all fermentations.
`Growth media
`The basic medium used for all fermentations contained (g.L-1 ):
`(NH4)zS04, 10; Na2HP04, 2.16; KH2P04, 0.64; NaCI, 5; citnc acid, 0.2; FeS04.1H20, 0.2;
`MgS04.1H20, 0.2; kanamycin, 0.01; thiamine, 0.1; trace element solution, 1mL.L-1.
`The medium also contained the following trace elements (g.L -I):
`CaCI2, 0.001; H3B03, 0.004; MnCI2AH20, 0.002; ZnS04.1H20, 0.002; CuS04.5H20, 0.0004;
`CoCI2.6H20, 0.0004; NaMo04.2H20, 0.0002.
`Galactose was added as described in the text.
`All chemicals were obtained from the Sigma Chemical Company Ltd. The kanamycin A potency was a
`minimum of 750 J.lg.mg-1. Thiamine was present as the hydrochloride.
`
`Fermentation
`Inoculum
`An inoculum of 6.25% by volume, grown from plate colonies for 12 hours at 37°C, 200 rpm in an orbital
`incubator, was added to the fed-batch fermentations.
`Fed-batch
`Fed-batch fermentations were performed in a Chemap 14 L fermenter, with the monitoring and control of
`standard fermentation variables including pH to 7 using 3 M sodium hydroxide and temperature to 37°C. The
`culture was grown aerobically with dissolved oxygen tension (DOT) always greater than 20 %. Galactose (500
`g.L -I) was fed according to algorithms in a process control system described elsewhere (Gregory et al. 1993a,
`1994).
`Continuous culture
`The continuous fermentation was performed in an LH 2000 series 2 litre fermenter. The fermentation medium
`used was that described above, supplemented with galactose (5 g.L-1 ). The dilution rate was varied to set a
`number of different steady states allowing a minimum of 3-5 pot volumes before each steady state was sampled.
`At least three samples were taken at each steady state, and each sample was separated by at least half a pot
`volume ie. 750 mL diluent. In the experiment in which the effect of acetate on specific growth rate was
`determined, acetate was added as sodium acetate in concentrations 2, 5 and 8 g.L-1 to the medium. In the
`experiment in which the effect of acetate on amylase production was determined, acetate was added as a single
`pulse of 7.4 g.L-1 sodium acetate in 15 mL water.
`
`Assay methods
`Determination of maximum specific growth rate (J.lmax)
`The dilution rate (D) was increased to approximately twice the expected value of J.lmax• and the absorbance of the
`fermentation broth was measured at 600 nm every 15 minutes. If the natural logarithm of the optical density is
`plotted against time, the slope (m) is related to J.lmax according to equation I
`J.lmax = D + m (I)
`
`Off-line monitoring and HPLC
`Fermentations were sampled at intervals for measurement of their absorbance at 600 nm, their dry weight and
`their a-amylase activities. Galactose and acetate concentrations were measured in the samples after their
`separation on an HPLC system supplied by LDC and fitted with a refractive index detector. The analytes were
`eluted isocratically from a Biorad Aminex HPX-87H column with a mobile phase comprising 4mM H2S04 at a
`flow rate of 0.65 mL.min-1 at 50°C. The peak areas of acetic acid and galactose were compared to those from
`standard solutions.
`
`892
`
`BEQ 1038
`Page 2
`
`
`
`On-line monitoring and control
`Galactose and acetate were also measured automatically in the fed-batch fermentations using an on-line HPLC
`system which aseptically removes and centrifuges a sample before injecting it onto an HPLC column (Turner et
`al. 1993). The monitoring system was serially linked to a computer running LabView. This is a process
`control system which can program a feeding strategy and alter it in response to the on-line HPLC results
`(Gregory et al. l993b). The system was programmed to feed galactose to cells during fed-batch fermentations at
`exponential rates to give constant specific growth rates of 0.1, 0.2 and 0.4 h-1. Such rates can be obtained
`provided the cells are not overfed, the galactose input does not exceed the oxidative capacity of the cells, and the
`feed-pump is accurately calibrated and the initial biomass known. Galactose and acetate were monitored on-line
`throughout the. experiments to ensure that the cells were not overfed. If either was detected then the algorithm
`was revised with a new lower estimate of biomass, thus reducing the galactose feed rate.
`a-amylase assay
`The assay for a-amylase was based on the decolourisation of starch-iodine complex. A sample (1 mL) of
`fermentation broth was centrifuged; the supernatant was assayed directly for any extracellular leakage of a(cid:173)
`amylase, and the pellet, after treatment, was assayed for the bound periplasmic enzyme. The latter was released
`with 0.2 mL lysozyme buffer (sucrose 20% w/v, lmM EDTA, 0.2 M tris buffer adjusted to pH7.5 and 0.5
`mg.mL-1 chicken egg white lysozyme obtained from Sigma) in which the pellet was resuspended at 22°C.
`After 10 minutes, 0.2 mL distilled water was added, and finally, after a further 10 minutes, the sample was
`centrifuged. The assays should be performed within 2 hours.
`The a-amylase assay was adapted by Dr Carol French in the department of Biochemistry at UCL from the
`method described by Blanchin-Roland and Massen (1989). Samples of the supernatants were diluted to 0.5 mL
`with 15 mM.sodium phosphate buffer adjusted to pH 5.8, and were mixed with 0.5 mL of the same buffer
`containing 0.5 % soluble starch (obtained from Sigma). The experimental samples were incubated at 500C.
`Portions (50 j.JL) were removed at timed intervals between 0.5 and 15 minutes and were mixed with l mL of a
`solution of iodine in potassium iodide (100 mL 2 % Kl mixed with 0.2 mL 4.4% Kl containing 2 2% 12).
`The absorbances of the resulting mixtures were read at 620 nm.
`A plot of A620 versus time will be linear if the correct sample dilution is used. Typical slopes range from 0.0 l
`to 0.035 absorbance units per minute. One unit of a-amylase activity corresponds to a decrease in absorbance of
`l unit per minute.
`
`RESULTS
`
`The effect of specific erowth rate on a-amylase production
`
`Continuous culture
`The rate at which recombinant E. coli cells produce a-amylase is dependent on the growth rate
`of the organism, which, in continuous culture, is equal to the dilution rate. The fastest
`production rate occurs at about 0.29 h-1 (figure 1) which is intermediate amongst the dilution
`rates tested. It is also well below the maximum specific growth rate (Jl
`) above which
`max
`washout occurs (0.53 h-1). Indeed this latter value is increased to about 0.65 h-1 if the cells are
`grown at a higher dilution rate just below washout for an extended period.
`
`Experiments at increasing dilution rates also show how fast the cells can grow without
`exceeding their oxidative capacity. As the rates are increased, this limit is characterised by their
`excretion of acetate when all the galactose fed to the culture is still consumed. When J.lmax is
`0.53 h-1 this limit is reached at a growth rate of about 0.5 h-1.
`
`Fed-batch culture
`The galactose feed system was programmed with a control algorithm set to achieve specific
`growth rates ofO.l, 0.2 and 0.4 h-1 (see Methods). The actual specific growth rates estimated
`from optical density or dried biomass data, were held between 0.1 and 0.4 h-t (table 1). The
`rate of production of amylase was greatest when the specific growth rate was controlled at 0.2
`
`893
`
`BEQ 1038
`Page 3
`
`
`
`h-1 (table l) and this high level of amylase production was maintained throughout the
`fermentation (figure 2).
`
`It is difficult to compare rates of production in the fed-batch fermentations in which the cell
`concentration changes, with those in continuous culture where the concentration is fixed.
`However, at a biomass concentration in the fed-batch culture of about 2.5 g.L-1, which is
`comparable with that in the continuous culture when it is fed with nutrients containing 5 g.L-1
`galactose, the maximum rates of amylase production appear to be similar in the two
`fermentations (compare figure l and table 1), although the precise specific growth rate at which
`the maximum occurs may be lower in the fed-batch fermentation.
`
`The effect of acetate on cell arowth
`
`When a continuous culture of E. coli JM107 is diluted with feed at a rate sufficient to wash out
`• When these
`the cells, the presence of acetate in the feed medium reduces the measured J.L
`max
`experiments are conducted with normal medium, J.L
`is about 0.65 h-1 (table 2). However, if
`max
`the dilution medium contains 5 g.L-1 acetate J.L
`is reduced by about 12% while at 8 g.L-1,
`max
`the reduction is over 50%. A similar amount of sodium as sodium chloride had no effect,
`implying that the effects were due to the acetate alone.
`
`The effect of acetate on a-amylase production
`
`Acetate in high concentrations is known to reduce the production of recombinant proteins in E.
`coli. Since it is a metabolite which the organism tends to excrete at high growth rates it seemed
`that its effect might not be direct, but linked to growth. When some fed-batch fermentations of
`E. coli JM 107 were fed an excess of galactose, the concentration of acetate rose as high as 7.4
`g.L-1. When this concentration was added as a single pulse to a continuous culture whose
`dilution rate was 0.3 h-1, it unexpectedly caused an increase in the rate of cell growth, and after
`some 40 minutes, a low level of galactose {up to 0.1 g.L- 1) appeared in the medium (figure 3).
`The transitory fall in the growth rate which apparently occurs immediately after adding the
`acetate is probably due to the shock to the cells. It seems that then the cells temporarily used
`the acetate as a source of carbon for growth. After 5 hours, the acetate was either consumed or
`was diluted from the medium and the growth rate returned to its preset value (figure 3).
`
`In contrast to its effect on growth, the added acetate almost completely inhibited the production
`of a-amylase; the amount of enzyme fell almost as rapidly as the theoretical rate of decline due
`to dilution of the culture (figure 4). Although the increase in growth rate to about 0.45 h- 1
`would decrease the rate of synthesis of the recombinant protein, it would still be some 35 % of
`synthesis at a growth rate of 0.3 h-1 (figure I) and it might be more if the effect were only
`transient. Moreover, the production of a-amylase continues as before once the acetate
`concentration has fallen below about 1 g. L-1 even though the growth rate remains above 0.4
`h-1 (figures 3 and 4). We believe that this demonstrates a direct effect of acetate on a-amylase
`production which is not mediated through its effect on the rate of cell growth.
`
`894
`
`BEQ 1038
`Page 4
`
`
`
`Actual specific growth rate (optical density)
`Actual specific growth rate {dry cell weight)
`Specific production rate of ex-amylase
`(units.mL-l.min-1 .(gDCW.L-lyl.h-1)
`
`Controlled specific growth rate h-1
`0.1
`0.2
`0.4
`0.09
`0.18
`0.41
`0.1
`0.2
`0.33
`0.09
`3.4
`1.2
`
`Table 1. Specific growth rates set and those obtained based on dry biomass and optical density data in controlled
`fed-batch fermentations. Specific ex-amylase production rates are given for each growth rate.
`
`ACETATE
`CONCENTRATION (g.L -I)
`0
`2
`5
`8
`
`llmax (h-I)
`
`0.655
`0.67
`0.58
`0.303
`
`%REDUCTION
`IN Jlmax
`0
`0
`11.6
`52.6
`
`Table 2. Summary of experiment in which different concentrations of acetate (as sodium acetate) were added to
`the feed medium in washout experiments in continuous culture. In each case, cells were growing at IJ.max when
`the medium was supplemented with acetate.
`
`. ~·
`
`; -7 2.5
`~ s::.
`g.,.· 20
`~-7 0
`"8~ as 1.5
`., -
`.. c:
`>- E 10
`~-;· 0
`.g-g
`&l ~ 0.5
`a.c:
`"'2.
`
`0.0
`
`0.,
`
`0.3
`0.2
`0.~
`dilution rate (h'1 )
`
`0.5
`
`Figure l. Specific amylase production rates of
`E. coli JMI07 at different dilution rates for a
`continuous culture.
`
`0
`
`2
`
`8
`6
`4
`dry Cell weight (g.l ' 1
`)
`
`10
`
`12
`
`Figure 2. Amylase activity versus dry cell weight
`of samples taken during three fed-batch cultures.
`J1S were controlled at 0.1, 0.2 and 0.4 b'1
`•
`
`.r s.
`if
`.. 1
`
`_..J... _
`
`_..J... _
`
`_..J... _ _,_ _ _,f-1--.i
`
`.... i
`
`0.21-_
`•
`
`~ 0
`....
`.:-
`.9'
`0 ~6
`
`-5
`
`5
`0
`time after acetate addition (hours)
`
`20
`
`Figure 3. Acetate and galactose profiles of continuous
`culture after addition of acetate dose (7.4 g.L-1
`).
`Plotted specific growth rate was calculated from the
`optical density increase and the known dilution rate
`of the culture.
`
`20
`5
`0
`-5
`time alter acetate addition (hours)
`
`Figure 4. Amylase profile of a continuous culture
`
`after a dose of acetate (7.4 g.L-1) had been added
`to the fermenter. Also plotted on graph is the amylase
`profile that would result due to dilution effects alone
`if no further amylase had been produced by the cells
`after the acetate addition.
`
`895
`
`BEQ 1038
`Page 5
`
`
`
`ACKNOWLEDGEMENTS
`
`The authors would like to thank Dr John Ward and Dr. Carol French for providing the E. coli
`strain, the method for cell fractionation, the adapted a-amylase assay and useful discussions.
`This project was supported jointly by the Biotechnology and Biological Sciences Research
`Council and the Engineering and Physical Science Research Council through the IRC in
`Biochemical Engineering at University College and the IRC in Process Systems Engineering,
`Imperial College London. The support of both centres is gratefully acknowledged.
`
`REFERENCES
`Bahri SM, Ward 1M (1990) Cloning and expression of a-amylase gene from Streptomyces thermoviolaceus
`CUB74 in Escherichia coli JMI07 and S. lividans TK24. J Gen Microbioll36: 811-818.
`Blanchin-Roland, S. and Masson, J-M. (1985) Protein secretion controlled by a synthetic gene in Escherichia
`coli.. Protein Eng. 2:473-480.
`Brown SW, Meyer H-P, Fiechter A (1985) Continuous production of human leucocyte interferon with
`Escherichia coli and continuous cell lysis in a two stage chemostat. Appl Microbiol Biotechnol 23: 5-9.
`Fu J, Wilson DB, Shuler ML (1993) Continuous high level production and excretion of a plasmid encoded
`protein by Escherichia coli in a two stage chemostat. Biotech Bioeng 4: 937-946.
`George HA, Powell AL, Dahlgren ME, Herber WK, Maigetter RZ, Burgess BW, Stirdivant SM, Greasham R
`( 1992) Physiological effects of TGFa-PE40 expression in recombinant Escherichia coli JMl 07. Biotech Bioeng
`40:437-445.
`Gregory ME, Turner C (1993a) Open-loop control of specific growth rate in fed-batch cultures of recombinant
`Escherichia coli. Biotechnol Tech 7 (12):889-894.
`Gregory ME, Turner C (1993b) Use of on-line HPLC analysis in the control of fed-batch cultures of E. coli
`producing recombinant protein. Proceedings of the 6th. European Conference on Biotechnology. Florence. June
`1993.Vol.2, ppTUl46.
`Gregory ME, Keay PJ, Dean P, Bulmer M, Thornhill, N (1994) A visual programming environment for
`bioprocess control. J Biotechnol 33(2):233-241.
`Holms WH (1986) The central metabolic pathways of Escherichia coli: relationship between flux and control at
`a branch point, efficiency of conversion to biomass and excretion of acetate. Current Topics in Cellular
`Regulation 28:69-105.
`Koh BT, Nakashimada U, Pfeiffer M, Yap MGS (1992) Comparison of acetate inhibition on growth of host and
`recombinant E. coli Kl2 strains. Biotechnol Letts 14 (12): 1115-1118.
`Luli GW and Strohl WR (1990) Comparison of growth, acetate production and acetate inhibition of Escherichia
`coli strains in batch and fed-batch fermentations. Appl Env Microbiol 56 (4): 1004-1011.
`Nancib N, Boudrant J (1992) Effect of growth rate on stability and gene expression of a recombinant plasmid
`during continuous culture of Escherichia coli in a non-selective medium. Biotechnol Letts 14 (8): 643-648.
`Pan JG, Rhee JS, Lebeault JM (1987) Physiological constraints in increasing biomass concentration of
`Escherichia coli B in fed-batch culture. Biotechnol Letts 9 (2): 89-94.
`Park, S. and Ryu, D.D.Y. (1990) Effect of operating parameters on specific production rate of a cloned gene
`product and performance of recombinant fermentation process. Biotech Bioeng 35: 287-295.
`Sun W-J, Lee C, George HA, Powell AL, Dahlgren ME, Greasham R, Park C-H (1993) Acetate inhibition on
`growth of recombinant E. coli and expression of fusion protein TGFa-PE40. Biotechnol Letts 15 (8): 809-814.
`Turner C, Thornhill NF, Fish NM (1993) A novel method for the on-line analysis of fermentation broth using a
`sampling device, microcentrifuge and HPLC. Biotechnol Tech 7 (1): 19-24.
`Yang X-M (1992) Optimisation of a cultivation process for recombinant protein production by Escherichia coli.
`J Biotechnol 23: 271-389.
`
`896
`
`BEQ 1038
`Page 6