VO . 39 NO.1
`
`21 FEBRUARY 1995
`
`ISSN 0168-1656
`JBITD4 39 (1) 1 98 (1995)
`
`rnal of
`J
`I I technol gy
`
`E SEVIER
`
`BEQ 1008
`Page 1
`
`

`
`© 1995 Elsevier Science B.V.
`
`All rights reserved.
`
`0168-1656/95/$09.50
`
`No part of this publication may be reproduced , stored in a retrieval system or transmitted in any form or by any mean s, electronic, mechani(cid:173)
`cal, photocopying, recording or otherwise, without the prior written permission of the Publisher, Elsevier Science B.V., Copyright and
`Permissions Department, P.O . Box 521, 1000 AM Amsterdam, The Netherlands.
`
`Special regulations for Authors. Upon acceptance of an article by the journal, the author(s) will be asked to transfer copyright of the article to
`the Publisher. 1l1is transfer will ensure the widest possible dissemination of information.
`
`Special regulations for readers in the USA. This journal has been registered with the Copyright Clearance Center, Inc. Consent is given for
`copying of art icles for personal or internal use, or for the personal use of specific clients. This consen t is given on the condi ti on that the
`copier pay th rough the Center the per-copy fee stated in the code on the first page of each article for copying beyond that permitted by
`Sections I 07 or I 08 of the US Copyright Law. The appropriate fee should be forwarded with a copy of the first page of the article to the
`Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; Tel. (508) 7508400; Fax (508) 7504744. If no code
`appears in an article, the author has not given broad consent to copy and pern1ission to copy must be obtained directly from the author(s). The
`fee indicated on the firs t page of an article in this iss ue will apply retroactively to all art ic les published in the journal, regardless of the year of
`publication. This consent does not extend to other kinds of copying, such as for genera l distribution, resale, advertising and promotion pur(cid:173)
`poses, or for creating new collective works. Special written pern1ission must be obtained from the Publisher for such copying.
`
`No responsibility is assumed by the Pub lisher for any injury and/or damage to persons or property as a matter of products liability, negli(cid:173)
`gence or otherwise, or from any use or operation of any methods, products, instruction s or ideas contained in the material here in.
`
`Although all advertising materia l is expected to conform to ethi cal (medical) standards, inclusion in this publication does not constitute a
`guaran tee or endorseme nt of the quality or value of such product or of the claims made of it by its manufacturer.
`
`Elsevier Science B. V.
`Fax (+31-20)4853258; Tel. (+31-20)48534 12
`
`Postal Address:
`Journal of Biotechnology
`Elsevier Science B. V.
`P.O. Box 1527
`1000 BM Amsterdam
`The Netherlands
`
`Courier Service Address:
`Journa l of Biotechnology
`Elsevier Science B. V.
`Molenwerf I
`1014 AG Amsterdam
`The Netherlands
`
`This joumal is in the ADONIS Service, whereby copies of individual artic les can be printed out from CD-ROM on request. An ex planatory
`Ieanet can be obtained by writing to ADON IS B.V ., P.O. Box 839, 1000 AV Amsterdam, The Netherlands.
`8 The paper used in this pub lication meets the requirements of ANSIINISO 239.48- 1992 (Pennanence of Paper).
`
`Printed in 1l1e Netherlands
`
`BEQ 1008
`Page 2
`
`

`
`ELSEVIER
`
`Journal of Biotechnology 39 (1995) 59-65
`
`journal of
`biotechnology
`
`Simple fed-batch technique for high cell density cultivation of
`Escherichia coli
`
`D.J. Korz 1, U. Rinas, K. Hellmuth, E.A. Sanders, W.-D. Deckwer *
`
`GBF National Research Center for Biotechnology, Biochemical Engineering Division, Mascheroder Weg I, 38124 Braunschweig,
`Germany
`
`Received 31 August 1994; accepted 4 November 1994
`
`Abstract
`
`A simple fed-batch process for high cell density cultivation of Escherichia coli TG 1 was developed. A
`pre-determined feeding strategy was chosen to maintain carbon-limited growth using a defined medium . Feeding
`was carried out to increase the cell mass concentration exponentially in
`the bioreactor controlling biomass
`accumulation at growth rates which do not cause the formation of acetic acid (JL < ILcri,). Cell concentrations of 128
`and 148 g per I dry cell weight (g 1- 1 DCW) were obtained using glucose or glycerol as ca rbon source, respectively.
`
`Keywords: Escherichia coli; Fed-batch; Cultivation, high cell ~~ n sity
`
`1. Introduction
`
`Escherichia coli is still the most important host
`organism for recombinant protein production. To
`maximize the volumetric productivities of bacte(cid:173)
`rial cultures it is important to grow E. coli to high
`cell concentrations. Preventing the accumulation
`of toxic levels of acetic acid is the main task to
`achieve high cell concentrations in the bioreactor.
`Growth-inhibiting acidic by-products of incom(cid:173)
`plete substrate oxidation such as acetic acid are
`produced in response to OX'Ygen
`limitation or
`excess carbon. Different strategies to grow E. coli
`
`-~Corresponding author.
`
`Present address: MAT Miill- und Abfalltechnik GmbH,
`Monchstr. 11, 70191 Stullgart, Germany
`
`to high cell densities in fed-batch cultures have
`been reviewed recently (Riesenberg, 1991; Yee
`and Blanch, 1992). The highest cell concentra(cid:173)
`tions of non -recombinant E. coli reported in the
`literature were obtained by feeding solid glucose
`via a special apparatus into a pressurized bioreac(cid:173)
`tor (134 g l- 1 DCW; Matsui et al., 1989) or by
`removing inhibitory substances by dialysis (174 g
`l- 1 DCW; Markl et al., 1993). Using feed-back
`control strategies for glucose feeding to maintain
`the dissolved oxygen in a certain range (Cutayar
`and Poillon, 1989; Mori et al., 1979) and to keep
`the specific growth rate approximately constant
`(Riesenberg et al., 1991) cell densities of 110-125
`g 1- 1 DCW have been obtained.
`All these strategies require considerable exper(cid:173)
`imental effort. For instance, in the high cell den(cid:173)
`sity cultivation proposed by Riesenberg et al.
`
`0168-1656j 95 j$09.50 © 1995 Elsevier Science B.V . All rights rese1ved
`SSD! 0168-1656(94)00 143-X
`
`BEQ 1008
`Page 3
`
`

`
`60
`
`D.J. Korz eta/. j lou mal of Biotechnology 39 ( 1995) 59-65
`
`(1991) control of growth was realized by a p0 2
`control loop (by variation of glucose feeding) and
`a f..L control loop (by variation of impeller speed),
`while the actual f..L was calculated from the off-gas
`composition. Therefore, we report here on a sim(cid:173)
`ple fed-batch technique using liquid feed medium
`and a pre-determined feeding rate to maintain
`carbon-limited growth during the fed-batch pro(cid:173)
`cess at growth rates which do not cause the
`formation of acetic acid (f..L < f..Lcrit). Thus, no so(cid:173)
`phisticated feed-back control to prevent the accu(cid:173)
`mulation of toxic levels of acetic acid or removal
`of acetic acid by dialysis is necessary to grow E.
`coli to cell densities of up to 130 and 150 g I -I
`DCW using glucose and glycerol, respectively, as
`the carbon and energy source.
`
`2. Materials and methods
`
`2.1. Bacterial strain
`
`The organism used in this study was Es(cid:173)
`cherichia coli TG1 (DSM 6056) (Riesenberg et
`a!., 1991; Carter et a!., 1985).
`
`2.2. Medium preparation
`
`The medium was prepared essentially as de(cid:173)
`scribed by Riesenberg et a!. (1991). The composi-
`
`4
`
`tion of the batch and feed medium is given in
`Table 1. For preparation of 2.5 I batch medium
`(NH 4 ) 2 HP04 , KH 2 P04 , citric acid, EDTA and
`trace elements were dissolved in 2.3 I distilled
`water in the bioreactor, pH was adjusted to 6.3
`with 5 N NaOH and the solution was sterilized
`·for 30 min at 121 oc. Stock solutions of MgS0
`and glucose or glycerol were each sterilized sepa(cid:173)
`rately for 30 min at 121 oc. Thiamine was steril(cid:173)
`ized separately by filtration. After cooling, all
`solutions including antifoam reagent SP1 (Th.
`Goldschmidt AG, Essen, Germany) were com(cid:173)
`bined and the pH was adjusted to 6.7 with aque(cid:173)
`ous NH 3 (25 % wjw) prior to inoculation. All
`components for the feed medium were sterilized
`separately and mixed afterwards. The following
`stock solutions were prepared: glucose (826.0 g
`1- 1), MgS04 · 7H 20 (1.0 g ml- 1), EDTA (8.4 mg
`ml - 1
`) and trace elements per ml: CoCI 2 · 6H 20
`(2.5 mg), MnCI 2 · 4H 20 (15.0 mg), CuCI 2 · 2H 20
`(1.5 mg), H 3 B0 3 (3.0 mg), Na 2 Mo04 · 2H 20 (2.5
`(13.0 mg) and
`mg), Zn(CH 3C00) 2 · 2H 2 0
`Fe(III)citrate (12.5 mg). Glucose, glycerol and
`MgS04 were sterilized for 30 min at 121°C. EDTA
`and trace elements were sterilized by filtration.
`Stock solutions of thiamine (10 mg ml - 1) were
`prepared directly prior to use (sterilized by filtra·
`tion). Antifoam reagent SP1 was dissolved in
`ethanol (50% v jv).
`
`Table 1
`Medium composition
`
`Components
`
`Glucose (glycerol) a
`KH 2 P04
`(NH 4 ) 2 HP04
`MgS04 · 7H 2 0
`Citric acid
`EDTA
`CoCI2 · 6H 20
`MnC1 2 · 4H 20
`CuCI2 · 2H 20
`H 3B0 3
`Na 2 Mo04 · 2H 20
`Zn(CH 3C00)2 · 2H 2 0
`Fe(III)citrate
`Thiam ine · HCI
`Antifoam SPl
`
`Batch med ium (per I)
`
`Feeding solut ion (per I)
`
`25.0 (30.0) g
`13.3 g
`4.0 g
`1.2g
`1.7g
`8.4 mg
`2.5 mg
`15.0 mg
`1.5 mg
`3.0 mg
`2.5 mg
`13.0 mg
`100.0 mg
`4.5 mg
`0.1 ml
`
`795 .0 (1021.0) g
`
`20.0 g
`
`13.0 mg
`4.0 mg
`23.5 mg
`2.5 mg
`5.0 mg
`4.0 mg
`16.0 mg
`40.0 mg
`
`a Initial concentrations of glucose or glycerol for precultures were 15.0 g 1- 1
`
`BEQ 1008
`Page 4
`
`

`
`D.J. Korz e/ a/. / Joumal of Biotechnology 39 ( 1995) 59-65
`
`61
`
`given biomass concentration the actual feed rate
`of C substrate is given by
`
`)
`iJ-(t)
`ms(t) = F(t)SF(t) = - - + m V(t)X(t)
`(
`Yx ; s
`
`(1)
`
`where 111 s is the mass flow of substrate (g h - 1
`) , F
`the volumetric feeding rate (I h - I), SF the con(cid:173)
`centration of the substrate in the feeding solution
`(g 1- 1
`) , f-1- the specific growth rate (h - 1
`), y x; s the
`biomass/ substrate yield coeffici ent (g g - 1
`) , m
`the specific maintenance coefficient (g g - 1 h - 1),
`X the biomass concentration (g 1- 1
`) and V the
`cultivation volume (!). In a fed-batch system which
`is essentially variable in volume the following
`growth equation applies
`
`d(XV)
`---=iJ-XV
`dt
`Assuming f-1- as time invariant one obtains an
`integration when starting the feeding at time t F
`
`(2)
`
`2.3. Culture conditions
`
`Precultures
`The first preculture (10 ml batch medium in
`100-ml shake flask) was inoculated with a single
`colony of E. coli TG1 from M9 minimal medium
`agar plates and incubated on a rotary shaker at
`30oC for 10-14 h. The second precultures (four
`precultures, each 100 ml batch medium in 1000-ml
`shake flasks) were
`inoculated with
`the first
`preculture (1% v jv) and incubated on a rotary
`shaker at 30°C for 10-12 h. Cells were harvested
`by centrifugation, resuspended in 60 ml batch
`medium (8-times concentrated) and used for in(cid:173)
`oculation of the main batch culture.
`
`Batch culture and controls
`Main cultivations were carried out at 28oC in a
`5-l bioreactor (Type Biostat MD; B. Braun Dies(cid:173)
`set Biotech GmbH, Melsungen, Germany)
`equipped with an additional cooling device. The
`initial batch culture conditions were as follows:
`I
`initial culture volume= 2.5 I, air flow= 2.5
`
`min - 1, oxygen flow = 0.1 I min - 1
`, stirrer speed =
`500 min - 1
`. Thermal mass flow meters (Brooks
`Instruments B.V., Veenendaal, Netherlands) were
`used for mixing air and oxygen. The dissolved
`oxygen was maintained at 20% of air saturation
`by increasing the stirrer speed. pH was controlled
`at 6.7 by addition of aqueous NH 3 (25 % w jw).
`Control of pH and dissolved oxygen was carried
`out by the digital control unit (DCU) of the
`bioreactor. The concentrations of oxygen and car(cid:173)
`bon dioxide in the exhaust gas were determined
`by paramagnetic and infrared gas analysis sys(cid:173)
`tems, respectively (Maihak, Hamburg, Germany).
`
`Fed-batch culture and controls
`After consumption of the initial glucose or
`glycerol and metabolic by-products (e.g., acetic
`acid) indicated each by an increase of the dis(cid:173)
`solved oxygen concentration, the fed -batch phase
`was started. Feeding was carried out to allow the
`volumetric cell mass concentration to increase
`exponentially. In the C-limited fed-batch culture
`without significant product formation the C sub(cid:173)
`strate is solely used for growth and maintenance.
`Hence, for a desired specific growth rate f-1- and a
`
`X(t)V(t) =X1 F~Fe~< 1 - 1 F>
`Thus, by introducing Eq. 3 into Eq. 1 the
`substrate mass feeding rate for a constant specific
`.$rowth rate (f-1- se ) follows as
`111 (t) = (~ +111)V X e~'"'(t-IF)
`Yx ; s
`
`S
`
`IF
`
`IF
`
`(3)
`
`(4)
`
`where the yield coefficient y x; s was set to 0.5
`and 0.45 for glucose and glycerol, respectively,
`and 111 = 0.025 g g h - 1 was used for both sub(cid:173)
`strates. Eq. 4 applies to large-scale reactors where
`the volume increase is predominantly governed
`by the feeding rate, and volume changes due to
`sampling and pH control are negligible. However,
`in lab-scale fed-batch reactors the cultivation vol(cid:173)
`ume does not only depend on the feeding rate
`but also on sample volume, sampling frequency
`and ammonia addition. Such volume changes may
`be significant and are not negligible in small scale
`when balancing the overall process. Therefore,
`we found it more convenient to calculate the
`time-dependent feeding rate 111's(t) with the ac(cid:173)
`tual cultivation volume V(t) instead of ~ F . Thus
`(5)
`
`m's(t) =ms(t)V(t)/V';F
`
`BEQ 1008
`Page 5
`
`

`
`62
`
`D.J. Korz eta/. / Journal of Biotech11ology 39 ( 1995) 59- 65
`
`(6)
`
`J..l(f) =
`
`However, when using Eq. 5 for the feeding
`rate the specific growth rate !L is no more con(cid:173)
`stant. It can be shown that
`V( t)
`1
`J..l set + - - In - -
`VF
`t- t F
`The deviation of JL(t) from IL se t increases with
`increasing fed-batch time but is usually less than
`20%
`towards the end of the cultivation. The
`changes in !L as a result of the feeding profile
`given by Eq. 6 are not important as long as it is
`guaranteed that JL < IL crit> i.e., below the critical
`specific growth rate above which by-products like
`acetate are formed which inhibit growth. For the
`E. coli strain applied in this study IL cri t is about
`0.17 h - 1 but may be less at high cell densities
`andjor if other limitations occur.
`The feeding solution and the aqueous NH 3
`(25 % w jw) were each placed on a balance in
`order to allow the determination of the time-de-
`
`pendent volume of the culture broth (densities of
`glucose or glycerol feeding solutions and aqueous
`NH 3 were 1.3 g cm - 3
`, 1.23 g cm - 3 and 0.91 g
`em - 3, respectively). The bioreactor was not placed
`on a balance in order to minimize disturbance
`caused by stirring and sampling. Changes in cul(cid:173)
`ture volume caused by sampling were corrected
`manually by subtracting sample volume from the
`volume of the culture medium. The dissolved
`oxygen was maintained at 20% of air saturation
`by increasing the stirrer speed andjor blending
`air with pure oxygen. Foam was suppressed, when
`necessary, by the addition of antifoam reagent
`SPl. DCU, mass flow meters, balances, feeding
`pumps and exhaust gas analysis systems were
`interfaced to a VME-bus microcomputer using
`UBICON (Universal Bio-Process Control System)
`software (Bellgardt e t al., 1992). In addition to
`control functions carried out by the DCU (tem(cid:173)
`pe rature, pH, dissolved oxygen concentration by
`
`Batch
`
`~ = ~ max
`
`Fed-batch
`11 = 11 set
`1b
`
`1a
`
`35
`
`100
`
`250 -
`
`Batch
`
`11:;:: IJ max
`
`c
`
`Fed-batch
`11 = 11 set
`1b
`
`;f.
`
`N
`0
`X
`
`80
`
`60
`
`40
`
`20
`
`200
`
`~
`§ 150
`0
`0
`0
`
`100
`
`50
`
`0
`
`.,
`c .E
`-
`
`N
`0
`·>
`
`10
`
`2000
`
`D
`
`~ 1500
`
`6
`
`~ 1000
`4 ~
`
`500
`
`t. I
`
`to
`
`N
`0
`0
`
`"
`
`stirrer
`
`6
`
`'7c e
`
`2.0
`
`1.6
`
`0>
`
`-
`-o
`·c;
`"' 1.2
`u
`~
`u 0.8
`<(
`0
`
`0.4
`
`2
`
`-
`
`0>
`ni
`·c
`0
`E
`E
`<(
`0
`
`0.0
`
`10
`
`15
`
`20
`
`25
`
`30
`
`0
`
`10
`
`15
`
`20
`
`25
`
`30
`
`Time, h
`
`Time, h
`
`Fig. 1. High cell -de nsity cultivation of E. coli TG 1 usin g glucose as carbon source. After unlimited growth durin g batch mode
`(fLnmx = 0.45 h - t), fed-batch mod e was started with a desired growth rate of /L set = 0.17 h - t (phase la), after 3 h of fed-batch
`cultiva tion th e desired growth rate was reduced to ILset = 0.14 h -
`t (phase lb). (A) Time-course of DCW (g 1- t ), mass now of
`glucose into th e bioreactor (g h - t) and glucose conce ntrat io n (g 1- t) in th e cell-free culture medium. (B) T ime-course of acetic
`acid ; phosphate (Po]- ) and ammonium (NH t} concentrations (g 1- 1) in th e cell-free culture medium. Arrows indica te the
`addition of 7.2 gPo ] - , 2.7 g NH t and 20 mg thiamine (1) and 3.0 g NH t (2) to the culture medium. (C) Time-course of optical
`density (00600 ) and concentrations of oxygen and ca rbon dioxide in th e ex haust gas(% ). Arrows indica te the time wh en the air
`now rate was increased. (D) Time-course of stirrer speed (rpm) and air and oxygen now rates (I min - I).
`
`I
`
`-
`0>
`
`10.0
`~
`0
`0
`0
`
`1.0
`
`0.1
`
`I
`~-/ 15
`
`30
`
`25
`
`20
`
`-
`0>
`.,-
`~
`g
`10 a
`•
`
`200
`
`150
`
`100
`
`50
`
`"'
`
`0>
`
`· E
`
`16
`
`-
`0> 12
`i
`J!1
`5l-
`0
`"' 0.
`
`8
`
`0
`
`BEQ 1008
`Page 6
`
`

`
`D.l. Korz eta/. / lou mal of Biotechnology 39 ( 1995) 59- 65
`
`63
`
`changing stirrer speed), UBICON was used to
`control the mass flow meters (control of dissolved
`oxygen concentration by blending air with pure
`0xygen), to calculate the time-dependent volume
`of the culture medium, to control the substrate
`feeding pump (concentrations of glucose or glyc(cid:173)
`erol feeding solutions were 0.61 g g - 1 and 0.83 g
`g-I, respectively) and for data acquisition (ex(cid:173)
`haust gas analysis) and processing (e.g., calcula(cid:173)
`tion of oxygen uptake rates).
`
`2.4. Analytical methods (off-line)
`
`Cell growth was followed by measurement of
`the optical density at a wavelength of 600 nm
`(Novaspec II, Pharmacia LKB, Freiburg, Ger(cid:173)
`many; 1 00600 corresponds to 0.52 g I - 1 DCW).
`In addition, dry cell weights (DCW) were deter(cid:173)
`mined from 1-ml aliquots of culture medium col(cid:173)
`lected in balanced 1.5-ml centrifugation tubes.
`Cell pellets were collected by centrifugation for 3
`
`min at 3300 X g , resuspended in distilled water,
`centrifuged again and dried at 40°C under vac(cid:173)
`uum until constancy of weight. Test kits from
`Boehringer-Mannheim GmbH (Germany) were
`used to analyze the concentrations of glycerol
`and acetic acid in the medium. A glucose ana(cid:173)
`lyzer (Yellow Springs, OH, USA) and an ammo(cid:173)
`nium electrode (Type Orion 95-12; Colora, Lorch,
`Germany) were employed to analyze the concen(cid:173)
`trations of glucose and ammonium, respectively.
`The concentration of phosphate was analyzed
`according to a modified procedure described m
`the German Standard Methods (1983).
`
`3. Results and discussion
`
`The results of the high cell density cultivation
`using glucose as carbon source are shown in Fig.
`1. After unlimited growth during the batch mode
`(JJ- = P-maJ, glucose-limited growth at reduced
`
`Batch
`
`IJ=IJ max
`
`Fed-batch
`1-1 = 1-1 sel
`
`100
`
`80
`
`60
`
`40
`
`20
`
`""
`
`N
`0
`><
`
`300
`
`c
`
`E
`~ 200
`0
`0
`0 100
`
`~ 1500
`;,;
`-~ 1000
`~
`
`500
`
`N
`0
`·>
`
`r-- - - - - .
`
`50
`
`X 0 2 40
`
`30
`
`20
`
`10
`
`6
`
`""
`
`N
`0
`u
`><
`
`"7c:
`E
`
`~
`·>
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`Time, h
`
`200
`
`0>
`
`~ 150
`"
`~ 100
`E
`
`50
`
`100.0
`
`en 10.0
`~
`0
`0
`0
`
`1.0
`
`Batch
`
`1-1 = 1-1 max
`
`Fed-batch
`
`1-l = 1-l sel
`
`30
`
`. .
`
`..
`..
`•
`. ...
`. .....
`
`.
`
`-
`"'
`20 e ..
`6 •
`
`10
`
`~
`
`-"
`
`'
`ni ·c
`0
`E
`E
`<>:
`<>
`
`0
`
`~ 16
`
`0>
`Qj 12
`~
`~ 8
`
`0 " ..
`
`0
`
`·-
`"' -o
`'i}
`"'
`.g
`1l
`<>:
`0
`
`2 -o<>~---«J:l(X))._
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`Time, h
`
`Fig. 2. High cell density cultivation of E. coli TGl us ing glycerol as carbon source. After unlimited growth during batch mode
`(J.I.max = 0.29 h - 1
`), fed-batch mode was started with a desired growth rate of 1-Lsct = 0.12 h - t (phase 1). (A) Time-course of DCW (g
`1- t), mass flow of glycerol into the bioreactor (g h - 1) and glycerol concentration (g 1- 1) in the cell-free culture medium. (B)
`Time-course of acetic acid; phosphate (Po] - ) and ammonium (NH ! ) concentrations (g 1- 1) in the cell-free culture medium . Each
`arrow indica tes the addition of 4.4 gPO]- and 1.6 g NH ! to the culture medium. (C) Time-course of optical density (00600 ) and
`concentrations of oxygen and carbon dioxide in the exhaust gas (%). (D) Time-course of stirrer speed (rpm) and air and oxygen
`flow rates (I min - 1 ).
`
`BEQ 1008
`Page 7
`
`

`
`64
`
`D.J. Korz el a/. I lou mal of Biotechnology 39 ( 1995) 59-65
`
`growth rates (/J-sel = const. < 1L cri1) was realized by
`increasing the substrate feed during the fed-batch
`mode as described in Materials and methods
`(Fig. 1A). Feeding was started after acetic acid -
`produced as by-product during batch mode - was
`consumed (Fig. 1B). Exponentia l increase in the
`volumetric cell mass concentration was obsetved
`for 14 h or 2.6 doubling times of the biomass
`concentration during the fed-batch phase 1 (Fig.
`1A). Cell growth stopped (phase 2) after deple(cid:173)
`tion of phosphate. Phosphorylation of glucose as
`a prerequisite of glucose uptake became improb(cid:173)
`able and, consequently, glucose accumulated in
`the culture medium (Fig. 1A). The concentration
`of acetic acid started to increase in fed -batch
`phase 2, however, it did not exceed 0.3 g 1- 1 in
`the cell-free culture medium (Fig. lB). The in(cid:173)
`crease of the optical density of the culture and
`the concentrations of oxygen and carbon dioxide
`in the exhaust gas are shown in Fig. 1C. Oxygen
`and air flow rates and the stirrer speed are de(cid:173)
`picted in Fig. 10.
`The results of the high cell density cultivation
`using glycerol as carbon source are shown in Fig.
`2. After unlimited growth during the batch mode
`(j.L = /J-max), the fed-batch phase 1 with glycerol(cid:173)
`limited growth at a reduced growth rate (/J-sel =
`canst. < 1L cri1 )
`followed (Fig. 2A). Feeding was
`started after glycerol was consumed. Acetic acid
`was not produced during batch mode due to the
`lower maximum growth rate supported by glyc(cid:173)
`erol as carbon source (Fig. 2B). Exponential in(cid:173)
`crease in the volumetric cell mass concentration
`was obsetved for 18 h or 3.1 doubling times of the
`biomass concentration (phase 1). At the end of
`
`the fed-batch process, the actual specific growth
`rate decreased and the concentration of acetic
`acid started to increase (phase 2). Acetic acid
`accumulated up to 3.3 g 1- 1 in the cell-free cul(cid:173)
`ture medium (Fig. 2B). As a result of using aque(cid:173)
`ous NH 3 for pH control, the concentration of
`ammonium also increased during phase 2 of the
`fed-batch process (Fig. 2B). Cell growth stopped
`completely after 22 h of fed -batch cultivation.
`Again, phosphate was identified as the limiting
`substrate (Fig. 2B). The time-course of the opti(cid:173)
`cal density and the concentrations of oxygen and
`carbon dioxide in the exhaust gas are shown in
`Fig. 2C. Oxygen and air flow rates and the stirrer
`speed are presented in Fig. 20. A summary of
`the results of high cell density cultivations using
`glucose or glycerol as carbon source is shown in
`Table 2.
`Using the described culture conditions the crit(cid:173)
`ical desired growth rate which did not cause the
`accumulation of acetic acid was approx. 0.17 h- 1
`with either glucose or glycerol as carbon source.
`However, a desired growth rate of 0.12-0.14 h- 1
`was chosen during the fed -batch process in order
`to keep the desired gi·owth rate definitely below
`the critical value and to consider the increasing
`tendency to excrete acetic acid at high cell mass
`concentrations. Increased acetate formation at
`high cell mass concentrations may be caused by
`other by-products which accumulate to inhibitory
`concentrations and may enhance acetate excre(cid:173)
`tion and growth cessation at elevated cell concen(cid:173)
`trations (e.g., the color of the cell-free culture
`broth is turning brown with increasing biomass
`concentration).
`
`Table 2
`Fed-batch cultivation
`
`Carbon source
`
`Desired growth rate 1-L sc t (h - I)
`Final biomass X (g 1- 1)
`Yield coeffi cie nt (carbon) Y x; s (g g - 1)
`C0 2 formation rate qc02 (g g - 1 h - I)
`0 2 upt ake rate q0 2 (g g - 1 h - 1)
`Respiratory quotient RQ
`
`Glucose
`
`0.14
`128
`0.506
`0.205 ± 0.01
`0.131 ± 0.02
`1.04 ± 0.02
`
`Glycerol
`
`0.12
`148
`0.43
`0.18 ± 0.1
`0.163 ± 0.1
`0.77 ± 0.1
`
`BEQ 1008
`Page 8
`
`

`
`D.J. Korz eta/. / lou mal of Biotechnology 39 (1995) 59-65
`
`65
`
`Applying this simple feeding strategy, final
`biomass concentrations obtained using glucose
`and glycerol as carbon source correspond to 58%
`and 67%, respectively, of the maximum theoreti(cid:173)
`cal cell mass concentration with respect to the
`viscosity of the culture medium (the viscosity of
`the culture medium for E. coli increases steeply
`above 200 g 1- 1 DCW and the culture medium
`looses f!uidity ·at around 220 g l- 1 DCW (Mori et
`al., 1979)). The higher biomass concentration ob(cid:173)
`tained with glycerol as carbon source is mainly
`caused by the higher concentration of glycerol in
`the feeding solution. The concentration of glu(cid:173)
`cose in the feeding solution cannot be increased
`further because of the limited solubility of glu(cid:173)
`cose. However, using this fed-batch technique
`and supplementary feeding of phosphate it should
`be possible to increase the final cell mass concen(cid:173)
`tration in the bioreactor even further. This fed(cid:173)
`batch technique is insensitive to short-term dis(cid:173)
`turbances in the dissolved oxygen concentration
`caused by the addition of antifoam reagent mak(cid:173)
`ing this pre-determined feeding strategy superior
`to feed-back control strategies for carbon source
`feeding to maintain the dissolved oxygen in a
`certain range. It is sufficient to meet the oxygen
`requirements of the microorganisms (Table 2)
`and to maintain the dissolved oxygen concentra(cid:173)
`tion around 20% of air saturation by increasing
`the stirrer speed andjor blending air with pure
`oxygen. Off-gas analysis systems are not required
`to grow E. coli to extreme high cell densities.
`In addition to biomass production, this fed(cid:173)
`batch technique can be applied for production of
`recombinant proteins. Modifying the equation
`used for calculating the substrate mass flow rate
`as described in Materials and methods, this fed(cid:173)
`batch protocol can be used to maintain a constant
`specific growth rate during the fed -batch process
`allowing the investigation of growth rate effects
`on host-vector interactions in E. coli (Hellmuth
`et al., 1994).
`
`Acknowledgements
`
`We gratefully acknowledge the excellent tech(cid:173)
`nical assistance of Martina Schreiner. This work
`was supported in part by the Fonds der Chemi(cid:173)
`schen lndustrie (to UR).
`
`References
`
`Bellgardt, K.-H. , Gabel, T. , Gollmer, K. , Liehr, E., Nothnagel,
`J., Posten C. and Schulze, W. (1992) Ubicon - ein deze n(cid:173)
`trales Prozel3leitsystem fur die Biotechnologie. BioEngi(cid:173)
`neering 8, 27-33.
`Carter, P., Bedouelle, H . and Winter, G . (1985) Improved
`oligonucleotide site directed mutagenesis using Ml3 vec(cid:173)
`tors. Nucleic Acids Res. 13, 4431-4443.
`Cutayar, J.M. and Poillon, D. (1989) High cell density culture
`of E. coli in a fed-batch system with dissolved oxygen as
`substrate feed indicator. Biotech no!. Lett. 11 , 155-160.
`Hellmuth, K. , Korz, D.J., Sanders, E .A. and Deckwer, W.-D.
`(1994) Effect of growth rate on stability and gene expres(cid:173)
`sion of recombinant plasmids during continuous and high
`cell density cultivation of E. coli TGl. J . Biotechnol. 32,
`289-298.
`Mark!, H. , Zenneck, C., Dubach , A.Ch. and Ogbonna, J.C.
`(1993) Cultivation of Escherichia coli to high cell densities
`in a dialysis reactor. Appl. Microbial. Biotechnol. 39, 48-
`52.
`Matsui, T., Yokota, H., Sato, S., Mukataka, S. and Takahashi,
`J. (1989) Pressurized culture of Escherichia coli for a high
`...
`concentration. Agric. Bioi. Chern. 53, 2115-2120.
`Mori, H., Yano, T., Kobayashi, T. and Shimizu, S. (1979) High
`density cultivation of biomass in fed-batch system with
`DO-stat. J. Chern. Eng. 12, 313-319.
`Riesenberg, D. (1991) High cell-density cultivation of Es(cid:173)
`cherichia coli. Curr. Opinion Biotechnol. 2, 380-384.
`Riesenberg, D., Schulz, V., Knorre, W.A. , Pohl, H.-D., Korz,
`D ., Sanders, E.A., Rol3, A . and Deckwer, W.-D. (1991)
`High-cell density cultivation of Escherichia coli at con(cid:173)
`trolled specific growth rate. J. Biotechnol. 20, 17-28.
`Yee, L. and Blanch, H.W. (1992) Recombinant protein ex(cid:173)
`pression in high cell density fed-batch cultures of Es(cid:173)
`cherichia coli. BiojTechnology 10, 1550-1556.
`German Standard Methods for the Examination of Water,
`Waste Water and Sludge. (1983). Anions (group D); deter(cid:173)
`mination of phosphorus compounds (D 11). VCH, Wein(cid:173)
`heim.
`
`BEQ 1008
`Page 9