Biotechnology Advances 23 (2005) 345 – 357

www.elsevier.com/locate/biotechadv

Research review paper

Growing E. coli to high cell density—A historical

perspective on method development
Joseph Shiloacha,*, Rephael Fassb
a
Biotechnology Unit, Bldg. 14A Rm. 173, NIDDK, NIH Bethesda, MD 20892-5522, USA
b
Department of Biotechnology, IIBR, Ness-Ziona 74100, Israel
Received 25 January 2005; received in revised form 31 March 2005; accepted 11 April 2005

Abstract

E. coli is the major bacterial platform for expressing simple heterologous proteins. Growing E.
coli to high densities has been the subject of numerous studies since the early 1970s, exploring the
limits of bacterial culture density in order to achieve maximum productivity. Research strategies
were focused on improving the cultivation techniques, manipulating the bacteria’s physiology or
both. As a result, batch, fed batch and dialysis fermentation techniques had been developed. These
growth strategies, together with optimization of media composition and the application of
molecular biology methods, made it possible to grow E. coli to cell densities of up to 190 g/l (dry
weight), while avoiding media precipitation and preventing acetate accumulation. Additional
research on the effects of heterologous protein biosynthesis on signal transduction, proteolysis and
post transcription events in E. coli may improve its productivity.
Published by Elsevier Inc.

Keywords: E. coli; Growth strategies; Acetate excretion; High density

Contents

1. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
2. Cell density limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347

* Corresponding author. Tel.: +1 301 496 9719; fax: +1 301 451 5911.
E-mail address: yossi@nih.gov (J. Shiloach).

0734-9750/$ - see front matter. Published by Elsevier Inc.

doi:10.1016/j.biotechadv.2005.04.004
346 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

3. Factors that affect high-density growth of E. coli . . . . . . . . . . . . . . . . . . . 347

3.1. Oxygen considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
3.2. Medium considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
3.3. Acetate inhibition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
4. Growth techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
4.1. Fed batch technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 350
4.2. Dialysis fermentation techniques . . . . . . . . . . . . . . . . . . . . . . . . 352
5. Using molecular biology to improve high-density E. coli growth . . . . . . . . . . . 352
5.1. Strategy to reduce acetate accumulation . . . . . . . . . . . . . . . . . . . . 352
5.2. Avoiding stress response . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
6. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355

1. Introduction

Growing E. coli to high density is currently the method of choice for the production of
recombinant proteins, mainly because of the high volumetric productivity associated with
this method. Exploring the growth limits of microorganisms in general and E. coli in
particular, engaged industrial microbiologists many years before it was possible to convert
E. coli to a bproduction machineQ for heterologous proteins. Today, techniques to obtain
the highest possible density of productive E. coli in submerged cultures are well
developed.
Current commercial products obtained from E. coli cultures include mainly
recombinant proteins from prokaryotic and eukaryotic sources which are considered to
be low-volume–high-value products (Riesenberg et al., 1990). In addition, recent
advancements in metabolic engineering made it possible to use E. coli as a platform to
produce high-volume–low-value products. Products such as poly-hydroxy-butyrate,
succinic acid, octanoic acid, aromatic compounds, ethanol, acetone and styrene oxide are
a few examples of the latter (Lee, 1996). For high-volume–low-values products, high
cell density and high volumetric yield are essential conditions for economical feasibility.
For low-volume–high-value products high cell density significantly reduces capital
investment and operation costs of the GMP production facilities. This reduction in cost
is achieved due to reduction in size of the fermentation equipment, upstream utilities
such as purified water, purified steam, clean air supply and clean room environments.
The size of downstream process units such as centrifuges, micro and ultra filtration
devices, and purification apparatus is also being reduced. Hence, obtaining high-density
cultures and improving the volumetric productivity is a major objective of any E. coli-
based process.
Early studies on high cell density growth of aerobic gram-negative bacteria, including
E. coli were performed either to investigate the limits of bacterial growth in liquid
cultures (Hestrin et al., 1943; Gorelick et al., 1951; Vinet and Fredette, 1951; Tyrrell et
al., 1957; Gerhardt and Gallup, 1963) or to obtain large quantities of exponentially grown
E. coli needed for biochemical studies (Bauer and Shiloach, 1974; Shiloach and Bauer,
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 347

1975; Bauer and Ziv, 1976). During the 1980s, when much information on the genetics
and physiology of the bacterium accumulated and E. coli became the obvious organism
of choice for recombinant protein production (Lee, 1996), much more emphasis was put
on its high-density growth. Since then, numerous methods to obtain high-density cultures
have been developed, each aiming at providing means to bypass the physiological
constrains that prevent bacteria from growing to the limit of physical barriers between
solid state and liquid suspension of cells.

2. Cell density limit

Earlier work attempting to achieve concentrated cultures of bacteria were done

between 1943 and 1953 following the observation that high initial substrate
concentration is a major obstacle for achieving high-density growth of bacteria in
liquid media. Hestrin et al. (1943); Gorelick et al. (1951); and Vinet and Fredette (1951)
propagated bacteria using a cellophane sac to separate the culture from a large nutrient
supply. This idea was developed further by Tyrrell et al. (1957) who used a biphasic
system to obtain bheavy-cream-looking Serratia marcescens culturesQ with cell
concentrations of 6
1011 viable cells/ml—corresponding to 15.5 dry cell weight (g/l
dcw). The biphasic systems consisted of a highly aerated flask containing a rich medium
agar layer (playing the role of the bnutrient sacQ), covered by liquid rich medium at an
optimal solid / liquid ratio of 4 : 1. A few years later in 1963, Gallup and Gerhardt
(Gerhardt and Gallup, 1963; Gallup and Gerhardt, 1963) grew S. marcescens, trying to
create, in their words, bvirtually unlimited population density in liquid cultureQ. They
implemented a dialysis system to effectively remove low molecular weight molecules
from the growth medium and deliver a fresh supply of nutrients, including oxygen, to
the growing culture. They were the first to develop the concept of dialysis fermentation,
which, according to them, bcould be scaled up to any desired levelQ. Using supplemental
feeding in reservoir-dialysis, they grew cultures for 48 h achieving a cell density of
5.6
1012 viable bacteria/ml (148 g/l dcw) and a partial cell volume of 50% (Gallup and
Gerhardt, 1963). The cell density obtained by Gallup and Gerhardt is very close to the
maximal possible density of an E. coli liquid culture, which today is estimated to be
about 200 g/l dcw. According to Lee (1996), E. coli culture fluidity is lost when the dry
cell weight is higher than 220 g/l. Shiloach and Bauer (1975) mentioned in 1974, that
b. . .the limit of yield of dry bacterial cells per liter is well below the value of 250 g
which is the approximate amount of dry materials in packed cells obtained in a tube type
centrifugeQ. Märkl et al. (1993) and Lee and Lee (1996) reported E. coli cell density of
175 g/l dcw. Nakano et al. (1997) and Fuchs et al. (2002) reported a record E. coli cell
density of 190 g/l dcw.

3. Factors that affect high-density growth of E. coli

The idea that the only obstacle to unlimited growth of bacteria in liquid culture is
the medium liquidity is based on the assumption that bacteria would continue to
348 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

divide as long as their nutritional requirements are met and no toxic or inhibitory
factors are excreted into the medium by the bacterial community throughout the course
of growth. The last three decades witnessed a continuous effort, conducted by many
research groups, to develop growth media, aeration techniques, monitoring devices,
controlled fermentation systems, and culturing strategies for the purpose of achieving
high-density cultures of E. coli and other bacteria. All these means were designed to
establish growth conditions that keep every single cell in a growing E. coli culture
bsatisfied and happyQ. Potential inhibitory factors are described below.

3.1. Oxygen considerations

Results published in 1965 by Herbert et al. (1965), and later confirmed by Bauer
and coworkers in 1974–1976 (Bauer and Shiloach, 1974; Shiloach and Bauer, 1975;
Bauer and Ziv, 1976), indicate that the yield-constant for oxygen of an E. coli w
strain growing on glucose as a carbon source is close to 1 (1 g of oxygen is needed
for the production of 1.06 g of E. coli biomass). Since an average fermentation system
at 25 8C can supply 1.5 M O2 l/h when aerated with pure oxygen, E. coli culture
could theoretically grow up to a cell density of 350 g/l dcw at a growth rate of 0.2
h 1, provided all other nutritional requirements are met. This value is above the 250
g/l dcw limit based on fluidity. Therefore, if pure oxygen supply is available the
oxygen transfer rate in liquid cultures would not become a limiting factor.

3.2. Medium considerations

The growth of E. coli can be limited by other nutritional requirements including

carbon, nitrogen, phosphorus, sulfur, magnesium potassium iron, manganese, zinc,
copper and some growth factors (Gunsalus and Stanier, 1960). When growing E. coli
to low density, all the required nutrients can be added initially into the basal broth. The
popular complex Luria Bertani (LB) broth allows the growth of E. coli in a
temperature-, pH-, and oxygen-controlled environment up to a cell density of 1 g/l dcw.
To accommodate the nutritional requirements of denser cultures, concentrations of
media components must be increased and phosphorus, sulfur and trace elements must
be added. Unfortunately, most media ingredients required for growth become inhibitory
to E. coli when added at high concentrations. It is well established (Riesenberg et al.,
1991) that nutrients such as glucose at a concentration of 50 g/l, ammonium at 3 g/l,
iron at 1.15 g/l, magnesium at 8.7 g/l, phosphorus at 10 g/l and zinc at 0.038 g/l
inhibit E. coli growth. A defined medium that contains the maximum non-inhibitive
concentration of nutrients allows growth of E. coli to a cell density of about 15 g/l dcw
(Lee, 1996).
Another concern is the precipitation of media components that can hamper the
adequate supply of nutrients and interfere with the fermentation process and the
monitoring devices. Precipitates can also affect downstream recovery and purification
operations. Precipitation occurs when non-soluble complexes of divalent metal-
ammonium phosphates, magnesium phosphates and other metal phosphates (Dean,
1979) are formed. The relative concentration of different anions (phosphates, sulfates
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 349

and chloride) and cations (sodium, calcium, ammonium and magnesium) in a medium
would determine if precipitation occurs during media preparation and sterilization.
Precipitation may also form during the fermentation process due to production of
organic acids and carbon dioxide.
Another concern is the osmotic pressure and conductivity caused by ion concentration
in the growth media that may affect membrane potential and may activate different stress
mechanisms that induce a decrease in growth rate or termination of the growth cycle
(Winzer et al., 2002). Using non-precipitating concentrations of most needed ions in a
defined medium allows the growth of E. coli to a cell density of 1 g/l dcw (Curless et
al., 1996). Curless et al. (1996) suggested eliminating precipitation problems by using
slow release of phosphate ions into the growth medium from a non-soluble poly-
phosphate glass.
From the above data, it is clear that a well-designed medium and feeding
strategies are needed for high-dense growth of E. coli. In 1975 Shiloach and Bauer
(1975) designed a semi-defined medium that is based on nutritional requirements of
E. coli described by Herbert et al. (1965) and consisted of potassium phosphate,
ammonium phosphate, magnesium sulfate, glucose, yeast extract and a solution of
trace elements. Phosphorous magnesium precipitates were avoided by including the
magnesium salts in the glucose feed. This medium supported the fast exponential
growth of E. coli to a cell density of 54 g/l dcw. Neidhardt et al. (1974) developed
a defined minimal nutrient medium containing only simple inorganic salts and a
defined carbon source. To avoid the increase in the culture volume due to the
addition of nutrients, Matsui et al. (1989) improved Neidhardt’s method by adding
solid glucose powder to a growing culture achieving cell density of 134 g/l dcw.
Riesenberg et al. (1990) added citric acid, EDTA and thiamine-HCl to the defined
medium (containing only simple inorganic salts and a defined carbon source). Using
fed batch technique that maintained linear growth pattern, he added feeding solution
containing 80% glucose, 2% MgSO4 and trace elements and achieved 110 g/l dcw.
Korz et al. (1995) used this medium to obtain even a higher cell density by
substituting glucose with glycerol in the feed (glycerol’s solubility is higher).
Macaloney et al. (1996) used a complex medium, similar to the LB medium
supplemented with glycerol, to grow E. coli to a cell density of 86 g/l. These
defined and complex media were designed and used as part of gradual feeding
strategies that were found to be the solution to growth inhibition and for improving
media utilization. In summary, continuous or intermittent feeding can satisfy the
increasing demand for energy, carbon and nitrogen sources by an exponentially or
linearly growing E. coli culture.

3.3. Acetate inhibition

The need for adequate nutrient supply served as the obvious motivation to develop
sophisticated feeding techniques for a variety of microorganisms. The strongest
motivation to develop the fed batch and dialysis methods to achieve higher density of
E. coli culture was to decrease acetate accumulation, since high acetate concentration
can inhibit growth and recombinant protein production (Majewski and Domach, 1990;
350 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

Han et al., 1992; Konstantinov et al., 1990; Luli and Strohl, 1990; Bentley et al., 1990;
Ko et al., 1995; Shimizu et al., 1991).
Acetate metabolism and its effect on E. coli were studied extensively during the
last three decades. In 1982, Doelle et al. (1982) described acetate excretion under
aerobic conditions, resulting from excess carbon source, especially glucose. In 1984,
Meyer et al. (1984) defined conditions for acetate formation in a continuous culture of
E. coli grown on defined and complex media. Between 1985 and 1990, Zabriskie and
Arcuri (1986), El-Mansi and Holms (1989) and Bauer et al. (1990) attributed acetate
accumulation to rapid growth on glucose. They defined conditions under which the
control of carbon flux in E. coli leads to lower excretion of acetate. Majewski and
Domach (1990) cited overloading of the TCA cycle by fast oxidation through
glycolysis as the main reason for acetate accumulation. In 1992, Han et al. (1992)
attributed acetate accumulation to both saturation of the TCA cycle and saturation of
the electron transport process. Chen and Bailey (1993) showed that acetate could act
as an uncoupler between phosphorylation and electron transport via the TCA cycle.
The accumulated information on acetate metabolism indicates that acetate may have
detrimental effect on process performance. This was reported in 1990 by Konstantinov et
al. (1990), Luli and Strohl (1990), Bentley et al. (1990), and later by Johnston et al.
(2002), Han et al. (1992), Ko et al. (1995) and Shimizu et al. (1991). During the early
1990s, it became a general consensus that acetate accumulation above 2 g/l in the growth
media slows down E. coli growth, may stop biomass build up and may inhibit
recombinant-protein biosynthesis. As a result, propagation techniques that keep acetate
concentration below 1.5–2 g/l were adopted and further developed.

4. Growth techniques

Two major growth techniques were developed for high-density growth, fed batch and
dialysis. The high cell densities reported for various versions of these techniques are
presented in Table 1.

4.1. Fed batch technique

One way to lower acetate accumulation is to reduce the growth rate by supplying the
glucose slowly. Riesenberg suggested in 1991 (Riesenberg, 1991) that by implementing
this strategy, the TCA cycle can handle all the acetyl CoA produced by the glycolytic
pathway, thus eliminating acetate formation.
Several fed batch methods for carbon source feeding that can maintain low acetate
levels were developed. These methods are based on mathematical models that describe
growth patterns and the expected demand for nutrients. The models are based on the
generally accepted average value of E. coli yield on glucose (0.5 g/g) and the elemental
analysis of E. coli cells (C : N : P : S ratio). Using such models, an optimal concentration
of all components and feed rates can be calculated, and the carbon feed becomes
limiting and determines the growth rate. Keeping the growth rate below a critical value
for acetate accumulation allows the culture to grow up to the limits set by other
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 351

Table 1
Highest cell densities by various propagation techniques of E. coli
Propagation technique Basic medium C source Yield; g dcw/l Reference
Fed batch
1. Exponential growth Semi-defined Glucose 54 Shiloach and Bauer,
Medium yeast 1975
Extract

2. Carbon source Defined Solid glucose 134 Neidhardt et al.,

defended Minimal medium 1974;
linear growth Matsui et al., 1989

Defined Glucose 104 Riesenberg et al.,

Minimal medium Citric acid 1990
Defined Glycerol 148 Korz et al., 1995
Minimal medium Glucose 128

Protein hydrolyzate Glycerol 84 Macaloney et al.,

and yeast extract 1996
3. Slow linear growth Defined minimal Low glucose 145 Horn et al., 1996
to keep acetate medium Glycerol
concentration close
to zero

Dialysis
1. Membrane dialysis Against complete Glycerol 174 Märkl et al., 1993
reactor growth medium
Against basal medium Glycerol 190 Nakano et al., 1997
2. bNutrient-splitQ Against buffer salt Separate glycerol 150 Ogbonna and Märkl,
feeding solution feeding 1993

parameters. Typical equation to determine the desired flow rate was proposed by Lee
(1996).
   
l l
Ms ðt Þ ¼ F ðt ÞSF ðt Þ ¼ þ m X ðt ÞV ðt Þ ¼ þ m X ðt0 ÞV ðt0 Þexp½lðt  t0 Þ
YX =S YX =S

where: M s Mass-flow rate of the carbon source (g/h); F Feed flow rate (l/h); S F Carbon-
substrate concentration in the feed (g/l); X Cell concentration (g/l dcw); m Specific
maintenance coefficient (g/g dcw/h); V Culture volume (l); t 0 Time of feeding start;
t Process time; l Specific growth rate (l/h); Y X / S Cell yield on carbon substrate (g/g).
Such an equation may be used to pre-determine linear or exponential addition rates of
nutrients by connection of feed pumps to a controller. The controller calculates the amount
of carbon needed and adjusts the feed rate accordingly (Riesenberg and Guthke, 1999;
Lee, 1996).
Another feeding approach is the use of direct and indirect feedback control systems.
Indirect feedback control is based on on-line monitoring of parameters such as pH,
dissolved oxygen, CO2 evolution rate and cell concentration (optical density, NAD, ATP).
352 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

Direct feedback is based on monitoring the concentration of the major carbon substrate for
controlled addition of nutrients (Bauer and Ziv, 1976; Konstantinov et al., 1990;
Riesenberg et al., 1991; Korz et al., 1995; Sakamoto et al., 1994; and Turner et al., 1994).
Horn et al. (1996) obtained a density of 145 g/l dcw E. coli by keeping close control on the
concentration of all major nutrients (such as glucose, nitrogen and phosphate). They kept
the dissolved oxygen at above 20% saturation during the whole course of the culture
growth. The feedback control was designed to keep the glucose concentration non-
limiting, below 1.5 g/l. The culture grew at its maximal growth rate at 26 8C with no
accumulation of acetate.

4.2. Dialysis fermentation techniques

Dialysis fermentation is another way to overcome the inhibitory effect of acetate and
other nutrients and to obtain high-density growth of E. coli. Dialysis is defined as the
separation of solute molecules by their unequal diffusion through a semi-permeable
membrane based on a concentration gradient. Gallup and Gerhardt first demonstrated this
concept in 1963 (Gallup and Gerhardt, 1963) for the growth of S. mercedens. Several
groups of investigators adopted the use of dialysis fermentation for E. coli culture to
remove acetate from the growth medium, to minimize precipitation and to prevent
substrate inhibition (Pörtner and Märkl, 1998). Two configurations of vessel arrangement
were proposed for dialysis reactors: 1) two-vessel reactor consisting of a culture reactor
and a medium reservoir connected by a dialysis device; 2) a single-vessel dialysis reactor
consisting of two chambers (e.g. 5 l outer chamber and a 2 l inner chamber) separated by a
dialysis membrane. The single-vessel arrangement is less preferable because it is difficult
to sterilize, and sensitive to mechanical stress and oxygen limitation. Record high cell
densities were obtained by Märkl et al. (1993) and Nakano et al. (1997) by using
membrane dialysis reactors to grow E. coli to 174 and 190 g/l dcw, respectively. However,
since growth medium was used as the dialyzing fluid, large amounts of glycerol were
wasted in the effluent. Ogbonna and Märkl (1993) proposed to reduce glycerol use by
implementing a bnutrient-splitQ, a feeding technique in which glycerol was added directly
to the propagation chamber and the dialysis was done with a buffered salt solution.

5. Using molecular biology to improve high-density E. coli growth

5.1. Strategy to reduce acetate accumulation

The direct approach to overcome acetate inhibition was taken up by several researchers
who implemented a variety of recombinant strategies to decrease acetate biosynthesis,
increase acetate utilization and divert glycolysis to other pathways.
One approach used E. coli mutants with defects in the acetate biosynthesis pathway
enzymes, specifically in the phosphotransacetylase (pta) and acetate kinase (ack). Contiero
et al. (2000) generated a collection of mutants of E. coli w3100, each lacking acetate
associated enzymatic activity, and used them to prove that all aspects of acetate
metabolism play an integral role in growth of E. coli to high cell densities. Bauer et al.
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 353

(1990) used pta mutant E. coli K, to obtain improved cell density and IL-2 production.
Diaz-Ricci et al. (1991) suggested that ack and pta mutants grow better than E. coli K in
high glucose batch fermentation. Kakuda et al. (1994) showed that although ack E. coli
mutant growth on glucose was impaired, acetate secretion and uptake were still high,
indicating that a pathway, other than the Pta–Ack, functions in the control of excess carbon
flow in the mutants.
An opposite approach to minimize acetate accumulation was to enhance acetate
utilization by the growing culture. Acetate is metabolized via acetyl CoA, the glyoxylate
shunt pathway and the TCA cycle (Oh et al., 2002). The enzymes involved are normally
repressed when glucose is present. It was logical to assume that mutants lacking this
repression would consume the acetate along with the consumption of glucose. Farmer and
Liao (1997) were successful in reducing acetate concentrations four-fold by redirecting the
carbon flux through phospenolpyruvate and the glyoxlate shunt pathway. They used
recombinant strains with over-expression of the phospenolpyruvate carboxylase gene that
lacked one of the glyoxylate shunt repressor peptide fadR. Shiloach et al. (1996) found
that the widely used E. coli B (BL21) accumulated significantly less acetate than E. coli K
(JM109). It was hypothesized that the glyoxylate shunt is active in the low acetate
producer and is inactive in the high acetate producer. This hypothesis was later confirmed
by metabolic flux analysis (Van de Walle and Shiloach, 1998) and by 13C NMR/MS
analysis (Noronha et al., 2000).
Another strategy to prevent acetate accumulation was to divert part of the flow of
glucose in the cell from the TCA cycle. Dedhia et al. (1994) used induced glycogen
overproduction in ack–pta-double mutants to divert part of the glucose phosphate flow
from the glycolysis–TCA pathway and eliminated acetate accumulation in the growth
medium. Using this strategy, they achieved higher biomass yield and reduced pyruvate
accumulation as compared to un-induced cultures. Aristidos et al. (1994) diverted the
carbon flow from pyruvate to aceto-lactate and acetoin by cloning the Baccilus subtilis
acetolactate synthase in E. coli. Similarly, Bermejo et al. (1998) cloned a Clostridial
constitutive acetone operon in E. coli K, thus continuously activating a process that
converts acetate to acetone which was stripped off by aeration and agitation due to its high
volatility. In 1994 Chou et al. (1994) used methyl alpha-glucoside (alpha-MG), a
metabolically inert glucose analog that shares the same phosphotransferase system and
acts as a nontoxic competitive inhibitor, to modulate glucose uptake and subsequently
reduce drastically the acetate accumulation rate in the medium.

5.2. Avoiding stress response

Cells in high-density cultures may become stressed rapidly due to: lack of nutrients,
elevated osmotic pressure caused by feeding, and induction of heterologous protein
expression. An example of response to stress is the filamentation phenomena in E. coli,
which occurs when foreign proteins are expressed. This phenomena is more important in
cases of constitutive products. The expression of foreign proteins enhances the biosyn-
thesis of the repressor of the cell division protein FtsZ and was found to hamper the
productivity of a fermentation process for poly-hydroxybutyrate. Wang and Lee (1998)
overcame the effect of the repressor by inserting a gene for over-expressing FtsZ. The
354 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

continuous expression of FtsZ allowed unconditional cell division and consequently, high-
density growth and high poly-hydroxybutyrate accumulation.
A more traditional approach to overcome the stress response obstacle was taken by
Weikert et al. (1997) who isolated and selected mutants more suitable for high-density
growth. These mutants have a higher specific growth rate, increased biomass yield,
reduced secretion of overflow metabolites, and increased resistance to a variety of adverse
conditions including heat shock, osmotic stress and nutrient deprivation. A three fold
increase in expressing Bacillus stereothermophilus á amylase was observed from the
mutant CWML2:pCSS4-p (Weikert et al., 1998).

6. Concluding remarks

Growing bacteria to the theoretical maximal cell density started as a challenging

scientific endeavor during the midst of the 20th century. It became a practical when
advancements in molecular biology made it possible to express heterologous proteins in
bacterial hosts, mainly in E. coli. The purpose of growing E. coli to high density is to
achieve the highest possible volumetric productivity of recombinant proteins. Most of the
various techniques that had been developed towards this goal were designed to allow
uninterrupted linear growth at a slow growth rate, and to avoid precipitation of salts. The
key physical element in high-density fermentation systems is adequate oxygen supply to
the growing culture, which can be achieved only by supplementing or substituting the air
supply with pure oxygen. But even with pure oxygen, it is possible to obtain high-density
E. coli growth only if either temperature, availability of other nutrients, or both limit the
growth rate. Exponential growth of high-density E. coli culture cannot exceed 50 g/ l dcw
(Bauer and Shiloach, 1974; Shiloach and Bauer, 1975; Bauer and Ziv, 1976); on the other
hand, slow growing cultures can grow to densities as high as 190 g/l dcw, provided key
metabolic obstacles are resolved (Nakano et al., 1997). The main physiological obstacle
for high-density growth was found to be acetate accumulation. This problem was dealt
with by developing sophisticated feeding strategies and/or dialysis protocols, and by
adopting metabolic engineering approaches.
Evaluation of growth methods is based on the overall productivity of the process, which
is a function of the specific production (gram product/gram biomass) and the total amount
of biomass. When dealing with recombinant protein production, the overall productivity
depends on numerous factors such as plasmid stability, promoter response to inducer, post-
transcriptional inhibition events, and post-translation inhibition caused by proteolysis and
improper folding. Several works demonstrated that it is difficult to predict the yield and
stability of the final product in high-density cultures (Weikert et al., 1998; Curless et al.,
1996), yet others found no effect of culture density on the final product concentration or its
quality (Ryan et al., 1996; Zabriskie et al., 1987; Shin et al., 1997).
Downstream processing considerations are also important in choosing the growth
method and the desired final cell density. High-density cultures may hamper the operation
of micro- and ultra-filtration units due to high concentration of antifoam-derived surface-
active agents. Cell lysis products, which are more common in cases of high density, may
contribute to product impurities and can interfere with the purification process.
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 355

Economical considerations play a significant role when large-scale operations are

considered. These include longevity of the process, loss of nutrients, safety and cost of
working with pure oxygen and pressurized fermentors, mechanical load on agitation
system, efficiency of cooling, heating and mass transfer, sensing and probing limitations
and equal distribution of ambient conditions in the whole culture.
A close examination of past publications on high-density techniques for E. coli cultures
reveal a wide array of final cell concentrations ranging from 20 to 190 g/l dcw. This
significant variability can be attributed to the fact that the growth of different strains may
be controlled by parameters not completely understood such as stress response and
quorum sensing in E. coli (Winzer et al., 2002). Better understanding of these parameters
may contribute to the development of improved growth methods and recombinant protein
expression.

Acknowledgements

The authors would like to thank D. Livant for careful reading of the manuscript and
L. Trinh for assisting with the preparation of the manuscript.

References

Aristidos AA, San KY, Bennett GN. Modification of central metabolic pathway in Escherichia E coli to reduce
acetate accumulation by heterologous expression of the Bacillus subtilis acetolactate synthase gene.
Biotechnol Bioeng 1994;44:944 – 51.
Bauer KA, Ben Bassat A, Dawson M, De la Puente VT, Neway JO. Improved expression of human interleukin-2
in high cell density fermentor cultures of Escherichia coli k-12 by a phosphotransacetylase mutant. Appl
Environ Microbiol 1990;56:1256 – 302.
Bauer S, Shiloach J. Maximal exponential growth rate and yield of E coli obtainable in a bench-scale fermentor.
Biotechnol Bioeng 1974;16:933 – 41.
Bauer S, Ziv E. Dense growth of aerobic bacteria in a bench-scale fermentor. Biotechnol Bioeng 1976;18:81 – 94.
Bentley WE, Mirjalili N, Andersen DC, Davis RH, Kompala DS. Plasmid-encoded protein: the principal factor in
the bmetabolic burdenQ associated with recombinant bacteria. Biotechnol Bioeng 1990;35:668 – 81.
Bermejo LL, Welker NE, Papoutsakis ET. Expression of Clostridium acetobutylicum ATCC 824 genes in
Escherichia coli for acetone production and acetate detoxification. Appl Environ Microbiol 1998;
64:1079 – 85.
Chen R, Bailey JE. Observations of an aerobic, growing Escherichia coli metabolism using an on-line nuclear
magnetic resonance spectroscopy system. Biotechnol Bioeng 1993;142:215 – 21.
Chou CH, Bennett GN, San KY. Effect of modified glucose uptake using genetic engineering techniques on high-
level recombinant protein production in Escherchia coli dense cultures. Biotechnol Bioeng 1994;44:952 – 60.
Contiero J, Beatty C, Kumari S, DeSanti CL, Strohl WR, Wolfe A. Effects of mutations in acetate metabolism on
high-cell-density growth of Escherichia coli. J Indus Microbiol Biotechnol 2000;24:421 – 30.
Curless C, Baclaski J, Sachdev R. Phosphate glass as a phosphate source in high cell density Escherichia coli
fermentations. Biotechnol Prog 1996;12:22 – 5.
Dean JA. Lang’s handbook of chemistry. 12th ed. New York (NY)7 McGraw-Hill; 1979. p. 5.7 – .12.
Dedhia NN, Hottinger T, Bailey JE. Overproduction of glycogen in Escherichia coli blocked in the acetate
pathway improves cell growth. Biotechnol Bioeng 1994;44:132 – 9.
Diaz-Ricci JC, Regan L, Baily JE. Effect of the alteration of the acetic acid synthesis pathway on the fermentation
pattern of Escherichia coli. Biotechnol Bioeng 1991;38:1318 – 24.
356 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357

Doelle HW, Ewings KN, Hollywood NW. Regulation of glucose metabolism in bacterial systems. In: Fiechter A,
editor. Adv Biochem Eng8 vol 23. N.Y.7 Springer Verlag; 1982. p. 1 – 35.
El-Mansi EMT, Holms WH. Control of carbon flux to acetate excretion during growth of Escherichia coli in
batch and continuous culture. J Gen Microbiol 1989;135:2875 – 83.
Farmer WR, Liao JC. Reduction of aerobic acetate production by Escherichia coli. Appl Environ Microbiol
1997;63:3205 – 10.
Fuchs C, Koster D, Wiebusch S, Mahr K, Eisbrenner G, Märkl H. Scale-up of dialysis fermentation for high cell
density cultivation of Escherichia coli. J Biotechnol 2002;93:243 – 51.
Gallup DM, Gerhardt P. Dialysis fermentor systems for concentrated culture of microorganisms. Appl Microbiol
1963;11:506 – 12.
Gerhardt P, Gallup DM. Dialysis flask for concentrated culture of microorganisms. J Bacteriol 1963;86:919 – 29.
Gorelick AN, Mead DD, Kelly EH. The growth of bacteria in a charcoal–cellophane system. J Bacteriol
1951;61:507 – 13.
Gunsalus IC, Stanier RY, 1960. The bacteria, vol 1. New York7 Academic Press, Inc.; 1960. p. 14 – 6.
Han K, Lim HC, Hong J. Acetic acid formation in Escherichia coli fermentation. Biotechnol Bioeng
1992;39:663 – 71.
Herbert D, Phipps PJ, Tempest DW. The chemostat: design and instrumentation. Lab Pract 1965;14:1150 – 61.
Hestrin S, Avineri-Shapiro S, Aschner M. The enzymic production of levan. Biochem J 1943;37:450 – 6.
Horn U, Strittmatter W, Krebber A, Knupfer U, Kujau M, Wenderoth R, et al. High volumetric yields of
functional dimeric miniantibodies in Esherichia coli, using an optimized expression vector and high-cell-
density fermentation under non-limited growth conditions. Appl Microbiol Biotechnol 1996;46:524 – 32.
Johnston W, Cord-Ruwisch R, Cooney MJ. Industrial control of recombinant E. coli fed batch culture: new
perspectives on traditional controlled variables. Bioprocess Biosyst Eng 2002;25:111 – 20.
Kakuda H, Shiroishi K, Hosono K, Ichihara S. Construction of Pta–Ack pathway deletion mutants of Escherichia
coli and characteristic growth profiles of the mutants in a rich medium. Biosci Biotechnol Biochem
1994;58:2232 – 5.
Ko YF, Bentley WE, Weigand WA. The effects of cellular energetics on foreign protein production. Appl
Biochem Biotechnol 1995;50:145 – 59.
Konstantinov K, Kishimoto M, Seki T, Yoshida T. A balanced DO-stat and its application to the control of acetic
acid excretion by recombinant Escherichia coli. Biotechnol Bioeng 1990;36:750 – 8.
Korz DJ, Rinas U, Hellmuth K, Sanders EA, Deckwer WD. Simple fed-batch technique for high cell density
cultivation of Escherichia coli. J Biotechnol 1995;39:59 – 65.
Lee SY. High cell density culture of Escherichia coli. Trends Biotech 1996;14:98 – 105.
Lee Y, Lee SY. Enhanced production of poly (3-hydroxy butyrate) by filamentation suppressed recombinant
Escherichia coli in a defined medium. J Environ Polym Degrad 1996;4:131 – 4.
Luli GW, Strohl WR. Comparison of growth, acetate production, and acetate inhibition of Escherichia coli strains
in batch and fed-batch fermentations. Appl Environ Microbiol 1990;56:1004 – 11.
Macaloney G, Draper I, Preston J, Anderson KB, Rollins MJ, Thompson BG, et al. At-line control and fault
analysis in an industrial high cell density Escherichia coli fermentation, using NIR spectroscopy. Trans Inst
Chem Eng 1996;74:212 – 20.
Majewski RA, Domach MM. Simple constrained-optimization view of acetate overflow in E. coli. Biotechnol
Bioeng 1990;35:732 – 8.
Märkl H, Zenneck C, Dubach ACH, Ogbonna JC. Cultivation of Escherichia coli to high cell densities in a
dialysis reactor. Appl Microbiol Biotechnol 1993;39:48 – 52.
Matsui T, Yokota H, Sato S, Mukataka S, Takahashi J. Pressurized culture of Escherichia coli for a high
concentration. Agric Biol Chem 1989;53:2115 – 20.
Meyer HP, Leist C, Fiechter A. Acetate formation in continuous cultures of Escherichia coli K12 D1 on defined
and complex media. J Biotechnol 1984;1:355 – 8.
Nakano K, Rischke M, Sato S, Märkl H. Influence of acetic acid on the growth of Escherichia coli K12 during
high-cell-density cultivation in a dialysis reactor. Appl Microbiol Biotechnol 1997;48:597 – 601.
Neidhardt FC, Bloch PL, Smith DF. Culture medium for enterobacteria. J Bacteriol 1974;119:736 – 47.
Noronha SB, Yeh HJC, Spande TF, Shiloach J. Investigation of the TCA cycle and the glyoxylate shunt in E coli
BL21 and JM 109 using 13C-NMR/MC. Biotechnol Bioeng 2000;68:316 – 27.
 J. Shiloach, R. Fass / Biotechnology Advances 23 (2005) 345–357 357

Ogbonna JC, Märkl H. Nutrient split feeding strategy for dialysis cultivation of Escherichia coli. Biotechnol
Bioeng 1993;41:1092 – 100.
Oh MK, Rohlin L, Kao KC, Liao JC. Global expression profiling of acetate-grown Escherichia coli. J Biol Chem
2002;277:13175 – 83.
Pörtner R, Märkl H. Dialysis cultures. Appl Microbiol Biotechnol 1998;50:403 – 14.
Riesenberg D, Menzel K, Schulz V, Schumann K, Veith G, Zuber G, et al. High cell density fermentation of
recombinant Escherichia coli expressing human interferon alpha-1. Appl Microbiol Biotechnol 1990;
34:77 – 82.
Riesenberg D. High-cell density cultivation of E coli. Curr Opin Biotechnol 1991;2:380 – 4.
Riesenberg D, Guthke R. High cell density cultivation of microorganisms. Appl Microbiol Biotechnol
1999;51:422 – 30.
Riesenberg D, Schulz V, Knorre WA, Pohl HD, Korz D, Sanders EA, et al. High cell density cultivation of
Escherichia coli at controlled specific growth rate. J Biotechnol 1991;20:17 – 28.
Ryan W, Collier P, Loredo L, Pope J, Sachdev R. Growth kinetics of Escherichia coli and expression of a
recombinant protein and its isoforms under heat shock conditions. Biotechnol Prog 1996;12:596 – 601.
Sakamoto S, Iijima M, Matsuzawa H, Ohta T. Production of thermophilic protease by glucose-controlled fed-
batch culture of recombinant Escherichia coli. J Ferment Bioeng 1994;78:304 – 9.
Shiloach J, Bauer S. High yield growth of E coli at different temperatures in a bench scale fermentor. Biotechnol
Bioeng 1975;17:227 – 39.
Shiloach J, Kaufman J, Guillard AS, Fass R. Effect of glucose supply strategy on acetate accumulation, growth,
and recombinant protein production by Escherichia coli BL21(DE3) and Escherichia coli JM109. Biotechnol
Bioeng 1996;49:421 – 8.
Shimizu N, Fukuzono S, Harada Y, Fujimori K, Gotoh K, Yamazaki Y. Mass production of human epidermal
growth factor using fed-batch cultures of recombinant Escherichia coli. Biotechnol Bioeng 1991;38:37 – 42.
Shin CS, Hong MS, Bae CS, Lee J. Enhanced production of human mini-proinsulin in fed-batch cultures at high
cell density of Escherchia coli BL21(DE3) [pET-3aT2M2]. Biotechnol Prog 1997;13:249 – 57.
Turner C, Gregory ME, Thornhill NF. Closed-loop control of fed-batch cultures of recombinant Escherichia coli
using on-line HPLC. Biotechnol Bioeng 1994;44:819 – 29.
Tyrrell EA, Mac Donald RE, Gerhardt P. Biphasic system for growing bacteria in concentrated culture. J Bacteriol
1957;75:1 – 4.
Van de Walle M, Shiloach J. Proposed mechanism of acetate accumulation in two recombinant Escherichia coli
strains during high cell density fermentation. Biotechnol Bioeng 1998;57:71 – 8.
Vinet G, Fredette V. Apparatus for the culture of bacteria in cellophane tubes. Science 1951;114:549 – 50.
Wang F, Lee SY. High cell density culture of metabolically engineered Escherichia coli for the production of
poly(3-hydroxybutyrate) in a defined medium. Biotechnol Bioeng 1998;58:325 – 8.
Weikert C, Sauer U, Baily JE. Use of glycerol-limited long term chemostat for isolation of Escherichia coli
mutants with improved physiological properties. Microbiology 1997;143:1567 – 74.
Weikert C, Sauer U, Baily JE. An Escherichia coli host strain useful for efficient overproduction of secreted
recombinant protein. Biotechnol Bioeng 1998;59:386 – 91.
Winzer K, Hardie KR, Williams P. Bacterial cell to cell communication: sorry, can’t talk now—gone to lunch.
Curr Opin Microbiol 2002;5:216 – 22.
Zabriskie DW, Arcuri EJ. Factors influencing productivity of fermentation employing recombinant micro-
organisms. Enzyme Microb Technol 1986;8:705 – 16.
Zabriskie DW, Wareheim DA, Polansky ML. Effects of fermentation feeding strategies prior to induction of
expression of a recombinant malaria antigen in Escherichia coli. J Ind Microbiol 1987;2:87 – 95.