The effect of antifoam addition on protein production yields

Sarah J Routledge and Roslyn M Bill

School of Life and Health Sciences, Aston University, Aston Triangle, Birmingham, B4
7ET, UK

For correspondence:

Sarah J Routledge
Tel: +44 (0) 121 204 3168
E-mail: s.j.routledge@aston.ac.uk

Abstract
Pichia pastoris is a widely-used host for recombinant protein production. The foaming
associated with culturing it on a large scale is commonly prevented by the addition of
chemical antifoaming agents or antifoams. Unexpectedly, the addition of a range of
antifoams to both shake flask and bioreactor cultures of P. pastoris has been shown to
alter the total yield of the recombinant protein being produced. Possible explanations for
this are that the presence of the antifoam increases the total amount of protein being
produced and secreted per cell or that it increases the density of the culture. Antifoaming
agents may therefore have specific effects on the growth and yield characteristics of
recombinant cultures, in addition to their primary action as de-foamers.

Keywords: Antifoam, protein, foam destruction yield, kLa, viability

1. Introduction
When producing recombinant proteins on a large scale, growth in bioreactors is often the
most convenient option. One problem with such cultures, which are typically intensely
aerated and stirred, is the formation of foam. Foaming can lead to reduced process
productivity since the bursting of bubbles may damage cells and proteins (1) and result in
loss of sterility if the foam escapes the bioreactor (2). To prevent the formation of foam,
thermal methods (3), mechanical agitation, ultrasound or most often addition of chemical
antifoaming agents or antifoams (2) is employed.

Chemical antifoams can act as surfactants and generally consist of several components.
There are many different varieties which can usually be grouped as hydrophobic solids
dispersed in carrier oil, aqueous or water-based suspensions or emulsions, liquid single
component antifoams and solid antifoams (3, 4). There are thought to be several
mechanisms of foam dispersion by chemical antifoams: bridging-dewetting, spreading
fluid entrainment and bridging-stretching (5), but the precise details are not well
understood.

It is clear that antifoams are often added to bioreactors without considering the effects on
the bioprocess, the cells or the recombinant proteins being produced. This is surprising as
antifoams are known to affect the kLa or oxygen transfer rate of a system (6-14), the
ability of cells to secrete protein and can even be used at high concentrations to boost the
yield of a recombinant protein in shake flasks (15). However, different types of antifoam
at varying concentrations can have different effects. These properties should be taken into
account when choosing the most suitable antifoam for a particular process.

1.1. Antifoams
A wide variety of antifoams is available, with different compositions and properties. One
consideration when deciding on the type of antifoam to be used is whether it is suitable
for a particular process. Certain antifoams may not be suitable for some applications,
such as producing proteins for drug development, as not all are FDA approved. Some
antifoams may also affect downstream processing, such as silicone-containing antifoams,

which can coat equipment. Several antifoams should therefore be tested to determine
which is the most appropriate for the process.

1.2. Bartsch Test

When selecting an antifoam, its primary function as a de-foamer should be evaluated. A
simple Bartsch test (16, 17) can be performed to asses foam destruction properties. This
involves shaking the medium to be used in the process to induce foam formation. When
this is done in the presence of the antifoam of interest, it is possible to assess the ability of
the agent to destroy foam over time.

1.3. kLa
The kLa, or volumetric mass oxygen transfer coefficient, is a measure of how much
oxygen is transferred into the medium over a certain amount of time (14). The kLa of a
system can be influenced by several factors such as the properties of the medium
including its viscosity, the presence of organisms and their by-products. Additions to the
medium such as antifoams also have an effect (13, 14). It has been observed that low
concentrations of antifoam can reduce the kLa but at higher concentrations the kLa may
rise (10, 12). To ensure optimum oxygen transfer within a system, the effect of differing
concentrations of the antifoam to be used should be assessed.

1.4. Influence of antifoams on protein yield

Antifoams added to shake flask cultures at higher concentrations than normally used in
bioreactors have been shown to increase the yield of recombinant green fluorescent

protein (GFP) (15). Two groups of antifoams were identified in the study, depending on
their mode of action: one that resulted in improvements in biomass yields of the cultures
and one that enhanced protein production or secretion. This finding may provide a simple
method to increase productivity in recombinant protein production experiments. It also
provides much needed insight into how antifoams interact with yeast host cells.

1.5. Effect of antifoams on cell viability

A check should be done to ensure antifoams at high concentrations are not detrimental to
the viability of cells. Flow cytometry can be performed using shake flask samples stained
with propidium iodide. In this assay, dead cells are stained red.

2. Materials
2.1. Antifoams
Antifoams from different groups can be selected depending on process requirements e.g.

Antifoam A (Sigma), 30% emulsion of silicone polymer

Antifoam C (Sigma), 30% emulsion of silicone polymer

Antifoam J673A (Struktol), an alkoxylated fatty acid ester on a vegetable base

Antifoam P2000 (Fluka), a polypropylene glycol

Antifoam SB2121 (Struktol), a polyalkylene glycol

2.2. Bartsch test

1.

Culture medium required for the process e.g. BMMY medium for Pichia pastoris
cultures is composed of 1% yeast extract, 2% peptone, 100 mM potassium
phosphate pH 6.0, 1.34% YNB, 4 x 10-5% biotin, 0.5% methanol. Dissolve 10g
yeast extract and 20g peptone in 700 mL water and autoclave at 121C for 20 min.
After cooling to room temperature, add 100 mL 1 M potassium phosphate buffer
pH 6.0, 100 mL 10 YNB, 2 mL biotin and 100 mL 10 methanol. Store at 4C
with a shelf life of approximately 2 months.

2.

YNB: dissolve 134 g yeast nitrogen base with ammonium sulfate and without
amino acids in water to a total volume of 1L and filter sterilize. Store at 4C with a
shelf life of approximately 1 year.

3.

500 biotin (0.02%): dissolve 20 mg biotin in water to a total volume of 100 mL

and filter sterilize. Store at 4C with a shelf life of approximately 1 year.

4.

10 methanol (5%): mix 5 mL methanol with 95 mL water and filter sterilize. Store
at 4C with a shelf life of approximately 2 months.

5.

1 M potassium phosphate buffer pH 6.0: Combine 132 mL of 1M K2HPO4 with 868

mL KH2PO4 and set the pH to 6.0 using a pH meter and phosphoric acid. Autoclave
the solution and store at room temperature with a shelf life of over 1 year.

6.

300 mL graduated glass measuring cylinder.

7.

Parafilm to seal the measuring cylinder.

8.

Stopwatch.

2.3. kLa measurement

1.

Data logging equipment to record dissolved oxygen (DO) measurements rapidly,

such as a Picolog ADC-16 (Applikon) which records data every 300 ms from the
DO probe. This can be used with PicoLog software to display a plot of the output.
The DO is recorded in mV by this equipment and can be converted to % DO once
the mV readings at 0% and 100% DO have been determined.

2.

The medium used for kLa measurements should be the same as for a full bioreactor
run e.g. BMMY (see section 2.2).

3.

A bioreactor with control systems and equipment usually used for a full bioprocess,
including a DO probe.

2.4. Shake flask cultures of recombinant GFP

1.

An agar plate with colonies of the organism of interest e.g. P. pastoris X33 GFP
grown on YPD agar composed of 1% yeast extract, 2% peptone, 2% dextrose
(glucose), 2% agar. Dissolve 20g peptone and 10g yeast extract in water to a total
volume of 900mL and add 20g agar. Autoclave the solution at 121C for 20 min
and cool to room temperature before adding 100 mL 10 glucose and pouring
plates which should be stored at 4C.

2.

10 glucose: dissolve 200 g glucose in water to a total volume of 1 L and autoclave

at 121C for 20 min then cool to room temperature. Store at 4C.

3.

Medium to be used to set up a starter culture e.g. BMGY medium composed of 1%

yeast extract, 2% peptone, 100 mM potassium phosphate pH 6.0, 1.34% YNB,
4x10-5 % biotin, 1% glycerol. Dissolve 10 g yeast extract and 20 g peptone in water
to a total volume of 700 mL. Autoclave at 121C for 20 min. After cooling to room

temperature, add 100 mL 1 M potassium phosphate buffer pH 6.0, 100 mL 10

YNB, 2 mL biotin and 100 mL 10 glycerol. The final volume should be adjusted
to 1 L. Store at 4C with a shelf life of approximately 2 months.
4.

BMMY medium (see section 2.2).

5.

One 250 mL baffled shake flask and 18 100 mL non-baffled shake flasks for
antifoam evaluations, all autoclaved.

6.

Spectrophotometer to determine the optical density of cultures.

7.

100% methanol for inducing protein production.

2.5 Assessment of cell viability by flow cytometry

1. PBS pH 7.0: dissolve one 10 PBS tablet in 1 L water.
2. Light microscope.
3. Haemocytometer.
4. Flow cytometer e.g. Beckman Coulter (High Wycombe, UK) flow cytometer with
488 nm excitation from an argon-ion laser at 15 mW.
5. Propidium iodide (1mg mL-1) in water.
6. WinMDI software for data analysis.

3. Methods
3.1. Bartsch Test
1.

This method is adapted from Denkov et al (16).

2.

Fill a 300 mL graduated glass measuring cylinder with 100 mL medium.

3.

Pipette the antifoam to be tested into the medium to 0.01% v/v, i.e. 10L (see
Notes 1 and 2).

4.

Seal the cylinder using parafilm and shake up and down ten times at ambient
temperature to induce foam formation.

5.

Record the total volume of the system (the medium and foam combined), and the
volume of the medium alone every 30 s for 15 min (see Note 3).

6.

Determine the activity by subtracting the volume of the medium from the total
volume.

7.

Compare the effects of different antifoams to a control with no added antifoam to

select the most efficient agent. An example is shown in Figure 1.

3.2. Measurement of kLa

1.

The dynamic method of kLa measurement is based on the method outlined by

Bandyopadhyay and Humphrey (18). Here the experimental set-up is a 3 L glass
bioreactor (Applikon Biotechnology) containing BMMY medium in the absence of
cells.

2.

A working volume of 1 L medium is used and the bioreactor set up for conditions
typical of the required bioprocess, e.g. 1.0 L min-1 compressed air (60% O2), pH
6.0, 30C and 700 rpm (See Note 4).

3.

Calibrate the DO probe by measuring the DO in the bioreactor at these settings for
100 % (60 % O2 in compressed air will achieve 100 % DO), then flush with
nitrogen gas instead of air to obtain a 0 % DO reading.

4.

kLa measurements are carried out by starting at 100 % DO and flushing with
nitrogen until the DO drops to 0 %. Supply the system with air and the DO
gradually rises to 100 %, whereupon it is again flushed with nitrogen to reduce it to
0 % before reconnecting the air. The time points are recorded by the data logger
and the DO can be plotted versus time, as shown by Figure 2.

5.

If several concentrations of antifoam are to be tested, add them in a stepwise

manner to the bioreactor once the DO is at 100 %. Follow each addition by flushing
with nitrogen, as in step 4.

6.

The data generated for the upwards slope of the plot are used to calculate the kLa
with the following equation:

kLa (t2-t1) = ln c1, - c1,t1

c1, - c1,t2

Where t1 and t2 are consecutively-logged time points, c1, is the oxygen saturation
concentration and c1 is the oxygen concentration at each time point. An example
calculation can be found in the Notes section (see Note 5).
7.

c1, is calculated using the following constants:

XO2 (fraction of O2 in air) = 0.2095

MO2 (O2 partition coefficient) = 30

R (gas constant) = 8.3144 J K-1 mol-1

And the following variables:

10

8.

T (temperature) = 30C or 303 K

P (approximate pressure in a 3 L glass bioreactor) = 0.1 barg or 1.1 bar

C (concentration of all gas in the head space)

n (number of moles of gas in the head space)

V (volume of gas in the head space)

Combining (C = n/V) and the universal gas law (PV=nRT) gives C=P/RT,
allowing the concentration of air in the gas phase in the head space of the vessel,
CgAir, to be calculated.

9.

The concentration of O2 in the gas phase, CgO2, can then be calculated by

multiplying the concentration of gas in air by XO2.

10.

Dividing the gas concentration of O2 by MO2 gives the maximum liquid oxygen
saturation concentration, c1, at 100% DO.

11.

DO percentage values at a particular time point are converted to oxygen

concentrations by dividing by 100 and multiplying by c1,.

3.3 Analyzing the influence of antifoams on the yield of GFP in shake flasks
1.

Pick a single colony of P. pastoris X33 GFP from a YPD plate and inoculate a 250
mL baffled shake flask containing 50 mL of BMGY.

2.

Incubate at 30C, 220 rpm overnight.

3.

Measure the optical density at 595 nm using a spectrophotometer. A typical value

would be 20

11

4.

Calculate the volume of culture necessary to inoculate the required volume of

BMMY to achieve a final OD595 of 1. In this example, the required volume of
BMMY is 360 mL.

5.

Centrifuge this volume of culture at 5,530 g and discard the supernatant.

6.

Re-suspend the pellet in 360 mL BMMY and mix.

7.

Dispensed 20 mL culture into each of the 18 100 mL non-baffled shake flasks.

8.

Incubate the flasks with the desired concentration of antifoam at 30C, 220rpm. A
suggested experimental set-up is triplicate flasks for antifoams at 0%, 0.2%, 0.4%,
0.6%, 0.8% and 1.0%.

9.

Record the optical density of the samples at 0 h, 24 h and 48 h.

10.

After 24 h add 100% sterile methanol to 1% v/v i.e. 200 L to 20 mL, and continue
the incubation.

11.

At 48 h harvest the cultures.

12.

Analyze the protein content. For secreted GFP, centrifuge 1 mL samples at 18,625
g for 10 min and separate the pellet and supernatant. Analyze the supernatant
samples on a fluorescence plate reader (Figure 3).

3.4. Analyzing the effect of antifoams on the viability of cells

1.

Dilute samples from section 3.3 step 11 using PBS to obtain a concentration of 106107 cells mL-1. Use a haemocytometer and a light microscope.

2.

Stain 2 mL of each sample by adding 10 L propidium iodide.

3.

Analyze the samples using a flow cytometer. Propidium iodide fluorescence is

measured at 630 nm and GFP fluorescence at 525 nm.

12

4.

WinMDI software can be used to plot the data, as shown in Figure 4.

4. Notes
1.

Antifoams may be sterilized prior to use by autoclaving at 121C for 20 min,

although this is not necessary.

2.

Some antifoams are viscous and sticky and may take several seconds to be drawn
into a pipette tip of the required volume. Additionally they may stick to the inside
of the pipette tip and therefore must be dispensed fully by repeatedly drawing
medium into the tip and pipetting it out.

3.

Foam may reach different heights around the sides of the cylinder, therefore the
highest level of foam should be recorded.

4.

A lower impeller speed may be required before adding antifoam as the foam level
may become too high at higher speeds depending upon the type of medium.

5.

An example of a kLa calculation using the constants and variables from section 3.2
step 7 is given here:
a. Calculate the concentration of air in the gas phase at 100% DO:

C = P/RT

Therefore CgAir = 1 x 104/(8.3145 303)

= 3.97 mol m-3

b. Using the concentration of air in the gas phase, the concentration of O2 present
in the gas phase is calculated:

CgO2 = XO2 x CgAir

= 0.2095 x 3.97

13

= 0.83 mol m-3

c. Convert O2 in the gas phase to O2 in the liquid phase for the maximum liquid
saturation concentration:

c1, = CgO2/MO2
= 0.83/30
= 0.028 mol m-3 at 100% DO

d. Convert DO values at a given time point to oxygen saturation concentrations, Cl,

by dividing by 100 and multiplying by c1,:

e.g. oxygen saturation concentration at 72% DO

Cl at 72% DO = (72/100) x 0.028

= 0.02 mol m-3

e. Substitute the Cl values calculated at particular time points into the equation in
3.2 step 6 to calculate the kLa.

5. References
1.

Holmes, W. J., Smith R. and Bill R.M. (2006) Evaluation of antifoams in the
expression of a recombinant FC fusion protein in shake flask cultures of
Saccharomyces cerevisiae, Microb Cell Fact 5, 30.

2.

Varley, J., Brown, A., Boyd, R., Dodd, P. and Gallagher, S. (2004) Dynamic
multipoint measurement of foam behaviour for a continuous fermentation over a
range of key process variables, Biochem Eng J 20, 61-72.

3.

Hfer, R. Struktol Foams and Foam Control. In: Struktol; 2008.

14

4.

Joshi, K., Jeelani, S., Blickenstorfer, C., Naegeli, I. and Windhab, E. (2005)
Influence of fatty alcohol antifoam suspensions on foam stability, Colloids Surf A
263, 239-249.

5.

Denkov, N.D., Krastanka, M, Christova, C, Hadjiiski, A, Cooper, P. (2000)

Mechanisms of action of mixed solid-liquid antifoams: 3. Exhaustion and
reactivation, Langmuir 21, 8163-8619.

6.

Al-Masry, W. (1999) Effects of antifoam and scale-up on operation of bioreactors,

Chem Eng Process 38, 197-201.

7.

Arjunwadkar, S.J., Sarvanan, K., Kulkarni, P.R. and Pandit, A,B, (1998) Gasliquid mass transfer in dual impeller bioreactor, Biochem Eng J 1, 99-106.

8.

Calik, P., Ileri, N., Erdinc, B.I., Aydogan, N. and Argun M: (2005) Novel
antifoam for fermentation processes: fluorocarbon-hydrocarbon hybrid
unsymmetrical bolaform surfactant, Langmuir 21, 8613-8619.

9.

Koch, V., Rffer, H., Schgerl, K., Innertsberger, E., Menzel, H. andWeis J.
(1995) Effect of antifoam agents on the medium and microbial cell properties and
process performance in small and large reactors, Process Biochem 30, 435-446.

10.

Morao, A., Maia, C., Fonseca, M., Vasconcelos, J. and Alves, S. (1999) Effect of
antifoam addition in gas-liquid mass transfer in stirred fermenters, Bioprocess
Eng 20, 165-172.

11.

Koide, K., Yamazoe, S. and Harada, S. (1985) Effects of surface-active

substances on gas hold up and gas-liquid mass transfer in bubble column, J Chem
Eng Jpn 18, 287-292.

15

12.

Liu, H.-S., Chiung, W.-C. and Wang, Y.-C. (1994) Effect of lard oil and caster oil
on oxygen transfer in an agitated fermentor, Biotechnol Tech 8, 17-20.

13.

Yagi, H. and Yoshida, F. (1974) Oxygen absorption in fermenters - effects of

surfactants, antifoaming agents and sterilized cells, J Ferment Technol 52, 905916.

14.

Stanbury, P.F., Whittaker, A. and Hall, S.J. (1995) Principles of Fermentation

Technology, Second edn: Butterworth Heinemann.

15.

Routledge, S.J., Hewitt, C.J., Bora, N. and Bill, R.M. (2011) Antifoam addition to
shake flask cultures of recombinant Pichia pastoris increases yield, Microb Cell
Fact 10, 17.

16.

Denkov, N.D., Tcholakova, S., Marinova, K.G. and Hadjiiski, A. (2002) Role of
oil spreading for the efficiency of mixed oilsolid antifoams, Langmuir 18, 58105817.

17.

Bartsch, O.(1924) ber Schaumsysteme. Fortschrittsberichte ber Kolloide und

Polymere, 20, 1-49.

18.

Bandyopadhyay, P. and Humphrey, A.E. (1967) Dynamic measurement of the

volumetric oxygen transfer coefficient in fermentation systems, Biotechnol
Bioeng 9, 533-544.

16

Figure legends

Figure 1: A Bartsch test of foam volume over time for various antifoams in BMMY
medium. Foam volume over time was recorded for 0.01% v/v of each antifoam in
BMMY medium where n = 5. All antifoams were effective at foam destruction compared
with the control and most foam was destroyed within 1 min.

Figure 2: DO over time for a 2 L glass bioreactor containing 1 L BMMY medium. The
bioreactor was flushed with nitrogen to reduce the DO, followed by air. The DO was
recorded in mV with a data logger and PicoLog software.

Figure 3: Effect of antifoams on the total yield of GFP produced by 20 mL recombinant

P. pastoris X33 over 48 h. The yield of secreted GFP was influenced by antifoam A (A)
and J673A (B). The horizontal line represents the average optical density for each
concentration of antifoam and n = 9 for each point. The data were analyzed using a oneway ANOVA (P < 0.001) and a Dunnetts multiple comparison test where* = P 0.05,
** = P 0.01. Each of the antifoams increased the yield at higher concentrations than
would normally be used in bioreactors.

Figure 4: Viable cells without antifoam (A) and viable cells with 0.6% antifoam A (B).
Population A is made of events that are related to electronic and particulate noise and are
not cells. Population B are cells showing enhanced green fluorescence due to GFP and

17

quadrant C is where dead cells stained red with propidium iodide (PI) would be observed.
Antifoam A did not adversely affect the viability of the cells.

18

Control

Antifoam C

Antifoam A

SB2121

P2000

Mazu DF 204

J673A

15

13.5

12

10.5

Time (min)
9

7.5

4.5

1.5

80

60

40

20

Foam Volume (mL)

Figure 1

120

100

(mV)

Figure 2

(ms)

Figure 3

**

500

*
40

450

35

400

GFP ( g)

300

25

250

20

200

15

150

OD595

30

350

10

100
5

50
0

0
0

0.2

0.4

0.6

0.8

% Antifoam

500

**

450

35
30

400

GFP (g)

300

20

250
15

200
150

10

100
5

50
0

0
0

0.2

0.4

0.6
% Antifoam

0.8

OD595

25

350

Figure 4