Eng. Life Sci.

2015, 15, 195207

Pascal Baumann
Anna Osberghaus
Jrgen Hubbuch
Institute of Process Engineering
in Life Sciences, Section IV:
Biomolecular Separation
Engineering, Karlsruhe Institute
of Technology (KIT), Karlsruhe,
Germany

www.els-journal.com

Research Article

Systematic purification of salt-intolerant

proteins by ion-exchange chromatography:
The example of human -galactosidase A
Chromatography is an essential tool for purifying biopharmaceutical products. Many
processes are still developed based on traditional routines and empirical procedures.
Product losses are mostly due to insufficient optimization of purification setups
and product sensitivity to process conditions. In order to eliminate these shortcomings, a systematic strategy for the setup of ion-exchange chromatography is
presented, which considers both product stability as well as operational conditions.
The stagesa hybrid approach combining high-throughput screening and analytical small-scale chromatographyare as follows: (1) pH stability (short-term);
(2) pH stability (long-term), followed by a screening of additives to enhance protein
stability, if required; (3) analytical pH gradient chromatography for evaluation of
the operational pH window; and (4) salt stability (long-term) in the operational
pH window determined. The efficiency and straightforwardness of the strategy were
shown in a case study on capturing the human -galactosidase A enzyme. Following
the above procedure, the enzyme was found to be salt-unstable; a purification factor
of 13.2, a concentration factor of 4, and an overall yield of 84.3% were achieved.
The applied strategy allowed for a quick establishment of a dedicated capture step
at low salt concentrations under stable conditions by well-chosen prior screening
experiments.
Keywords: -Galactosidase A / High-throughput screening / Hybrid approach / Ion-exchange
chromatography / Salt-intolerance

Additional supporting information may be found in the online version of this article at
the publishers web-site

Received: October 16, 2014; revised: December 19, 2014; accepted: January 15, 2015
DOI: 10.1002/elsc.201400210

Introduction

Chromatographic procedures are commonly used to capture biological products from complex mixtures. The critical process
parameters determining purity and yield are the pH, salt type,
and salt concentration used during binding, wash, and elution
procedures. Ion-exchange chromatography (IEX) has become

Correspondence: Prof. Jurgen Hubbuch (juergen.hubbuch@kit.

edu), Institute of Process Engineering in Life Sciences, Section IV:
Biomolecular Separation Engineering, Karlsruhe Institute of Technology (KIT), Engler-Bunte-Ring 1, 76131 Karlsruhe, Germany
Abbreviations: 4-NP, 4-nitrophenol; 4-NPG, 4-nitrophenyl -dgalactopyranoside; AEX, anion-exchange chromatography; AU, absorption unit; CEX, cation-exchange chromatography; CV, column volume;
IEX, ion-exchange chromatography

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

the method of choice for capture steps due to the rather mild
process conditions and high binding capacities. However, even
moderate concentrations of salts being low in chaotropic or cosmotropic nature might affect protein integrity as was reported
in a variety of studies on therapeutically relevant proteins. The
granulocyte colony stimulating factor was found to form aggregates in the presence of 150 mM sodium chloride under physiological conditions in PBS [1, 2]. Analog aggregation behavior
was observed for the recombinant human tissue factor pathway
inhibitor [3]. A slight increase in sodium chloride concentration from 0.04 to 0.11 mM induces major insulin fibrillation
under acidic conditions [4]. RNAse A was found to be destabilized by moderate sodium chloride concentrations of 250 and
500 mM [5]. Lysozyme also showed a strong decrease in solubility at salt concentrations below 0.5 M [6].
A very simple approach to enhancing separation efficiency
of IEX chromatography is the initial determination of an
195

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

operational pH window exploiting the charge-pH dependency

of product and contaminants. In a heuristic approach, the binding pH in IEX is chosen to be 0.5 to 1 pH units above or
below the pI of the product [7]. Those pI values from literature are either calculated in silico based on molecular modeling
or determined experimentally using isoelectric focusing. Such
values are, however, mostly not identical to the actual product elution pH on a specific column. Charge distributions and
charge clusters on protein surfaces may lead to binding to the
adsorber even at a molecular net charge of zero [8]. Another
factor to be considered is the interference with the resin backbone, where nonelectrostatic forces, such as hydrophobic interactions, come into play [9]. Especially proteins with a neutral pI
(pI 68) tend to elute experimentally at much higher pH values
than expected [10,11]. Also posttranslational modifications such
as glycosylation patterns were shown to affect the elution pH of
proteins [12].
A deterministic approach to evaluating an operational pH
window applies pH gradient chromatography as an elegant analytical tool in IEX process development. Instead of relying on
pI values from literature, the elution pH values of products and
contaminants are determined experimentally on a specific column [10]. The salt gradient purification can then be designed
for a pH value of 0.5 apart from the experimental elution
pH. Previously, pH gradients for the determination of the pH
operational window in IEX chromatography were used in the
purification of a monoclonal antibody from a cell culture supernatant on an anion-exchange chromatography (AEX) and
cation-exchange chromatography (CEX) column [13]. Similar
approaches were applied for separating Nucleolin from an Sf9
insect cell extract by AEX and CEX chromatography [14] and the
anthrax-protective gene (PA) from an Escherichia coli cell lysate
by AEX chromatography [15].
However, protein stability and use of additives under typical
process conditions have not been considered so far. In fact, colloidal and conformational stability problems, namely aggregation and unfolding of proteins, will lead to product losses [16]. A
pH close to the pI of a protein might lead to aggregation due to a
reduction of repulsive intermolecular forces [17]. In contrast to
this, a pH value far from the pI leads to identical intramolecular
charges [17]. This can induce an unfolding of proteins. Another
factor to be considered is the amount and type of salt, as discussed above.
So far, protein stability and additive screenings have mostly
been applied for protein refolding and crystallization studies.
Lysozyme and carbonic anhydrase were stabilized by addition
of cyclodextrins and surfactants [18]. The native state of lactamase was maintained by adding sucrose to the buffer. Other
examples of additives are cofactors, amino acids (e.g. lysine, proline), several sugars, alcohols (e.g. glycerol), and other detergents
such as CHAPS [19]. Reducing agents, such as DTT in low concentrations, were applied as well [20]. Lysozyme is shown to
be stabilized by low molecular weight PEGs, detergents, such as
Tween 20, and different sugar molecules such as sorbitol and
sucrose [6].
For this reason, an integral strategy for optimizing an IEX
capture step combining product stability and operational parameters over a wide pH range needs to be considered. In the
following article, the term stability refers to the conservation
of enzymatic activity.
196

www.biotecvisions.com

A systematic strategy for the purification of salt-intolerant

proteins in IEX chromatography will be introduced in this article. The strategy is illustrated in Fig. 1A and consists of four consecutive stages, including screenings for short- and long-term pH
stability, salt tolerance, additives, and optimal pH for a pH-based
IEX procedure: In a first stage (1) pH stabilityshort-term,
the general pH stability of the product/feed stream is evaluated.
Conditions of instant denaturation or irreversible aggregation
can be excluded in the further process development. Product
exposure times to a certain chemical environment can, however,
be in the range of several minutes to hours. The identified pH region of stage 1 is further evaluated in a second stage (2) analyzing
product pH stabilitylong-term. If the product is found to be
unstable at this point, an additive screening follows. During this
stage, the procedures of stage 2 pH stabilitylong-term are repeated with different excipients to find out whether the product
can be kept in a native state for a predefined period time. In the
third stage (3), the crude feedstock is analyzed by pH gradient
chromatography to identify the operational pH window for IEX
chromatography. At this point, the mode of IEX chromatography needs to be chosen. In order to define the optimal elution
mode, screenings to determine (4) salt stabilitylong-term
in the operational pH window determined earlier are used to
test the effect of product exposure to different concentrations
of salt. If the protein is found to remain in a native state during salt exposure, traditional (5a) IEX chromatographysalt
elution mode can follow. For salt-unstable products, a lowsalt purification process based on product elution by pH shift
IEX chromatographypH elution mode (5b) is carried out.
Such elution setups are commonly used in affinity chromatography [21, 22]. In several cases, IEX chromatography applying pH
shifts for elution was used for separating protein isoforms (e.g.
PEGylated lysozyme, human transferrin) or chemically similar
variants (e.g. deaminated protein variants) or as an intermediate purification step of d-amino acid oxidase from Candida
guilliermondii [2326].
This case study focuses on the use of the systematic strategy
proposed for IEX for capturing human -galactosidase A from
Pichia pastoris. The enzyme consists of two 50 kDa subunits, is
naturally located in human cellular lysosomes [27], and is involved in the cellular detoxification mechanism by metabolizing
harmful glycolipids. The absence of -galactosidase A inside the
human body leads to an accumulation of glycolipids and major
cell damage (Fabry disease).

Materials and methods

2.1

Materials

Sample conditioning was performed using PD-10 desalting

columns (bed volume 8.3 mL) packed with Sephadex G-25
Medium resin (GE Healthcare Life Sciences, Sweden). Fractionations were carried out in 2 mL 96-well square deep well
plates (VWR, Germany). The enzyme assay was performed in 96well flat-bottom polypropylene microplates (Greiner Bio-One,
Germany). Absorption measurements were carried out in 96well flat-bottom UV-Star microplates (Greiner Bio-One). For
chromatography runs, MediaScout MiniChrom 550 columns
(Atoll, Germany) prepacked with 1 mL of the following resins

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

www.biotecvisions.com

Figure 1. Schematic illustration of the systematic strategy for the purification of salt-intolerant proteins in ion-exchange chromatography.
(A) Sequential order of screenings for an optimized IEX chromatography process. (B) pH gradient chromatography for determination of
the products elution pH. (C) IEX chromatographypH elution mode with two different elution setups of a step (a) or 5 CV gradient (b).

were purchased: Source 30 S, Source 30 Q (GE Healthcare Life

Sciences).
The buffer system used, allowing for the generation of
highly linear pH gradients, follows a description by Kroner
et al. [11]. For cation exchange, the buffer system uses the
following components: 10 mM sodium chloride, 11 mM
MES, 9.9 mM formic acid, and 13 mM acetic acid (Merck,
Germany), 8.7 mM 3-morpholino-2-hydroxypropanesulfonic
acid, 9.9 mM 4-(2-hydroxyethyl)-piperazine-1-(2-hydroxy)propanesulfonic acid, 4.6 mM N-[tris(hydroxymethyl)methyl]3-aminopropanesulfonic acid, 9.4 mM CHES, and 15.6 mM
CAPS (AppliChem, Germany). The anion-exchange buffer system is composed of the following components: 7.7 mM hydroxylamine, 9.8 mM methylamine, 6.99 mM 1,2-ethane-diamine,
13.47 mM 1,4-dimethylpiperazine (Merck), 5.8 mM Bis-Tris
(Molekula, UK), and 6.4 mM 1-methylpiperazine (SigmaAldrich, USA). The pH of the buffers was adjusted by titration
using hydrochloric acid or sodium hydroxide (Merck), respectively.
For the adapted 4-nitrophenol (4-NP) assay, the substrate
4-nitrophenyl -d-galactopyranoside (4-NPG) and the colorimetric product 4-NP were purchased from Sigma-Aldrich. As
a reaction medium, a 100 mM acetate buffer at pH 5 was prepared from acetic acid (Merck). For termination of the enzyme
reaction, disodium carbonate (Merck) was applied.
A supernatant from P. pastoris fermentation (strain NRRLY
11430), during which the enzyme human -galactosidase A
is produced extracellularly, as well as purified human


C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

-galactosidase A standard were obtained from BioIngenium

(Spain).

2.2

Instrumentation and software

For pH adjustment of all buffers, an HI-3220 pH meter (Hanna

Instruments, USA) was used. The instrument was calibrated
using high-precision standards from Hanna Instruments. For
the enzyme assay, an MUR 13 thermo-shaker with additional
lid heating (HLC BioTech, Germany) was used. For absorption
measurements, an Infinite M200 Reader (Tecan, Germany) was
applied and operated with I-control 1.9 (Tecan). The pH gradient chromatography and IEX chromatographypH elution
TM

purifier system opermode were carried out in an AKTA

ated with Unicorn 5.2 (GE Healthcare Life Sciences). The system
was equipped with a pump P-900, mixer M-925, UV detector
UV 900, motor valve INV-907, pH and conductivity monitoring
pH/C-900, and a fraction collector Frac-950 unit. Data processing and creation of figures were performed in MATLAB R2011 a
(MathWorks, USA).

2.3
2.3.1

Strategic process development

Adapted 4-NP assay

A necessary prerequisite for the application of the proposed

systematic strategy to the case study was a pH-insensitive assay
determining product concentration or activity.

197

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

2.3.1.1 Adapted assay procedure

For the determination of product titers, a 4-NPG enzyme activity
assay was adapted. The basic principles presented by Rezende
et al. [28] and Marin et al. [29] were modified and combined.
The adjusted 4-NPG protocol works as follows: 200 L of 2 mM
4-NPG in 100 mM acetate buffer (pH 5) is added to 50 L
sample and incubated at 37C and 400 rpm in a plate shaker
with temperature control. The enzyme reaction is stopped by
addition of 25 L of a 1 M sodium carbonate solution. The
concentration of accumulated 4-NP is determined in a plate
reader by 405 nm absorption measurements.

2.3.1.2 Assay validation

As fractions collected throughout the pH gradient intrinsically
exhibit a large variation of the pH value, special emphasis was
put on evaluating the sensitivity of the assay to pH changes of the
sample. The enzymatic activity of -galactosidase A is optimal in
the pH range from about 4.55 [3032]. In order to test whether
the buffering strength of the receiving buffering system200 L
of 100 mM acetate buffer at pH 5is sufficient to buffer samples,
the following procedure was applied in triplicate measurements:
2.5 mL of a sample from the P. pastoris supernatant containing
-galactosidase A was conditioned by buffer exchange in a PD-10
desalting column into 100 mM acetate buffer at pH 5. By mixing
100 L of the conditioned sample and 100 L of 4 mM 4-NPG
solution, a final 2 mM substrate concentration was reached. By
adding 50 L of pH gradient buffer in a pH range of 311 to
this basic mixture, a potentially observed decrease in activity can
be correlated directly to a strong pH fluctuation in the reaction
mixture.
Next to operating at the optimal pH for the enzyme reaction, care needs to be taken in the pH-dependent readout of
the colorimetric reaction. It was pointed out by Werner [33]
that the absorption coefficient of the cleaved colorimetric 4-NP
is strongly pH-dependent under acidic and neutral conditions.
Hence, terminating the enzyme reaction by adding 1 M sodium
carbonate solution is used to shift the pH to basic conditions
where 4-NP shows no pH dependence in absorption at 405 nm.
A total of 25 L of 1 M sodium carbonate solution was added
and the absorbance was determined at 405 nm.
The 405 nm signals obtained were then correlated to the effective enzyme activity by determination of released 4-NP per reaction time. For this purpose, a linear calibration of 125 g/mL
4-NP in AEX buffer, pH 11, was performed, leading to a linear
regression 4-NP of 0.0726 absorption unit (AU)/(g 4-NP/mL).
The molar enzyme activity Act (mol 4-NP/mL/h) of samples
was then calculated as shown in (Eq. (1)):

Act =

(E 405,sample E 405,blank ) 5.5

t 4NP M4NP

(1)

with E405 being the extinction of the sample/blank measured at 405 nm (AU), 5.5 being the dilution factor of the assay,
t denoting the assay duration (h), 4-NP denoting the slope of
the calibration (AU/(g 4-NP/mL)), and M4-NP being the molar
mass of 4-NP (139.11 g/mol).
198

2.3.2

www.biotecvisions.com

Screening procedure

The screening methods applied as shown in Fig. 1A are described

in detail in the following article.

2.3.2.1 pH stabilityshort-term
Short-term product stability was defined to be the stability of galactosidase A during a 5-min period under process conditions.
To analyze this stability, the following procedure was applied
in triplicate measurements: 5 L of purified -galactosidase A
standard in water was mixed with 45 L of the respective pH
gradient buffer in the range from pH 3 to 11 in steps of one pH
unit to cover the acidic and basic pH range. Product activity was
measured with the 4-NPG assay. The reduction of enzymatic
activity was defined to be a measure of the quantity of denatured
protein during 5-min exposure time.

2.3.2.2 pH stabilitylong-term
Long-term product stability was defined to be the stability of
-galactosidase A during a 120-min period under process conditions. To analyze this stability, the following procedure was
applied in triplicate measurements: The buffer composition of
the P. pastoris supernatants was adjusted to pH 4, 5, 6, 7, and
8 with PD-10 desalting columns. From each of these stock solutions, samples of 50 L were taken and the 4-NPG assay was
started after 0-, 20-, 40-, and 80-min exposure time, respectively.
The assay was stopped after 120 min by adding 1 M sodium carbonate. To counteract potential pH instability of the product, an
additive screening was carried out after the pH stability screenings. The experimental setup was identical to that of the pH
stabilitylong-term process, with additives being added to the
respective pH gradient buffers. Here, 100 mM sodium chloride
and 0.02 % Tween 20 were chosen.

2.3.2.3 pH gradient chromatography

TM

pupH gradient chromatography was performed in an AKTA

rifier system using MediaScout MiniChrom 550 columns (column volume (CV) 1 mL) packed with Source 30 S/Q resin. The
pH valuespHgradient,start and pHgradient,end defining the pH gradient were determined by the pH stabilitylong-term experiments described above. The buffer composition of the P. pastoris
supernatant was consequently adjusted to pH 8 for AEX and
pH 4 for CEX chromatography in PD-10 desalting columns. A
load of 500 L sample was applied and the volumetric flow rate
was set to 1 mL/min. The integrated pH electrode was two-point
calibrated for the start and end conditions of the pH gradient.
The general chromatographic procedure is shown in Fig. 1B. For
CEX experiments, the pH gradient was increased from pH 4 to
8. For AEX experiments, the pH was decreased from pH 8 to 4
(buffer species and concentrations are described above). Initial
equilibration was performed with 5 CV of the respective binding
buffer. Binding buffer (10 CV) was applied after sample application in the wash step. Subsequently, a 20 CV pH gradient was
applied to reach the final pH of the respective gradient. This pH
was then kept constant for another 3 CV. Finally, regeneration
was performed with a 4 CV 1 M sodium chloride step applied at
the end of the chromatographic procedure. Additionally, blank

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

runs were conducted to calculate buffer-independent UV signals. Throughout the chromatography run, 500 L fractions
were taken for product-specific analysis. A total of 50 L of each
fraction was evaluated using the adapted 4-NPG assay.
The operational pH window for product elution was defined
by a minimum of 10 % enzyme activity found in the respective fractions when compared to the maximum enzyme activity
found (see Fig. 1B). The framing pH values are referred to as
pHelution,start and pHelution,end .

2.3.2.4 Salt stabilitylong-term

The P. pastoris supernatant was adjusted to pHelution,start and the
respective buffer system using PD-10 desalting columns. For the
CEX setup, pH 4.3, and for the AEX setup, pH 7.5, was selected
based on the criteria determined above. The conditioned samples
were then mixed (dilution factor 2) with two-fold concentrated
salt solutions of the respective buffers, leading to final sodium
chloride concentrations of 0, 50, 100, 200, 300, and 500 mM. The
4-NPG assay was started after 0-, 20-, 40-, and 80-min exposure
time and stopped after 120 min by adding 1 M sodium carbonate.

2.3.2.5 IEX chromatography

As human -galactosidase A was found to be salt-unstable, a
processing mode applying a pH shift was used. The P. pastoris
samples were conditioned to pHgradient,start (compare Fig. 1C)
using the respective buffer components of the analytical pH
TM

purigradients. All experiments were conducted in an AKTA

fier system using MediaScout MiniChrom 550 columns packed
with Source 30 S/Q resin. A load of 1000 L sample was applied
and the volumetric flow rate was set to 1 mL/min. Spectroscopic
analysis was conducted using the 280 nm signal for total protein
and the 254 nm signal for nucleic acids. The chromatographic
procedure was conducted as follows (see Fig. 1C): The IEX column was equilibrated with 5 CV of binding buffer, followed
by sample application, and wash with another 5 CV of binding
buffer. A second wash procedure for 5 CV was run by a stepwise pH shift to pHelution,start , to elute contaminants close to the
products elution pH. Product elution within the operational pH
window was then performed in two different setups for both the
AEX and CEX chromatography columns: Step elution and gradient elution over 5 CV. Elution was followed by a 5 CV pHelution,end
wash and a 4 CV 1 M NaCl high-salt regeneration step. Analogously to the pH gradient screening experiments, blank runs
were conducted to calculate the buffer-independent UV signals.

2.3.3

Analysis of chromatographic procedures

Throughout the IEX chromatographypH elution mode run,

fractions of 1000 L were taken for product-specific analytics.

2.3.3.1 Mass balances

To investigate denaturation, protein remaining bound to the adsorber or formation of aggregates inside the chromatography column, recovery rates were calculated. The recovery of total protein
RTP was performed by dividing the integrated blanked 280 nm
signal from the entire chromatography run A280,run (sample

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.biotecvisions.com

injection up to the final salt step) by the integrated signal of a samTM

ple injected into the AKTA

system A280,sample (Eq. (2)). For this
purpose, the signals have to be in a linear measurement range.
The integration process was performed in MATLAB R2011 a
(MathWorks) using a trapezoidal numerical integration.
RTP =

A 280,run
A 280,sample

(2)

The recovery of activity RA was calculated considering the

enzymatic activity of -galactosidase A (Eq. (3)). The added
activities Actfraction of all fractions (sample injection up to the
final salt step) were divided by the total activity of the applied
sample Actsample taking the sample volumes Vfraction and Vsample
into account.

Actfraction Vfraction
(3)
RA = 
Actsample Vsample

2.3.3.2 Overall product yield

The product yields Y were calculated as the ratio between eluted
enzyme activity Actelution and applied sample Actsample taking the
sample volumes Velution and Vsample into account (Eq. (4)).

Actelution Velution
(4)
Y= 
Actsample Vsample

2.3.3.3 Purification factor

The product purity was calculated using specific activities as
defined by Eq. (5). The purification factor PF was defined to
be the ratio between the specific activity of the pooled elution
fraction and the applied sample.

Actelution / A 280,elution
(5)
PF = 
Actsample / A 280,sample

2.3.3.4 Qualitative nucleic acid factor

As a third criterion, the qualitative nucleic acid factor NF was
determined by the ratio of the 254 nm signal and 280 nm signal.
In contrast to measurements at 260 nm, DNA and RNA show
almost equivalent absorption at these wavelengths, whereas proteins exhibit a lower absorption at 254 nm and, hence, can be
distinguished better from nucleic acids [34]. The factor is defined to be the ratio between the integrated 254 nm signal A254
and 280 nm signal A280 (Eq. (6)).
NF =

A 254
A 280

(6)

By comparing the NF values of the applied sample and the

elution fractions, the qualitative reduction of nucleic acid levels
was determined.

2.3.3.5 Concentration factor

For capturing unit operations, not only the yield, purity, and
nucleic acid level reduction are important, but also the achieved
product concentration. The product concentration factor CF was
199

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

determined (Eq. (7)) to be the ratio between the pooled elution

activity Actpool and the applied sample Actsample .
CF =

2.3.4

Actpool
Actsample

(7)

Process simplification

2.3.4.1 Reduction of buffer species

Given the optimal system parameters determined according to
the procedure described above, the complex mixture of buffer
substances was simplified to a two-component system. The
buffer components were chosen in a buffering range of the
operational pH window defined by the values of pHelution,start
and pHelution,end (see Fig. 1C). For the case study, the binding
buffer was set to 10 mM acetate (logarithmic acid dissociation constant (pKa ) 4.76) at pH 4.8. The elution buffer (pH 6)
consisted of 10 mM MES (pKa 6.10). Sample binding occured
at pHelution,start , thus eliminating the first pH shift depicted in
Fig. 1C.

2.3.4.2 Product capture at the capacity limit

Product capture with a simplified two-buffer component system
was performed at the column capacity limit, which was determined to be a load of 20 CV at 10 % of product breakthrough.
A total of 20 mL of sample was applied and the chromatography
procedure was evaluated as described in Section 2.3.3.

Results and discussion

3.1

Adapted 4-NP assay

The successful adaptation of the 4-NPG reaction toward a

robust, reproducible, and widely pH-insensitive assay for galactosidase A was a necessary prerequisite for the proposed
development strategy shown in Fig. 1. To validate the suitability
of the chosen assay, several aspects relating to pH insensitivity
during reaction and readout as well as reaction kinetics were
evaluated. Supporting information, Fig. 1 illustrates the enzyme
activities obtained by assaying samples containing identical enzyme concentrations at different pH values. The highest 4-NPG
conversion rates were measured for samples in a pH range from
pH 3 to 5, where the pH shift to the conditions of the reaction
buffer is negligible. In general, an increase with sample pH led
to a decrease of measured enzyme activity. For the AEX buffer
system, an overall decrease in enzyme activity of 26.1% could
be observed over the complete pH range from pH 3 to 11. In
the pH region of pH 3 to 8, in which most of the stability and
screening experiments were conducted, the decrease was about
19.3%. The use of the CEX buffer system led to a decrease in
enzyme activity of 28.8% in the overall range from pH 3 to 11.
In the region of pH 3 to 8, the activity difference was determined
to be 15.6%. In summary, the chosen reaction buffer proved to
be a good compromise between sufficient buffer capacity and a
sufficiently low salt concentration to prevent salt-induced loss of
enzyme activity.
200

www.biotecvisions.com

The methodology can be used analogously for proteins that

can be assayed specifically, including hormones, enzymes, etc.
For most other proteins, antibody-based procedures could be
applied (e.g. ELISA) for product determination. Another fast
method for analyzing proteins is capillary gel electrophoresis
using, for example the Caliper Lab Chip GX II system. Here,
samples can be analyzed within approximately 3040 s of processing time.
Regardless of the subsequent analytics, all time-dependent
screening steps, such as the pH stabilityshort-term should
have a similar setup: After the predefined exposure time, a stabilizing buffering solution (analogously to the enzyme reaction buffer) should be added to stop the exposure to pH,
salt, etc. The remaining soluble and active fraction can then be
analyzed.

3.2
3.2.1

Screening procedure
pH stabilityshort-term

During the first stage, the pH stabilityshort-term of the

given target molecule was investigated to identify potential pH
limits of instant enzyme denaturation. pH values of 311 were
investigated for AEX and CEX buffer systems. Following 5-min
pH exposure to purified -galactosidase A standard, enzyme activities were determined using the assay. Figure 2A shows the
activities Act obtained scaled to the maximal activity determined
in each test set Actmax . Both the AEX and CEX buffer systems
show a similar trend: The maximal enzyme activity could be
found at pH 6. A stable process window regarding enzyme activity was found in the range from pH 4 to 8. In this region,
maximal variances of 12.8% (AEX buffers) and 10.1% (CEX
buffers) were detected. Buffer systems of pH values higher than
8 led to a significant decrease in converted 4-NPG. Enzyme
activities of 59.1% (AEX) and 64.1% (CEX) were already lost
at pH 3. Similar results were observed for the pH 9 samples.
Here, a loss in enzyme activity of 37.2% (AEX) and 42.0%
(CEX) was found. At pH 11 (the most basic pH investigated),
a maximal degradation of the product was detected, with decreases in activity of 82.4% (AEX) and 93.4% (CEX). Consequently, the appropriate region for all further experiments
was fixed to an operational pH window ranging from pH 4
to 8.

3.2.2

pH stabilitylong-term

During purification processes, product streams are exposed to

chemical environments much longer than 5 min. Figure 2BE
shows relative enzyme activities at pH levels between pH 4 and
8 and exposure times of 080 min. All activities are scaled to the
activity maximum determined in the respective setup. For the
CEX buffer system (Fig. 2B), a constant activity ratio of >0.965
was achieved at pH 57, even at 80-min exposure time. Samples
exposed to pH 4 and 8 showed constant activity ratios >0.871 at
exposure times of up to 20 min. Even at times of up to 80 min,
the activities were constantly above 68.3%. For the AEX buffer
systems (Fig. 2C), enzyme stability was achieved only at 5-min
exposure time (as shown in pH stabilityshort-term) and at
pH 56. At pH 8, an exposure time of 20 min already caused

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

www.biotecvisions.com

A
80
60
40
20
0

CEX buffer
AEX buffer

pH

10

40
20
0
4

pH

40

0
4

pH

80
60

40
40
20
0
4

80

Act/Actmax [%]

100

60

20
5

pH

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

100
80

60

60
40
40
20
0
4

a loss of activity of 71.5%. The CEX buffer systems were found

to retain the enzyme activity, while AEX buffers led to a strong
enzyme inactivation outside an operational pH window from
pH 5 to 6.
For the AEX buffer system, an additive screening was carried
out to stabilize human -galactosidase A. As a mild increase
in salt concentration is reported to lead to a stabilizing effect
by salting-in for some proteins [17], a first screening was conducted by adding 100 mM of sodium chloride to the AEX buffers
(Fig. 2D). The results, however, indicated no improvements in
enzymatic stability. A second screening was performed by adding
the detergent Tween 20 to the AEX buffer system. In this setup,
a constant activity of more than 91.1% was reached at pH 47
within the period investigated (Fig. 2E). Exposure to pH 8 revealed activities of more than 74.8% at exposure times of up to
20 min. Up to 80 min of exposure, the activities were constantly

C

20

E
80

Time [min]

60
40
20

20
5

80

20
5

pH

Act/Actmax [%]

Time [min]

60
40

100

60

Time [min]

80

60

80

Act/Actmax [%]

100

80

Act/Actmax [%]

11

Time [min]

Act/Actmax [%]

100

Figure 2. Product pH stability

short- and long-term in AEX and
CEX buffers. Short-term experiments were performed in the
range from pH 3 to 11, longterm experiments in the range
from pH 4 to 8 with product exposure times of 080 min before
stabilization by starting the 4NPG assay. Activities are scaled
to the maximal activity determined in each test set. (A) Product pH stabilityshort-term in
AEX and CEX buffers. The error
bars indicate the data variability
of the triplicate measurements.
(B) Product pH stabilitylongterm for the CEX buffer system without additional excipients. (C) Product pH stability
long-term for the AEX buffer system without additional excipients. (D) Product pH stability
long-term for the AEX buffer
system with 100 mM sodium
chloride added. (E) Product pH
stabilitylong-term for the AEX
buffer system with 0.02% Tween
20 detergent added.

above 63.8%. These results obtained for the AEX buffer system
containing Tween 20 were then comparable to the enzymatic stability obtained during pH stabilitylong-term experiments
for the CEX buffer system. Tween 20 prevented aggregation of
active product molecules, which must have occurred in the presence of AEX buffer substances over time. Hence, the additive
screening was capable of stabilizing the enzyme and finding a
suitable chemical environment for product purification. In general, suited additives can be found when looking at the final
formulation buffer or the cultivation conditions of the investigated product.

3.2.3

pH gradient chromatography

Analytical pH gradient chromatography was used for comparing contaminant and product elution profiles of the
201

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

B
8

15

10

CV
30
20
10
0

10

20

UV 280 nm [mAU]

pH

20

30

40

280 nm - 1
280 nm - 2
pH - 1
pH - 1

25

20

15

10

30
20
10
0

CV

complex P. pastoris feedstock. The 280 nm signals and the enzyme activities of the assayed fractions for both the CEX and
AEX (including Tween 20 as a stabilizer) pH gradient runs are
shown in Fig. 3. Taking the respective fraction sizes into account,
activities are given in nmol 4-NP/h in contrast to the units of
nmol 4-NP/mL/h used above.
During the CEX pH gradient chromatography (Fig. 3A),
a flowthrough of negatively charged material is seen during the
first 5 CV. When determining the enzymatic activity in fractions covering CV1 to CV10 , a total -galactosidase A activity
of 51.9 15.5 nmol 4-NP/h was found. This indicates that
the dynamic binding capacity was reached. At CV13 , the pH
gradient was started, leading to the elution of bound species.
The maximum elution peak with an UV 280 nm value of
9.5 mAU was found at CV19 and pH 5.3. After this peak, the
280 nm signal decreased until the end of the pH gradient at
CV35 and pH 8. Enzymatic activity was found in the region
from CV16 to CV26 and at pH values ranging from pH 4.6 to
6.7. A maximum of 39.2 2.2 nmol 4-NP/h was reached at
CV19.5 and pH 5.4. This results in an operational pH window
for CEX (declared as 10% of the activity peak maximum) from
pH 4.8 to 6. During the 1 M NaCl high-salt wash from CV35
to CV39 , a final 280 nm peak of 16.6 mAU was observed. Evaluation of the salt fraction showed no major residual enzyme
activity.
During the AEX pH gradient chromatography (Fig. 3B),
flowthrough of positively charged material is seen during the
first 5 CV. When determining the enzymatic activity in fractions from CV1 to CV10 , a total -galactosidase A activity of
7.8 1.2 nmol 4-NP/h was observed. In this setup, less enzyme was detected in the flowthrough compared to the CEX
setup. Although the pH gradient started at CV13 , the 280 nm
signal only started to increase at CV15 and pH 7.5. Beginning
at CV15 and pH 7.5, two local maxima at CV28 and pH 4.9 of
17.3 mAU and at CV30 and pH 4.5 of 7.0 mAU were reached.
Enzymatic activity was found in the region from CV16 to CV30
and at pH values ranging from pH 7.3 to 4.5, the maximum
being 19.1 1.6 nmol 4-NP/h at CV19.5 and pH 6.7. This

10

20

30

CV

40

pH

280 nm - 1
280 nm - 2
pH - 1
pH - 1

Activity
[nmol 4-NP/ h]

Activity
[nmol 4-NP/ h]

UV 280 nm [mAU]

25

202

www.biotecvisions.com

Figure 3. pH gradient chromatography of the P. pastoris

supernatant in duplicates. UV
280 nm signals and pH values
are plotted against column volumes. The corresponding fraction enzyme activities are shown
in a bar diagram below. (A) CEX
pH gradient chromatography on
a Source 30 S column (pH 4
8). (B) AEX pH gradient chromatography on a CEX Source 30
Q column (pH 84).

results in an operational pH window for AEX (declared as 10%

of the activity peak maximum) from pH 7 to 4.7. During the
1 M NaCl high-salt wash from CV35 to CV39 , a final 280 nm
peak of 93.0 mAU was observed. Evaluation of the salt fraction
showed a residual summed up enzyme activity of 51.6 5.7 nmol
4-NP/h.
In summary, the operational pH window for CEX was determined to extend from pHelution,start = 4.8 to pHelution,end =
6 and for AEX from pHelution,start = 7 to pHelution,end = 4.7.
During both chromatographic procedures using an analytical pH gradient, a broad range of product elution pH values
was obtained. This can most probably be explained by glycosylation and charge variants, as P. pastoris tends to hyperglycosylate products. Posttranslational modifications of proteins are reported to influence the elution pH of proteins
[12].

3.2.4

Salt stabilitylong-term

The traditional IEX chromatographysalt elution mode requires product stability in the presence of salt at the chosen
pH value. The results of the salt stabilitylong-term experiments with exposure times of 080 min are shown in Fig. 4.
Operational pH values of 7.5 (AEX) and 4.3 (CEX) were chosen and the salt strength was varied between 0 and 500 mM of
sodium chloride. All activities are scaled to the activity maximum determined in the respective setup. After 20 min, enzymatic activity decreased due to enzyme denaturation. This effect increased with increasing salt strength and interaction time.
When using the CEX setup (Fig. 4A), an 80-min product exposure to 50 mM sodium chloride already led to a loss in activity
of about 32.3%. In contrast to this, the AEX setup (Fig. 4B)
was much less influenced by lower added salt concentrations
of up to 200 mM sodium chloride over the entire time range.
The maximal product inactivation occurred at the highest salt
strength of 500 mM tested and an interaction time of 80 min. In
this setup, 50.3% (CEX) and 45.2% (AEX) of the product was
lost.

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

B
80

100

40
20
0
0

20
100

200

300

400

500

60
40
40
20
0
0

NaCl [mM]

3.2.5

80

60

Time [min]

Time [min]

60
40

Act/Actmax [%]

80

60

100

80

20
100

200

IEX chromatographypH elution mode

Based on the results of the salt stabilitylong-term experiments, a purification scheme using a salt gradient was
no option for the purification of human -galactosidase A.
As a consequence, a low-salt purification alternative by IEX
chromatographypH elution mode was applied. Figure 5A
D shows the results of the pH step and pH gradient elution
in the determined operational pH windows (see above). The
flowthrough fractions and the high-salt wash fractions will not
be discussed any further.
In the CEX setup (Fig. 5A and B), binding of the product
occured at pH 4 (pHgradient,start ). Following the binding step, the
pH was shifted to pH 4.8 (pHelution,start ). During the stepwise pH
shift, the 280 nm signal showed a well-defined peak at CV7 to
CV8 with an absorption maximum of 3537 mAU. Evaluation
of the fractions collected by the 4-NPG assay revealed minor
amounts of product. Thus, contaminants close to elution pH
of the product were removed. In the next step, elution of the
product was performed either by a pH shift in a step (Fig. 5A) or
by a 5 CV gradient (Fig. 5B) to pH 6 (pHelution,end ). During the
pH shift, the 280 nm signal increased in both setups, resulting in
a peak maximum of 33 mAU at CV11.5 for the step pH shift and
a lower 10 mAU peak maximum at CV14.5 for the 5 CV gradient
pH shift. The decreased 280 nm maximum at a delayed process
time in the 5 CV gradient pH shift setup can be explained by the
longer product elution due to an extended time for pH change.
The elution fractions were chosen to be CV11 to CV14 with a total
activity of 285.2 31.9 nmol 4-NP/h for the step and to be CV13
to CV18 with a total activity of 332.4 20.7 nmol 4-NP/h for the
5 CV gradient.
The binding of the product in the AEX setup (Fig. 5C and D)
was performed at pH 8 (pHgradient,start ). A stepwise shift to pH 7
(pHelution,start ) led to a 280 nm peak maximum of 1622 mAU at
CV7 to CV8 . No considerable amounts of product were detected
by the 4-NPG assay and only close contaminants were removed.
The second pH shift to pH 4.7 (pHelution,end ) for product elution
resulted in 280 nm double-peak maxima of 47 and 50 mAU at
CV13 and CV14 for the step pH shift and lower 25 and 25 mAU
peak maxima at CV17 and CV19 for the 5 CV gradient pH shift.
The lower 280 nm maxima at a delayed process time in the
gradient pH shift setup were discussed previously. The elution
fractions were chosen to be CV11 to CV14 with a total activity of
244.0 13.4 nmol 4-NP/h for the step and to be CV12 to CV18
with a total activity of 249.2 13.5 nmol 4-NP/h for the 5 CV
gradient.

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

300

400

NaCl [mM]

3.3
3.3.1

500

Act/Actmax [%]

www.biotecvisions.com

Figure 4. Product salt stabilitylong-term

in AEX and CEX buffers in the sodium
chloride concentration range of 0500 mM.
Product exposure times of 080 min were
evaluated before starting the 4-NPG assay.
Activities are scaled to the maximal activity
determined in each test set. (A) CEX buffer
system at a pH of 4.3 without additional excipients; (B) AEX buffer system at a pH of
7.5 with 0.02% Tween 20 detergent added.

Analysis of chromatographic procedures

Mass balances

Mass balances were determined, as was described in the experimental section. The results are shown in Table 1. For the CEX
setups, the recovery of activity RA was found to be 103.9 7.1%
and the recovery of total protein RTP was 94.1 0.3% for the
step elution. For the 5 CV gradient, similar results were observed
with an RA of 107.5 7.2% and RTP of 89.7 1.2%. For the
AEX setup, the RA was determined to be 98.0 1.1 and the RTP
was 90.6 3.7% for the step elution. The 5 CV gradient again
produced similar results with an RA of 100.8 0.6% and RTP
of 86.9 4.9 %. This indicates that for all processes, no significant product losses were observed. However, minor amounts
of impurities remained in the chromatography column or were
degraded and not detected.

3.3.2

Evaluation of the purification setups

For the evaluation of all purification setups, different factors

were considered (see Section 2.3.3). Yields Y, purification factors
PF, and qualitative nucleic acid factors NF are listed in Table
1. All CEX setups resulted in better yields of 83.9 10.6% for
step elution and 89.9 5.5% for 5 CV gradient elution. The
AEX setups produced yields of about 6070%. A similar trend
was found for the purification factors. The CEX setup resulted
in a PF of 10.9 1.6 for the step and 12.6 1.4 for the 5 CV
gradient. For the AEX setup, PF values of 4.4 0.0 and 4.9 0.1
were observed. Hence, they were smaller by factor of 23. As
regards the PF, gradient elution was found to be superior to
step elution in all experiments. This can be explained by the
broadened elution peak that allows for a more precise and
distinct product fractionation. Finally, the qualitative nucleic
acid factor NF was evaluated. Although the CEX setup was
capable of reducing nucleic acid levels by about 20%, AEX
chromatography resulted in NR values of above 1. This indicates
a qualitative increase in nucleic acid levels. pH dependence of
the buffer species or the influence of Tween 20 can be excluded,
as blank runs were conducted for all experiments. The increase
in NR was thus attributed to nucleic acid fragments attached
to the product or coeluting contaminants. When comparing
the purification performances of the four purification setups
investigated (Table 1), the IEX chromatographypH elution
mode on CEX resins reaches about 20% higher yields as well as
purification factors higher by a factor of 3 than the AEX setups. A
203

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

www.biotecvisions.com

Figure 5. IEX chromatographypH elution mode runs for the P. pastoris supernatant on CEX Source 30 S/Q columns in duplicates. UV
280/254 nm signals and pH values are plotted against column volumes. The corresponding fraction enzyme activities are shown in a bar
diagram below. (A) Step pH shift for product elution (CEX); (B) 5 CV gradient pH shift for product elution (CEX); (C) step pH shift for
product elution (AEX); (D) 5 CV gradient pH shift for product elution (AEX).

qualitative reduction of nucleic acid levels was achieved by

CEX purification only. The best purification setup for process
simplification was found to be the CEX 5 CV gradient elution
from pH 4.8 (pHelution,start ) to pH 6 (pHelution,end ).

3.4

Process simplification

The complex mixture of buffer substances was reduced to a twocomponent system using acetate (pKa 4.76) at pH 4.8 for binding
204

and MES (pKa 6.10) at pH 6 as the elution buffer. The column

capacity limit of 10% product breakthrough was determined to
be 20 CV of the loaded sample (data not shown). The 280/254 nm
signals and the enzyme activities of the assayed fractions of the
CEX product capture at the capacity limit are shown in Fig. 6.
The absorption plateau of 490 mAU from CV1 to CV21
describes the flowthrough after sample injection at pH 4.8
(pHelution,start ). The evaluation of the fractions by enzymatic
assay showed a total activity of 835.3 48.8 nmol 4-NP/h.

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eng. Life Sci. 2015, 15, 195207

www.els-journal.com

www.biotecvisions.com

Table 1. Recovery rates and quality factors for the different setups in the product capturing by IEX chromatographypH elution mode
Purification mode

RA (%)

CEX step
CEX gradient
AEX step
AEX gradient

103.9
107.5
98.0
100.8

RTP (%)

7.1
7.2
1.1
0.6

94.1
89.7
90.6
86.9

Y (%)

0.3
1.2
3.7
4.9

83.9
89.9
67.8
65.3

PF
10.6
5.5
0.8
1.1

10.9
12.6
4.4
4.9

NF

1.6
1.4
0.0
0.1

0.8
0.8
1.3
1.5

0.0
0.0
0.2
0.0

350
300
250

280 nm - 1
280 nm - 2
254 nm - 1
254 nm - 2
pH - 1
pH - 2

6
5.75

200

5.5

150

5.25

100

Activity
[nmol 4-NP/ h]

50

4.75

Run 1
Run 2

4000
3000
2000
1000
0

and is slightly dependent on the buffering species only. However,

it is important to keep the total buffering capacity constant.

6.25

pH

UV 280/ 254 nm [mAU]

RA and RTP denote the recovery of activity and total protein. Y is the yield of product in the purified elution fractions. PF and NF describe the purification
factor and the qualitative nucleic acid reduction, respectively.

10

20

30

40

CV

Figure 6. IEX chromatographypH elution mode runs for the P.

pastoris supernatant at the capacity limit on a CEX Source 30 S
column in duplicates. The complex CEX buffer system was simplified to a two-buffer system using acetate (pKa 4.76) at pH 4.8 and
MES (pKa 6.10) at pH 6. UV 280/254 nm signals and pH values
are plotted against column volumes. The corresponding fraction
enzyme activities are shown in a bar diagram below.

After the wash step for 8 CV at pH 4.8, the pH was shifted to

pH 6 (pHelution,end ) using a 5 CV gradient. The 280 nm signal
reached a peak maximum of 350 mAU at CV30 . The elution
fraction was chosen to be CV29 to CV33 (10% of the maximum
enzyme activity as shown in Fig. 1B) with a total activity of
9284.7 421.5 nmol 4-NP/h. During the 1 M NaCl high-salt
wash from CV38 to CV46 , a 280 nm peak with a maximum of
2500 mAU was observed. The evaluation of the fractions by
enzymatic assay shows a summed up -galactosidase A activity
of 456.4 59.3 nmol 4-NP/h. The recovery of activity RA was
determined to be 97.7 0.5% and the recovery of total protein
RTP was 92.8 3.0%. These results agree with the previous
findings presented in the product capture by pH shift section. A
yield of 84.3 1.2% and a purification factor of 13.2 0.3
were achieved. This agrees with the result obtained for the
complex buffer mixture within the standard deviations (Table
1). The qualitative nucleic acid factor NF was determined to
be 0.9 0.0. As an additional parameter, the concentration
factor CF was calculated (see experimental section) to be
4.12 0.08, indicating an increase in human -galactosidase A
concentration by more than a factor of 4 compared to the
applied sample. It was shown that a process simplification is
possible. The pH value is a rather on/off binding mechanism

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Concluding remarks

It was demonstrated that optimizing an IEX capture step by

product stability screenings and deterministic process development is a powerful tool for protein purification. A systematic
strategy for the purification of salt-intolerant proteins in IEX
was evaluated in a case study using human -galactosidase A
from P. pastoris. The 80-min time frame for the determination
of long-term stability can be varied. The users of this systematic
strategy can extend or shorten the exposure times according to
the expected process conditions.
During stages (1) and (2) pH stabilityshort-/longterm of the strategy proposed, pH boundaries of human galactosidase A were determined. For the AEX buffer system,
an additive screening had to be performed to ensure longterm product stability. For this purpose, the detergent Tween
20 was used. In a deterministic IEX purification process design,
the subsequent pH gradient chromatography (3) resulted in
the experimental elution pH of the product. As the enzyme
was found to be salt-unstable at its elution pH even at NaCl
concentrations of 0.1 M during stage (4) salt stabilitylongterm, IEX chromatographypH elution mode (5b) was the
purification method of choice. Different operational modes for
pH shift elution were compared, including different IEX modes
(CEX vs. AEX) and different setups of elution by pH shift (step
vs. 5 CV gradient). The best purification results were obtained
in a CEX operational mode with 5 CV gradient elution. Step
elution should be chosen especially for proteins that are known
to aggregate if eluted at the pH value corresponding to their experimental pI. The critical pH value is then exceeded for a short
period of time only.
In a simplified two-buffer component process, human galactosidase A was purified from a P. pastoris supernatant based
on the results of prior screening experiments. In this setup, a
purification factor of 13.2 and an overall yield of 84.3% were
achieved, which agreed very well with the experimental results
for the complex buffer mixture. The product was four times
concentrated and the level of nucleic acid content was qualitatively reduced. The results agreed with the experiments using
the complex buffer system and lower sample volumes. The IEX
chromatographypH elution mode makes IEX purification
processes accessible for salt-unstable products and, hence, for a
much broader spectrum of products.
205

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

The combination of protein stability tests and deterministic

screenings should now be extended to cover other purification
modes, such as hydrophobic interaction chromatography, affinity chromatography, and aqueous two-phase systems. This will
result in a toolbox for process development. For upstream and
formulation studies, such an approach may provide valuable
information for enhancing the efficiency of these processes.
The authors would like to acknowledge the financial support by the
German Federal Ministry of Education and Research (BMBF)
funding code 0316071Band the support by the industrial partner
BioIngenium (Spain). This research work is part of the project ERA
Net Euro Trans Bio - 6: Development of an integrated strategy for
a high-throughput process development platform on a micro-scale
format FORECAST. The authors bear the complete responsibility
for the content of the publication.
The authors have declared no conflict of interest.

References

[1] Krishnan, S., Chi, E. Y., Webb, J. N., Chang, B. S. et al., Aggregation of granulocyte colony stimulating factor under physiological conditions: Characterization and thermodynamic inhibition. Biochemistry 2002, 41, 64226431.
[2] Chi, E. V. A. Y., Krishnan, S., Kendrick, B. S., Chang, B. S. et
al., Roles of conformational stability and colloidal stability in
the aggregation of recombinant human granulocyte colonystimulating factor. Protein Sci. 2003, 12, 903913.
[3] Chen, B. L., Wu, X., Babuka, S. J., Hora, M., Solubility of
recombinant human tissue factor pathway inhibitor. J. Pharm.
Sci. 1999, 88, 881888.
[4] Brange, J., Andersen, L., Laursen, E. D., Meyn, G. et al., Toward
understanding insulin fibrillation. J. Pharm. Sci. 1997, 86, 517
525.
[5] Tsai, A. M., van Zanten, J. H., Betenbaugh, M. J., II. Electrostatic effect in the aggregation of heat-denatured RNase A and
implications for protein additive design. Biotechnol. Bioeng.
1998, 59, 281285.
[6] Berg, A., Schuetz, M., Dismer, F., Hubbuch, J., Automated
measurement of apparent protein solubility to rapidly assess
complex parameter interactions. Food Bioprod. Process 2014,
92, 133142.
[7] GE Healthcare, Ion Exchange Chromatography & Chromatofocusing: Principles and Methods, www.uni-ulm.de/uploads/
media/Ion exchange chromatography and Chromatofocusing
02.pdf, 2007.
[8] Hallgreen, E., Kalman, F., Farnan, D., Horvath, C. et al., Protein retention in ion-exchange chromatography: Effect of net
charge and charge distribution. J. Chromatogr. A 2000, 877,
1324.
[9] DePhillips, P, Lenhoff, A. M., Determinants of protein retention characteristics on cation-exchange adsorbents. J. Chromatogr. A 2001, 933, 5772.
[10] Ahamed, T., Nfor, B. K., Verhaert, P. D. E. M., VanDedem, G.
W. K. et al., pH-gradient ion-exchange chromatography: An
206

[11]

[12]

[13]

[14]

[15]

[16]
[17]

[18]
[19]

[20]

[21]

[22]

[23]

[24]

[25]

www.biotecvisions.com

analytical tool for design and optimization of protein separations. J. Chromatogr. A 2007, 1164, 181188.
Kroner, F., Hubbuch, J., Systematic generation of buffer systems for pH gradient ion exchange chromatography and their
application. J. Chromatogr. A 2013, 1285, 7887.
Zhu, K., Zhao, J., Lubman, D. M., Miller, F. R. et al., Protein pI
shifts due to posttranslational modifications in the separation
and characterization of proteins. Anal. Chem. 2005, 77, 2745
2755.
Ahamed, T., Chilamkurthi, S., Nfor, B. K., Verhaert, P. D. E.
M. et al., Selection of pH-related parameters in ion-exchange
chromatography using pH-gradient operations. J. Chromatogr.
A 2008, 1194, 2229.
Kroner, F., Hanke, A. T., Nfor, B. K., Pinkse, M. W. H. et al.,
Analytical characterization of complex, biotechnological feedstocks by pH gradient ion exchange chromatography for purification process development. J. Chromatogr. A 2013, 1311,
5564.
Bhut, B. V., Christensen, K. A, Husson, S. M., Membrane
chromatography: Protein purification from E. coli lysate using newly designed and commercial anion-exchange stationary
phases. J. Chromatogr. A 2010, 1217, 49464957.
Cromwell, M. E. M., Hilario, E., Jacobson, F., Protein aggregation and bioprocessing. AAPS J. 2006, 8, 572579.
Chi, E. Y., Krishnan, S., Randolph, T. W., Carpenter, J. F.,
Physical stability of proteins in aqueous solution: Mechanism
and driving forces in nonnative protein aggregation. Pharm.
Res. 2003, 20, 13251336.
Wang, W., Protein aggregation and its inhibition in biopharmaceutics. Int. J. Pharm. 2005, 289, 130.
Ericsson, U. B., Hallberg, B. M., DeTitta, G. T., Decker, N. et al.,
Thermofluor-based high-throughput stability optimization of
proteins for structural studies. Anal. Biochem. 2006, 357, 289
298.
Jancarik, J., Pufan, R., Hong, C., Kima, S.-H. et al., Optimum solubility (OS) screening: an efficient method to optimize buffer conditions for homogeneity and crystallization
of proteins. Acta Crystallogr. D Biol. Crystallogr. 2004, 60, 1670
1673.
Platis, D., Drossard, J., Fischer, R., Ma, J. K. et al., New downstream processing strategy for the purification of monoclonal
antibodies from transgenic tobacco plants. J. Chromatogr. A
2008, 1211, 8099.
Schwarz, A., Affinity purification of monoclonal antibodies,
in: P. Bailon, G. K. Ehrlich, W.-J. Fung, W. Berthold (Eds.),
Affinity chromatography: Methods and protocols 147, chapter
5, Humana Press, Totowa, NJ 2000, pp. 4956.
Choy, D., Creagh, L., Haynes, C., Improved isoelectric focusing
chromatography on strong anion exchange media via a new
model that custom designs mobile phases using simple buffers.
Biotechnol. Bioeng. 2013, 111, 552564.
Kang, X., Kutzko, J. P., Hayes, M. L., Frey, D. D., Monoclonal
antibody heterogeneity analysis and deamidation monitoring with high-performance cation-exchange chromatofocusing using simple, two component buffer systems. J. Chromatogr.
A 2013, 1283, 8997.
Gevorgyan, G. K., Davtyan, M. A., Hambardzumyan, A. A., A
simple purification procedure of D amino acid oxidase from
Candida guilliermondii. Microbiology 2012, 81, 498504.

C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.els-journal.com

Eng. Life Sci. 2015, 15, 195207

[26] Maiser, B., Kroner, F., Dismer, F., Brenner-Wei, G. et al.,

Isoform separation and binding site determination of monoPEGylated lysozyme with pH gradient chromatography. J.
Chromatogr. A 2012, 1268, 102108.
[27] Schiffmann, R., Murray, G. J., Treco, D., Daniel, P. et al., Infusion of -galactosidase A reduces tissue globotriaosylceramide
storage in patients with Fabry disease. Proc. Natl. Acad. Sci.
USA 2000, 97, 365370.
[28] deRezende, S. T., Guimaraes, V. M., deCastro, M., Felix, C. R.,
Purification and characterization of an -galactosidase from
Aspergillus fumigatus. Brazilian Arch. Biol. Technol. 2005, 48,
195202.
[29] Marin, S., Guynot, M. E., Sanchis, V., Ramos, A. J., Hydrolytic enzyme activities as indicators of fungal spoilage
in bakery products. J. Sci. Food Agric. 2003, 83, 685
691.
[30] Ioannou, Y. A., Zeidner, K. M., Grace, M. E., Desnick, R.
J., Human Alpha-galactosidase A: Glycosylation site 3 is


C

2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

[31]

[32]

[33]

[34]

www.biotecvisions.com

essential for enzyme solubility. Biochem. J. 1998, 332, 789

797.
Zeilinger, S., Kristufek, D., Arisan-Atac, I., Hodits, R. et al.,
Conditions of formation, purification, and characterization of
an alpha-galactosidase of Trichoderma reesei RUT C-30. Appl.
Environ. Microbiol. 1993, 59, 13471353.
Desnick, R. J., Dean, K. J., Grabowski, G., Bishop, D. F. et al.,
Enzyme therapy in Fabry disease: Differential in vivo plasma
clearance and metabolic effectiveness of plasma and splenic
alpha-galactosidase A isozymes. Proc. Natl. Acad. Sci. USA
1979, 76, 53265330.
Werner, V. A., Konformerspezifitat der Proteinphosphatase 2 A bei der Dephosphorylierung prolinspezifischer
Phosphorylierungsstellen. der Martin-Luther-Universitat
Halle-Wittenberg 2001. http://webdoc.sub.gwdg.de/ebook/w/
2002/pub/bio/02H052/prom.pdf.
Gagnon, P., Dissociation of antibodycontaminant complexes
with hydroxyapatite. Bioprocess J. 2011, 9, 1424.

207