1. Process Biochemistry 46 (2011) 101–107
Contents lists available at ScienceDirect
Process Biochemistry
journal homepage: www.elsevier.com/locate/procbio
Purification of rabbit polyclonal immunoglobulin G using anion exchangers
Rattana Wongchuphana,e
, Beng Ti Teyb,d
, Wen Siang Tanc,d
, Senthil Kumar Subramanianc
,
Farah Saleena Taipa
, Tau Chuan Linga,∗
a
Department of Process and Food Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400, Serdang, Selangor Darul Ehsan, Malaysia
b
Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400, Serdang, Selangor Darul Ehsan, Malaysia
c
Department of Microbiology, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, 43400, Serdang, Selangor Darul Ehsan, Malaysia
d
Institute of Bioscience, Universiti Putra Malaysia, 43400, Serdang, Selangor Darul Ehsan, Malaysia
e
Department of Chemistry, Faculty of Science and Technology, Suratthani Rajabhat University, 84100, Surat Thani, Thailand
a r t i c l e i n f o
Article history:
Received 25 February 2010
Received in revised form 21 July 2010
Accepted 22 July 2010
Keywords:
Polyclonal IgG
STREAMLINETM
DEAE
Rabbit serum
Albumin removal
Anion exchange adsorbents
Negative chromatography antibody
purification
a b s t r a c t
Negative chromatography antibody purification (N-CAP) using the weak anion exchanger STREAMLINETM
DEAE to extract impurities while retaining the target antibody is proposed as an effective method for the
recovery of antibody from rabbit serum. The effects of pH and initial protein concentration on the removal
of albumin were investigated. The optimal pH and initial protein concentration for the efficient removal
of albumin from rabbit serum were pH 8.0 and 0.5 mg/ml, respectively. Under optimal binding conditions,
DEAE successfully removed more than 90% of the albumin from rabbit serum with less than 20% IgG loss.
This process offered good polyclonal IgG yield of 80% with a purity of 83% and a purification factor of 5.5.
The use of a strong anion exchanger like STREAMLINETM
Q XL for albumin removal was also explored.
Under similar optimized conditions, albumin removal by Q XL was as high as 90%. However, IgG recovery
and purity were reduced to about 70% and 62%, respectively. Thus, N-CAP using the anion exchanger
DEAE removes albumin from rabbit serum and thereby offers an efficient means of purifying polyclonal
antibodies.
© 2010 Elsevier Ltd. All rights reserved.
1. Introduction
Protein A affinity chromatography, which is used widely in anti-
body purification, offers specificity and highly efficient product
capture. Protein A has been the standard antibody manufactur-
ing method during the past few decades due to its selectiveness,
optimal throughput, and its stability to be cleaned and re-used
[1,2]. In addition, it has been used for product capture in multistep
platforms [3]. However, since advances in cell culture technol-
ogy have been developed, this approach now has a number of
disadvantages. As the antibody titer in cell culture increases, the
downstream processes become a bottleneck and the product recov-
ery cost increases in line with the production scale. In such a case, it
may not be cost-effective to purify antibody using Protein A affinity
chromatography. Moreover, the low stability of Protein A is a limi-
tation in production scale antibody purification [2]. Therefore, due
to its high cost and concerns over low ligand stability after many
cycles of production, there is a genuine need for a replacement
for Protein A affinity chromatography [1]. Hence, the development
∗ Corresponding author. Tel.: +60 603 89466447; fax: +60 603 89464440.
E-mail address: ltc555@eng.upm.edu.my (T.C. Ling).
of a cost-effective alternative method for antibody purification is
needed to overcome these concerns.
As recently recommended by some adsorbent manufacturers,
negative chromatography antibody purification (N-CAP) can be
used for antibody purification from animal serum. Typically, this
requires passing the clarified antibody feedstock through a series
of negative chromatography columns to remove all the impuri-
ties. This approach could be an efficient way of overcoming the
high cost and the physical capacity limits of Protein A affinity chro-
matography. Ion exchange chromatography (IEC) is a simple, rapid
technique used in protein separation that depends on the differen-
tial ability of various proteins to be adsorbed onto the adsorbent
[4].
IEC is used commonly in protein purification due to its high bind-
ing capacity and cost-effectiveness [5–8]. Commercially available
ionic matrices for protein purification (including DEAE sepharose
and Q XL) are highly activated with a high density of ionic lig-
ands such as diethylaminoethyl and quaternary amine groups [9].
These matrices have high adsorption capacity towards proteins
even in the presence of Escherichia coli (E. coli) biomass and/or
cell debris [4,10,11]. Recently, Pessela and colleagues successfully
purified immunoglobulin G (IgG) from whey proteins concentrate
(WPC) by eliminating BSA with DEAE-agarose adsorbent [9]. BSA
was adsorbed strongly on highly activated DEAE-agarose, while
1359-5113/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.procbio.2010.07.023

2. 102 R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107
IgG was adsorbed on and desorbed from anion exchange adsor-
bents weakly activated with mild monoaminoethyl-N-aminoethyl
(MAMAE) ionic groups. Such weakly activated adsorbents have
been developed and used for the selective adsorption of only large
proteins such as ␤-galactosidase (Mr, 70 kDa for each monomer;
with a predominant tetrameric form) [4] and IgG (Mr = 150 kDa)
[9]. BSA was also found to be adsorbed onto weakly activated
adsorbents. This phenomenon is related to the tendency of BSA
to form aggregates either with itself or other proteins [4,12–14].
The mechanism of adsorption of these matrices is based on mul-
tipoint adsorption [9,15]. Aggregates of albumin complexes and
albumin–IgG complexes exhibit pseudo-affinity for different lig-
ands on either weakly or strongly activated adsorbents [4]. The
presence of albumin in animal serum can hinder the purification
of an antibody. Thus, albumin must be eliminated before purifying
antibodies with anion exchange adsorbents. Removal of albumin
from any antibody-containing source can be achieved via positively
charged adsorbents.
In the present study, anion exchange matrices were used to
purify polyclonal antibodies from rabbit serum by removing albu-
min in a batch binding system. Polyclonal antibodies against
hepatitis B core antigen (anti-HBcAg IgG) were generated in rab-
bits. The albumin removal capacity of two anion exchange matrices
(STREAMLINETM DEAE and STREAMLINETM Q XL) was evaluated.
The albumin-adsorbing capacity of DEAE sepharose was assessed
as functions of the pH of the medium and the initial protein con-
centration. The recovery of polyclonal IgG as well as its purity
was determined after the removal of albumin. The strong anion
exchanger Q XL, which is known to offer a higher binding capacity
for albumin, was also studied. The efficiency of both the weak and
strong anion exchangers were compared and discussed in terms of
yield and purity of the polyclonal antibodies recovered.
2. Materials and methods
2.1. Materials
The STREAMLINETM
DEAE and STREAMLINETM
Q XL adsorbents were purchased
from GE Healthcare (Uppsala, Sweden). DEAE, which is a modified sepharose matrix,
is a weak anion exchanger that provides about 0.13–0.21 mmol/ml of total ionic
adsorbent capacity. It has high binding capacity towards BSA, at least 40 mg/ml
adsorbent. On the other hand, the matrix structure of Q XL is based on highly cross-
linked 6% agarose beads modified by including a crystalline quartz core and bound
dextran. The strong Q ion exchange groups are then coupled to these dextran chains
through chemically stable ether bonds. This causes an increase in the effective inter-
acting volume as well as in the steric availability of the ligands for the substance to
be adsorbed. The Q XL adsorbent offers a total ionic capacity of 0.23–0.33 mmol/ml
adsorbent. The binding capacity of Q XL for BSA is up to 110 mg/ml adsorbent. Bovine
serum albumin (BSA) was purchased from Sigma (St. Louis, MO, USA). The isoelectric
point (pI) of BSA given in the product information is 4.7. Serum samples collected
from immunized rabbits were used as the source of polyclonal anti-HBcAg IgG.
2.2. Production of HBcAg in E. coli
E. coli W31101Q cells carrying the plasmid pTacpcore encoding full-length
HBcAg [16] were cultured in Luria Bertani (LB) broth containing ampicillin
(100 ␮g/ml). The bacteria were grown at 37 ◦
C until the OD600nm reached about
0.6, followed by adding isopropyl-␤-d-thiogalactopyranoside (IPTG; final concen-
tration of 0.5 mM) to induce HBcAg expression. Then, the cells were allowed to grow
for another 16–20 h. The cells were harvested by centrifugation at 3752 × g (Avanti
JLA-16.250 rotor, Beckman, USA) at 4 ◦
C for 15 min.
2.3. Purification of HBcAg by sucrose density gradient ultracentrifugation
Sucrose gradient density ultracentrifugation was employed for the purification
of HBcAg [17,18]. The cell pellets were resuspended in Tris–HCl buffer (50 mM, pH
8.0 containing 0.1% Triton X-100) and then lysed with lysozyme (0.2 mg/ml) in the
presence of MgCl2 (4 mM) and DNase (20 ␮g/ml) at room temperature for 2 h. The
cell suspension was clarified by centrifugation at 12,096 × g (Avanti JA-25.50 rotor,
Beckman, USA) for 20 min at 4 ◦
C, the HBcAg was precipitated with ammonium
sulphate (35% saturation) and it was centrifuged at 12,096 × g. The pellet was sus-
pended in dialysis buffer (50 mM Tris–HCl, pH 8.0 containing 150 mM NaCl) and
dialyzed in the same buffer (2 L) for 16 h at 4 ◦
C. The dialyzed sample was loaded on
a 8–40% continuous sucrose gradient and centrifuged at 210,053 × g (SW 41 rotor,
Beckman, USA) for 5 h at 4 ◦
C. The gradient was fractionated (0.5 ml per fraction)
and fractions containing HBcAg were concentrated with a Vivaspin concentrator
(300,000 MWCO, Vivascience, UK) by centrifugation at 3024 × g for several hours at
4 ◦
C.
2.4. Preparation of sera
New Zealand white rabbits were injected subcutaneously with HBcAg (250 ␮g;
0.5 ml) containing 0.5 ml of complete Freund’s adjuvant (Millipore, USA). Four weeks
after first injection, the rabbits were injected with a mixture of HBcAg and incom-
plete Freund’s adjuvant (Millipore, USA) in subsequent boosters. Two boosters were
administered every 2 weeks and the rabbits were bled 10 days after each booster.
Rabbit serum was collected by centrifugation at 2700 × g for 10 min. Sera containing
polyclonal anti-HBcAg IgG from rabbits were pooled as the antibody feedstock for
N-CAP.
2.5. Protein assays
2.5.1. The Bradford assay
The total protein content of serum protein solutions before and after albumin
removal was determined by the Bradford method [19]. BSA was used as a protein
standard.
2.5.2. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)
Serum protein solutions before and after albumin removal were separated on
10% polyacrylamide gel under non-reducing condition [20]. The gel was stained with
Coomassie Brilliant Blue (CBB) R-250 and destained with destaining solution (30%
methanol and 10% glacial acetic acid). The intensity of the protein band on SDS-PAGE
was determined using a Gel-Doc system [21,22] with some minor modifications.
2.5.3. Enzyme-linked immunosorbent assay (ELISA)
The amount of polyclonal anti-HBcAg IgG present in serum before and after the
albumin depletion process was quantified with ELISA [23]. A microtiter plate was
first coated with HBcAg (100 ng/well, 100 ␮l) overnight at 4 ◦
C. The coated plate
was first washed with TBST (50 mM Tris–HCl and 150 mM NaCl, pH 7.5 containing
0.05% Tween 20). The plate was then blocked with 10% milk diluent/blocking solu-
tion (KPL, USA) for 2 h. After washing, 100-␮l samples containing anti-HBcAg IgG
diluted in TBS buffer were added to the plate and then incubated at room tempera-
ture for 1 h. The plate was washed once more and anti-rabbit antibody conjugated
to alkaline phosphatase (KPL, USA; 1:5000) was loaded into the wells. The plate was
further incubated at room temperature for 1 h. After giving six washes with TBST, p-
nitrophenyl phosphate (pNPP 1 mg/ml, Sigma) in 1 M diethanolamine buffer (DEA,
pH 9.8 containing 0.5 mM MgCl2·6H2O) was added to the plate. The absorbance at
405 nm was measured using a universal microplate reader (Bio-Rad, USA). All mea-
surements were performed in quadruplicate. Mouse anti-HBcAg IgG monoclonal
antibody (Millipore, USA) was used as a standard.
2.6. BSA adsorption on DEAE sepharose
At pHs beyond its pI (4.7), BSA is negatively charged and is able to bind DEAE
sepharose carrying a positive charge. On the other hand, at pHs lower than its pI, very
little or no binding of BSA is observed. Thus, the binding capacity of DEAE sepharose
for BSA was tested in binding buffers at different pH ranging from 4.0 to 8.0 (25 mM
sodium acetate buffers pH 4.0–5.0; 25 mM sodium phosphate buffers pH 6.0–8.0).
The binding capacity of DEAE sepharose for BSA was estimated and compared to
that recommended by the manufacturer. BSA (10 mg) was allowed to bind to DEAE
sepharose adsorbent (0.25 ml) on a roller mixer until no more protein was adsorbed.
Absorbance readings at 280 nm were taken before and after binding. The amount of
BSA was determined using a U-2900 spectrophotometer (Hitachi, Japan). At 280 nm,
the extinction coefficient of BSA is 6.67 for a 1% solution (E1%
279
= 6.67). Therefore, the
extinction coefficient used to calculate the concentration of BSA solutions (1 mg/ml)
was fixed at 0.667 ml/(mg cm). The measurements were carried out in duplicate.
2.7. Albumin removal at various pHs
The effect of pH on albumin removal by DEAE sepharose was investigated at pHs
ranging from 5.0 to 8.0. Briefly, DEAE sepharose adsorbent (0.25 ml) was incubated
with a serum sample (0.75 mg/ml initial concentration, 10 ml) at various pHs. A
batch binding system was used with a universal bottle that was mixed well on a
roller mixer. A fraction was collected every 15 min for 1 h. Samples were analyzed
for total protein content with the Bradford assay. The intensity of the albumin and
IgG bands on SDS-PAGE was determined using a Gel-Doc system (Quantity One,
Bio-Rad, USA). The protein profile was used to assess the optimal pH for efficient
albumin extraction with minimal antibody binding.
2.8. Albumin removal at various initial protein concentrations
A serum sample (10 ml) was prepared in 25 mM sodium phosphate buffer, pH
8.0. Solutions with two initial protein concentrations (0.5 and 0.75 mg/ml) were

3. R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107 103
mixed separately with 0.25 ml DEAE sepharose or 0.1 ml Q XL. The binding continued
for 1 h on a roller mixer. A fraction was collected every 15 min and then analyzed
quantitatively by three protein assays. Total protein concentration was estimated
by the Bradford assay. The intensity of the albumin band was estimated using a Gel-
Doc system. ELISA was used to determine the quantity of polyclonal anti-HBcAg IgG
in the serum protein solution before and after albumin removal.
2.9. Calculations
Yield is stated as the percentage of polyclonal IgG in the albumin-depleted serum
divided by the initial amount of polyclonal IgG in the serum sample:
Yield =
Amount of IgG in albumin-depleted serum
Amount of IgG in serum sample
× 100% (1)
The purity (or relative quantity) of polyclonal anti-HBcAg IgG is expressed as
the ratio of the amount of polyclonal anti-HBcAg IgG to the total amount of protein
in the serum sample either before or after albumin removal:
Purity =
Amount of polyclonal anti-HBcAg IgG
Amount of total protein
(2)
The purification factor (PF) is calculated from the purity of polyclonal IgG in the
albumin-depleted serum divided by the purity of polyclonal anti-HBcAg IgG in the
serum sample:
Purification factor =
Purity of polyclonal IgG in albumin-depleted serum
Purity of polyclonal IgG in serum sample
(3)
3. Results and discussion
3.1. Adsorption of BSA on DEAE sepharose
DEAE sepharose 6BCL is a commercially available anion
exchanger that has high binding capacity for albumin (up to
40 mg/ml adsorbent). In this study, DEAE sepharose was incubated
with BSA solution under varying pHs ranging from acidic to basic.
Fig. 1 shows the percentage of BSA adsorbed onto DEAE sepharose
at various pHs. There was no adsorption of BSA onto the adsorbent
at pH 4.0. This can be explained by the fact that at pH below its
pI (around 4.7), BSA assumes a net positive charge, and there is an
electropositive repulsion between positively charged BSA and the
cationic surface of DEAE sepharose, an observation similar to one
presented earlier in the literature [5]. On the other hand, greater
than 80% BSA adsorption was observed at pHs higher than its pI.
At pH 8.0, the highest amount of BSA was bound to the DEAE
sepharose, 10.1 ± 0.1 mg (Fig. 1). At pHs beyond its pI (4.7), BSA is
negatively charged and is able to bind DEAE Separose, activated
with ionized cationic groups. It also has been reported that the
tertiary structure of BSA bound to DEAE sepharose is relatively
unaltered in the pH range between 5.0 and 8.0 at room temperature
[5].
The high adsorption of BSA onto DEAE sepharose is due to the
fact that this highly activated adsorbent has high-density ionic lig-
ands so that it can adsorb large quantities of protein in solutions
Fig. 1. Amount of BSA adsorbed onto DEAE sepharose (0.25 ml) as a function of pH
of the medium. The pH was varied as follows: 25 mM sodium acetate buffer at pH
4.0 and 5.0; 25 mM sodium phosphate buffer at pH 6.0–8.0. The results are from two
independent measurements.
and even in crude extracts of E. coli [24,25]. Serum albumins are
small and water-soluble proteins that are able to interact with
many different compounds or even other serum albumins, exhibit-
ing a tendency to form homo or hetero protein aggregates even at
low concentrations [15]. Hence, this multipoint protein–adsorbent
interaction forms the basis of the adsorption of these proteins to
most chromatographic adsorbents [4,9,11,26]. It has been noted
that protein aggregation in solutions is pH-dependent [27,28].
The result obtained here might reinforce the perception that the
BSA aggregation was induced strongly by attractive ionic forces
at the pHs beyond its pI (4.7) as it was adsorbed onto the DEAE
sepharose.
3.2. Effect of pH on serum albumin depletion
Initially, 10 mg of total protein in a serum sample at pH 8.0 was
subjected to 0.25 ml DEAE sepharose. It was found that about 80%
of the serum albumin was removed. The ELISA result, however,
showed that there was no improvement in the relative quantity of
anti-HBcAg IgG (data not shown) in the albumin-depleted serum.
This suggests that IgG molecules were removed along with the
albumin. Band intensity analysis showed that the albumin deple-
tion was associated with 25% IgG loss. About 75% of total serum
protein was adsorbed onto the adsorbent. Moreover, when the
amount of total serum protein was increased to 12.5 mg, reduced
albumin binding (from 80% to 50%) was observed (data not shown).
This smaller amount of bound albumin (30% less) corresponds to
a lower degree of albumin aggregation. Indeed, a previous study
showed that aggregate size can be controlled by the albumin
concentration [15]. This result suggests that initial protein concen-
tration is another important parameter that impacts the retention
of albumin by adsorbents. This parameter was further investigated
to improve albumin removal from the serum sample while retain-
ing more antibodies.
Therefore, rabbit serum albumin binding to DEAE sepharose was
assessed as a function of both pH and initial protein concentration,
and the latter was fixed at 0.75 mg/ml. Due to the fact that albumin
binding did not increase even when the binding time was increased
beyond 1 h, binding was performed within 1 h in this study.
After adsorption of serum proteins onto DEAE sepharose for 1 h
at the indicated pHs, unbound proteins were separated on 10% SDS-
PAGE and the levels of albumin and IgG were estimated by band
intensity analysis (see Fig. 2). Most of the albumin was removed
efficiently at pH 7.0 and pH 8.0. SDS-PAGE gels presented in Fig. 2a
and b show a qualitative analysis of rabbit serum proteins before
and after albumin removal under both pH conditions. With 1 h of
binding time, less serum albumin was expected to be removed at
pH 7.0 than at pH 8.0. However, although the majority of the albu-
min was removed under both conditions, the kinetics of albumin
binding were different. By comparing lane 5 of Fig. 2a with lane
3 of Fig. 2b, the band intensity of albumin at pH 7.0 after 1 h of
adsorption was similar to that at pH 8.0 after 30 min of adsorption.
Thus, albumin adsorption onto DEAE sepharose occurs with faster
kinetics at pH 8.0.
This observation was further supported by quantitative analy-
sis Fig. 2c shows the amount of albumin removed from the serum
sample (as determined by band intensity analysis) with 1 h of bind-
ing and at pH 7.0 and 8.0. At pH 8.0, rapid albumin depletion was
observed in which about 80% of the albumin was depleted from
the initial protein solution within the first 15 min. Approximately
89 ± 1% of the albumin was removed observed at pH 8.0 after 1 h of
adsorption, whereas at pH 7.0, 85 ± 1% of the albumin was removed
after 1 h. Fig. 2d shows the percentage of IgG loss at each time point
under different conditions. Band intensity analysis revealed that
IgG was adsorbed onto the DEAE sepharose at the two pHs with
different efficiencies. About 20% of the IgG was lost at pH 8.0, while

4. 104 R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107
Fig. 2. Non-reducing SDS-PAGE (10%) analysis of protein adsorption from rabbit serum onto DEAE sepharose as a function of pH of the medium. The adsorption conditions
were: 0.25 ml settled bed volume of DEAE sepharose; 0.75 mg/ml initial protein concentration; room temperature. Albumin removal was carried out in 25 mM sodium
phosphate buffers at (a) pH 7.0 and (b) pH 8.0 for 1 h. Lanes: (M) protein markers in kDa; (1) initial serum proteins; (2) serum proteins adsorbed in 15 min; (3) serum proteins
adsorbed in 30 min; (4) serum proteins adsorbed in 45 min; (5) serum proteins adsorbed in 60 min. Quantitative analysis of albumin removal (c) and total unbound IgG (d)
was performed based on the intensity of the bands. The results are from duplicate measurements.
up to 30% was lost at pH 7.0. Hence, pH 8.0 was better than pH 7.0
for removing albumin from rabbit serum with minimal IgG loss.
Based on these results, it can be concluded that rabbit serum
albumin is adsorbed rapidly onto highly activated DEAE sepharose
at pH 7.0. However, the binding is more efficient at pH 8.0. As
explained earlier, the extent of protein–protein interactions was
possibly highest at pH 8.0, and hence this was the optimal pH
for removing albumin. In addition, the IgG loss at this pH could
be explained by the formation of albumin–IgG complexes [9].
The greater loss of IgG at pH 7.0 could be explained by the fact
that the overall protein charge might be such that it allowed the
albumin–IgG interactions ensuing in the formation of large protein
complexes. As a result, the better IgG adsorption by DEAE sepharose
occurred at this pH [9]. As pH 8.0 is higher than the pI of rabbit IgG
(around 7.8) [29], it is likely that the formation of albumin–IgG
complexes at this pH is low. It should be noted that the aggregation
state of the protein may be altered as the effect of pH. Apparently,
at pH 8.0, there was minimal IgG loss and albumin elimination was
adequate.
3.3. Effect of initial protein concentration on albumin depletion
The initial protein concentration also significantly impacted the
quantity of albumin removed from rabbit serum (our earlier obser-
vation in Section 3.2). At high protein concentrations, albumin
removal by DEAE sepharose diminished. Hence, low initial pro-
tein concentrations (0.5 and 0.75 mg/ml) were finalized upon for
albumin removal. Table 1 shows that, at pH 8.0, the amount of
total protein bound to the DEAE sepharose decreased as the ini-
tial concentration of the serum sample increased. At the initial
protein concentration of 0.5 mg/ml, about 80% of the total pro-
tein was bound, while only 66% bound at 0.75 mg/ml. In other
words, when the initial protein concentration was low (0.5 mg/ml),
the total unbound protein was found to be only 20%. If all the
serum IgG remained in this 20% unbound total protein fraction,
then this adsorption condition at 0.5 mg/ml initial protein would
have been adequate for purifying antibodies. Quantitative analysis
in this regard was required and is discussed below.
To estimate the efficiency of albumin removal by DEAE
sepharose at pH 8.0, the SDS-PAGE gels shown in Fig. 2b
(0.75 mg/ml) and Fig. 3 (0.5 mg/ml) were analyzed and compared.
As illustrated in Fig. 2b, with the former condition, about 90% of the
albumin was adsorbed onto the adsorbent. The adsorption of IgG
was found to be about 20 ± 2% at that pH. According to the band
intensity analysis (Fig. 4a), at the initial protein concentration of
0.5 mg/ml, a dramatic reduction of albumin (82 ± 1%) was observed
within 15 min of binding. Further removal of albumin (95 ± 1%)
was obtained after 1 h. Similarly, rapid albumin adsorption onto
DEAE adsorbent was observed by Pessela et al. [9]. As stated previ-
ously, albumin tends to form self-aggregates. The aggregation rate
of BSA in solution decreases with increasing protein concentra-
tion [30]. The rise of the aggregation rate is associated with the
increase in water mobility. It was noted that greater than 90% of
the albumin was efficiently removed at initial protein concentra-
tions of both 0.5 and 0.75 mg/ml. At the former concentration, the
removal of albumin was accompanied by low IgG adsorption onto
the DEAE sepharose (about 16 ± 3%). In other words, 80% of the
rabbit polyclonal IgG was recovered and about 90% of the albumin
was removed at the low initial protein concentration. Therefore,
Table 1
Total protein adsorbed onto DEAE sepharose as a function of initial protein concen-
tration at pH 8.0.
Initial protein
concentration
(mg/ml)
Amount of total
protein bound to DEAE
sepharose (mg)
Amount of
total protein
unbound (%)
0.5 4.09 ± 0.04 19
0.75 5.15 ± 0.21 34

5. R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107 105
Fig. 3. Non-reducing SDS-PAGE (10%) analysis of proteins from rabbit serum on
DEAE sepharose. The adsorption conditions were: 0.25 ml settled bed volume of
DEAE sepharose; 0.5 mg/ml initial protein concentration, pH 8.0; 1 h at room tem-
perature. Lanes: (M) protein markers in kDa; (1) initial serum proteins; (2) proteins
adsorbed in 15 min; (3) proteins adsorbed in 30 min; (4) proteins adsorbed in
45 min; (5) proteins adsorbed in 60 min.
both initial protein concentrations seem to be good for albumin
depletion and antibody yield, but these parameters are somewhat
better at the lower protein concentration (0.5 mg/ml). Contamina-
tion with unbound albumin, however, would affect the purity of
the IgG. It was found that less than 5% of the albumin remained
in albumin-depleted serum at the initial protein concentration of
0.5 mg/ml, while more than 5% of it remained in the solution at
0.75 mg/ml. Hence, the purity of the recovered IgG was better at
0.5 mg/ml than at 0.75 mg/ml initial protein concentration.
The adsorption was performed for 1 h because no further
improvement in purity was observed as the binding time was pro-
longed. Fig. 4b shows the relative amount of polyclonal anti-HBcAg
IgG after adsorption onto DEAE sepharose as a function of initial
protein concentration. It was found that 83 ± 4% pure polyclonal
anti-HBcAg IgG was obtained at the initial protein concentration of
0.5 mg/ml. On the other hand, at 0.75 mg/ml, there was no improve-
ment in the relative quantity of polyclonal IgG, which remained
steady at about 30%. These results indicate that the rabbit poly-
clonal IgG was purified to about 80% purity by using the DEAE anion
exchanger to remove albumin at low initial protein concentrations.
By eliminating the major contaminant (albumin) with DEAE
sepharose, a high yield of IgG antibodies was obtained from whey,
as reported earlier by Pessela and colleagues [9]. The loss of
IgG molecules was explained by the formation of some type of
albumin–IgG complex [9]. The retention of some albumin after
considerable albumin removal affects the purity of an antibody as
well as its yield. Notably, DEAE sepharose showed a good capacity
for purifying polyclonal IgG by eliminating albumin at low serum
protein concentrations, giving 83% purity and 82% recovery. For
a scaled-up process, the ratio of total protein to DEAE sepharose
adsorbent can be fixed to 5 mg/0.25 ml. However, different serum
sources and batches might alter the efficiency of albumin removal
and hence the purity of the antibody. Based on our experience,
IgG should be at 15% or higher in the serum for efficient antibody
purification via the negative chromatography approach.
3.4. Capacity of Q XL for albumin depletion
The capacity of the STREAMLINETM Q XL strong anion exchanger
was also tested under similar optimized conditions for albumin
removal. Initially, according to the maximum binding capacity of Q
XL, 10 mg of total protein in serum sample was prepared in pH 8.0
binding buffer for binding with 0.1 ml Q XL. The adsorbent showed
inefficient albumin removal (<50%) at 1.0 mg/ml total protein, in
contrast to that of the single protein binding system in which ∼95%
of the BSA was adsorbed onto Q XL (data not shown). These results
Fig. 4. (a) Serum albumin adsorption (estimated by band intensity analysis) onto
DEAE sepharose as a function of initial protein concentration. Two initial protein
concentrations (0.5 and 0.75 mg/ml) of serum samples were prepared individually
in 25 mM sodium phosphate buffer at pH 8.0. (b) Relative quantity of polyclonal
anti-HBcAg IgG (determined by ELISA) adsorbed onto DEAE sepharose as a function
of initial protein concentration. The adsorption conditions were: 0.25 ml settled bed
volume of DEAE sepharose; 1 h at room temperature. The results are from triplicate
measurements.
suggest that the presence of a protein in different environments
that include other components (such as serum) affects its adsorp-
tion onto the ionic exchange adsorbent. Since the total protein
concentration has an influence on protein aggregation [15,32], the
binding capacity of Q XL for the removal of albumin was carried out
at the lower initial protein concentration (0.5 mg/ml).
Fig. 5a shows the SDS-PAGE analysis of total protein before and
after serum albumin was removed in a time course (1 h). Albu-
min was removed efficiently up to 89 ± 1% as estimated by band
intensity analysis. However, the loss of IgG was as high as 27 ± 1%.
Fig. 5b shows the relative quantity of polyclonal anti-HBcAg IgG
before and after albumin removal as determined by ELISA. IgG was
adsorbed as well as albumin at the mentioned time points, resulting
in antibodies only 62% pure. These results show that albumin was
removed efficiently by Q XL (about 90%), but only 73 ± 1% of the IgG
was recovered. Q XL demonstrated binding capacity for albumin as
high as that of DEAE sepharose (90–95% of the albumin removed).
The adsorption of albumin and IgG molecules onto the highly acti-
vated adsorbent Q XL occurs due to the formation of aggregates in
solution. In fact, albumin is capable of interacting with numerous
different species of molecules [12–14]. Albumin aggregates as well
as the albumin-IgG aggregates are adsorbed quickly onto Q XL, as
reported earlier [9].
A comparative analysis of albumin removal using DEAE
sepharose and Q XL is summarized in Table 2. Under similar con-
ditions, greater than 90% of the albumin was removed by DEAE
sepharose within 1 h. The loss of IgG in the presence of DEAE
sepharose adsorbent was less than 20%. Thus, the purity of the IgG
antibody in the serum after albumin removal by DEAE sepharose

6. 106 R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107
Fig. 5. (a) Non-reducing SDS-PAGE (10%) analysis of protein adsorption from rabbit
serum onto Q XL. Lanes: (M) protein markers in kDa; (1) initial serum proteins; (2)
proteins adsorbed in 15 min; (3) proteins adsorbed in 30 min; (4) proteins adsorbed
in 45 min; (5) proteins adsorbed in 60 min. (b) The relative quantity of anti-HBcAg
IgG adsorbed onto Q XL as determined by ELISA. The adsorption conditions were:
0.10 ml settled bed volume of Q XL; 0.5 mg/ml initial protein concentration, pH 8.0;
1 h at room temperature. The results are from triplicate measurements.
Table 2
A comparison of STREAMLINETM
DEAE and STREAMLINETM
Q XL adsorbents in neg-
ative chromatography antibody purification.
Polyclonal IgG DEAE sepharose Q XL
Yield (%) 82 ± 3 68 ± 3
Purity in serum sample 0.15 0.15
Purity after albumin removal 0.83 0.62
Purification factor 5.5 4.1
was higher than that for Q XL. In summary, DEAE sepharose is bet-
ter than the strong anion exchanger Q XL for purifying polyclonal
IgG via negative chromatography in terms of both yield and purity;
the former brings about a yield of 80% with a purification factor 5.5.
4. Conclusions
Negative chromatography, antibody purification using
STREAMLINETM DEAE anion exchanger was used to remove
contaminants from rabbit serum. The pH of the binding medium
and the initial protein concentration were optimized. Under
the optimized conditions, DEAE sepharose efficiently removed
albumin and other contaminants from rabbit serum. Depletion
of albumin increased the purity of the polyclonal IgG (83%) and
resulted in recovery as high as 80% at pH 8.0 with 1 h of binding
time. The protein/adsorbent ratio of 5.0 mg total protein and
0.25 ml DEAE sepharose was very useful for scaling-up the process.
The method proposed here for removing albumin is simple and
inexpensive. Furthermore, it is a single-step purification that
can be easily scaled up. This method might be applicable for
antibody purification from other animal sera as well. The use of
STREAMLINETM DEAE anion exchanger offers a simple method
for purifying polyclonal antibodies from serum, which achieved
by eliminating contaminating proteins with the N-CAP approach.
Furthermore, DEAE sepharose is superior to Q XL for the removal
of albumin and thus for augmenting the purity and recovery
of IgG.
Acknowledgements
The authors acknowledge Suratthani Rajabhat University, Thai-
land and Grants 02-01-04-SF0808 and RUGS 91937 for financial
support. The authors thank Suet Lin Chia and Ayele Taddese for
technical support in the immunization work and Khai Wooi Lee for
the purification of HBcAg.
References
[1] Low D, O’Leary R, Pujar NS. Future of antibody purification. J Chromatogr B
2007;848:48–63.
[2] Hober S, Nord K, Linhult M. Protein A chromatography for antibody purification.
J Chromatogr B 2007;848:40–7.
[3] Giovannoni L, Ventani M, Gottschalk U. Antibody purification using membrane
adsorbers. BioPharm Int 2008;21:48–52.
[4] Pessela BCC, Munilla R, Betancor L, Fuentes M, Carrascosa AV, Vian A, et al.
Ion exchange using poorly activated supports, and easy way for purification of
large proteins. J Chromatogr A 2004;1034:155–9.
[5] Goheen SC, Hilsenbeck JL. High-performance ion-exchange chromatography
and adsorption of plasma proteins. J Chromatogr A 1998;816:89–96.
[6] Lin D-Q, Fernández-Lahore HM, Kula M-R, Thömmes J. Minimising
biomass/support interactions in expanded bed adsorption processes: a
methodological design approach. Bioseparation 2001;10:7–19.
[7] Qi Y, Yan Z, Huang J. Chromatography on DEAE ion-exchange and Protein G
affinity columns in tandem for the separation and purification of proteins. J
Biochem Biophys Methods 2001;49:263–73.
[8] Huang Y, Bi L, Zhou W, Li Y, Wang Y, Ma G, et al. Improving recovery of recom-
binant hepatitis B virus surface antigen by ion-exchange chromatographic
supports with low ligand density. Process Biochem 2006;41:2320–6.
[9] Pessela BCC, Torres R, Batalla P, Fuentes M, Mateo C, Fernández-Lafuente R, et
al. Simple purification of immunoglobulins from whey proteins concentrate.
Biotechnol Prog 2006;22:590–4.
[10] Chase HA, Machielse B, Naveh D. Characterization of the adsorption of
immunoglobulin M onto Q Sepharose Fast Flow ion exchangers. Bioseparation
1997;7:47–55.
[11] Fuentes M, Pessela BCC, Mateo C, Munilla R, Guisán JM, Fernández-Lafuente
R. Detection and purification of two antibody–antigen complexes via selec-
tive adsorption on lowly activated anion exchangers. J Chromatogr A
2004;1059:89–94.
[12] Slade CJ. Molecular (or bio-) imprinting of bovine serum albumin. J Mol Catal
B: Enzym 2000;9:97–105.
[13] Militello V, Verti V, Leone M. Conformational changes involved in ther-
mal aggregation processes of bovine serum albumin. Biophys Chem
2003;105:133–41.
[14] Vaiana SM, Emanuele A, Palma-Vittorelli MB, Palma MU. Irreversible formation
of intermediate BSA oligomers requires and induces conformational changes.
Proteins: Struct Funct Bioinf 2004;55:1053–62.
[15] Fuentes M, Pessela BCC, Mateo C, Palomo JM, Batalla P, Fernández-Lafuente R,
et al. Adsorption behavior of bovine serum albumin on lowly activated anionic
exchangers suggests a new strategy for solid-phase proteomics. Biomacro-
molecules 2006;7:1357–61.
[16] Tan WS, Dyson MR, Murray K. Hepatitis B virus core antigen: enhancement of
its production in Escherichia coli, and interaction of the core particles with the
viral surface antigen. Biol Chem 2003;384:363–71.
[17] Dyson MR, Murray K. Selection of peptide inhibitors of interactions involved
in complex protein assemblies: association of the core and surface antigens of
hepatitis B virus. Proc Natl Acad Sci USA 1995;92:2194–8.
[18] Tan WS, Dyson MR, Murray K. Two distinct segments of the hepatitis B virus
surface antigen contribute synergistically to its association with the viral core
particles. J Mol Biol 1999;286:797–808.
[19] Bradford MM. A rapid and sensitive method for the quantitation of micro-
gram quantities of protein utilizing the principle of protein dye binding. Anal
Biochem 1976;72:248–54.
[20] Laemmli UK. Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature 1970;227:680–5.
[21] Tey BT, Yong KH, Ong HP, Ling TC, Ong ST, Tan YP, et al. Optimal conditions for
hepatitis B core antigen production in shaked flask fermentation. Biotechnol
Bioprocess Eng 2004;9:374–8.
[22] Ng MYT, Tan WS, Abdullah N, Ling TC, Tey BT. Direct purification of recombi-
nant hepatitis B core antigen from two different pre-conditioned unclarified
Escherichia coli feedstocks via expanded bed adsorption chromatography. J
Chromatogr A 2007;1172:47–56.
[23] Engvall E, Perlmann P. Enzyme-linked immunosorbent assay (ELISA) quantita-
tive assay of immunoglobulin G. Immunochemistry 1971;8:871–4.

7. R. Wongchuphan et al. / Process Biochemistry 46 (2011) 101–107 107
[24] Mateo C, Abian O, Fernández-Lafuente R, Guisán JM. Reversible enzyme
immobilization via a very strong and nondistorting ionic adsorption on
support-polyethylenimine composites. Biotechnol Bioeng 2000;68:98–105.
[25] Pessela BCC, Fernández-Lafuente R, Fuentes M, Vian A, Garcia JL, Carrascosa
AV, et al. Reversible immobilization of a thermophilic ␤-galactosidase via ionic
adsorption on PEI-coated Sepabeads. Enzyme Microb Technol 2003;32:369–74.
[26] Fuentes M, Mateo C, Pessela BCC, Guisán JM, Fernández-Lafuente R. Purifica-
tion, stabilization and concentration of very weak protein–protein complexes:
shifting the association equilibrium via complex selective adsorption on lowly
activated supports. Proteomics 2005;5:4062–9.
[27] Middaugh CR, Volkin DB. Protein solubility. In: Ahern TJ, Manning MC, editors.
Stability of protein pharmaceuticals. Part A: chemical and physical pathways of
protein degradation. Pharmaceutical biotechnology, vol. 2. New York: Plenum
Press; 1992. p. 109–34.
[28] Chi EY, Krishnan S, Randolph TW, Carpenter JF. Physical stability of proteins in
aqueous solution: mechanism and driving force in nonnative protein aggrega-
tion. Pharm Res 2003;20:1325–36.
[29] Bagchi P, Birnbaum SM. Effect of pH on the adsorption of immunoglobulin
G on anionic poly(vinyltoluene) model latex particles. J Colloid Interface Sci
1981;83:460–78.
[30] Bulone D, Martorana V, San Biagio PL. Effects of intermediates on aggregation
of native bovine serum albumin. Biophys Chem 2001;91:61–9.
[32] Jordan GM, Yoshida S, Terao T. The aggregation of bovine serum albumin in
solution and in the solid state. J Pharm Pharmacol 1994;46:182–5.