See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/267928547

Preparation and evaluation of dextran-grafted

agarose resin for hydrophobic charge-
induction chromatography

ARTICLE in JOURNAL OF CHROMATOGRAPHY A · NOVEMBER 2014

Impact Factor: 4.26 · DOI: 10.1016/j.chroma.2014.10.014

CITATION DOWNLOADS VIEWS

1 126 88

4 AUTHORS, INCLUDING:

Tao Liu Dong-Qiang Lin

Zhejiang University Zhejiang University
2 PUBLICATIONS 1 CITATION 117 PUBLICATIONS 1,198 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Tao Liu

Retrieved on: 09 September 2015
 Journal of Chromatography A, 1369 (2014) 116–124

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Preparation and evaluation of dextran-grafted agarose resin for

hydrophobic charge-induction chromatography
Tao Liu a , Dong-Qiang Lin a,c,∗ , Hui-Li Lu a , Shan-Jing Yao b,c
a
State Key Laboratory of Chemical Engineering, Department of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, China
b
Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Department of Chemical and Biological Engineering, Zhejiang University,
Hangzhou 310027, China
c
Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300072, China

a r t i c l e i n f o a b s t r a c t

Article history: Hydrophobic charge-induction chromatography (HCIC) is a new and effective technology for antibody
Received 27 June 2014 separation. In the present work, HCIC resin MMI-B-XL was prepared with dextran-grafted agarose gel
Received in revised form 2 October 2014 as the matrix and 2-mercapto-1-methyl-imidazole (MMI) as the functional ligand. The preparation pro-
Accepted 6 October 2014
cedures were optimized, and the maximum ligand density could reach as high as 200 ␮mol/g gel. The
Available online 17 October 2014
adsorption isotherms and kinetics on new resins were investigated with human immunoglobulin G (hIgG)
as the model protein, which were compared with non-grafted HCIC resin MMI-B-6FF. It was found that
Keywords:
the saturated adsorption capacity (Qm ) increased with the increase of ligand density for MMI-B-XL. More-
Dextran-grafted agarose resin
Hydrophobic charge-induction
over, the effective diffusivity (De ) could be dramatically enhanced with the increase of ligand density for
chromatography MMI-B-XL, and the De for MMI-B-XL with the ligand density of 200 ␮mol/g gel was 18–40 times higher
Human immunoglobulin G than that for MMI-B-6FF. The breakthrough experiments indicated that new resins with the ligand den-
Ligand density sity of 200 ␮mol/g gel could be used for high superficial velocity and high dynamic adsorption could be
Adsorption capacity obtained. The results indicated that dextran-grafted layer on the resin could increase the ligand density,
Adsorption kinetics enhance the mass transport in the pore, and improve the dynamic adsorption at high velocity, which
showed a potential application for large-scale antibody purification.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction functional ligand is 4-mercaptoethyl-pyridine (MEP) with the pKa

Hydrophobic charge-induction chromatography (HCIC) has
been developed in recent years for improving protein separation
[1,2]. The basic properties of HCIC are the pH-dependent behavior
of mixed-mode ligands that combine the hydrophobic, thiophilic,
hydrogen bonding and electrostatic interactions together. The tar-
get protein could be adsorbed on the uncharged ligand at the
neutral pH mainly by the hydrophobic interaction, and desorbed by
the electrostatic repulsion interaction between target protein and
charged ligand at acidic condition. Normally, HCIC adsorbents could
capture target protein over a broader ionic strength range com-
pared to ion-exchange resins, which was called “salt-independent”
or “salt-tolerance” properties [3–5]. MEP HyperCel is a typical
commercial HCIC adsorbent developed by Pall Corporation. The
∗ Corresponding author at: Department of Chemical and Biological Engineering,
Zhejiang University, Hangzhou 310027, China. Tel.: +86 571 87951982;
fax: +86 571 87951982.
E-mail address: lindq@zju.edu.cn (D.-Q. Lin).
of 4.8, and the matrix is cellulose [6]. HCIC has been applied to
the purification of antibodies and other proteins, which showed a
potential as the cost-effective alternative to Protein A chromatog-
raphy due to high adsorption capacity and reusability, mild elution
condition and convenient Clean-In-Place process [7].
Combining several molecular interactions with specially
designed ligands [8–10], HCIC resins usually show high selectiv-
ity on the adsorption of target protein compared to traditional
ion-exchange chromatography (IEC) and hydrophobic interac-
tion chromatography (HIC) [11]. However, due to the relatively
weak hydrophobic interactions between target protein and
ligand, the dynamic binding capacities (DBC) of HCIC resins
are limited, especially for low retention time, which strongly
impedes the large-scale application of HCIC for protein produc-
tion. In recent years, polymer-functionalized IEC adsorbents have
been developed in order to improve the performance of pro-
tein separation [12]. Dextran-grafted agarose resins such as SP
Sepharose XL, Streamline Q XL and Capto S have been commercial-
ized successfully [13,14]. Besides, a new directly grafting charged
polymer poly(ethylenimine)-modified Sepharose FF has also been

http://dx.doi.org/10.1016/j.chroma.2014.10.014
0021-9673/© 2014 Elsevier B.V. All rights reserved.
 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
developed [15–19]. These polymer-grafted resins show significant
enhancements on DBC, which reflects in both adsorption equilib-
rium and mass transport properties [20–26]. The improvements
could be attributed to the following reasons: a solid or homo-
geneous diffusion and the specific architecture of the stationary
phase [27], increased adsorption sites and more ligand exposed
to three-dimensional binding space [28], an overall higher charge
density or a shortening of the path length for diffusion [29], “bucket
brigade effect” (protein passed from one polymer chain to the
next) in the grafted layer [30], and the electrostatic coupling of
diffusion fluxes [31]. In addition, the flexibility of the grafted poly-
mer could improve diffusion process of protein in the pores to a
great extent, as a result, polymer-grafted resins have higher uptake
rates on the target protein [32]. Until now, most efforts have been
made on the improvements of IEC resins. The effects of dextran
layer on protein adsorption to dextran-grafted HCIC adsorbents
have also been reported by Yu et al. [33]. However, the saturated
adsorption capacity (Qm ) and effective pore diffusion coefficient
(De ) of dextran-grafted HCIC adsorbents did not change obviously,
which might be attributed to the low ligand density of dextran-
grafted adsorbents used in their work. Obviously, ligand density
determines the strength of the hydrophobic and electrostatic inter-
action between resins and target protein, which might have a great
influence on the adsorption equilibrium and adsorption kinetics
[34–37]. The effects of ligand density on protein adsorption to HCIC
adsorbents have also been reported by Lu et al. [38]. It was found
that the Qm and De increased with the increase of ligand density.
In general, the polymer modification (such as dextran grafting) on
the internal porous surface and the control of ligand density on the
resins are two effective ways to improve the target binding and
consequently the DBC for HCIC process.
In the present work, with 2-mercapto-1-methyl-imidazole
(MMI) as the typical HCIC ligand developed in our previous
work [39], new resin based on the dextran-grafted agarose gel
would be prepared to improve the protein binding and DBC. The
matrix was activated by allyl bromide (AB), brominated by N-
bromosuccinimide (NBS) and then coupled with MMI ligand to
prepare new dextran-grafted HCIC resins, named as MMI-B-XL.
The reaction conditions of activation, bromination and ligand cou-
pling would be optimized, and a series of new resins MMI-B-XL
with different ligand densities would be prepared. With human
immunoglobulin G (hIgG) as the model protein, the adsorption
isotherms and adsorption kinetics would be investigated with
series of MMI-B-XL resins and non dextran-grafted HCIC resins
MMI-B-6FF based on 6% crosslinked agarose gel would be used as
the comparison. The influences of dextran-grafted layer and ligand
density on the Qm , the dissociation constant (Kd ), De and DBC would
be discussed.
2. Materials and methods
2.1. Materials
The dextran-grafted agarose gel prepared with 6% agarose and
grafted with dextran (Mw = 140,000), named Bestarose XL, was
purchased from Bestchrom Bio-Technology Co., Ltd. (Shanghai,
China). Non-grafted agarose gel with 6% agarose, named Bestarose
6FF, was also purchased from Bestchrom Bio-Technology Co.,
Ltd. (Shanghai, China) as the comparison. Allyl bromide (AB)
and mercaptoacetic acid (MMA) were purchased from Sinopharm
Chemical Reagent Co., Ltd. (Shanghai, China). 2-Mercapto-1-
methylimidazole (MMI) and N-bromosuccinimide (NBS) were
purchased from Aladdin (Shanghai, China). Human ␥-globulin
which is human normal immunoglobulin G (IgG > 98%), was
obtained from Merck KGaA (Darmstadt, Germany). hIgG for
117
H2C
Br
CH
OH Allyl Bromide (AB) O CH2
Matrix (B-XL) NaOH in 20% DMSO Br
N
O O
50% acetone
NBS
Br+
O CH2 CH CH2
CH3
HS N
pH 10.0
N
MMI
CH3
O CH2CH CH2 S N
OH
N
Dextran-grafted HCIC resin (MMI-B-XL)
Fig. 1. Preparation scheme of the dextran-grafted HCIC resins, MMI-B-XL.
intravenous injection (IgG > 96%), was purchased from RAAS Blood
Products Co., Ltd. (Shanghai, China). Other reagents were of analyt-
ical reagent grade and purchased locally.
2.2. Preparation of HCIC resins
The preparation routes of HCIC resins were based on the pre-
viously reported methods [39,40] with some modifications, and
the scheme is shown in Fig. 1. The dextran-grafted agarose gel
was activated by allyl bromide (AB). Then the activated matrix was
brominated by N-bromosuccinimide (NBS). Finally the brominated
matrix was coupled with MMI ligand. Typically, 2 g drained agarose
gel was mixed with AB (0.8 ml, 9.3 mM) and sodium hydroxide
(0.41 g, 10.3 mM) in 20% dimethyl sulfoxide (DMSO) solution in a
25 ml triangular flask. The reaction was continuously agitated at
180 rpm and 30 ◦ C for 24 h. After being washed with ethanol and
deionized water, the allyl-activated gel was obtained and named as
AB-B-XL, and then AB-B-XL was brominated with 3 molar excess of
NBS over allyl groups in 50% acetone solution at 180 rpm for 1 h at
30 ◦ C. Finally, the brominated gel was mixed with 3 molar excess of
MMI over allyl groups in 1 M carbonate buffer (pH 10.0) at 180 rpm
and 30 ◦ C for 24 h. The resin was washed with deionized water and
stored in 20% (v/v) ethanol. The dextran-grafted HCIC resins pre-
pared were named as MMI-B-XL. As the comparison, the HCIC resins
prepared with non-grafted agarose gel, Bestarose 6FF, using similar
procedure were named as MMI-B-6FF.
The activated density on AB-B-XL and the ligand density on
MMI-B-XL were determined by the titration method [2,39]. To mea-
sure the activated density, 1 g drained activated resin was mixed
with 120 ␮l mercaptoacetic acid (MMA), 1 ml deionized water and
25 mg ammonium persulfate, then incubated at 60 ◦ C for 8 h with
a water bath shaker (180 rpm). After that, the resin was washed
with deionized water, 0.1 M NaOH and 0.1 M HCl solution, then
drained and transferred to a vial. Thereafter, 5 ml 0.5 M NaCl solu-
tion was added into the vial, and the mixture was titrated with
0.1 M Tris solution to pH 6.4. For the measurement of ligand density,
1 g drained MMI-B-XL was washed with 0.1 M NaOH solution and
deionized water, then drained and transferred to a vial. Thereafter
118 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
5 ml 0.5 M NaCl solution was added into the vial, and the mixture
was titrated with 0.1 M HCl solution to pH 3.8.
The activated density on AB-B-XL and the ligand density on
MMI-B-XL were calculated and expressed in ␮mol/g gel as follows:
Titer × 0.1
D= (1)
Adsorbent
where D are the activated density on AB-B-XL or the ligand density
on MMI-B-XL, Titer is the titer of 0.1 M Tris solution (␮l). Adsorbent
is the mass of resin (g).
2.3. Adsorption equilibrium experiments
The adsorption isotherms of hIgG (Merck KGaA, Germany) on
MMI-B-XL and MMI-B-6FF resins with different ligand densities
were determined by batch adsorption equilibrium experiments
[37,38]. The resins were pre-equilibrated with 20 mM sodium phos-
phate buffer (pH 7.0) and then drained. 0.04 g drained resin was
added to 0.8 ml 20 mM sodium phosphate buffer (pH 7.0) con-
taining different concentrations of hIgG (0.5–10 mg/ml) in 2 ml
centrifuge tube. The mixture was agitated in a thermomixer
(1000 rpm) at 25 ◦ C for 5 h. After the adsorption reached equilib-
rium, the resins were separated by centrifugation (4000 × g) and
decanting, and the protein concentrations in the supernatant were
determined at 280 nm with a spectrophotometer (NanovueTM Plus
Spectrophotometer, GE Healthcare, Uppsala, Sweden). The amount
of adsorbed protein was calculated according to mass balance. The
adsorption isotherm was fitted by the Langmuir equation,
Qm · C ∗
Q∗ = (2)
Kd + C ∗
where Q* and C* are the equilibrium adsorption capacity of resin
(mg/g gel) and the equilibrium protein concentration in liquid
phase (mg/ml), respectively. Qm is the saturated adsorption capac-
ity (mg/g gel) and Kd is the dissociation constant (mg/ml).
2.4. Adsorption kinetics experiments
The adsorption kinetics of hIgG (Merck KGaA, Germany) on
MMI-B-XL and MMI-B-6FF resins with different ligand densities
were measured by the batch method. Generally, the resins were
pre-equilibrated with 20 mM sodium phosphate buffer (pH 7.0),
then 2 g resin was added into 40 ml protein solution (0.7 mg/ml
hIgG in 20 mM sodium phosphate buffer at pH 7.0) in a 50 ml
flask. The suspension was magnetically stirred and the temperature
was kept at 25 ◦ C. The protein solution was pumped out and flow
through a 2 ␮m stainless filter at a flow rate of 20 ml/min. The real-
time concentration of protein in the suspension was monitored at
280 nm by a UV monitor (UV Detector K-2600, Knauer, German).
The adsorption kinetics curves were fitted by the pore diffusion
model (PDM) as described in the previous work [16,37,38,41–43].
The total effective diffusivity De was obtained for MMI-B-XL and
MMI-B-6FF with different ligand densities. PDM is one of the most
widely used models to describe the mass transfer for the adsorption
of protein onto porous adsorbents [44].
2.5. Column breakthrough experiments
Column breakthrough experiments of hIgG for MMI-B-XL-200
(ligand density 204 ± 7 ␮mol/g gel) and MMI-B-6FF-140 (ligand
density 139 ± 5 ␮mol/g gel) were performed in a Tricorn column
5/100 (5 mm × 100 mm). In this part, hIgG for intravenous injec-
tion was chosen as the loading protein. All experiments were
performed at 25 ◦ C in 20 mM sodium phosphate buffer at pH 7.0.
After the agarose gels were packed into the column, the column
was equilibrated with 10 column volumes of the equilibrium buffer
until a stable baseline of absorbance at 280 nm was reached. Then
2.0 mg/ml hIgG solution prepared in 20 mM sodium phosphate
buffer (pH 7.0) was loaded at the superficial linear velocity of 100,
200 and 300 cm/h. The protein concentration in the effluent was
monitored by absorbance at 280 nm with a UV monitor (UV Detec-
tor K-2600, Knauer, German). The loading was continued until the
protein concentration at the outlet reached at least 70% of the load-
ing protein concentration. Then the column was eluted with acetate
buffer (pH 4.0). Finally, the column was regenerated with 0.1 mol/l
NaOH and re-equilibrated with 20 mM sodium phosphate buffer
(pH 7.0). DBC at 10% breakthrough, Q10% (mg/ml settled gel), was
calculated as follows:
 V10%
C0 · 0
(1 − C/C0 )dV
Q10% = (3)
Vs
where C and C0 are the protein concentration of outlet fluid and
initial fluid, respectively. V10% is the loading volume at 10% break-
through and Vs is the settled volume of the gel.
3. Results and discussion
3.1. Activation of dextran-grafted agarose gel
Epichlorohydrin is often used for activating hydroxyl, thiol, and
amine groups. However, some side reactions (e.g., crosslinking and
epoxide hydrolysis) would cause low activation density on the cel-
lulose or agarose-based matrix [45]. It was found that AB could
effectively enhance the activation densities of matrix. Under alka-
line conditions, the bromide group of AB is active, and allyl group is
relatively inert, which would restrict the crosslinking and hydrol-
ysis [46]. So the activation efficiency with AB is higher than that
with epichlorohydrin. In addition, it was found that the addition of
DMSO in the reaction system would promote the mass transfer and
result in increasing the activation efficiency [40].
The activation reaction of the matrix should be operated under
alkaline conditions [39,40,46]. The appropriate addition of NaOH
could induce nucleophilic substitution reaction between AB and
hydroxyl groups on the matrix, but the excess of NaOH would accel-
erate the hydrolysis of allyl groups and lower activation efficiency
[39,40]. Therefore, the appropriate amount of NaOH addition was
investigated firstly. The activation density as a function of the molar
ratio of NaOH to AB is shown in Fig. 2. It was found that the maxi-
mum activation densities reached about 340 ␮mol/g gel when the
NaOH/AB molar ratio was at the range of 1.05–1.3. When the molar
ratio was lower than 1.05 or above 1.3, a dramatical reduction of
the activation density occurred due to the reasons as mentioned
above. The result was similar to that reported by Xia et al. [39], in
which the activated density could reach the maximum when the
molar ratio of hydroxide to AB was about 1.1. In the present work,
the optimal molar ratio of NaOH to AB for the activation reaction
was also set as 1.1.
The amount of activation agent AB in the reaction system also
played an important role in the activation density of the matrix
used. As shown in Fig. 2, the activation density of AB-B-XL increased
roughly linearly with the increase of AB addition when the amount
of AB was below 0.4 ml/g gel. When the amount of AB higher than
0.4 ml/g gel, the activation reaction reached saturation and the acti-
vation density was about 340 ␮mol/g gel. The results indicated that
the activation level could be controlled by the amount of AB addi-
tion, which would be useful to prepare series of resins with different
ligand density.
Fig. 2 also shows the activation density as the function of reac-
tion time. The activation reactions were carried out at 30 ◦ C, the AB
addition at 0.4 ml AB/g gel and NaOH addition at the [OH− ]/[AB]
of 1.1. The results indicated that the activation density could reach
 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
Fig. 2. Activation density on the AB-B-XL as the function of hydroxide to AB, AB to
matrix and the reaction time, respectively. The activation reactions were performed
at 30 ◦ C in 20% DMSO solution.
the maximum value (340 ␮mol/g gel) in 24 h. Therefore, 24 h was
chosen as the suitable reaction time for the activation.
3.2. Bromination and ligand coupling
After the activation, the relatively inert allyl group should be fur-
ther activated to form the bromohydroxypropyl groups and then for
ligand coupling [39,40,46]. NBS has been often used as the bromi-
nating agent [47]. The amount of NBS has significant effects on
the ligand density of the resin prepared [39,40]. Fig. 3 shows the
ligand density of MMI-B-XL as the function of the molar ratio of
NBS to allyl group on AB-B-XL. AB-B-XL with the activation density
of 340 ␮mol/g gel was used, and the brominated gel was further
coupled with MMI ligand at 3 molar ratio of MMI to allyl group
on AB-B-XL. It was found that the ligand density increased signif-
icantly with the increase of the amount of NBS addition when the
molar ratio of NBS to allyl groups was below 3.0. After the molar
ratio of NBS to allyl groups reached 3.0, the ligand density kept
about 200 ␮mol/g gel. Fig. 3 also shows the coupling density of
MMI ligand on the brominated gel as the function of molar ratio
of MMI to ally groups. It was found that 3-fold mole of MMI to ally
Fig. 3. Ligand densities of MMI-B-XL as the function of molar ratio of NBS and MMI
to the allyl groups on AB-B-XL, respectively. The activation density of AB-B-XL used
was 340 ␮mol/g gel. The bromination reactions were performed at 30 ◦ C for 1 h in
50% acetone solutions. The ligand coupling reactions were performed at 30 ◦ C for
24 h in 1 M sodium carbonate buffer (pH 10.0).
119
Fig. 4. Comparison on the ligand densities of MMI-B-XL and MMI-B-6FF with dif-
ferent AB addition during the activation procedure.
groups was enough for the coupling reaction. The ligand density of
MMI on MMI-B-XL prepared could reach 200 ␮mol/g gel.
3.3. Comparison on dextran-grafted and non-grafted HCIC resins
Based on the results mentioned above, it could be found that the
most important factor to control the activation efficiency and con-
sequently ligand density on dextran-grafted agarose gel, Bestarose
XL, was the amount of AB addition. The activation and ligand cou-
pling on non-grafted agarose gel, Bestarose 6FF, have also been
done with the same methods. Similar trends were found. As we
know, the gels, Bestarose XL and Bestarose 6FF, are prepared with
same concentration of agarose gels (6%), and the difference is only
on the dextran-grafted layer on Bestarose XL. Fig. 4 compares the
ligand density of MMI-B-XL and MMI-B-6FF as the function of AB
addition in order to investigate the effects of dextran-grafted layer.
For two resins, the coupling efficiency of MMI ligand kept similar,
about 55–70%. However, the ligand density showed some differ-
ences. When AB addition was at the range of 0–0.3 ml/g gel, the
ligand density on MMI-B-XL was slightly higher than that on MMI-
B-6FF. When AB addition amount was as high as 0.4 ml/g gel, the
ligand density on MMI-B-XL improved significantly. Compared to
MMI-B-6FF, about 43% improvement on the ligand density of MMI-
B-XL was found, increasing from about 140 to 200 ␮mol/g gel. The
reason might be that the existence of dextran-grafted layer on the
internal pores of agarose gel could provide more hydroxyl groups
for the activation and ligand coupling, meanwhile those hydroxyl
groups on the grafted dextran could expose freely to the pore space
and be reacted easily [12]. When the amount of AB in the solution
is enough, ligand density on the dextran-grafted agarose gel could
be improved significantly.
In general, new dextran-grafted HCIC resin MMI-B-XL prepared
in this work had higher ligand density compared to traditional
non dextran-grafted resins, which indicated that new resins have
a potential to improve the protein binding. For the further adsorp-
tion studies, five ligand densities of MMI-B-XL resins were chosen
as 62 ± 4, 91 ± 5, 119 ± 5, 138 ± 5 and 204 ± 7 ␮mol/g gel, named
MMI-B-XL-60, MMI-B-XL-90, MMI-B-XL-120, MMI-B-XL-140 and
MMI-B-XL-200, respectively. For the comparison, non-grafted
MMI-B-6FF resins with similar ligand densities were used as 62 ± 4,
91 ± 3, 118 ± 4 and 139 ± 5 ␮mol/g gel, named MMI-B-6FF-60,
MMI-B-6FF-90, MMI-B-6FF-120, and MMI-B-6FF-140, respectively.
3.4. Adsorption isotherms of hIgG
hIgG was used as the model protein in the present work to inves-
tigate the adsorption behavior of new HCIC resins. The pKa of MMI
120 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
Fig. 5. Adsorption isotherms of hIgG on MMI-B-XL (a) and MMI-B-6FF (b) with
different ligand densities at pH 7.0.
ligand is around 5.2, which shows non-charge at neutral condition
[38,39,48]. The pI of hIgG is at the range of 6.5–10.0 [49–51]. So
hIgG would be adsorbed on the MMI ligand-based resins at pH 7.0
by the hydrophobic interactions between ligand and target protein
[38,48].
The adsorption isotherms of hIgG on MMI-B-XL with five ligand
densities as mentioned above were determined at pH 7.0 and the
Langmuir equation was used to fit the experimental data. Non-
grafted HCIC resins, MMI-B-6FF with four ligand densities were
used as the comparison. The adsorption isotherms were shown in
Fig. 5. The correlated parameters (Qm and Kd ) are shown in Fig. 6(a)
and (b). The results indicated that ligand density and dextran-
grafted layer played a significant role in the adsorption of hIgG,
which would be discussed in detail as follows.
3.4.1. Effect of ligand density
As shown in Fig. 6(a), it is obvious that with the increase of ligand
density on MMI-B-XL, Qm increased obviously and Kd decreased
rapidly. For MMI-B-XL, with the ligand density increasing from 60
to 200 ␮mol/g gel, Qm increased linearly from 12.5 to 107.5 mg/g
gel while Q* at 2 mg/ml of IgG concentration increased from 2.9
to 55.4 mg/g gel. The results were consistent with the previous
work with non-grafted HCIC resins [38]. In addition, it could also
be found that Kd decreased sharply with the increase of ligand den-
sity on MMI-B-XL, as shown in Fig. 6(b). Especially, when ligand
density increased from 90 to 120 ␮mol/g gel, Kd decreased from
5.5 to 3.5 mg/ml on MMI-B-XL. When ligand density reached the
maximum, Kd were about 1.9 mg/ml for MMI-B-XL. The results
indicated that the increase of ligand density could dramatically
enhance the affinity between MMI ligand and hIgG [38]. However, it
should be noted that the values of Kd were higher than those of ion-
exchangers reported by Yu et al. [33]. For example, the Kd values of
Fig. 6. Comparison on the adsorption capacities (a) and Kd (b) of hIgG with MMI-B-
XL and MMI-B-6FF as the function of ligand density.
porcine IgG onto the SP-Sepharose resins were between 0.008 and
0.048 mg/ml [33]. This might be originated from the ligand–protein
interactions, mainly hydrophobic interactions in HCIC, were rela-
tively weaker than the electrostatic attraction of IEC [38].
The results on MMI-B-XL were similar to the data reported
by Bak et al. [52] for the commercial non-grafted HCIC resin,
MEP HyperCel, with Qm of 81.5 mg/g and Kd of 2 mg/ml for rabbit
immunoglobulin. In our previous work [5,38,39], the adsorption of
bovine IgG and chicken IgY by MEP HyperCel and MMI-based resins
had also been reported. Qm of bovine IgG with both MEP HyperCel
and MMI-based resins at pH 7.0 were about 160 mg/g gel. At pH 7.0,
Qm of chicken IgY with MMI-based resins was 137.6 mg/ml, which
were higher than 106.7 mg/ml with MEP HyperCel. The results
indicated that MMI-based resins showed good adsorption ability
compared with other HCIC resins.
3.4.2. Effect of dextran-grafted layer
Some differences on Qm and Kd values between dextran-grafted
MMI-B-XL and non-grafted MMI-B-6FF could also be found in Fig. 6.
When the ligand density was relatively low, below 140 ␮mol/g gel
in the present work, Qm of hIgG on MMI-B-XL was slightly lower
than that of MMI-B-6FF with the same ligand density. The results
were consistent with the data reported by Yu et al. [33]. Yu et al.
attributed the phenomenon to the collapse of dextran-grafted layer
on the resins, which was caused by hydrophobic ligand immo-
bilized onto dextran-grafted layer stucking together at pH 8.0.
However, the present work showed some differences. When the
ligand density was high enough, such as MMI-B-XL-200, Qm and
Q* at 2 mg/ml of IgG concentration were improved to 107.5 and
55.4 mg/g gel, improving about 24% and 14% compared to 87.0 and
48.6 mg/g for MMI-B-6FF-140, respectively.
 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
Fig. 7. Adsorption kinetics curves of hIgG on MMI-B-XL and MMI-B-6FF with dif-
ferent ligand densities at pH 7.0.
3.5. Adsorption kinetics of hIgG
The mass transport and diffusion of protein in the resins have
a critical effect on the dynamic adsorption behavior. The matrix
structure and ligand density could influence the adsorption rates
significantly. In the present work, the adsorption kinetics of hIgG
onto MMI-B-XL and MMI-B-6FF resins with series of ligand densi-
ties were determined in the sodium phosphate buffer (pH 7.0). The
pore diffusion model was used to correlate the experimental data,
the adsorption kinetics curves are shown in Fig. 7 and De values
were calculated and shown in Fig. 8.
3.5.1. Effect of ligand density
For the dextran-grafted HCIC resin, MMI-B-XL, De values
increased dramatically with increasing ligand density. For example,
De values increased from 0.32 × 10−11 to 6.89 × 10−11 m2 /s (about
21.5-fold) when ligand density increased from 90 to 200 ␮mol/g.
This behavior was different from those reported by Bowes and
Lenhoff [53]. For dextran-grafted IEC resins, the electrostatic attrac-
tion forces between protein and ligands are so strong, and the
protein molecules may bind to the exterior area of resin initially,
consequently the mass transport restriction for the coming protein
would exist as the combined influences of the steric hindrance of
bound protein and the electrostatic repulsion of the charged pro-
tein [16,35,37,38,50,53]. The restriction would become stronger
with the increase of ligand density, and thus reducing De values,
as shown by Bowes and Lenhoff [53]. However, for HCIC resins,
the diffusion restriction was mainly caused by weak hydropho-
bic interaction. The fairly weak hydrophobic interactions between
protein and HCIC ligands could be strengthened with the increase
of ligand density, and thus De values would be improved [38]. As
Fig. 8. Comparison on the De of hIgG with MMI-B-XL and MMI-B-6FF as the function
of ligand density.
121
Table 1
Comparison on the Q10% , Q10% /Q* and Q10% /Qm of hIgG with MMI-B-XL-200 and MMI-
B-6FF-140 at different superficial velocities.
HCIC resin U (cm/h) Q10% (mg/ml) Q10% /Q* (%) Q10% /Qm (%)
MMI-B-XL-200 100 26.8 44.0 22.5
200 19.7 32.3 16.6
300 13.0 21.4 11.0
MMI-B-6FF-140 100 14.7 28.7 16.4
200 6.9 13.5 7.7
300 5.1 10.0 5.7
reported by Yu et al. [33], De values did not increase obviously
with ligand density on dextran-grafted HCIC resins, which could
be attributed to the relative low ligand density used in their work
(lower than 66 ␮mol/g gel). As in present work, the De value was
only 0.02 × 10−11 m2 /s with the ligand density of 60 ␮mol/g gel.
Besides, it is necessary to discuss the effects of ligand density
on MMI-B-XL in details. When the ligand density increased from
90 to 120 ␮mol/g, De value increased slowly from 0.32 × 10−11
to 0.60 × 10−11 m2 /s. When the ligand density increased to
140 ␮mol/g, De value increased continuously to 1.38 × 10−11 m2 /s.
When ligand density reach 200 ␮mol/g, De value increased sharply
to 6.89 × 10−11 m2 /s. As we know, the increase of ligand density
would lead to the resins adsorbing more target protein, which
might cause stronger steric effect. On the other hand, high ligand
density could strengthen the hydrophobic interactions between
ligand and hIgG, resulting in the less surface binding resistance.
For MMI-B-XL resins, due to the relative weak hydrophobic bind-
ing force (Kd > 2 mg/ml), there might be a balance of adsorption
and desorption processes, and the surface binding resistance would
dominate the intraparticle mass transfer. As a result, De of hIgG onto
MMI-B-XL was improved dramatically with the increase of ligand
density. When the ligand density was up to 200 ␮mol/g, the flexible
dextran chain would form an extended three-dimensional grafted
layer, which might cause MMI ligand fully exposed to outer space of
pore, and the hydrophobic interaction would be strengthened fur-
ther. In addition, the dextran chains could become closer, leading
to facilitated transfer of adsorbed protein by swing of the flexible
chain. All factors mentioned above resulted in the rapid increase of
De on MMI-B-XL at high ligand density.
3.5.2. Effect of dextran-grafted layer
As shown in Fig. 8, it could also be found that De value for
dextran-grafted resin MMI-B-XL was significantly higher than non-
grafted resin MMI-B-6FF. When the ligand density was at the range
of 90–140 ␮mol/g, De value for MMI-B-XL was greater by 88–273%
compared to MMI-B-6FF with the same ligand density. When the
ligand density reached the maximum, the De value for MMI-B-XL-
200 was about 18-fold higher than MMI-B-6FF-140. The results
were consistent with the data reported previously, in which De val-
ues for IgG on ion exchangers grafted with dextran or sulfonated
polymer were improved 2–31-folds compared to the non-grafted
resins [21,23,25,33].
3.6. Dynamic binding capacities of hIgG
The DBC of hIgG onto MMI-B-XL-200 and MMI-B-6FF-140 resins
were measured at the linear velocity of 100, 200 and 300 cm/h at pH
7.0. Table 1 compared the DBC at 10% breakthrough (Q10% ), and two
parameters, Q10% /Q* and Q10% /Qm . At the linear velocity of 100 cm/h,
the values of Q10% for MMI-B-XL-200, MMI-B-6FF-140 were 26.8
and 14.7 mg/ml, respectively. With the increase of linear velocity,
Q10% values of both resins decreased. However, Q10% for MMI-B-XL-
200 dropped more slowly. For example, at linear velocity 200 cm/h,
122 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124

Fig. 9. Mechanistic schemes on the effects of dextran-grafted layer and ligand density on the adsorption of hIgG. The black lines represent agarose matrix, the red lines
represent dextran chains and spacer arms, and the blue triangles represent MMI ligands. The green dotted lines represent the interaction between ligand and hIgG. (a)
MMI-B-XL and MMI-B-6FF with low ligand density; (b) MMI-B-XL and MMI-B-6FF with high ligand density; (c) “chain delivery” effect of MMI-B-XL. (For interpretation of
the references to color in this figure caption, the reader is referred to the web version of this article.)

the Q10% for MMI-B-XL-200 was 19.7 mg/ml, increasing by 185.5%
compared to the MMI-B-6FF-140.
The ratios of Q10% to the equilibrium adsorption capacities Q* at
loading hIgG concentration of 2 mg/ml and the saturated adsorp-
tion capacities Qm were used to evaluate the dynamic adsorption
efficiency of the resins during the chromatographic process. As
listed in Table 1, at 100 cm/h superficial velocity, the Q10% value
of MMI-B-6FF-140 was only 28.7% of Q* and 16.4% of Qm , while
the Q10% value of MMI-B-XL-200 was 44.0% of Q* and 22.5% of Qm ,
which increased by 53.3% and 37.2% compared to MMI-B-6FF-140,
respectively. The results were similar to the data of dextran-grafted
IEC adsorbents reported by Shi et al. [54], in which the Q10% /Q*
of lysozyme on the dextran-grafted resin was improved 32% com-
pared to non-grafted resin. With the increase of linear fluid velocity,
the values of Q10% /Q* and Q10% /Qm for both resins decreased. How-
ever, compared to MMI-B-6FF-140, the fluid velocity had weaker
influences on the dynamic adsorption for MMI-B-XL-200. For
example, at linear velocity 200 cm/h, the Q10% /Q* and Q10% /Qm for
MMI-B-XL-200 were 32.3% and 16.6%, respectively, increasing by
139.3% and 115.6% compared to the MMI-B-6FF-140, respectively.
In this work, the hydrophobic binding force between MMI lig-
ands and hIgG was relatively weak, so the equilibrium adsorption
capacity Q* at low protein concentration (2 mg/ml) is lower than the
saturated adsorption capacity Qm obviously. Besides, the intraparti-
cle mass transport resistance also plays an important role on the
dynamic binding capacity Q10% of hIgG. As a result, Q10% was much
lower than Qm . In our previous work, at 100 cm/h superficial veloc-
ities, the Q10% /Qm of chicken IgY on MEP-based and non-grafted
MMI-based resins were only 12.6% and 8.9%, respectively. In this
work, the Q10% /Qm of hIgG on the grafted MMI-B-XL resin was
about 22.5%. The results demonstrated that the dextran-grafted
HCIC resins MMI-B-XL had higher DBC, especially for high fluid
velocity, which show a good potential for the large-scale separation
processes.
3.7. Discussion
As mentioned above, the dextran-grafted layer and the ligand
density influenced both the adsorption equilibrium and adsorption
kinetics significantly. An assumption scheme is shown in Fig. 9 to
explain the effects of dextran-grafted layer and ligand density on
the adsorption behaviors of MMI-B-XL resins. Some discussions are
provided as follows.
Firstly, the ligands on dextran-grafted MMI-B-XL resins are cou-
pled on the stretched dextran grafted layer while the ligands on
non-grafted MMI-B-6FF resins are directly coupled on the pore
surface [12,20–26,33,38], so the surface ligand density on MMI-
B-XL is relatively lower than that on MMI-B-6FF at same ligand
density based on the volume or mass of resin beads, as shown
in Fig. 9(a) and (b). As we know, for HCIC, high ligand density is
necessary for high adsorption capacity due to the relative weak
hydrophobic interactions between ligand and target protein [38].
Therefore, the adsorption ability could not be improved effectively
on MMI-B-XL compared to MMI-B-6FF when ligand density was
low (60–140 ␮mol/g gel), as shown in Fig. 9(a). When ligand den-
sity on MMI-B-XL resin was improved to 200 ␮mol/g gel, as shown
in Fig. 9(b), the significant increase of ligand density would shorten
the distance between adjacent ligands, meanwhile the presence
of the dextran-grafted layer made more MMI ligands expose to
the porous space of pore and thus provide higher accessibility
to bind hIgG [12,16,20,28,54]. It would enhance the attraction
forces between ligand and target protein and thus result in a
rapid increase of adsorption capacity on MMI-B-XL-200. However,
the pattern was different from the dextran-grafted IEC resins, in
which the adsorption ability of the dextran-grafted IEC resin was
obviously higher than that of non-grafted IEC resins at same ligand
densities [20,28,33]. The difference was mainly attributed to that
the electrostatic attraction interactions are obviously stronger
than the hydrophobic interactions provided by HCIC ligands [38].
Therefore, for dextran-grafted HCIC resin MMI-B-XL, higher ligand
density (as 200 ␮mol/g gel) is essential to provide the sufficient
attractive interactions to capture the target protein.
Secondly, when ligand density is at a low level (<120 ␮mol/g),
for MMI-B-XL, the ligand numbers in the exterior space of pore
were not enough for the rapid binding of hIgG [16,28] as shown
in Fig. 9(a). So the De value of MMI-B-XL resins increases slowly
with the increasing ligand density, and would keep at a low level
(<0.60 m2 /s). However, the De value of MMI-B-XL resins would also
 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
be higher than that of the non-grafted MMI-B-6FF resins because
the ligand distributing on the dextran-grafted layer could bind the
target protein more easily [12,24,54,55]. When the ligand density
is as high as 200 ␮mol/g, the presence of the dextran-grafted layer
would shorten the path of mass transport significantly, as shown
in Fig. 9(b). The increase of ligand density would also improve the
accessibility of the binding sites. As a result, the adsorption rates
could be enhanced substantially when the ligand density reaches
enough high level.
Thirdly, for the dextran-grafted resins, the “chain delivery”
effect would be a kind of additional surface diffusion (protein pass-
ing from one polymer chain to the next), which could improve the
mass transport further [16,17,30]. As shown in Fig. 9(c), due to
the flexibility of dextran-grafted layer, the relatively weak bind-
ing force between MMI ligands and IgG molecule would promote
the occurrence of the “chain delivery” effect. This would be another
reason to improve dramatically the De values of MMI-B-XL resins
compared to that of MMI-B-6FF resins. The “chain-delivery” effect
is related to the ligand density. When the ligand density of MMI-B-
XL resin is lower than 120 ␮mol/g gel, the ligands on the adjacent
dextran-grafted layers are located scatteringly and far away from
each other as shown in Fig. 9(a), limiting the happening of the “chain
delivery” effect [16]. When the ligand density is high enough (such
as 200 ␮mol/g gel), the ligands between adjacent chains becomes
closer as shown in Fig. 9(b), which caused the significant “chain
delivery” effect by swing of the flexible dextran chain and thus
resulted in the remarkable increase of the De values of MMI-B-XL
resin with high ligand density [16,30].
The discussions mentioned above are just some assumption to
explain the effects of dextran-grafted layer and ligand density on
the adsorption behaviors of MMI-B-XL resins. More studies are
needed to reveal the fact of the binding mechanism for dextran-
grafted HCIC resins. However, in general, high ligand density is
necessary for exploiting fully the advantages of dextran-grafted
layer and improving the saturated adsorption capacity, adsorption
rate, and then dynamic binding capacity. In our previous researches,
non-grafted MMI-based resins have been used to separate some
immunoglobulins from different resources [38,48,56–58]. It was
found that IgY could be separated from chicken plasma with the
purity of 92.6% and a recovery of 95.0% by MMI-based resin, while
the recovery of chicken IgY from egg yolk was only 58.4% with the
purity of 92.0% by MEP-based resin [56,57]. IgY could also be sep-
arated efficiently from goose plasma with high purity of 98.6% and
a recovery of 85.0% by MMI-based resins [48]. The results demon-
strated that MMI-based resins had good adsorption selectivity, and
could be used for antibody purification. The coming work would
be focused on the separation performance and the applications of
new dextran-grafted MMI-based resins.
4. Conclusions
A novel dextran-grafted HCIC resin, MMI-B-XL, was prepared
by coupling MMI ligand onto dextran-grafted agarose beads. The
reaction conditions of activation, bromination and ligand coupling
were optimized, and the ligand density could reach 200 ␮mol/g,
improving 43% compared to 140 ␮mol/g for the non-grafted MMI-
B-6FF resins. With hIgG as the model protein, the adsorption
isotherms and adsorption kinetics were investigated to evaluate
the ability of protein adsorption on new HCIC resins with differ-
ent ligand densities. High adsorption capacities of hIgG were found
for MMI-B-XL-200 at pH 7.0. With the increase of ligand den-
sity, Qm increased while the Kd decreased sharply, which indicated
higher affinity for the resin with higher ligand density. In addition,
MMI-B-XL-200 had higher adsorption capacities for hIgG compared
to MMI-B-6FF-140. For adsorption kinetics, it was found that the
123
De increased with increasing ligand density and the presence of
dextran-grafted layer improved significantly the mass transport
in the pore space. It was found that De of MMI-B-XL-200 was
18-fold higher than that of non-grafted resin MMI-B-6FF-140. In
addition, the Q10% , Q10% /Q* and Q10% /Qm on the MMI-XL-200 were
greater by 82–186%, 53–139% and 37–116%, respectively, com-
pared to MMI-6FF-140 at linear fluid velocity of 100–300 cm/h. The
results indicated that new dextran-grafted HCIC resins MMI-B-XL
with high ligand density had a potential application for antibody
purification. The separation selectivity and the applications of new
dextran-grafted MMI-based resins would be focused in the coming
work.
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (21036005 and 21276228), and the Zhejiang
Provincial Natural Science Foundation of China (LR12B06003).
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.chroma.
2014.10.014.
References
[1] S.C. Burton, D.R.K. Harding, Hydrophobic charge induction chromatography:
salt independent protein adsorption and facile elution with aqueous buffers, J.
Chromatogr. A 814 (1998) 71–81.
[2] S.C. Burton, D.R.K. Harding, Preparation of chromatographic matrices by free
radical addition ligand attachment to allyl groups, J. Chromatogr. A 796 (1998)
273–282.
[3] S.C. Burton, D.R.K. Harding, Salt-independent adsorption chromatography: new
broad-spectrum affinity methods for protein capture, J. Biochem. Biophys.
Methods 49 (2001) 275–287.
[4] G.E. Hamilton, F. Luechau, S.C. Burton, A. Lyddiatt, Development of a mixed
mode adsorption process for the direct product sequestration of an extra-
cellular protease from microbial batch cultures, J. Biotechnol. 79 (2000)
103–115.
[5] H.-F. Tong, D.-Q. Lin, X.-M. Yuan, S.-J. Yao, Enhancing IgG purification from
serum albumin containing feedstock with hydrophobic charge-induction chro-
matography, J. Chromatogr. A 1244 (2012) 116–122.
[6] T. Arakawa, M.F. Sugai, K. Tsumoto, Y. Kita, H. Sato, D. Ejima, MEP HyperCel
chromatography II: binding, washing and elution protein, Express Purif. 71
(2010) 168–173.
[7] W. Schwartz, D. Judd, M. Wysocki, L. Guerrier, E. Birck-Wilson, E. Boschetti,
Comparison of hydrophobic charge induction chromatography with affinity
chromatography on protein A for harvest and purification of antibodies, J.
Chromatogr. A 908 (2001) 251–263.
[8] D.-Q. Lin, H.-F. Tong, H.-Y. Wang, S.-J. Yao, Molecular insight into the
ligand–IgG interactions for 4-mercaptoethyl-pyridine based hydrophobic
charge-induction chromatography, J. Phys. Chem. B 116 (2012) 1393–1400.
[9] D.-Q. Lin, H.-F. Tong, H.-Y. Wang, S. Shao, S.-J. Yao, Molecular mechanism
of hydrophobic charge-induction chromatography: interactions between the
immobilized 4-mercaptoethyl-pyridine ligand and IgG, J. Chromatogr. A 1260
(2012) 143–153.
[10] R.-Z. Wang, D.-Q. Lin, H.-F. Tong, S.-J. Yao, Molecular insights into the binding
selectivity of a synthetic ligand DAAG to Fc fragment of IgG, J. Mol. Recogn. 27
(2014) 250–259.
[11] S.C. Burton, N.W. Haggarty, D.R.K. Harding, One step purification of chymosin
by mixed mode chromatography, Biotechnol. Bioeng. 56 (1997) 45–55.
[12] A.M. Lenhoff, Protein adsorption and transport in polymer-functionalized ion-
exchangers, J. Chromatogr. A 1218 (2011) 8748–8759.
[13] A. Staby, I.H. Jensen, Comparison of chromatographic ion-exchange resins: II.
More strong anion-exchange resins, J. Chromatogr. A 908 (2001) 149–161.
[14] Y. Yao, A.M. Lenhoff, Pore size distributions of ion exchangers and relation to
protein binding capacity, J. Chromatogr. A 1126 (2006) 107–119.
[15] L.L. Yu, Y. Sun, Trace adsorption of positively charged proteins onto Sepharose
FF and Sepharose FF-based anion exchangers, J. Chromatogr. A 1253 (2012)
105–109.
[16] L.L. Yu, S.P. Tao, X.N. Dong, Y. Sun, Protein adsorption to poly(ethylenimine)-
modified Sepharose FF: I. A critical ionic capacity for drastically enhanced
capacity and uptake kinetics, J. Chromatogr. A 1305 (2013) 76–84.
[17] L.L. Yu, Y. Sun, Protein adsorption to poly(ethylenimine)-modified Sepharose
FF: II. Effect of ionic strength, J. Chromatogr. A 1305 (2013) 85–93.
124 T. Liu et al. / J. Chromatogr. A 1369 (2014) 116–124
[18] Y. Hong, N. Liu, W. Wei, L.L. Yu, G. Ma, Y. Sun, Protein adsorption to
poly(ethylenimine)-modified Sepharose FF: III. Comparison between different
proteins, J. Chromatogr. A 1342 (2014) 30–36.
[19] N. Liu, L.L. Yu, Y. Sun, Protein adsorption to poly(ethylenimine)-modified
Sepharose FF: IV. Dynamic adsorption and elution behaviors, J. Chromatogr.
A 1362 (2014) 218–224.
[20] M.C. Stone, G. Carta, Protein adsorption and transport in agarose and dextran-
grafted agarose media for ion exchange chromatography, J. Chromatogr. A 1146
(2007) 202–215.
[21] M.C. Stone, Y.Y. Tao, G. Carta, Protein adsorption and transport in agarose
and dextran-grafted agarose media for ion exchange chromatography: effect
of ionic strength and protein characteristics, J. Chromatogr. A 1216 (2009)
4465–4474.
[22] C. Harinarayan, J. Mueller, A. Ljunglöf, R. Fahrner, J.V. Alstine, R.V. Reis, An
exclusion mechanism in ion exchange chromatography, Biotechnol. Bioeng.
95 (2006) 775–787.
[23] E.X.P. Almodovar, Y.Y. Tao, G. Carta, Protein adsorption and transport in cation
exchangers with a rigid backbone matrix with and without polymeric surface
extenders, AIChE J. 27 (2011) 1264–1272.
[24] Y.Y. Tao, G. Carta, Rapid monoclonal antibody adsorption on dextran-grafted
agarose media for ion-exchange chromatography, J. Chromatogr. A 1211 (2008)
70–79.
[25] B.D. Bowes, A.M. Lenhoff, Protein adsorption and transport in dextran-modified
ion-exchange media. II. Intraparticle uptake and column breakthrough, J. Chro-
matogr. A 1218 (2011) 4698–4708.
[26] Y.Y. Tao, G. Carta, G. Ferreira, D. Robbins, Adsorption of deamidated antibody
variants on macroporous and dextran-grafted cation exchangers: I. Adsorption,
J. Chromatogr. A 1218 (2011) 1519–1529.
[27] J. Thömmes, Investigations on protein adsorption to agarose-dextran compos-
ite media, Biotechnol. Bioeng. 62 (1999) 358–362.
[28] B.D. Bowes, H. Koku, K.J. Czymmek, A.M. Lenhoff, Protein adsorption and trans-
port in dextran-modified ion-exchange media. I: Adsorption, J. Chromatogr. A
1216 (2009) 7774–7784.
[29] E. Müller, Comparison between mass transfer properties of weak-anion-
exchange resins with graft-functionalized polymer layers and traditional
ungrafted resins, J. Chromatogr. A 1006 (2003) 229–240.
[30] Y.Y. Tao, G. Carta, G. Ferreira, D. Robbins, Adsorption of deamidated antibody
variants on macroporous and dextran-grafted cation exchangers: II. Adsorp-
tion, J. Chromatogr. A 1218 (2011) 1530–1537.
[31] Y.Y. Tao, E.X.P. Almodovar, G. Carta, G. Ferreira, D. Robbins, Adsorption kinetics
of deamidated antibody variants on macroporous and dextran-grafted cation
exchangers. III. Microscopic studies, J. Chromatogr. A 1218 (2011) 8027–8035.
[32] J. Hubbuch, T. Linden, E. Knieps, A. Ljunglöf, J. Thömmes, M.R. Kula, Mechanism
and kinetics of protein transport in chromatographic media studied by confocal
laser scanning microscopy: Part I. The interplay of sorbent structure and fluid
phase conditions, J. Chromatogr. A 1021 (2003) 93–104.
[33] L.L. Yu, Q.H. Shi, Y. Sun, Effect of dextran layer on protein uptake to dextran-
grafted adsorbents for ion-exchange and mixed-mode chromatography, J. Sep.
Sci. 34 (2011) 2950–2959.
[34] J. Fogle, N. Mohan, E. Cheung, J. Persson, Effects of resin ligand density on
yield and impurity clearance in preparative cation exchange chromatography.
I. Mechanistic evaluation, J. Chromatogr. A 1225 (2012) 62–69.
[35] K. Wrzosek, M. Gramblicka, M. Polakovic, Influence of ligand density on anti-
body binding capacity of cation-exchange adsorbents, J. Chromatogr. A 1216
(2009) 5039–5044.
[36] A. Franke, N. Forrer, A. Butté, B. Cvijetić, M. Morbidelli, M. Jöhnck, M. Schulte,
Role of the ligand density in cation exchange materials for the purification of
proteins, J. Chromatogr. A 1217 (2010) 2216–2225.
[37] H.-L. Lu, D.-Q. Lin, M.-M. Zhu, S.-J. Yao, Protein adsorption on DEAE ion-
exchange resins with different ligand densities and pore sizes, J. Sep. Sci. 35
(2012) 3084–3090.
[38] H.-L. Lu, D.-Q. Lin, D. Gao, S.-J. Yao, Evaluation of immunoglobulin adsorption
on the hydrophobic charge-induction resins with different ligand densities and
pore sizes, J. Chromatogr. A 1278 (2013) 61–68.
[39] H.-F. Xia, D.-Q. Lin, L.-P. Wang, Z.-J. Chen, S.-J. Yao, Preparation and evaluation
of cellulose adsorbents for hydrophobic charge induction chromatography, Ind.
Eng. Chem. Res. 47 (2008) 9566–9572.
[40] D. Gao, S.-J. Yao, D.-Q. Lin, Preparation and adsoription behavior of a cellulose-
based, mixed-mode adsorbent with a benzylamine ligand for expanded bed
applications, J. Appl. Polym. Sci. 107 (2008) 674–682.
[41] W.D. Chen, X.Y. Dong, Y. Sun, Analysis of diffusion models for protein adsorption
to porous anion-exchange adsorbent, J. Chromatogr. A 962 (2002) 29–40.
[42] D. Gao, D.-Q. Lin, S.-J. Yao, Protein adsorption kinetics of mixed-mode adsor-
bent with benzylamine as functional ligand, Chem. Eng. Sci. 61 (2006)
7260–7268.
[43] H.-F. Xia, D.-Q. Lin, Z.-M. Chen, S.-J. Yao, Salt-promoted adsorption of an anti-
body onto hydrophobic charge-induction adsorbents, J. Chem. Eng. Data 55
(2010) 5751–5758.
[44] M.T. Tyn, T.W. Gusek, Prediction of diffusion coefficients of proteins, Biotechnol.
Bioeng. 35 (1990) 327–338.
[45] J.A. Scoble, R.K. Scopes, Assay for determining the number of reactive groups
on gels used in affinity chromatography and its application to the optimisation
of the epichlorohydrin and divinylsulfone activation reactions, J. Chromatogr.
A 752 (1996) 67–76.
[46] S.C. Burton, D.R.K. Harding, High-density ligand attachment to brominated
allyl matrices and application to mixed mode chromatography of chymosin,
J. Chromatogr. A 775 (1997) 39–50.
[47] D.J. Porter, A.T. Stewart, C.T. Wigal, Synthesis of a bromohydrin: an experiment
demonstrating Markovnikov addition, J. Chem. Educ. 72 (1995) 1039–1040.
[48] H.-F. Tong, D.-Q. Lin, Y. Pan, S.-J. Yao, A new purification process for goose
immunoglobulin IgY(Fc) with hydrophobic charge-induction chromatogra-
phy, Biochem. Eng. J. 56 (2011) 205–211.
[49] E.J. Suda, K.E. Thomas, T.M. Pabst, P. Mensah, N. Ramasubramanyan, M.E.
Gustafson, A.K. Hunter, Comparison of agarose and dextran-grafted agarose
strong ion exchangers for the separation of protein aggregates, J. Chromatogr.
A 1216 (2009) 5256–5264.
[50] K. Wrzosek, M. Polakovic, Effect of pH on protein adsorption capacity of strong
cation exchangers with grafted layer, J. Chromatogr. A 1218 (2011) 6987–6994.
[51] N. Forrer, A. Butté, M. Morbidelli, Chromatographic behavior of a polyclonal
antibody mixture on a strong cation exchanger column. Part I: adsorption
characterization, J. Chromatogr. A 1214 (2008) 59–70.
[52] H. Bak, O.R.T. Thomas, J. Abildskov, Lumped parameter model for prediction
of initial breakthrough profiles for the chromatographic capture of antibodies
from a complex feedstock, J. Chromatogr. B 848 (2007) 131–141.
[53] B.D. Bowes, A.M. Lenhoff, Protein adsorption and transport in dextran-modified
ion-exchange media. III. Effects of resin charge density and dextran content on
adsorption and intraparticle uptake, J. Chromatogr. A 1218 (2011) 7180–7188.
[54] Q.H. Shi, G.D. Jia, Y. Sun, Dextran-grafted cation exchanger based on super-
porous agarose gel: adsorption isotherms, uptake kinetics and dynamic protein
adsorption performance, J. Chromatogr. A 1217 (2010) 5084–5091.
[55] E.X.P. Almodovar, Y. Wu, G. Carta, Multicomponent adsorption of monoclonal
antibodies on macroporous and polymer grafted cation exchangers, J. Chro-
matogr. A 1264 (2012) 48–56.
[56] H.-F. Tong, D.-Q. Lin, S.-J. Yao, Purification of immunoglobulin IgY with
hydrophobic charge induction chromatography, CIESC J. 62 (2011) 1574–1580.
[57] H.-F. Xia, D.-Q. Lin, Z.-M. Chen, S.-J. Yao, Purification of immunoglobulin of
egg yolk with hydrophobic charge induction chromatography: comparison of
operation modes with packed bed and expanded bed, Sep. Sci. Technol. 47
(2012) 2366–2372.
[58] H.-F. Xia, D.-Q. Lin, Z.M. Chen, S.-J. Yao, Influences of ligand structure and pH
on the adsorption with hydrophobic charge induction adsorbents: a case study
of antibody IgY, Sep. Sci. Technol. 46 (2011) 1957–1965.