Investigation of Adsorption–Desorption Dynamism of

Bovine Serum Albumin on Crosslinked N,Nⴕ-

Diethylaminoethyl Dextran Microbeads: Solution Phase

Hatice Kaplan Can, Ali Güner

Faculty of Science, Department of Chemistry, Division of Polymer Chemistry, Hacettepe University, Beytepe 06532,
Ankara, Turkey

Received 26 April 2004; accepted 19 November 2004

DOI 10.1002/app.21949
Published online 14 December 2005 in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: New hydrogel microspheres based on for pH 6.9. The adsorption capacity factor and the adsorp-
crosslinked dextran containing N,N⬘-diethylaminoethyl tion equilibrium constant were obtained and mathematical
(DEAE) groups with different chemical structures have been modeling of adsorption, adsorption rate constants, and max-
used in adsorption– desorption studies. Bovine serum albu- imum adsorption were determined. Swelling kinetics of
min (BSA) is frequently used in biophysical and biochemical crosslinked DEAE dextran and optimum ionic strength, pH,
studies. BSA has a well-known primary structure that has and mass of hydrogel were also investigated. Desorption
been associated with binding of many different categories of studies were finally determined under optimum medium
small molecules. Both adsorption kinetics and equilibrium conditions. © 2005 Wiley Periodicals, Inc. J Appl Polym Sci 99:
isotherms for the adsorption of BSA on crosslinked DEAE 2288 –2299, 2006
dextran have been determined experimentally. These were
only slightly dependent on the initial concentration of BSA
but were considerably affected by the pH of the medium. Key words: DEAE dextran microbeads; adsorption; pro-
The results fitted the Freundlich–Langmuir isotherm model teins; UV–vis spectroscopy; polysaccharides

INTRODUCTION of purifying to homogeneity many biochemical sub-

stances from dilute mixtures containing a large num-
Industry has a significant interest in the design, opti-
ber of proteins exhibiting small differences in solubil-
mization, and control of large-scale affinity adsorption
ity, charges, and molecular size. This separation tech-
systems that are to be employed in the purification of
nique utilizes small differences in biological activities
biologically active macromolecules for use as pharma-
of molecules to achieve the purification of proteins.
ceuticals or in other applications where the purifica-
Often, very high levels of purity as well as high prod-
tion of the product is a very important consideration.
uct yields may be obtained in a single purification
Certain fundamental mechanisms underlying the af-
step. The chromatographic separation of proteins is
finity adsorption separations have been identified and
important not only for analysis but also in such large-
constitutive expressions that may be used to quantify
scale applications, as in the food and the drug indus-
these mechanisms and their effects have been sug-
tries.4 Bioaffinity chromatography is a highly specific
gested and constructed.1,2 Many conventional separa-
separation technique used for the isolation and puri-
tion processes (i.e., gel filtration, ion-exchange, precip-
fication of biomolecules. However, some limitations
itation, ultrafiltration, electrophoresis) and their com-
exist such as biological instability and leakage desorp-
binations (multistep purification procedures) may not
tion conditions, and cost may limit its applications. A
be applicable to purification of proteins from complex
number of ion exchangers for protein separation have
biochemical mixtures.3 However, biospesific adsorp-
been developed to use for analysis, and a large num-
tion (affinity chromatography) seems to have the
greatest potential to succeed in the very difficult task ber of chromatograms have been presented to show
that they are useful for protein separation. When these
ion exchangers are applied to large-scale chromato-
Correspondence to: A. Güner ([email protected]). graphic separation, it is necessary to investigate in the
Contract grant sponsor: Hacettepe University, Research detail the equilibria, kinetics, and dynamics of the
Foundation; contract grant number: 01 01 001 007. system.5 Yamamoto et al. examined the adsorption of
Contract grant sponsor: Turkish National Technical and bovine serum albumin (BSA) on crosslinked N,N⬘-
Research Council (TUBITAK) BAYG Group.
diethylaminoethyl (DEAE) dextran (a weakly basic
Journal of Applied Polymer Science, Vol. 99, 2288 –2299 (2006) dextran-type ion exchanger) decreases with decreas-
© 2005 Wiley Periodicals, Inc. ing pH.6 Tsou and Graham showed the effect of the
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2289

concentration of NaCl (ce) in the BSA solution on the K2HPO4-KHPO3 for pH 7–7.5). The ionic strength of
adsorption isotherm of BSA on DEAE at pH 6.9.7 They the medium was adjusted by NaCl. The initial concen-
showed that the isotherm for ce ⫽ 0 is much more tration of BSA solution in aqueous phase was varied
favorable than that for ce ⫽ 1%. However, these re- between 0.1 and 0. 01 g/100 mL.
ports are fragmentary and systematic experimental
investigations for adsorption isotherms of protein Adsorption studies
have not been reported. Yoshida et al. presented ex-
perimental equilibrium isotherms of adsorption of Adsorption experiments were carried out at a constant
BSA on crosslinked DEAE dextran ion exchanger us- temperature of 25°C. Microbeads were contacted with
ing Na⫹ and Cl⫺ ions.5 BSA is the most abundant the BSA solution and gently mixed. The amount of
protein in blood plasma. It has many important phys- BSA adsorbed on the particle was measured after a
iological functions that contribute significantly to col- certain time and until it was confirmed that there was
loid osmotic pressure and aid in the transport, distri- no further adsorption. The adsorption capacity was
bution, and metabolism of many endogenous and ex- determined by measuring the initial and final concen-
ogenous substances.8 –13 trations of BSA within the reservoir at ␭ ⫽ 279 nm
In this work, adsorption kinetics, equilibrium iso- using a Schimadzu-100 double beam UV–vis spectro-
therms for adsorption of BSA on crosslinked DEAE photometer. Using the formula the adsorbed phase
dextran, have been determined experimentally. The concentration of BSA was calculated with
adsorption isotherm model was found as Freundlich–
Langmuir isotherm model. Adsorption capacity factor q ⫽ V共c 0 ⫺ c e兲/W, (1)
and adsorption equilibrium constant were obtained
where c0 and ce are the initial and equilibrium concen-
and mathematical modeling of adsorption, adsorption
trations of BSA in liquid phase (mg/mL), respectively,
rate constant, and maximum adsorption was per-
q denotes the microbeads-phase concentration of BSA
formed. Adsorption mechanism of a BSA on
(mg BSA/g of crosslinked DEAE dextran microbeads),
crosslinked DEAE dextran was estimated according to
and W is the mass of the microbeads (g) and V is the
the experimental results of the adsorption. It was
volume of the solution (mL).
proved both experimentally and theoretically that pro-
tein is adsorbed by electrostatic attraction, ion-ex-
change, hydrophobic interaction, and/or hydrogen Desorption studies
bonding. Besides the adsorption mechanism, swelling Desorption of the adsorbed BSA from the crosslinked
kinetics of crosslinked DEAE dextran and optimum DEAE dextran microbeads was studied in a reservoir.
ionic strength, optimum pH, and optimum mass of The microbeads carrying different amounts of BSA
hydrogel were investigated. Desorption studies and were placed with in the desorption medium contain-
the desorption ratio of the system were determined for ing 1.0M NaCl at medium pH and an amount of BSA
optimum medium conditions. was released in 5 h. Desorption ratios for BSA were
calculated by use of the expression:
EXPERIMENTAL
Desorption ratio (%) ⫽ (amount of BSA released)/
Chemicals
(amount of BSA on the microbeads) ⫻ 100. (2)
Commercial DEAE dextran microbeads were used
(Sephadex A-50, Pharmacia Fine Chemicals).
Swelling studies
Swelling percentage value volumetric measurements
were made to measure the microgels. For this purpose
a constant weight of crosslinked DEAE dextran micro-
gels was left into the volumetric vessel and the incre-
ment of the volume was followed as the function of
time. Microgels were immersed in distilled water at
room temperature. The percentage of swelling (%)
was determined by the equation
(BSA was purchased from Sigma (lyophilized, frac-
tion V). Adsorption studies were performed in a BSA S共%兲 ⫽ V swollen共mL兲/Vdry共mL兲 ⫻ 100, (3)
reservoir by constant stirring rate at120 rev/min. The
pH of the adsorption medium was changed in the where Vdry is the initial volume of the microgels at t
range pH 4.5–7.2 using different buffer systems. (0.1M ⫽ 0 and Vswollen is the volume of the swollen micro-
CH3COONa-CH3COOH for pH 4 – 6 and 0.1M gels at time t.
2290 CAN AND GÜNER

TABLE I followed by the gradual binding of the latter to BSA as

Properties of Crosslinked DEAE Dextran Microbeads observed in the SEM micrographs (Fig. 1). There is a
(Sephadex A-50)5
small difference volume swelling between the SEM
Crosslinked DEAE dextran (Sephadex A-50) micrographs and the percentage swelling value. This
Description: Weakly basic anion exchanger
can be explained as experimental error, such as drying
Functional group: Diethylaminoethyl (DEAE) of the microbeads.
Diameter in watera (mm) A fundamental relationship exists between the level
Free: 0.2364 of swelling of a crosslinked polymer in a solvent and
Saturated by BSA: 0.2182 the chemical nature of both the polymer and the sol-
Diameter of dry particlea (mm): 0.0851
Water content (% wt): 94.1
Effective pH range: 2–9
pK of X-DEAE dextran group: 9.5
Density
True (kg/m3): 1690
Apparent (kg/m3): 604
Porosity: 0.963
Total capacity: 3.5 ⫾ 0.5 mg/g
Available capacity
Albumin (MW ⫽ 67,000): 7566 mg/g
Hemoglobin (MW ⫽ 69,000): 5000 mg/g
Ferritin (MW ⫽ 440,000): 74 mg/g
a
Average value of 50 particles.

Scanning electron microscopy (SEM) studies

To observe the surface topography of the crosslinked
DEAE dextran microspheres, we took scanning elec-
tron micrographs of the gold-coated samples with a
scanning electron microscope (Raster Electronen Mi-
croscopy, Leitz-AMR-1000, Germany).

RESULTS AND DISCUSSION

Properties of crosslinked DEAE dextran
microbeads
An ideal microbead for protein separation must have
the following properties: high hydrophilicity and low
nonspecific adsorption, fairly large pore size and nar-
row pore distribution, chemical and mechanical resis-
tance, as well as enough reactive functional groups.
Crosslinked DEAE dextran microgels have an appro-
priate distribution, good mechanical stability, and ad-
sorption capacity.4 Table Isummarizes the physical
properties of crosslinked DEAE dextran used in this
study as obtained commercially.
The SEM micrographs of the dry, swollen, and ad-
sorbed forms of microbeads can be seen in Figure 1.
This figure shows the surface structure of microbeads
in dry, swollen, and adsorbed BSA forms. Crosslinked
DEAE dextran microbeads have a spherical shape and
a smooth surface. These surface characteristics of the
microspheres favor better adsorption of the BSA as a
result of an increase in surface area. Moreover,
crosslinked DEAE dextran had initially a uniform and
spherical form with smooth surface characteristics. Figure 1 Representative SEM micrographs of crosslinked
After swelling and adsorption, the smooth surface of DEAE dextran in the form of (a) dry, (b) swollen, and (c)
the spheres was rapidly covered with BSA molecules, adsorbed BSA.
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2291

Figure 2 Swelling (%) –time curve of the crosslinked DEAE dextran microsphere in water.

vent. Swelling of the three-dimensional network struc- Adsorption studies

ture in a suitable solvent is the most important param-
Effect of mass of microspheres on adsorption
eter (other especially important parameters are mass
and volume swelling).14,15 To determine the optimum mass of the microspheres,
Charged groups attached to the polymeric network several masses of crosslinked DEAE dextran microgels
structure play an essential role in swelling properties. were used for the BSA adsorption studies. By increas-
Several attempts to establish the water content of mi- ing the mass of the microbeads, the adsorption of BSA
crospheres have been described in the literature, e.g., increased up to a certain value; however, when very
by determination of the difference in total volume of large amount of microspheres were employed the ad-
wet and dry microspheres or the difference in total sorption could not be measured. The optimum mass of
weight either with or without blotting. In another crosslinked DEAE dextran microspheres was esti-
study, the wet microspheres were filtered and the mated to be 0.005 g (Fig. 3).
change in filter mass was determined. Another ap-
proach for determination of the solvent content of
Adsorption kinetics
microspheres is based on their size upon swelling.16,17
The increase in size can be determined using static and Figure 4 implies the curves of adsorption kinetics ob-
dynamic light scattering techniques as well as with tained in medium at pH 6.9. High adsorption rates
microscopic methods. The percentage swelling of the were observed at the beginning of the adsorption, and
dextran microbeads can be derived by the difference saturation values were then gradually achieved in
between volume in the dry and swollen states. about 5 h. This may be because of the decrease in the
The presence of the hydrogen bonding groups in the BSA concentration in the reservoir with time due to
polymer matrix causes the network to be less swollen, adsorption. As expected; when BSA concentration in
compared to a less polar structure. Figure 2 shows the the mobile phase (i.e., the aqueous phase) decreases
swelling kinetics of the crosslinked DEAE dextran in (which also corresponds to an increase in the station-
water. Equilibrium swelling of crosslinked DEAE dex- ary phase, i.e., the sorbent microbeads) the driving
tran value for microspheres was approximately 105%. force (i.e., the concentration difference between the
The equilibrium degree of swelling of ionic hydrogels two respective phases) decreases, which brings about
is controlled by three major factors: mixing of the a decrement in the adsorption rate.
polymer with a swelling medium (miscibility), the Based on the adsorption kinetics of BSA/dextran
elastic-refractive force exerted on the network, and the system, it may be concluded that the protein adsorp-
ionic pressure generated from the mobile counterions tion can be observed on anion- and cation-exchange
onto the charged ion in network. Because of the low resins and semiplateau values vary with the experi-
equilibrium swelling value of the crosslinked DEAE mental conditions; the semiplateau or kink cannot be
dextran microbeads, this matrix can be easily used in observed in the isotherm and is not directly related to
the column separation applications.18 the surface coverage of protein molecules on resin.
2292 CAN AND GÜNER

Figure 3 Effect of mass of crosslinked DEAE dextran on the BSA adsorption at constant BSA concentration 0.1 g/100 mL,
temperature 25°C; and total protein concentration 10 mL.

This effect probably reflects a conformational alter- matrix (crosslinked DEAE dextran) should be a mini-
ation of the BSA molecules adsorbed. mum for high specificity. As shown Figure 5, increas-
ing BSA concentration in solution results in an in-
crease in the BSA amount adsorbed by the mi-
Effect of initial BSA concentration on adsorption
crobeads, reaching a plateau at protein concentration
A number of data on the adsorption isotherm thus far of 0.2 mg/mL (0.005 g/25 mL).
observed indicate that the higher the bulk concentra- The BSA adsorption capacity of the microbeads varies
tion, the greater the amount of protein adsorbed, between 44.2 and 13.6 mg/g, implying that the increase
showing a saturation curve. Figure 5 shows the ad- in the BSA adsorption capacity is due to the increase in
sorption of BSA on the crosslinked DEAE dextran the BSA concentration in the initial state. Adsorption
microbeads. Note that one of the main requirements in capacity values of crosslinked DEAE dextran mi-
bioaffinity chromatography is the specificity of the crobeads are given in Table II. This can be explained in
sorbent. The nonspecific interaction with the carrier terms of specific interactions (electrostatic, hydrophobic,

Figure 4 Crosslinked DEAE dextran microbeads–BSA adsorption kinetics curves at different BSA concentration (at 25°C pH 6.9).
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2293

Figure 5 Dependence of the adsorption on initial concentration of the bovine serum albumin (BSA) solution on the
adsorption.

H-bonding, and salt bridge) between microbeads and oropolymer containing anion exchange and cation ex-
the BSA molecules. The cationic groups of DEAE dex- change groups. Chase24 was able to reach 14 mg
tran microbeads interact with BSA molecules from BSA/g with Cibacron Blue F3GA attached to Sepha-
which it can be concluded that the BSA molecules were rose CL-6B. Compared with these results it is clear that
adsorbed on the microbeads via functional groups. crosslinked DEAE dextran microbeads exhibit a high
Different sorbents with varying adsorption capaci- albumin adsorption capacity in the present system.
ties were reported in the literature for albumin ad-
sorption. Nigel et al.20 used Sepharose CL-6B-200 as
the carrier matrix and they reported BSA adsorption Effect of pH
capacities around 1–3 mg BSA/g gel. Denizli et al.21 Proteins are best known as amphoteric molecules ca-
obtained 35 mg BSA/g adsorption capacity using pable of existing in solution as both cations (pH⬍pI)
Cibacron Blue F3GA attached to PVA particles. Boyer and anions (pH⬎pI); their negative charge increases
and Hsu22 used Sepharose beads carrying different with an increase in pH and its positive charge in-
amounts of Cibacron Blue F3GA and reported adsorp- creases with a decrease in pH. Adsorption capacities
tion values up to 55.9 mg BSA/g polymer. McCreath of the proteins increase with a decrease in pH. Pro-
et al.23 found a 9.7 mg HSA/g equilibrium adsorption teins can be defined as copolymers of some 22 differ-
capacity with the PVA-coated particulate perflu- ent amino acids of varying hydrophobicity; as a con-
sequence, proteins are amphiphilic molecules. For ion-
exchange resin containing weak acid or weak base
TABLE II groups, the degree of the dissociation of groups is
Adsorption Capacities of the DEAE Dextran Microbeads dependent on the pH of the solution; the number of
for Protein Adsorption at Constant BSA Concentrations
0.1– 0.03 g/100 mL, pH 6.9, at 25 °C; Total Protein dissociated DEAE group decreases with the increases
Volume is 10 mL of pH. The adsorption process involves the transfer of
a protein molecule from solution to resin surface, and
qeq
BSA concentration (mg BSA/g DEAE dextran the environmental change upon adsorption results in
(g/100 mL) microspheres) change in the structure of the proteins.
Table III represents the effect of pH on the equilib-
0.10 46.2
rium adsorption of protein. In all cases, the highest
0.09 41.2
0.08 37.8 adsorption was observed at pH 4.72 and 7.2. Signifi-
0.07 34.4 cantly, lower adsorption capacities were obtained in
0.06 26.2 more acidic or more alkaline pH regions. In all inves-
0.05 24.4 tigated cases, the maximum adsorption of BSA was
0.04 19.4
observed at pH 4.8, which is the isoelectric point of
0.03 13.6
BSA for DEAE dextran. The maximum adsorption was
2294 CAN AND GÜNER

TABLE III in a variation of the hydrophobic interactions between

Effect of pH on Bovine Serum Albumin (BSA) the protein and the resin matrix.
Adsorption from Microspheres at Constant BSA
Concentration 0.1 g/100 mL, at 25 °C; Total Protein
(iv) The increase of ionic strength may cause the net-
Volume is 10 mL work of the resin to shrink, reducing the porosity of the
resins and hence the availability of the binding sites.
Adsorption capacity
of the microbeads
There is clearly a need to develop a model to incor-
qeq (mg BSA/g X- porate the effects of ionic strength on protein adsorp-
Medium pH DEAE dextran gel) tion, but it is very difficult to develop a theoretical
model for the prediction of such a complex effect.25
4.50 0.80
4.72 1.60
The effects of the ionic strength (IS) (adjusted by add-
4.92 0.80 ing NaCl) on BSA adsorption are presented in Figure
5.98 0.80 6, which shows that the adsorption capacity decreases
6.80 0.80 with increasing ionic strength of the adsorption me-
6.90a 46.2 dium. It should be noted that in the experimental data
7.20 1.60
set, the salt concentration refers to sodium chloride
a
pH 6.9 is the nonbuffered protein solution. added to the solution. The actual ionic strength will be
slightly higher than that generated by the sodium
chloride due to additional ions provided by cationic
observed at pH 6.9, which is nonbuffered protein so- diethyl amino ethyl groups of the dextran resins and
lution. At pH 6.9 the value was used hereafter as the the fact that the stock protein is likely to have some
medium pH of the protein solution. The effect of pH salts in it. The “background” effect of the ionic
on the adsorption can be explained as the electrostatic strength provided by the additional ions is, however,
repulsion effects of the charges groups. The decrease constant and the increase in the ionic strength occurs
in the BSA adsorption capacity can be attributed to as the NaCl concentration increases.
electrostatic repulsion effects between the identically The effects of the IS (adjusted by adding NaCl) on
charged groups at higher and lower pH levels. At the BSA adsorption are presented in Figure 6, which
isoelectric points, proteins have no net charge and, shows that the adsorption capacity decreases with
therefore, the maximum adsorption from solution is increasing ionic strength of the adsorption medium.
usually observed at this point. When the negative In addition, given the small amount of salts that
charges of the BSA molecule increase, the high pH may possibly be brought into the solution with the
value for the number of fixed groups needed to neu- stock protein, it is reasonable to ignore the influence of
tralize the negative charges of one BSA molecule in- the background effect. The adsorption capacity of BSA
creases. Consequently, the adsorption capacity decreases by an amount 13% at 0.1 g/100 mL BSA
changes according to upper and lower pH limits. Be- (46.2% in the absence of salt reducing to 33.4% at the
cause the dissociation of anionic groups of the BSA minimum salt concentration) the adsorption capacity
molecule is completed at a pH value within the alka- drops rapidly as the salt concentration is increased.
line region, the adsorption capacities increase the max- The decrease in the adsorption capacity with the in-
imum values as of BSA/crosslinked DEAE dextran creasing ionic strength can be attributed to the de-
and its dependence on pH [(pH 4.72 (approximately creased attraction between cationic amine group of
isoelectric point of BSA) and pH 7.2 (anionic character dextran and protein molecules.
of BSA)] are shown Table IV. According to the data in When the salt concentration increases in the adsorp-
Table IV an increase in initial concentration of BSA can tion medium, this can cause cation– cation interactions
cause also increment the adsorption capacity of the
microbeads.
TABLE IV
Adsorption Capacity of BSA/Crosslinked DEAE Dextran
Microbeads Dependent on the pH Values
Effect of the ionic strength
Initial BSA qeq (mg BSA/g X- qeq (mg BSA/g X-
The effect of ionic strength on the protein adsorption concentration DEAE dextran gel) DEAE dextran gel)
onto ion exchange/affinity adsorbents is a very com- (g BSA/100 mL) at pH 4.72 at pH 7.20
plex phenomenon. It could be considered from differ-
0.10 2.60 3.20
ent aspects as follows. 0.09 2.50 2.40
(i) The salt counterions compete against the protein 0.08 1.60 1.60
ions for binding sites. 0.07 1.50 1.60
(ii) The salt coions shield the protein ions and un- 0.06 0.80 0.80
charged binding sites from each other. 0.05 0.76 1.60
0.04 2.44 2.40
(iii) The change of ionic strength changes the folding
0.03 0.80 1.60
and configuration of the protein molecules, resulting
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2295

Figure 6 Effect of ionic strength on BSA adsorption on the crosslinked DEAE dextran microspheres at constant BSA
concentration 0.1 g/100 mL, temperature 25°C, pH 6.9, and total protein concentration 10 mL.

with the sodium salt, which leads to low protein ad- and many different isotherm models, such as the lin-
sorption. The distortion of existing salt bridges in the ear Langmuir, the bi-Langmuir, the Fowler, the Redli-
presence of sodium salt also contributes to the low ch–Peterson, the Javanovic, the S-shaped isotherm, the
protein adsorption at high ionic strength. Tsou et al.26 Freundlich, and the Freundlich–Langmuir isotherms,
showed the effect of the concentration of NaCl in the have been proposed for the adsorption of solutes in a
BSA solution on the isotherm for the adsorption of liquid solution onto solid surfaces.27 Most of these
BSA on DEAE dextran at pH 6.9. models are essentially empirical although theoretical
derivations have been accomplished in some cases.
Among all, this model is probably the most popular
Adsorption equilibrium isotherms
due to its simplicity and its good agreement with the
The most common presentation of the adsorption data experimental data. Modeling of the adsorption pro-
is the adsorption isotherm, where, at constant temper- cesses on crosslinked DEAE dextran microbeads was
ature, the amount of sorbate (qeq) is plotted against the realized by applying different adsorption isotherms.
steady-state concentration ceq of the sorbate in the For determining equilibrium isotherms of BSA/
bulk solution after adsorption. Fully measured ad- crosslinked DEAE dextran microbeads Freundlich–
sorption isotherms provide a convenient method for Langmuir isotherm models were used.
determining whether an adsorption process can be
k1
treated as reversible. Investigations on simple model
systems, consisting of a well-characterized protein, a 共X-DEAE Dextran) ⫹ Cl ⫺ ⫹ BSA N
k2
well-characterized sorbent, and an aqueous solvent
containing only simple ions, have provided the most complex ⫹ Cl ⫺ (4)
reliable and meaningful data on the thermodynamics
of protein adsorption. In most cases, the system is After applying the Langmuir and Freundlich ad-
unbuffered so that pH changes and proton transfer sorption isotherm models, these two models were not
reactions can be monitored as a function of adsorbed valid for the interpretation of adsorption kinetics of
amounts. Results from such studies are the basis for the system. Thus, a third model was employed for the
most of the general principles that are thought to investigation of adsorption equilibrium trend.
govern the protein adsorption process; unfortunately, The third adsorption isotherm was fitted to the
even for these simple systems, a complete picture has Langmuir-Freundlich isotherm using a nonlinear re-
not yet emerged. Nevertheless, resolution of the inter- gression method, the same as in the case of previous
actions governing adsorption in these model systems studies.28 Although the Freundlich–Langmuir equa-
will provide the only reliable foundation for under- tion (Fig. 7) can be used to model adsorption cooper-
standing protein adsorption in more complex, techno- ativity, the combination of the Langmuir and the
logically important processes. For protein adsorption Freundlich expressions is more often applied on a
at solid/liquid interfaces, some model system studies strictly empirical basis, since the extra parameter (n)
2296 CAN AND GÜNER

Figure 7 Freundlich–Langmuir isotherm model for adsorption of BSA on crosslinked DEAE dextran microspheres.

ensures an improved correlation with the equilibrium t/q t ⫽ 共1/k 2 q eq

2
兲 ⫹ 共1/qeq兲t, (6)
data compared to the simple Langmuir isotherm,
where k2 (g mg min⫺1) is the rate constant of pseudo
q eq ⫽ qmax共K c1/n
d eq 兲/共1 ⫹ K c1/n
d eq 兲, (5) second-order adsorption and qeq denotes the amounts
of adsorption at equilibrium (mg g⫺1) and t is time
where qmax is the plateau adsorbed amount (maxi- (min). The kinetic model based on concentration is
mum amount of adsorption) and KD is the adsorption represented in Figure 8, which displays a second-
constant. In the Langmuir isotherm, Kd is the ratio of order kinetic. According to this model two different
the adsorption to the desorption rate constant. In the behaviors can be proposed. Due to the first assump-
Langmuir–Freundlich isotherm this definition does tion, it is considered that the empty sites of the micro-
not apply but its meaning is almost the same; a high spheres are first filled with BSA molecules, after which
value of Kd represents faster rates of adsorption over they move closer toward each other due to the elec-
desorption. Figure 7 represents the Langmuir–Freun- trostatic interactions among the DEAE dextran, pro-
dlich isotherm. From the isotherm of the BSA/ tein, and water molecules (anionic– cationic, hydro-
crosslinked DEAE dextran microspheres Kd and qmax phobic– hydrophilic), subsequently followed by fur-
values were as high as 7.9 mg/mL and 56.81 mg/g, ther binding of other BSA molecules in several layers
respectively (R2 ⫽ 0.987). onto the microspheres previously covered with BSA.
The latter explanation can be interpreted by the con-
centration of BSA molecules. As the BSA concentra-
tion increases, the adsorption capacity of the micro-
Mathematical modeling of the adsorption
spheres also increases where the adsorption rate is
To examine the controlling mechanism of the adsorp- very high at the beginning of kinetics (due to the
tion process of such a mass transfer and chemical empty sites of microspheres) and at high BSA concen-
reaction, kinetic models are used to test the experi- trations (Table V). However, with proceeding time,
mental data. The large number and different chemical the adsorption rate slows down since the adsorption
groups on the BSA (e.g., –COOH, -NH2, ⫽ NH, -SH, of the latter approaching BSA molecules becomes
-OH) imply that there are many types of protein– more difficult for the microspheres and consequently
DEAE dextran microgels interactions. Kinetic models the adsorption rate constant should increase to pro-
including the pseudo first-order and pseudo second- vide further adsorption of approaching BSA mole-
order equations can be used in this case assuming the cules. The theoretical qeq values, which can be seen
measured concentrations. The first-order rate equation from Table V, estimated from the kinetic modeling
of the Lagergren is one of the most widely used for the gave significantly close values to that of the experi-
sorption of a solute from a liquid solution. For deter- mental values, and correlation coefficients were also
mining kinetic modeling of adsorption the following found to be higher (greater than 0.995). All these re-
equation was used,29 sults suggest that the crosslinked DEAE dextran mi-
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2297

Figure 8 Mathematical modeling of adsorption kinetics of BSA on crosslinked DEAE dextran microspheres using adsorp-
tion capacity at 25°C, pH 6.9, and total protein volume 10 mL.

crospheres of BSA system are described with the conditions of the solution in which the surface is im-
pseudo second-order kinetic model, based on the as- mersed. With respect to reversibility of adsorption/
sumption that the rate-limiting step may be adsorp- desorption of proteins, distinction should be made
tion involving the electrostatic interactions among the among reversibility toward dilution of the solution,
DEAE dextran, protein, and water molecules as (i) changes in pH and ionic strength, addition of other
hydrophobic forces, (ii) electrostatic interactions, and types of surface-active substances, and exchange
(iii) hydrogen bonding through sharing or exchange of against dissolved proteins. The usual procedure for
electrons between sorbent and sorbate, providing the desorption is to remove the adsorbed protein rapidly
best correlation of the data. by increasing the salt concentration or changing the
pH. However, under conditions of reversible adsorp-
tion, as evidenced by a Langmuir isotherm, desorption
Desorption studies
can also be achieved (although more slowly) merely
The adsorption of proteins on solid surfaces is often by removing protein from the feed solution, i.e.,
described as irreversible since, after attachment is es- switching back to a solution of buffer and NaCl.31,32
tablished, which has usually reached a final value However, in the Freundlich–Langmuir isotherm sys-
within an hour, subsequent replacement of the protein tem, high protein recovery can be obtained by increas-
solution by pure solvent, as a rule, does not lead to ing the pH and/or the ionic strength of the eluent.
any significant desorption on a time scale of hours or Desorption experiments were carried out in a reser-
even days.30 However, it does not mean that protein voir. Albumin adsorbed crosslinked DEAE dextran
molecules remain attached to the surface whatever the microspheres were placed in the reservoir and NaCl

TABLE V
Adsorption Kinetic Constants and Adsorption Capacity of the System of Crosslinked
DEAE Dextran Microspheres in Different Concentration of BSA Solutions
k2
BSA concentration Experimental qeq Theoretical qeq (g mg⫺1 min⫺1)
(g/100 mL) (mg BSA/g microspheres) (mg BSA/g microspheres) ⫻104

0.10 46.2 48.9 1.59

0.09 41.2 43.9 1.61
0.08 37.8 40.5 2.58
0.07 34.4 36.3 2.20
0.06 26.2 28.2 3.46
0.05 24.4 25.7 5.56
0.04 19.4 20.3 7.29
0.03 13.6 13.9 39.5
2298 CAN AND GÜNER

Figure 9 Desorption ratio of the percentage of BSA from crosslinked-DEAE dextran microspheres.

solution (0.2M NaCl) was added. The amount of albu- crobeads cannot be desorbed from the microbeads.
min released in 1 h was determined. The desorption One may simply imply that only those BSA molecules
ratio for BSA was calculated using the equation attached to the surface in the upper layers have been
Desorption ratio (%) ⫽ (amount of BSA desorbed/ desorbed and not those in the lower layers or in the
amount of BSA adsorbed microbeads) ⫻ 100. empty sites of microspheres.
It can be concluded that the adsorption may be
categorized within a chemisorption process since the
CONCLUSIONS
maximum desorption value reaches only 42.7% where
the recovered amount access seems to be very low Unlike small molecules, protein possesses specific sec-
(Table IV). One may simply imply that only those BSA ondary and higher levels of structure. These higher lev-
molecules attached to the surface in the upper may els of structure are maintained by relatively weak non-
have been desorbed and not those in the lower layers covalent interactions. Distribution of these weak interac-
or in the empty sites of microspheres. The trend vali- tions can be caused by factors such as temperature (heat
dated by the Freundlich–Langmuir isotherm may also or cold), pH, salt, shear, surface interactions, and freeze-
be another solid proof for classification of our adsorp- drying. Results in the desorption parts indicate that bo-
tion behavior as a chemical–physical adsorption pro- vine serum albumin attachment to the crosslinked DEAE
cess. Figure 9 shows the difference between the ad- dextran microbeads occurs rapidly at medium pH. De-
sorbed amount of BSA and the desorbed amount of sorption occurs with NaCl but the adsorption process is
BSA from the crosslinked DEAE dextran. As can be not fully reversible. These results support the interpre-
seen from Table VI; approximately 20 mg BSA adsorp- tation that bovine serum albumin has a moderate affinity
tion capacity/g crosslinked DEAE dextran mi- for crosslinked DEAE dextran.

TABLE VI
Desorption Percentage (%) of BSA from Crosslinked DEAE Dextran Microspheres
Amount of Amount of
BSA solution Desorption ratio adsorbed BSA desorbed BSA BSA amount on gel
(g/100 mL) (BSA %) (mg BSA/g gel) (mg BSA/g gel) (mg BSA/g gel)

0.10 42.4 46.2 19.6 26.6

0.09 42.7 41.2 17.6 23.6
0.08 24.2 37.8 9.20 28.6
0.07 23.0 34.4 7.90 26.5
0.06 22.7 26.2 6.00 20.3
0.05 4.10 24.4 1.00 23.0
0.04 4.70 19.4 0.90 18.5
0.03 2.10 13.6 0.30 13.3
.
ADSORPTION–DESORPTION DYNAMISM ON DEAE MICROBEADS 2299

The authors thank Dr. Handan Yavuz and Dr. Yücel Şahin 17. Hennink, W. E.; van Nostrum, C. F. Advanced Drug Delivery
for their help. Reviews 2002, 54, 13.
18. Kawaguchi, H. Prog Polym Sci 2000, 25, 1171.
19. Boudy, V.; Voute, N.; Pradeau, D.; Chaumeil, J. C. Int J Pharm
References 2002, 239, 13.
20. Nigel, M.; Lindner, R.; Jeffcoat, R.; Lowe, C. R. J Chromatogr
1. Dabrowski, A. Adv Colloid Interface 2001, 93, 135. 1989, 473, 227.
2. McCoy, M. A.; Liapis, A. I. J Chromatogr 1991, 548, 25. 21. Tuncel, A.; Denizli, A.; Purvis, D.; Lowe, C. R.; Pişkin, E. J Chro-
3. Bruce, L. J; Chase, H. A. Chem Eng Sci 2001, 56, 3149. matogr 1993, 634, 161.
4. Denizli, A., Köktürk, G.; Yavuz, H.; Pişkin, E. J Appl Polym Sci 22. Boyer, P. M.; Hsu, J. T. Chem Eng J 1992, 47, 241.
1999, 74, 2803. 23. McCreath, G. E.; Owen, R. O.; Nash, D. C.; Chase, H. A. J Chro-
5. Yoshida, H.; Nishihara, H.; Kataoka, T. Separ Sci Technol 1999, matogr A 1996, 773, 73.
29, 2227. 24. Chase, H. A. J Chromatogr 1984, 297, 179.
6. Yamamoto, S.; Nakanishi, K.; Matsuno, R.; Kamikuto, T. Bio- 25. Lan, Q.; Bassi, A. S.; Zhu, J. X.; Margaritis, A. Chem Eng J 2001,
technol Bioeng 1983, 25, 1373. 81, 179.
7. Tsou, H. S.; Graham, E. E. AlChE J 1959, 1985, 31. 26. Tsou, H. S.; Graham, E. E. AlChE J 1959 1985, 31.
8. Norbert, W. Fundamentals of Clinical Chemistry, Saunders: 27. Norde, W.; Haynes, C. A. Thermodynamics of protein adsorp-
London, 1976. tion; In Brash, J. L., Wojciechowski, P. W., Eds., Interfacial
9. Gelamo, E. L.; Tabak, M. Spectrochim Acta A 2000, 56, 2255. Phenomena and Bioproducts, Dekker: New York, 1996; pp.
10. Nichifar, M.; Cristea, D.; Carpov A. Int J Biol Macromol 2000, 28, 27–54.
15. 28. Yoon, J. Y.; Kim, J. H.; Kim, W. S. Colloids Surf B Biointerfaces
11. Chang, C.; Lenhoff, A .M. J Chromatogr A 1998, 827, 281. 1998, 12, 15.
12. Hlady, V.; Buijs, J. Curr Opin Biotech 1996, 7, 72. 29. McKay, G.; Ho, Y. S. Process Biochem 1999, 34, 451.
13. Turkova, J. Affinity Chromatography; Elsevier: The Nether- 30. Taniguchi, T.; Duracher, D.; Delair, T.; Elassari, A.; Pichot, C.
lands, 1993. Colloids Surf B Biointerfaces 2003, 29, 53.
14. Kaplan, H.; Güner, A. J Appl Polym Sci 2000, 78, 994. 31. Leaver, G.; Howell, J. A.; Conder, J. R. J Chromatogr 1992, 590,
15. Kaplan, H.; Güner, A. Adv Polym Technol 2000, 19, 210. 101.
16. Stenekes, R. J. H.; Hennink, W. E. Int J Pharm 1999, 189, 131. 32. Conder, J. R.; Hayek, B. O. Biochem Eng J 2000, 6, 225.