Elucidation of intraparticle mass transfer mechanisms in protein chromatography is essential for process design. This study investigates the differences of adsorption and diffusion parameters of basic human fibroblast factor 2 (hFGF2) in a simple (purified) and a complex (clarified homogenate) feed solution on the grafted agarose-based strong cation exchanger Capto S.
Trang 1Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/chroma
Markus C Berga, Jürgen Beckb, Alex Karnerb, Kerstin Holzerb, Astrid Dürauera, b,
Rainer Hahna, b, ∗
a Austrian Center of Industrial Biotechnology, Muthgasse 18, Vienna 1190, Austria
b Department of Biotechnology, Institute of Bioprocess Science and Engineering, University of Natural Resources and Life Sciences Vienna, Muthgasse 18,
Vienna 1190, Austria
a r t i c l e i n f o
Article history:
Received 30 March 2022
Revised 17 June 2022
Accepted 18 June 2022
Available online 19 June 2022
Keywords:
Pore diffusion
Solid diffusion
Grafted media
Ion exchange chromatography
hFGF2
a b s t r a c t
Elucidationofintraparticlemasstransfermechanismsinproteinchromatographyisessentialforprocess design.Thisstudyinvestigatesthedifferencesofadsorptionanddiffusionparametersofbasichuman fi-broblastfactor2(hFGF2)inasimple(purified)and acomplex(clarified homogenate)feedsolution on thegraftedagarose-basedstrongcationexchangerCaptoS.Microscopicinvestigationsusingconfocallaser scanningmicroscopyrevealedslowerintraparticlediffusionofhFGF2intheclarifiedhomogenate com-paredtopurifiedhFGF2.Diffusiveadsorptionfrontsindicatedastrongcontributionofsoliddiffusionto theoverallmasstransferflux.Proteinadsorptionmethodssuchasbatchuptakeandshallowbedaswell
as breakthroughcurve experimentsconfirmeda40-foldreductionofthemass transferflux forhFGF2
inthehomogenatecomparedtopurehFGF2.Theslowermasstransferwasinducedbycomponentsof theclarifiedhomogenate.Essentially,theincreased dynamicviscositycausedbyahigherconcentration
ofdsDNAand membranelipidsintheclarifiedhomogenatecontributedtothisdecreaseinmass trans-fer.Moreover,bindingcapacityforhFGF2wasmuchlowerintheclarifiedhomogenateandsubstantially decreasedtheadsorbedphasedrivingforceformasstransfer
© 2022TheAuthors.PublishedbyElsevierB.V ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/)
1 Introduction
Mass transfer of proteins in chromatographic media is a slow
diffusional process Typically, different transport resistances appear
in adsorption systems with porous particles [ 1, 2] Firstly, proteins
must pass through a stagnant film around the particle before diffu-
sional transport into the particle itself can occur Properties of the
liquid phase surrounding each particle affect these external mass
transfer mechanisms Concentration differences across the bound-
ary layer and the film thickness define the impact of film resis-
tance [2–4] The type of diffusion mechanism is defined by the
structural properties of the chromatographic resin and buffer con-
ditions as well as the protein characteristics Pore diffusion is dom-
inant for macro-porous resins which can be fully penetrated by a
protein of a size smaller than the pore structure During pore diffu-
sion, surface attachments and detachments can take place repeat-
∗ Corresponding author at: Department of Biotechnology, Institute of Bioprocess
Science and Engineering, University of Natural Resources and Life Sciences Vienna,
Muthgasse 18, Vienna 1190, Austria
E-mail address: rainer.hahn@boku.ac.at (R Hahn)
edly The main driving force of this mechanism is controlled by the solute concentration gradient in the liquid phase of the pore Solid diffusion, on the other hand, occurs for proteins in the adsorbed state In contrast to pore diffusion, detachment does not typically take place for adsorbed proteins This phenomenon enables an en- hanced mass transfer caused by an absorbed protein concentration gradient It is important to mention that other attributes such as narrow pores or pores blocked by adsorption of large molecules can lead to hindered diffusion [5]
Pore diffusion coefficients between 10 −6– 10 −8cm ²/s and solid diffusion coefficients ranging from 10 −8 to 10 −10 cm ²/s have been reported for various proteins [ 1, 6] The order of magnitude de- pends on the properties of the protein and the mobile phase as well as characteristics of the stationary phase [ 1, 7, 8] Besides con- ventional macro-porous chromatography resins, polymer grafted media are frequently used for protein purification [2] These special chromatographic materials contain polymer chains grafted onto the particle surface Such modifications enhance binding capacity
as well as mass transfer, as compared to conventional media [9– 16] According to Tao et al [8], effective pore diffusion coefficients that are greater than the diffusivity for proteins in free solution
https://doi.org/10.1016/j.chroma.2022.463264
0021-9673/© 2022 The Authors Published by Elsevier B.V This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )
Trang 2can be achieved on dextran grafted resins for ion exchange pro-
cesses This result may be explained by a transport enhancement
facilitated by an electrostatic driving force or a coupling transport
of proteins and ions [ 2, 9]
The mass transfer mechanisms of single or two component sys-
tems on ion exchange media as well as proteins on affinity resins
(such as protein A) have been previously investigated [ 1, 7, 17–21]
However, little data is available on diffusional adsorption processes
of proteins in complex solution such as clarified bacterial ho-
mogenates [ 22, 23] In general, most isoelectric points of E coli host
cell proteins are in the acidic range [24] This enables purifica-
tion of positively charged recombinant proteins on cation exchang-
ers with hardly any coeluting process-related impurities if binding
conditions are adjusted properly In principle, this high selectivity
should allow the determination of protein mass transport parame-
ters with single component adsorption models
The aim of the present study was to elaborate the differences
in mass transfer mechanisms of purified basic human fibroblast
growth factor 2 (hFGF2) and a clarified homogenate containing
hFGF2 overexpressed in E coli This basic protein exhibits an iso-
electric point of 9.6 at a molecular size of 17.2 kDa in monomeric
form [ 25, 26] and is therefore well suited for the purification by
cation exchangers For all our main investigations, the strong cation
exchange resin Capto S from Cytiva (Uppsala, Sweden) was chosen
as the stationary phase This resin is based on an agarose back-
bone grafted with dextran inclusions [ 8, 9] Since host cell-related
impurities such as DNA and membrane lipids are present in the
clarified homogenate, dynamic viscosity is increased significantly
compared to the pure protein solution which affects the diffu-
sion mechanisms [ 27, 28] For determining parameters describing
the mass transfer of the protein of interest, batch and packed bed
methods were performed
2 Theory
2.1 General model for pore and solid diffusion
A general model for mass transfer for batch adsorption assum-
ing parallel mass transfer of pore and solid diffusion mechanisms
within the bead pores as given by [29]:
∂q
∂t = 1
r2
∂
∂r
r2
D e∂c
∂r+D s∂q
∂r
(1)
with boundary conditions:
r=0∂c
r=r p D e∂c
∂r+D s∂q
and material balance:
dC
dt = −3k f V M
r p V
C − c|r=r p
= −V M
V
d ¯q
where q is the solute concentration in particle, t is the time, the
particle radial coordinate, D ethe effective pore diffusivity, D sis the
effective solid diffusivity , is the solute concentration in the pore
fluid, p is the particle radius, C the solute concentration in the
bulk fluid, C 0is the initial protein concentration, V Mis the volume
of the particles, V is the bulk solution volume and ¯q is the particle-
average solute concentration A reduction of the general model to
the pore diffusion model can be obtained when D s = 0 and vice versa for the solid diffusion model when D e = 0
For column adsorption operations Eq.(1d)is replaced by:
εb∂C
∂t +(1−εb) ∂¯q
∂t +u∂C
∂z=εb D L∂2C
with boundary conditions:
z=0u C F=uC− εb D L∂C
z=L∂C
where εb is the extra-particle void fraction, is the bed length coordinate, D L is the axial dispersion coefficient and L is the bed length
2.2 Analytical solutions for film and pore diffusion control
For a rectangular isotherm, an analytical solution for batch ad- sorption and film and pore diffusion control has been obtained by Teo and Ruthven [30]:
C0
q max
D e t
r p2 =
1− 1
Bi
with:
I1 = 1
6λ ln
λ3+η3
λ3+1
λ+1
λ+η
3
λ √3 tan
−1
2η−λ
λ√3
− tan−1
2−λ
λ√3
(3a)
I2=31ln
λ3+η3
λ3+1
(3b)
V C0
(3c)
λ=1
1/3
(3d)
Bi= k f r p
D e
(3f)
where q max is the maximum binding capacity, Bi is the Biot num- ber, and F is the frictional approach to equilibrium
The solutions for film mass transfer control are
C
C0 =exp
−3k f
r p
V M
V t
(4)
and
q
q m=1 1− exp
−3k f
r p
V M
V t
(5)
For shallow-bed adsorption operations the solution is [29]:
εp D e t
r p2
C0
q m =1
2−1 3
1− 1
Bi
F−1
2(1− F)2/3 (6)
Trang 3A constant pattern solution for film and pore diffusion in col-
umn adsorption described by Weber et al [31]is given as:
N pore( τ1− 1)= √15
3tan
−1 2η− 1
√ 3
−15
2 ln
1+η+η2
−1 3 + 5
Bi
ln
1−η3
+1
− 5π
2√
where
τ1=
ut
L −εb
N pore= 15(1−εb)D e L
ur2
p
(7b)
η=
1− C
C F
1/3
(7c)
=(1−εb)q max
where u is the superficial velocity, τ1 is the dimensionless time
and C F is the feed protein concentration The film mass transfer
coefficient k f for packed bed operation can be obtained from the
following correlation:
Sh=1.15
Re
ε
1/2
where Sh is the Sherwood number, Re is the Reynolds number and
Sc is the Schmidt number
2.3 Analytical solution for film and solid diffusion control
For solid diffusion control an analytical solution is given by
[32]:
F =1−π62
∞
n=1
1
n2exp
−n2π2D s t
r p2
(9)
Eq.(9)is valid for intraparticle mass transfer control if δ≥ 1
δ is the diffusion resistance parameter which determines the
controlling mechanism and is described as:
δ=1
5
k f r p
D s
C0
q max
(10)
An analytical solution for the column adsorption for external
mass transfer and solid diffusion has been obtained by Yoshida
et al [33]:
C
C0 =1
δexp
τ−ξ+δ− 1−1
δ
for
τ−ξ ≤ −δ+1+1δ− ln
1+δ δ
(11)
C
C0 =1− δ
1+δexp −τ+ξ−δ+1+
1
δln
1+δ δ
/δ
for
τ−ξ ≥ −δ+1+1
δ− ln
1+δ δ
(12)
when δ≥ 1, and:
C
C0 =exp( τ−ξ− 1)
for
C
C0 =1− δ
1+δexp{[−τ+ξ+1− ln(1+δ )]/δ}
for
when δ≤ 1
τ is the dimensionless time and ξ is the bed length parameter These parameters are defined as:
τ=3r k f
p
C0
q max
t−εL u
(15)
and
ξ =3(1−ε )
r p
k f L
3 Materials and method
3.1 Sample preparation
hFGF2 was produced in Escherichia coli according to the proce- dure published by Sauer et al [25] BL21 cells containing hFGF2 were resuspended in 50 mM Na 2HPO 4/NaH 2PO 4 pH 6.5 to yield
in a 60 gCDM/L suspension The suspension was homogenized at 700/70 bar for 2 passages on a Niro Soavi PANDAPlus 20 0 0 (GEA, Parma, Italy) Removal of cell debris was achieved by centrifuga- tion and a filtration step on a 0.2 μm sterile filter (Fluorodyne EX EDF, PALL, Dreieich, Germany)
Purification of hFGF2 was performed on a 10 mL Capto S col- umn The filtered homogenate (100 mL) was loaded and eluted using a linear gradient ranging from 0 to 1 M NaCl The protein eluate was buffer exchanged via a PD-10 desalting column (Cytiva, Uppsala, Sweden) into 50 mM Na 2HPO 4/NaH 2PO 4, 30 mM NaCl
pH 6.5 Both the clarified homogenate as well as the purified and buffer exchanged protein solutions were used to conduct the ex- periments as described in the subsequent sections A hFGF2 con- centration of 2 mg/mL was chosen for all experiments unless oth- erwise stated
3.2 Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)
SDS–PAGE was performed in an XCell SureLock TM Mini-Cell Electrophoresis System (Thermofisher Scientific, Dreieich, Ger- many) with an EPS 301 power supply (Cytiva, Uppsala, Sweden) A NuPage 4–12% BIS-Tris gel (Thermofisher Scientific, Dreieich, Ger- many) was used for separation All samples were 10-fold diluted
in 50 mM Na 2HPO 4/NaH 2PO 4 pH 6.5 prior to application onto the gel Separation was carried out according to the manufacturer’s di- rections after 5 μL of 4x sample buffer (Thermofisher Scientific, Dreieich, Germany), 2 μL of DTT, and 13 μL of the diluted sample aliquots were combined and an aliquot (14 μL) was loaded onto the gels The separated proteins were fixed with a solution con- taining 50% ethanol and 10% glacial acetic acid Protein bands were stained with Coomassie blue G 250 for 30 min Destaining was performed with a solution containing 25% ethanol and 8% glacial acetic acid Destained gels were scanned on an Epson Perfec- tion V700 Photo scanner (EPSON, Meerbusch, Germany) SeeBlue TM
Plus2 pre-stained protein standard (7 μL) (Thermofisher Scientific, Dreieich, Germany) was used as the molecular mass marker on each gel
3.3 Adsorption isotherms
A 50 mM Na 2HPO 4/NaH 2PO 4 30 mM NaCl pH 6.5 buffer was used for slurry preparation of the Capto S resin (Cytiva, Uppsala,
Trang 4Sweden) Specific amounts of slurry were mixed with fixed vol-
umes of either clarified homogenate consisting of hFGF2 or puri-
fied hFGF2 In general, the protein concentration of hFGF2 in both
matrices was set to 2–2.1 mg/mL The resin-protein mixtures were
incubated in 2 mL vials at room temperature for 24 h on an end-
over-end rotator (Stuart SB3, Cole-Parmer, Stone, Staffordshire, UK)
at 13 rpm Afterwards, all samples were filtered through 0.22 μm
syringe filters (Millex-GV PDFV, Merck Millipore, Darmstadt, Ger-
many) The filtrate was then quantified via HPLC on a Propac WCX-
10 (Thermofisher Scientific, Dreieich, Germany) with the dimen-
sions 4 mm x 250 mm A linear gradient over 2.5 CV ranging from
0 to 1 M NaCl in 100 mM Na 2HPO 4/NaH 2PO 4 pH 6.5 was applied
at a volumetric flow rate of 1 mL/min The amount of bound hFGF2
per mL resin was determined by mass balance calculation Equilib-
rium constant K Land maximum binding capacity q maxwere deter-
mined by fitting the data to the Langmuir model
3.4 Adsorption kinetics
For all batch uptake experiments, a 10 mL glass beaker was
filled with 5 mL protein solution The slurry of Capto S in 50 mM
Na 2HPO 4/NaH 2PO 4 (0.078–0.15 mL) was added and stirred on a
magnetic stirrer at 300 rpm Samples (150 μL) of the protein-resin
suspension were drawn at the predefined time points and immedi-
ately filtered through 0.22 μm syringe filters The filtrate was then
quantified analogously as described for adsorption isotherms The
film mass transfer coefficient k f was determined by a correlation
for external mass transfer [ 1, 4] Data obtained via adsorption ki-
netics were fitted to pore and solid diffusion models by applying
Eqs.(3)–(5)and (9)
Experiments at increased dynamic viscosity were conducted by
adding specific amounts of glycerol (Thermofisher Scientific, Dreie-
ich, Germany) to the protein solutions Further studies with pure
lysozyme as a model protein as well as Toyopearl Gigacap S (Tosoh,
Griesheim, Germany) and SP Sepharose FF (Cytiva, Uppsala, Swe-
den) as alternative stationary phases were performed analogously
3.5 Breakthrough curve experiments
Columns packed with Capto S and SP Sepharose FF resin were
purchased from Repligen (Weingarten, Germany) Breakthrough
curves (BTC) were recorded on columns with 0.2 mL and 1.0 mL
column volume corresponding to column dimensions of 0.5/1 and
0.5/5 cm A 50 mM Na 2HPO 4/NaH 2PO 430 mM NaCl pH 6.5 buffer
was used for column equilibration Residence time for all BTC ex-
periments varied between 1 and 10 min The breakthrough for
hFGF2 in the clarified homogenate was monitored by quantifying
collected fractions via HPLC on a WCX column as described above
Analytical solutions including external mass transfer were used to
fit all breakthrough profiles for obtaining solid and pore diffusion
coefficients Eqs.(7)and ( (11)–(14))
3.6 Shallow bed adsorption
Capto S resin (5 μL) was transferred into a HR 5/50 column
(Cytiva, Uppsala, Sweden) Silica beads (20 μL of 100 μm beads)
were added for emulating the conditions of a fixed bed [34] Prior
to sample application the column was equilibrated with 50 mM
Na 2HPO 4/NaH 2PO 4 30 mM NaCl pH 6.5 Protein solutions were
pumped onto the column in a circular manner at a flow rate of
1 mL/min Elution of bound protein after specific timepoints was
achieved by applying a step gradient of 1.0 M NaCl in the equi-
libration buffer for 1 min The amount of protein bound was de-
termined by using a linear equation describing the correlation of
the absorbance signal at 280 nm and the protein concentration
Adsorption profiles were fitted to pore and solid diffusion models Eqs.(6)and ( (9))
3.7 Confocal laser scanning microscopy
Fluorescent labeling of hFGF2 was performed with Rhodamine Red TM – X dye (Thermofisher Scientific, Dreieich, Germany) Prior
to labeling, hFGF2 was transferred into 500 mM bicarbonate pH 8.5 buffer by using a PD-10 desalting column labeling dye was added into a 5 mL vial in a 1:3 dye: protein molar ratio and in- cubated for 1 h wrapped in aluminum foil Unbound dye was then removed by using a desalting column A labeling ratio of 0.28 was obtained Labelled hFGF2 was then added to the protein solution in
a 1:40 mass ratio Adsorption kinetic experiments were performed with the labelled protein solutions as described above Moreover, microscopy images were recorded with a 40x dry objective and a resolution of 512 × 512 pixels at a frequency of 400 Hz
4 Results and discussion
4.1 Purification of hFGF2
Based on the purification procedure published by Sauer et al [25], a cation exchange chromatography system was selected to capture the basic hFGF2 To increase capture efficiency the poly- mer grafted cation exchanger Capto S was chosen instead of CM Sepharose FF and buffer concentration was reduced Clarified ho- mogenate containing hFGF2 was loaded onto a Capto S column In total, 40 mg of hFGF2 were loaded per mL resin Fig 1A shows the absorbance and conductivity profiles of the capture step A gaussian-shaped elution peak was obtained without any visible shoulders Analysis of the collected fractions by SDS-PAGE revealed
a highly pure eluate fraction as shown in Fig.1B, lane 14 No ad- ditional bands other than the one of monomeric hFGF2 were de- tected Since most host cell components of E coli strain BL21 are negatively charged at the chosen buffer conditions hardly any im- purities were captured and co-eluted in the desorption step In- creasing concentrations of hFGF2 could be detected in the flow through fractions at the end of the loading step (lanes 3–13) Ap- parently, hFGF2 was loaded beyond the maximum binding capac- ity of Capto S Nevertheless, replacing CM-Sepharose FF by Capto S substantially increased capacity, and even more importantly, purity
of the eluate was very high Due to this high selectivity, modeling
of the adsorption was simplified and competitive adsorption mod- els, as described for hFGF2 capture by Kołodziej et al [20], were not required
4.2 Adsorption isotherm
A first estimation of the binding capacity was done by evalu- ating the capture run profile Hence, for obtaining accurate values for the maximum binding capacity q max as well as for the equi- librium constant K L adsorption isotherms were established from batch uptake measurements Both purified hFGF2 and clarified ho- mogenate containing hFGF2 were investigated For pure hFGF2,
50 mM Na 2HPO 4/NaH 2PO 4 30 mM NaCl pH 6.5 was chosen as the buffer system which exhibited a comparable conductivity as deter- mined for the homogenate (8 mS/cm) Fig.2shows the adsorption isotherms including a fit of the Langmuir adsorption model to the experimental data Pure hFGF2 yielded in a maximum binding ca- pacity of 100 mg/mL whereas a q maxof 24 mg/mL was determined for hFGF2 in the clarified homogenate Additionally, a 32-fold re- duction of the equilibrium constant K L was obtained for the clari- fied homogenate containing hFGF2
Trang 5Fig 1 Capture run of clarified homogenate containing hFGF2 on Capto S (CV = 1 mL) (A) 40 mg hFGF2 per mL column volume were loaded A linear gradient from 0 to
1 M NaCl in 50 mM Na 2 HPO 4 /NaH 2 PO 4 pH 6.5 over 4 CV was applied to elute the protein (B) SDS-PAGE under reduced conditions with Coomassie staining of load, flow through and eluate fractions of Capto S run shown in (A) Lane 1 molecular weight marker SeeBlue; Lane 2 clarified homogenate/load containing hFGF2 (black); Lane 3 to
13 corresponding to the flow through fractions (blue); Lane 14 represents the eluate fraction (green) Note that hFGF2 has a molecular weight of 17.2 kDa (For interpretation
of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig 2 Adsorption isotherms of purified hFGF2 (A) and clarified homogenate containing hFGF2 (B) on Capto S Absorbed concentration q was plotted against the fluid phase
concentration C Solid curves represent fits with the Langmuir model
4.3 Adsorption kinetics
To gain an insight into the mass transfer mechanism, investi-
gation by confocal laser scanning microscopy (CLSM) of the intra-
particle mass transport was performed In general, pore diffusion
and steep isotherms result in a shrinking core behavior with vis-
ible sharp fronts In contrast, solid diffusion is distinguished by
diffusive profiles and fast progression to the center of the parti-
cle As previously shown by Beck et al [35], pore diffusion only
leads to diffusive fronts if the isotherm is extremely shallow ( K L <
0.1) Fig.3 shows the CLSM images of pure hFGF2 (A) as well as
hFGF2 in clarified homogenate (B) Both systems exhibit diffusive
fronts and progression to the particle center that occurred rapidly,
suggesting a significant contribution of solid diffusion to the mass
transfer A difference in the mechanism between the two feed so-
lutions is not apparent in this analysis, but it can clearly be seen
that mass transfer of hFGF2 in the clarified homogenate was sub-
stantially reduced
Adsorption kinetics was analyzed via batch uptake in a stirred
tank as well as in a shallow bed adsorption system [34] Experi-
mental adsorption kinetics at 2.0 mg/mL are shown in Fig 4 In
agreement with the isotherm measurements, binding capacity for hFGF2 in the clarified homogenate was much lower compared to pure hFGF2 In addition, for both experimental methods, batch up- take and shallow bed adsorption, a significant decrease of the ad- sorption kinetics of hFGF2 on Capto S in the clarified homogenate compared to purified hFGF2 was confirmed For quantitative anal- ysis of the uptake data, we followed a stepwise approach con- sidering different contributions to the overall mass transport For that purpose, we used established correlations and structural pa- rameters that had been previously determined Analytical solu- tions for batch uptake are available for both mass transfer mech- anisms, pore and solid diffusion as described in the theory sec- tion These solutions have been derived assuming a rectangular adsorption isotherm For pure hFGF2, the isotherm is highly favor- able but not for the clarified homogenate containing hFGF2 All ex- periments were performed at an hFGF2 concentration of 2 mg/mL which yielded a constant separation factor of ∼ 0.15 for the clar- ified homogenate To investigate if the analytical solution would lead to a significant error, we compared the analytical to the nu- merical solution (Fig A2 in supportive information) We found only minor differences and as such the analytical solutions can be used
Trang 6Fig 3 CLSM measurements of the adsorption process of purified hFGF2 (A) and clarified homogenate containing hFGF2 (B) on Capto S hFGF2 was labelled with Rhodamine
Red-X Measurement points from left to right: 30 s, 60 s, 120 s, 300 s, 600 s and 1800 s of incubation
Fig 4 Adsorption of purified hFGF2 ( ) and clarified homogenate containing hFGF2 (o) via batch uptake (A) and Shallow bed adsorption (B) on Capto S hFGF2 concentration
was 2.0 mg/mL for both model systems Solid lines represent a fit with the solid diffusion model ( Eq (9) )
as a good estimate From a mechanistic point of view, a pore or
solid diffusion model assuming only a single mass transfer mecha-
nism is not fully physically representative, as both diffusion mech-
anisms contribute to the overall mass transfer Despite that, exper-
imental data can be well approximated with both models and the
derived diffusion coefficients can be considered as apparent diffu-
sion coefficients Adsorption data shown in Fig.4were fitted with
both models yielding basically identical curves For clarity, only the
solid diffusion model is plotted
For the case of pore diffusion, an ab initio estimation of the ef-
fective pore diffusion coefficient D e can be derived from character-
istic parameters of the solute and the stationary phase, respectively
[ 1, 32]:
D e=ψp∗εp D0
In Eq.(17)ψp is the hindrance parameters for pore diffusion,
εp the intraparticle porosity, D 0 the free diffusivity of the protein
and τ is the tortuosity factor for intraparticle diffusion
The hindrance parameter ψpis proportional to the ratio of the
pore and protein radius and can be calculated by [ 1, 36]:
ψp = 0 865 ∗(1 −λm)2
∗
1 − 2 1044 ∗λm+ 2 089 ∗λ3
m∗ 0 984 ∗λ5
m
(18)
where λmis the ratio of pore over protein radius
The free diffusivity D 0 of hFGF2 can be estimated by the Tyn-
Gusek equation [37] if the dynamic viscosity and temperature of
the solution are known An intraparticle porosity of 0.74 and a tor-
tuosity of 1.5 have previously been determined for Capto S [38]
Calculated parameters are listed in Table1 Using Eq.(17)D e was
estimated as 3.9 × 10−8 cm ²/s for pure hFGF2 The calculated
pore diffusivity is reasonably low which is mainly attributed to the small pore radius of 5 nm Effectively, this is a truly theoreti- cal value that only would apply in the case of sole pore diffusion Calculated D e was then compared to effective pore diffusion coef- ficients derived from experimentally performed batch uptake and shallow bed adsorption systems Again, a maximum binding capac- ity q maxof 100 mg hFGF2/mL CaptoSwas obtained for the pure protein solution After fitting the experimental data to the corresponding model, an effective pore diffusivity of ˙D eof 1.4 ± 0.2× 10−6cm ²/s
was determined, which was 34-fold higher compared to the D e
calculated from Eq.(17) In accordance with the CLSM profiles, a strong contribution of solid diffusion to mass transfer is supposed
We followed the approach of Hunter et al [29], interpreting the ef- fective D e determined with the pore diffusion model as an overall mass transfer flux, which can be written as
˙
D e=D e+D s
q max
Eq.(19)considers the coherence of two individual mass transfer mechanisms where the first term represents pore diffusion and the second term solid diffusion As can be seen from Eq.(19), the effect
of solid diffusion depends on the magnitude of the solid diffusion coefficient D sand the driving force expressed as the ratio of q max
over the initial protein concentration C [2]
To further investigate the individual solid diffusion contribu- tion, the experimental data were fitted to a solid diffusion control model ( Eq.(9)) A solid diffusion coefficient D sof 1.7 ± 0.1× 10−8
cm ²/s was determined for pure hFGF2 via batch uptake and shal- low bed adsorption Inserting all parameters ( D e , D s , q max and C ) into Eq.(19)yields ˙D e,calcof 1.4 × 10−6cm ²/s This value is identi-
cal to the overall effective pore diffusivity ˙D e,exp obtained from the experimental data and the fitting to the pore diffusion model
Trang 7Table 1
Summary of the adsorption kinetics data obtained for purified hFGF2 and hFGF2 in the clarified homogenate via shallow bed adsorption and batch uptake experiments λm is the ratio of protein and pore radii ψ p is the hindrance parameters for pore diffusion and τ is the tortuosity D 0 is the solution diffusivity The overall effective pore diffusivity ˙D e,calc was calculated by Eq (19) The reported values for ˙D e,exp and D s are arithmetic means and standard deviations determined from batch uptake and shallow bed experiments
Fig 5 Breakthrough curves of purified hFGF2 (A) and clarified homogenate containing hFGF2 (B) on a Capto S column (CV = 1 mL) with 3 min residence time Filled circles
( ●) and filled squares ( ) represent experimental data at residence time of 3 min Dotted lines represent prediction with pore diffusion model Eq (7) ) and dashed lines represent prediction with solid diffusion model ( Eqs (11) –( (14) ) using parameters determined by batch uptake kinetics
As with pure hFGF2, we conducted a quantitative analysis of the
mass transfer for hFGF2 in the clarified homogenate Due to higher
dynamic viscosities of the solution an even lower hypothetical D e
of 7.0 × 10−9 cm ²/s was estimated Again, the quantitative analysis
was compared to the effective diffusion coefficients derived from
experimental data In contrast to pure hFGF2, a lower solid dif-
fusion contribution was determined, which was due to the lower
maximum binding capacity and the lower D sof 3.2 × 10−9 cm ²/s
Consequently, under these conditions, the overall mass transfer
flux was drastically reduced to ˙D e,exp = 3.3 × 10 −8cm ²/s This value
corresponded to half of ˙D e,calcwhich was based on the hypothetical
D ecalculated from Eq.(17)
A summary of the adsorption kinetics data obtained can be
found in Table1
4.4 Breakthrough curves
Predictions for column operation were made based on mass
transfer parameters derived from batch uptake kinetics The cal-
culated BTCs were then compared to experimental data of pure
hFGF2 Fig.5A) and hFGF2 in clarified homogenate ( Fig 5B) Ana-
lytical solutions for column adsorption are available for both mass
transfer mechanisms ( Eqs (7)and ( (11)–(14)) In Fig 5A, the ex-
perimental breakthrough curve for pure hFGF2 is plotted against
predictions using D s = 1.6 × 10−8 cm ²/s for solid diffusion model
and ˙D e,exp = 1.2 × 10−6 cm ²/s for the pore diffusion model Film
mass transfer coefficients were calculated from the correlation as
given by [4] It is evident that both models yielded almost identi-
cal breakthrough profiles However, both models captured only the
first 30% of the experimental BTC and predicted faster mass trans-
fer as occurred experimentally In contrast, breakthrough of hFGF2
in clarified homogenate was very well predicted by both models
using D s = 3.2 × 10 −9 cm ²/s and ˙D e,exp = 3.3 × 10 −8 cm ²/s, re-
spectively ( Fig.5B) Further experimental runs were conducted at
varying residence times ( Fig 6) Predictions based on the mass
transfer parameters as used above yielded essentially the same
results with strong overprediction for pure hFGF2 For the clari-
fied homogenate, the experimentally observed early breakthrough
of hFGF2 at a residence time of 1 min was also well captured by the model ( Fig.6B) To verify if this overprediction was specific for pure hFGF2 on Capto S, a series of supporting experiments was carried out First, lysozyme adsorption on Capto S was studied at the same experimental conditions Lysozyme has a molecular mass
of 14.3 kDa and an isoelectric point (11.0) comparable to hFGF2 Thus, adsorption properties were expected to be similar In fact, almost identical behavior in terms of batch uptake rate ( Fig 7A) and protein breakthrough including overprediction by the model was observed ( Fig 7B) Mass transfer parameters are included in Table2
To further investigate this phenomenon, adsorption of pure hFGF2 was studied on two other cation exchangers: Toyopearl Gi- gacap S has similar properties as Capto S, but with a polymethacry- late backbone and grafted surface extenders functionalized with sulfopropyl modifiers The batch uptake curve and fitting of the experimental data yielded almost the same diffusion coefficients
as obtained for the Capto S resin, confirming the very fast up- take rates on grafted media Diffusion coefficients are listed in Table2and the uptake curve is provided in the supportive infor- mation (Fig A3) We also investigated a macro-porous resin with- out grafted polymers, SP Sepharose FF ( Fig.8) The batch uptake was substantially reduced, and the diffusion coefficients obtained from the fit were one order of magnitude lower as compared to Capto S Correspondingly, BTCs were less steep compared to those obtained on the grafted media and exhibited a stronger depen- dence on velocity The profiles were well predicted by the pore dif- fusion model for the most part of the mass transfer zone A sum- mary of the diffusivity values obtained for all methods and sta- tionary phases used can be found in Table2 Adsorption of basic proteins on SP Sepharose FF has been investigated by several re- searchers Dziennik et al [39] have shown that the diffusion co- efficients for lysozyme varied considerably with buffer conditions For higher ionic strengths, protein uptake was comparably fast as
on grafted media Martin et al [40]showed that the uptake rates were also dependent on the protein characteristics, as shown for
Trang 8Fig 6 (A) Breakthrough curve (BTC) of purified hFGF2 on Capto S column (CV = 0.2 mL) with varying residence time Filled circles ( ●) represent residence time of 1 min
Residence time of 3 min is represented as open circle (o) (B) Breakthrough curve of clarified homogenate containing hFGF2 on Capto S column (CV = 1 mL) with varying residence time Filled squares ( ) represent residence time 1 min Residence time of 10 min represented as open squares ( ) Analytical solution for solid diffusion by Yoshida
et al Eqs (11) –( (14) ) used for data prediction is shown as dashed lines
Fig 7 Adsorption of lysozyme on Capto S (A) Solid diffusion model used for fitting data Eq (9) ) Breakthrough curve of lysozyme on Capto S (CV = 0.2 mL) with varying residence time (B) Filled circles ( ●) represent residence time of 1 min Residence time of 3 min is represented as open circle (o) Predictions via analytical solution for solid diffusion by Yoshida et al ( Eqs (11) –( (14) ) are shown as dashed lines
Table 2
Summary of all mass transfer coefficients obtained for purified hFGF2 and clarified homogenate containing hFGF2 and additional experiments with lysozyme
Mass transfer coefficient (cm ²/s) Method/Sample
Batch uptake on
SP Sepharose FF
1.5E-7 2.6E-9
lysozyme and cytochrome c The latter had a diffusion coefficient
that was almost one order of magnitude higher than lysozyme al-
though the molecular weight and the isoelectric point are very
similar Apparently, the amino acid composition and the binding
strength can have an important impact on the adsorption proper-
ties Taking all these factors and variations into account, the uptake
rates determined for hFGF2 on SP Sepharose FF in this work seem
very reliable
Furthermore, we investigated if extra-column contributions
could be responsible for deviations of the experimental BTCs and
the model predictions shown in Figs.5, 6A and 7B We performed
BTCs under non-binding conditions and bypassing the column As
shown in Fig A1 in supportive information, extra-column effects
could not be responsible for the discrepancy as the upper part of
the BTCs reached the plateau rapidly without any tailing or wash- out effects Bowes and Lenhoff[41]investigated protein adsorption
on grafted media including Capto S Under weak binding condi- tions, and more pronounced for smaller proteins, they observed a slow approach to equilibrium reflected by a tailing behavior in the upper part of the BTC They concluded that a rearrangement of the initially bound proteins on the dextran layer could be responsible for the observed effects All of these studies are indicative of such
a behavior of pure hFGF2 and lysozyme on grafted media which made the prediction of BTCs challenging Tao et al [8]predicted BTCs of a mAb on Capto S by a solid diffusion model However, since mAb has a molecular mass of 150 kDa, solid diffusion coeffi- cients were 3-fold lower than those we determined for the smaller proteins hFGF2 and lysozyme In contrast, when the uptake rates
Trang 9Fig 8 (A) Batch uptake kinetics of purified hFGF2 on SP Sepharose FF Pore diffusion model for rectangular isotherms used for fitting data ( Eq (3) ) Breakthrough curve experiments with hFGF2 on SP Sepharose FF (CV = 0.2 mL) (B) Filled circles ( ●) represent residence time of 1 min Residence time of 3 min is represented as open circles (o) Predictions via analytical solution for pore diffusion ( Eq (7) ) are shown as dashed lines
Fig 9 Solid diffusion coefficients of purified hFGF2 in solutions adapted to different dynamic viscosities by adding glycerol (A) Solid diffusion coefficients of purified hFGF2
adapted to different dynamic viscosities against maximum binding capacities (B) Addition of 23 mg/mL BSA represented as (x) Dashed line represents D s determined for purified hFGF2 without additives and dotted line represents D s of hFGF2 in clarified homogenate Analytical solution for solid diffusion ( Eq (9) ) was used for fitting the experimental data (C) Dynamic viscosities of different glycerol concentrations ( ), correlation of dynamic viscosity and clarified homogenates with varying cell dry mass concentrations ( ●) as well as dynamic viscosity of purified hFGF2 represented as dotted line
were lower, as was the case for hFGF2 in the clarified homogenate,
model predictions were accurate ( Fig.5B) It can be assumed that
the rearrangement effect, if occurring at all, had little impact on
the shape of the BTC
Finally, we addressed the question of the cause for the reduced
mass transfer rate in the clarified homogenate We identified that
the lower driving force, which is a major factor for solid diffusion-
based mass transfer, was due to the low capacity The question re-
mains: why is the capacity so much lower compared to pure hFGF2
since Capto S was very selective for hFGF2 and the eluate frac-
tion was highly pure? This outcome indicates very little competi-
tion from host cell proteins for binding, although the homogenate
contained around 25 mg/mL of proteins To further prove the as-
sumption that host cell proteins do not interfere with the mass
transfer rates, we spiked 23 mg BSA/mL to the pure hFGF2 so-
lution and performed batch uptake experiments on Capto S The
uptake data were then fitted with the solid diffusion model As
can be seen in Fig 9, D s values and q max were identical to pure
hFGF2 without BSA spiked This suggests that other impurities are
responsible for the reduced mass transfer E coli homogenates con-
tain numerous low molecular weight compounds in high concen-
trations which potentially were competing with hFGF2 adsorption
On one hand the identification and further quantification of these
compounds is very difficult On the other hand, even if the nature
and concentration of these compounds were known, the resulting
multi-component adsorption system would be extremely complex
Furthermore, we investigated if the reduced D s was caused by a
general decrease of diffusivity due to viscosity effects Therefore,
specific amounts of glycerol were added to pure hFGF2 solutions
to emulate the dynamic viscosity of hFGF2 in clarified homogenate Batch uptake kinetics were recorded and q max and D swere deter- mined Fig.9shows the decrease of the solid diffusivity with in- creasing glycerol concentration A glycerol addition of 44%, which corresponded to the viscosity of clarified homogenate, yielded the same D s value as obtained for hFGF2 in the clarified homogenate Also, in this case, q max was not affected Overall, the glycerol spik- ing experiments supported the explanation that the reduction of
D 0is affected by higher viscosity In combination with the reduced driving force caused by lower q max , the overall lower diffusivity of
hFGF2 in the homogenate was plausible Moreover, process design for a preparative capture step can easily be performed with either model using the respective diffusion coefficients
5 Conclusion
In general, mass transfer on the grafted chromatography medium Capto S is very fast due to enhanced solid diffusion trans- port We have shown that mass transfer of a protein from a clari- fied homogenate is significantly reduced when compared to a feed solution of pure protein The reduction of the transport rate from the clarified homogenate was caused by two factors: (1) higher viscosity intrinsically results in lower diffusion coefficients and (2) the solid diffusion driving force is much reduced at lower max- imum binding capacities, which was shown to be the case for hFGF2 in the clarified homogenate From an engineering point of view, both diffusion models, the pore diffusion with an apparent
Trang 10pore diffusion coefficient as well the solid diffusion model, were
able to describe the experimental data The models can easily be
applied for process design calculations and scale-up It remains un-
clear which components are responsible for the reduced binding
capacity, as other host cell proteins do not adsorb This factor will
be investigated in further studies
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared to
influence the work reported in this paper
CRediT authorship contribution statement
Markus C Berg: Conceptualization, Methodology, Formal anal-
ysis, Investigation, Data curation, Writing – original draft, Visu-
alization Jürgen Beck: Investigation, Data curation Alex Karner:
Methodology, Formal analysis, Investigation Kerstin Holzer:
Methodology, Formal analysis, Investigation Astrid Dürauer: Con-
ceptualization, Investigation, Methodology, Writing – review &
editing, Writing – original draft Rainer Hahn: Conceptualization,
Investigation, Methodology, Writing – original draft
Acknowledgment
The COMET center: acib: Next Generation Bioproduction is
funded by BMK, BMDW, SFG, Standortagentur Tirol, Government
of Lower Austria und Vienna Business Agency in the framework
of COMET Competence Centers for Excellent Technologies The
COMET-Funding Program is managed by the Austrian Research Pro-
motion Agency FFG We thank the colleagues M Martinetz, N
Hammerschmidt, S Krahulec, M Graf and C Brocard of our com-
pany partner BI RCV for their scientific input in fruitful continuous
discussions Furthermore, the authors thank Dr Monika Debreceny
from the Imaging Center and the Doctoral School “BioProEng” for
their support
Supplementary materials
Supplementary material associated with this article can be
found, in the online version, at doi: 10.1016/j.chroma.2022.463264
References
[1] G Carta, A Jungbauer, Protein chromatography, 2010 10.1002/9783527630158
[2] R Hahn, Methods for characterization of biochromatography media, J Sep Sci
(2012), doi: 10.10 02/jssc.20120 0770
[3] F.T Sarfert, M.R Etzel, Mass transfer limitations in protein separations us-
ing ion-exchange membranes, J Chromatogr A 764 (1997) 3–20, doi: 10.1016/
S0 021-9673(96)0 0894-1
[4] P.M Armenante, D.J Kirwan, Mass transfer to microparticles in agitated
systems, Chem Eng Sci 44 (1989) 2781–2796, doi: 10.1016/0 0 09-2509(89)
85088-2
[5] J Gutenwik, B Nilsson, A Axelsson, Effect of hindered diffusion on the ad-
sorption of proteins in agarose gel using a pore model, J Chromatogr A 1048
(2004) 161–172, doi: 10.1016/j.chroma.2004.07.052
[6] W.M Saltzman, R Langer, Transport rates of proteins in porous materi-
als with known microgeometry, Biophys J 55 (1989) 163–171, doi: 10.1016/
S0 0 06- 3495(89)82788- 2
[7] A Matlschweiger, P Fuks, G Carta, R Hahn, Hindered diffusion of proteins in
mixture adsorption on porous anion exchangers and impact on flow-through
purification of large proteins, J Chromatogr A 1585 (2019) 121–130, doi: 10
1016/j.chroma.2018.11.060
[8] Y Tao, G Carta, G Ferreira, D Robbins, Adsorption of deamidated antibody
variants on macroporous and dextran-grafted cation exchangers: II Adsorption
kinetics, J Chromatogr A 1218 (2011) 1530–1537, doi: 10.1016/j.chroma.2011.01
050
[9] A.M Lenhoff, Protein adsorption and transport in polymer-functionalized ion-
exchangers, J Chromatogr A 1218 (2011) 8748–8759, doi: 10.1016/j.chroma
2011.06.061
[10] P Pereira Aguilar, T.A Schneider, V Wetter, D Maresch, W.L Ling, A Tover,
P Steppert, A Jungbauer, Polymer-grafted chromatography media for the pu- rification of enveloped virus-like particles, exemplified with HIV-1 gag VLP, Vaccine 37 (2019) 7070–7080, doi: 10.1016/j.vaccine.2019.07.001
[11] R Liu, Q Shi, Protein retention in dextran-grafted cation exchange chromatog- raphy: the influence of pHs, counterions and polymer structure, Chin J Chem Eng 28 (2020) 1904–1910, doi: 10.1016/j.cjche.2020.01.006
[12] 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, doi: 10.1016/j.chroma.2009.09.014 [13] L Yu, Y Sun, Recent advances in protein chromatography with polymer-grafted media, J Chromatogr A 1638 (2021), doi: 10.1016/j.chroma.2020.461865 [14] 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, doi: 10.1016/j.chroma.2011.08.039
[15] L.L Yu, S.P Tao, X.Y Dong, Y Sun, Protein adsorption to poly(ethylenimine)- modified Sepharose FF: I A critical ionic capacity for drastically enhanced ca- pacity and uptake kinetics, J Chromatogr A 1305 (2013) 76–84, doi: 10.1016/j chroma.2013.07.014
[16] S.H.S Koshari, N.J Wagner, A.M Lenhoff, Effects of resin architecture and pro- tein size on nanoscale protein distribution in ion-exchange media, Langmuir
34 (2018) 673–684, doi: 10.1021/acs.langmuir.7b03289 [17] R Hahn, P Bauerhansl, K Shimahara, C Wizniewski, A Tscheliessnig, A Jung- bauer, Comparison of protein a affinity sorbents: II Mass transfer properties, J Chromatogr A 1093 (2005) 98–110, doi: 10.1016/j.chroma.2005.07.050 [18] J.T McCue, G Kemp, D Low, I Quiñones-García, Evaluation of protein-a chromatography media, J Chromatogr A 989 (2003) 139–153, doi: 10.1016/ S0 021-9673(03)0 0 0 05-0
[19] L.K Kimerer, B Niu, T.M Pabst, W Zhai, A.K Hunter, G Carta, Chromato- graphic and adsorptive behavior of a bivalent bispecific antibody and asso- ciated fragments, J Chromatogr A 1648 (2021) 462181, doi: 10.1016/j.chroma 2021.462181
[20] O Khanal, V Kumar, F Schlegel, A.M Lenhoff, Estimating and leveraging pro- tein diffusion on ion-exchange resin surfaces, Proc Natl Acad Sci USA 117 (2020) 7004–7010, doi: 10.1073/pnas.1921499117
[21] J Liang, G Fieg, Q.H Shi, Y Sun, Single and binary adsorption of proteins on ion-exchange adsorbent: the effectiveness of isothermal models, J Sep Sci 35 (2012) 2162–2173, doi: 10.10 02/jssc.20120 0101
[22] M Kołodziej, D.G Sauer, J Beck, W.K Marek, R Hahn, A Jungbauer, A Dürauer,
W Pi ˛atkowski, D Antos, Scale up of a chromatographic capture step for a clar- ified bacterial homogenate – influence of mass transport limitation and com- petitive adsorption of impurities, J Chromatogr A 1618 (2020), doi: 10.1016/j chroma.2020.460856
[23] W.K Marek, D Sauer, A Dürauer, A Jungbauer, W Pi ˛atkowski, D Antos, Pre- diction tool for loading, isocratic elution, gradient elution and scaling up of ion exchange chromatography of proteins, J Chromatogr A 1566 (2018) 89–
101, doi: 10.1016/j.chroma.2018.06.057 [24] A.J Link, K Robison, G.M Church, Comparing the predicted and observed properties of proteins encoded in the genome of Escherichia coli K-12, Elec- trophoresis 18 (1997) 1259–1313, doi: 10.1002/elps.1150180807
[25] D.G Sauer, M Mosor, A.C Frank, F Weiß, A Christler, N Walch, A Jungbauer,
A Dürauer, A two-step process for capture and purification of human basic fi- broblast growth factor from E coli homogenate: yield versus endotoxin clear- ance, Protein Expr Purif 153 (2019) 70–82, doi: 10.1016/j.pep.2018.08.009 [26] C.G Decker, Y Wang, S.J Paluck, Fibroblast growth factor 2 dimer with super- agonist in vitro acitivity improves granulation tissue formation during wound healing, Biomaterials 81 (2016) 157–168, doi: 10.1016/j.biomaterials.2015.12 003.Fibroblast
[27] B Balasundaram, D Nesbeth, J.M Ward, E Keshavarz-Moore, D.G Bracewell, Step change in the efficiency of centrifugation through cell engineering: co- expression of Staphylococcal nuclease to reduce the viscosity of the bioprocess feedstock, Biotechnol Bioeng 104 (2009) 134–142, doi: 10.1002/bit.22369 [28] C.C Miller, The Stokes-Einstein law for diffusion in solution, Proceedings of the Royal Society of London Series A, Containing Papers of a Mathematical and Physical Character 106 (740) (1924) 724–749, doi: 10.1098/rspa.1924.0100 [29] A.K Hunter, G Carta, Protein adsorption on novel acrylamido-based polymeric ion exchangers II Adsorption rates and column behavior, J Chromatogr A 897 (20 0 0) 81–97, doi: 10.1016/S0 021-9673(0 0)0 0865-7
[30] W.K Teo, D.M Ruthven, Adsorption of water from aqueous ethanol, Ind Eng Chem Process Des 25 (1986) 17–21
[31] T.W Weber, R.K Chakravorti, Pore and solid diffusion models for fixed-bed ad- sorbers, AIChE J 20 (1974) 228–238, doi: 10.10 02/aic.69020 0204
[32] R.H Perry, D.W Green, J.O Maloney, Perry’s Chemical Engineers’ Handbook, 7th edition, McGraw Hill, 1997
[33] H Yoshida, T Kataoka, D.M Ruthven, Analytical solution of the breakthrough curve for rectangular isotherm systems, Chem Eng Sci 39 (1984) 1489–1497, doi: 10.1016/0 0 09-2509(84)80 0 07-X
[34] R Hahn, A Tscheliessnig, A Zöchling, A Jungbauer, Shallow bed adsorption: theoretical background and applications, Chem Eng Technol 28 (2005) 1241–
1251, doi: 10.10 02/ceat.20 050 0221 [35] J Beck, E von Lieres, N Zaghi, S Leweke, G Carta, R Hahn, Patterns of pro- tein adsorption in ion-exchange particles and columns: evolution of protein concentration profiles during load, hold, and wash steps predicted for pore and solid diffusion mechanisms, J Chromatogr A 1653 (2021), doi: 10.1016/j chroma.2021.462412
... gels The separated proteins were fixed with a solution con- taining 50% ethanol and 10% glacial acetic acid Protein bands were stained with Coomassie blue G 250 for 30 Destaining was performed... by using a PD-10 desalting column labeling dye was added into a mL vial in a 1:3 dye: protein molar ratio and in- cubated for h wrapped in aluminum foil Unbound dye was then removed by using a... Adsorption kineticsTo gain an insight into the mass transfer mechanism, investi-
gation by confocal laser scanning microscopy (CLSM) of the intra-
particle mass transport