вход по аккаунту


Fractionation of Proteins Using Ultrafiltration Developments and Challenges.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process. 13(1/2), pp. 121-136, 2005.
Fractionation of Proteins Using
Ultrafiltration: Developments and
Y. Wan’, R. Ghosh2and Z. Cui”
Dept of Engineering Science, University of Oxford, Parks Road,
Oxford OX1 3PJ, UK
Dept of Chemical Engineering, McMaster University, Hamilton,
Ontario L8S 4L 7, Canada
In recent years ultrafiltration has attracted significant interest as a potential
technology for fractionating proteins. Although traditionally thought to be suitable
for size-based separation with high-throughput but relatively low-resolution, recent
research has demonstrated that it is possible to signijicantly increase the selectivity in
protein fractionation using ultrafiltration while still maintaining its inherent highthroughput. This paper reviews recent developments in the area of protein
pactionation using ultrafiltration, with focus on strategies employed to improve
selectivity of separation. Several new techniques, technological developments, and
their applications for selective ultrajiltration of proteins are also briejly introduced.
With the rapid growth in the field of biotechnology, the need for reducing production
costs coupled with projected increases in batch size requires development of protein
purification processes with high-throughput as well as hgh-resolution. Ultrafiltration
has the potential to meet both these challenges. It is very well suited to the processing
of biological products since it operates at relatively “mild” conditions and involves no
phase change or addition of chemicals. Thls minimizes the extent of denaturation,
deactivation, andor degradation of the highly labile biological molecules. A
significant advantage of ultrafiltration is the high throughput of product, whch is
attractive in biotechnological processes. Moreover, it is much less expensive than
other separation and purification methods, e.g. chromatography, and is easy to operate
and to scale up. All of these make it very promising for biotechnological applications.
* Author for correspondence (‘
Y. Wan, R. Ghosh andZ. Cui
The most common applications of ultrafiltration in downstream processing are
protein concentration (i.e. solvent removal), buffer exchange and desalting, virus
removal and clarification. In recent years, there has been growing interest in the use
of ultrafiltration for fractionating complex protein mixtures, commonly encountered
in many biotechnological, food and biomedical applications. However, protein
fractionation using ultrafiltration remains a technical challenge, and ultrafiltrationbased processes have not been commercialised to any significant extent. In most
cases, for the proteins to be separated having similar sizes and consequently obtaining
sufficient selectivity using available membranes is not a trivial matter.
Early attempts to use membrane technology for fractionation of protein mixtures
were often unsuccessful due to the fairly limited selectivity offered by the membrane
systems at the chosen operating conditions. This led to the myth that effective protein
separation could only be obtained for proteins differing by at least a factor of ten in
terms of molecular weight, a rule of thumb that is still used quite extensively in the
initial phases of development of ultrafiltration-based protein separation processes [ 11.
Nakatsuka and Michaels [2] reviewed the possible reasons for poor separation
obtained in protein ultrafiltration. These include:
1. Broad pore size distributions in the currently available membranes.
2. The effect of concentration polarization (or bulk mass transfer).
3. Protein adsorption within the porous structure of the membrane.
4. The formation of a protein deposit on the upper surface of the membrane.
5. Protein-protein interactions in the bulk solution andor membrane.
The above mentioned reasons also contribute to membrane fouling, another major
problem restricting the applications of membrane separation processes in general. It is
clear that the success of protein fractionation using ultrafiltration relies both on the
minimisation of membrane fouling and on obtaining sufficiently high selectivity of
separation. Since selectivity is a rather crucial factor for protein fractionation, current
research and development efforts have been directed towards selectivity
improvement. Recent work has shown that it is possible to obtain effective separation
of proteins having fairly similar molecular weight by adjusting the solution conditions
such as pH and salt concentration, andor by manipulating the system hydrodynamics
(e.g. by using gas sparging). This paper reviews recent developments in protein
fractionation using ultrafiltration with emphasis on strategies employed for improving
selectivity. These include:
(i) Selection of membrane and membrane surface modification.
(ii) Optimisation of operating conditions such as pH and salt concentration.
(iii) Concentration polarization control.
(iv) Fouling control.
New techniques and technological innovations employed in ultrafiltration to facilitate
process optimisation, and to increase the selectivity of separation, are also briefly
Selective Ultrafiltration of Protein
(i) Esfects of Membrane Type and Membrane Surface Modijication
As protein ultrafiltration is affected by interactions between the protein molecules and
the membrane, selection of the correct membrane is very important for the success of
Fractionation of Proteins Using Ultrafiltration: Developments and Challenges
a protein fractionation process. The selectivity of separation may also be improved by
manipulating the surface characteristics of the membrane.
Ohno et al. [3] obtained relatively good separation of albumin (MW 69,000,
isoelectric point of 4.7) and immunoglobulins (MW 155,000, isoelectric point is
pH = 7) using 100 kDa MWCO polyethersulfone membrane when the solution pH
was adjusted to between 4 and 5, the salt concentration was maintained below 0.2 M
and the total protein concentration was kept below 40 g/l. In contrast, Higuchi et al.
[4]found that no significant albumin-IgG separation could be obtained at pH 5 using
either native or surface modified 200 kDa MWCO polysulfone membranes under any
operating conditions. Zhang and Spencer [5] reported that when formed-in-place
microporous titanium dioxide membranes were used, high separation factors
(i.e. > 100) using dilute solutions of bovine gamma globulin (IgG) and BSA (<1 g/l)
could be obtained at pH near 8.3, without added salts. These experimental results
clearly demonstrate the significance of membrane selection in protein fractionation
Surface modification or pre-treatment of membranes can significantly affects
permeate flux and protein transmission, and these surface modifications may be
realized by using chemical reagents, irradiation and pre-adsorption of protein.
Millesime et al. 16, 71 investigated the fractionation of BSA and lysozyme
(MW 14100) using unmodified and chemically modified ultrafiltration membranes
(with polyvinyl imidazole, PVI). They found with the unmodified membrane that the
selectivity was dependent on ionic strength, while with the modified membrane
selectivity was less sensitive to ionic strength. Rabiller-Baudry et al. [8] studied the
ultrafiltration of mixed protein solutions of lysozyme and lactoferrin with chemically
modified inorganic membranes (MWCO 300 m a , pore radius 14 nm) with either
pyrophosphate (PP, anionic) or ethylenediamine (EDA, cationic) groups. They found
that lugher selectivity was obtained with the modified membranes than with the
original membrane. Ehsani and Nystrom [9] studied the filtration of myoglobin and
BSA using modified and unmodified polysulfone membranes. The modification of the
membrane was carried out by U V irradiation and protein pre-adsorption. Higher
permeate flux and protein transmission was observed with the pretreated membranes.
These researchers also demonstrated that modified polysulfone membrane treated
with either BSA or myoglobin or a mixture of them, showed unproved selectivity in
the fractionation of BSA and myoglobin. Nakatsuka et al. [2] demonstrated that
lugher transmission of myoglobin was obtained when filtering a solution of
myoglobin through polysulfone membrane pretreated with myoglobin than when
using a “clean” membrane. Ghosh and Cui [ 101 studied the fractionation of BSA and
lysozyme using 50 kDa MWCO polysulfone membranes and found that surface
pretreatment by myoglobin adsorption was effective for enhancing the transmission of
lysozyme through the membrane. An increase in transmission between 20-60% was
observed. Moreover, this membrane pretreatment was found to be effective for
maintaining permeate flux and selectivity.
(ii) Effects of p H and Salt Concentration
Ultrafiltration has traditionally been regarded as a pressure-driven, size-based
membrane process. In the past, separation with ultrafiltration was mainly focused on
Y. Wan, R. Ghosh and Z. Cui
using purely steric sieving effects, but in recent years much more effort has been
made to exploit electrostatic interactions among solutes and membrane pores in order
to improve the selectivity of protein separation. Electrostatic interactions are of
particular interest in many membrane systems, since most solutes and membranes are
charged (usually negatively) in aqueous solution around neutral pH. The presence of a
negative charge on the solute or pore wall causes depletion in the concentration of the
negative co-ions in the immediate vicinity of the charged surface and an increase in
the concentration of the positive counterions. This region is referred to as the
electrical double layer (see Figure 1). The characteristic thickness of the double layer
is the Debye length (also called the ion atmosphere radius), which depends on the
concentration of dissolved ions. With increasing salt concentration, the Debye length
decreases (often referred to as charge screening or shielding). On the other hand, the
charge properties of proteins and membrane are strongly affected by pH. Therefore,
the solution pH and salt concentration will significantly affect the electrostatic
interactions, and manipulating these parameters can probably provide opportunities
for improvement in selectivity. Many researchers have demonstrated that pH and salt
concentration indeed play critical roles in protein fractionation since both of these
parameters can significantly affect protein-protein and protein-membrane interactions,
in terms of electrostatic double layer effect [I] and electrostatic repulsion effects [ 101.
-- a
a a
a a
-- 8 a
= a
a a
Figure 1. Schematic of electrostatic double layer.
Fractionation of Proteins Using Ultrafiltration:Developments and Challenges
In the fractionation of albumin and immunoglobulins using ultrafiltration,
profound effects of pH and salt concentration have been reported by many
researchers. Higuchi et al. [4] reported the separation of BSA and y-globulin (IgG)
using surface-modified polysulfone membranes (MWCO 200 kDa) in batch type
ultrafiltration experiments. They found that separation could not be obtained at pH
5.0. However, at pH 8.0, the rejection of BSA was almost loo%, whereas a small
amount of y-globulin could still pass through the membrane (however, the rejection of
y-globulin was always greater than 96%). The authors attributed t h s “reverse
selectivity” (with higher throughput of the larger IgG) to the different interactions of
the BSA and IgG with hydrophobic and hydrophlic segments on the surface of the
modified membrane. Considering the very low transmission of IgG, these membranes
would not be of significant commercial interest.
Zhang and Spencer [5] also utilised this “reverse selectivity” to fractionate BSA
and IgG using formed-in-place microporous titanium dioxide membranes and
achieved high selectivity of separation at pH near 8.3 for dilute solutions of bovine
IgG and BSA ( 4 gA), without added salt. BSA selectivity of about 2 to 6 was
obtained at pH 4.9 with very low salt concentration, with the separation decreasing
significantly as salt concentration increased. Thls pH dependence was explained in
terms of the change in protein mobility within the membrane pores due to the effects
of pH on the electrostatic interactions between the proteins and membranes.
Pujar and Zydney [ 111 showed that the transmission of BSA through a 100 kDa
MWCO polyethersulfone membrane decreased by nearly 2 orders of magnitude as the
NaCl concentration was reduced from 150 mM to 1.5 mM. Saksena and Zydney [ 11
examined the effect of solution pH and ionic strength on the transport behaviour of
BSA and IgG in single component and binary ultrafiltration using OMEGA 100 kDa
and 300 kDa polyethersulfone membranes in a stirred cell. They found that selectivity
varied with pH and salt concentration. When the 100 kDa membrane was used, very
high selectivity (values from 30-50) could be obtained at pH 4.8 with a 1.5 mM NaCl
solution. An increase of the salt concentration to 150 mM NaCl resulted in a sharp
reduction of selectivity; the selectivity dropped to 2.5 due to the higher transmission
of IgG. At pH 7.4 the sieving coefficients for both IgG and BSA were low, and hence
low selectivity was obtained. When the 300 kDa MWCO membrane was used, no
significant separation could be obtained at pH 4.8. However, data at pH 7.4 indicated
that reverse selectivity (selectivity values ranging from 2-3) could be obtained at low
salt concentration due to the decreased sieving coefficient of BSA under these
conditions. The transmission of BSA was also significantly lower than that obtained
in single protein (BSA) ultrafiltration. These were explained in terms of the
electrostatic contributions to both bulk and membrane transport of the protein
molecules as well as protein-protein interactions. Nel et al. [12] studied the effects of
solution properties (IgG concentration, solution pH, ionic strength) on solute and
permeate flux in BSA/IgG ultrafiltration with 100 kDa MWCO cellulosic membranes
in a batch cell. It was observed that the solution properties had significant effects
when BSA was the only component. The presence of IgG could also affect the
transport of BSA. At pH 7.4 and 0.15 M NaCl concentration, the rejection of BSA
was increased even by very low concentrations of IgG. This phenomenon was
explained in terms of protein-protein interaction and steric hindrance.
Y. Wan, R. Ghosh and Z. Cui
Li et al. [13] studied the fractionation of HSA/IgG using 100 kDa tubular PVDF
membranes with or without gas sparging at different operating conditions. In this
study reversed selectivity was utilised to separate human IgG fiom HSA. It was
observed that the selectivity of separation could be dramatically improved by proper
adjustment of solution pH and salt concentration, with or without gas sparging. At pH
4.7 no separation of the two proteins could be observed. At relatively hgher pH, in
the range 7.0 to 8.5, a reverse selectivity of HSNIgG was obtained. The effect of salt
concentration was also examined. Under recommended optimal conditions with gas
sparging, almost complete separation of the two proteins was achieved.
The significance of pH and salt concentration can also be found in many other
protein ultrafiltration systems. Ingham et al. [ 141 reported the separation of BSA and
lysozyme with Amicon PM-30 membrane. A diafiltration mode of operation was
adopted in their study, in which lysozyme was very slowly washed out of the system
while quantitative retention of BSA was assumed. The rate at which the lysozyme
was removed was determined by the salt concentration, and increased with the
increase in salt concentration due to electrostatic interaction between BSA and
lysozyme. Iritani et al. [15] also studied the fractionation of lysozyme and BSA with
30 kDa MWCO polysulfone membrane and found that solution pH and salt
concentration had a significant effect on the transmission of lysozyme. Ghosh and Cui
[lo] examined the effects of pH on fractionation of BSA and lysozyme by
ultrafiltration through a 50 kDa MWCO polysulfone membrane. They found that the
selectivity of lysozyme/BSA separation for dilute mixtures of these two proteins was
very pH dependent and vaned fiom 3.3 at pH 5.2 to 220.0 at pH 8.8.
Nakao et al. [16] discussed the effects of the electrostatic charge of solutes on
transmission using negatively or positively charged ultrafiltration membranes. It was
observed that both negatively and positively charged membranes strongly rejected
charged molecules. In other words, the maximum transmission of each protein was
found at its corresponding isoelectric point (PI). In the separation of cytochrome C
(MW 12,400, PI 9.0) and myoglobin (PI 7.0), it was also shown that rejection and flux
were low near the isoelectric point, in all cases, and high when the protein had the
same charge as the membrane. Therefore, it might be possible to separate proteins
with different PIS by setting the solution’s pH near the PI of the protein desirable in
the permeate and the other protein is likely to be rejected on account of being
charged. This strategy was utilized by Nystrom et al. [17] for the fractionation of
binary protein solutions: myoglobin/BSA, ovalbumidmyoglobin, l a c t o f e d S A ,
conalbumdmyoglobin, myoglobin/lysozyme, and ternary protein solution:
conalbumin-ovalbumin-lysozyme.The proteins used in their study were globular with
molecular weights between 1500 and 80,000 and isoelectric points between 4.6 and
11. Various ionic strengths were used to show the effect of electrostatic shielding and
the significance of manipulating physicochemical parameters (i.e. pH and ionic
strength) in protein ultrafiltration was clearly illustrated. The effect of solution pH on
separation of human serum albumin (MW 68,5OO)/cytochrome C and ribonuclease
(MW 13,700)/ hemoglobin (MW 64,500) mixtures were discussed by Sudareva et al.
[18]. It was observed that the separation attained with the binary mixtures was
dramatically reduced compared to that predicted from single protein experiments.
This was attributed to the formation of complexes arising fiom electrostatic
interactions between the positively and negatively charged proteins in the bulk
Fractionation of Proteins Using Ultrafiltration: Developments and Challenges
solution. Nakatsuka and Michaels [2] studied the fractionation of BSA and myoglobin
(MW 17,500) using 30 kDa MWCO non-sorptive regenerated cellulose membrane
and 30 kDa MWCO sorptive polysulfone membrane. The selectivity was found to be
relatively independent of salt concentration but decreased as the solution pH was
increased from 4.8 to 6.8. van Eijndhoven et al. [19] investigated the separation of
hemoglobin (PI 6.85) and albumin (PI 4.7), proteins with essentially identical
molecular weights but different charge characteristics, using both batch and
continuous filtration systems. A selectivity of more than 70 was obtained with
100 kDa MWCO PES membranes by adjusting the pH around the isoelectric point of
hemoglobin (pH 7) and the salt concentration to 0.0023 M. Lower selectivity was
observed at higher salt concentrations (i.e. 0.016 and 0.1 M).
Balakrishnan and Agarwal [20, 211 discussed the transmission of lysozyme,
myoglobin and ovalbumin, and the fractionation of very dilute simulated mixtures:
o v a l b d l y s o z y m e and myoglobidlysozyme using a vortex flow ultrafiltration unit,
fitted with 100 kDa MWCO polyacrylonitrile membrane. Experimental results
showed that variations in the feed solution properties, particularly the ionic strength
and pH, could dramatically alter the protein transmission profiles. A selectivity of up
to 30 could be obtained for the fractionation of lysozyme/ovalbumin at pH 6.8 with
0.05 M NaC1, and a selectivity of more than 3 was also achieved for the separation of
lysozyme and myoglobin at pH 6.8 with 0.15 M NaCl. Ehsani et al. [22] studied the
fractionation of chicken egg white (CEW) proteins using modified and unmodified
50 kDa MWCO polysulfone membranes. Adjustment of pH and salt concentration
was used to manipulate the electrostatic interactions to prevent lysozyme from going
through the membrane, and thus obtain lysozyme-free ovalbumin. The hghest
enrichment of ovalbumin was obtained at pH 4.8 when no salt was added to the feed.
More recently, Ghosh and Cui [23] examined the separation of lysozyme from CEW
by ultrafiltration with 25 and 50 kDa MWCO polysulfone membranes. With both
types of membranes, pH could significantly affect the permeate flux and protein
transmission; higher permeate flux and lysozyme transmission were observed at
hgher pH. A feed solution pH of 11.O seemed to be best suited for fractionation. A
strategy for purification of lysozyme using two sequential stages of ultrafiltration and
50 and 25 kDa MWCO membranes was further examined. Under recommended
operating conditions, a productivity of 21,874 units m-2s-' of lysozyme with 99.2%
purity was obtained.
The effects of pH, salt concentration and calcium chelation on permeate flux and
protein transmission have been reported by Mehra and Donnelly [24] for whey
protein ultrafiltration through a 100 kDa MWCO membrane. Fractionation of low
molecular weight proteins (e.g. P-lactoglobulin and a-lactalbumin) from high
molecular weight proteins (e.g. BSA, lactoferrin and immunoglobulins) was best
achieved at pH 8.0. Kawasaki et al. [25] reported the fractionation of K-casein
glycomacropeptide from whey protein concentrate. The separation was based on the
pH-dependent apparent molecular weight for h s compound. The role of ionic
interactions between peptides and the membrane for fractionation of tryptic
hyholysates of beta casein was examined by Nau et al. [26]. The factors examined
included the solution pH, salt concentration and electrical potential across the
membrane. Peptides having the same charges as that of the membrane were found to
have lower transmission than expected from a size exclusion model, while peptides
Y. Wan, R. Ghosh and Z. Cui
having opposite charge were found to have higher transmission than expected. Also
the charge properties of both peptides and membrane could be changed by varying the
solution pH and salt concentration. Zydney [271 examined the potential for the
application of membrane systems to whey protein fractionation. It was concluded that
appropriately designed membrane systems, whch effectively exploit the effects of
solution pH and ionic strength, can be used for the commercial-scale production of
purified whey protein products.
(iii) ConcentrationPolarization Control
Concentration polarization (bulk mass transfer) has been shown to have a significant
effect on' permeate flux and selectivity in ultrafiltration processes. Increase in
concentration polarization can result in lowering of permeate flux, and in addition
may promote protein adsorption and membrane fouling. The rule of thumb is that
lower selectivity is normally observed at higher degree of concentration polarization.
Therefore, control and management of concentration polarization is important from
the point of view of both selectivity and permeate flux. As in heat transfer, any
technique which interrupts the formation of a continuous boundary layer on the
transfer surface and enhances mixing of the fluid across the flow channel, is likely to
reduce concentration polarization, and hence enhance permeate flux and selectivity.
Various techniques for improvement of system hydrodynamics have been
investigated, including vortex mixing [20, 21, 28, 291, tube inserts [30-321, pulsatile
flow [33], and gas sparging [34-381. Most of these techniques have been proved to be
effective in enhancing the permeate flux in ultrafiltration. However, relatively fewer
results are available of their effects on the selectivity of protein fractionation.
Increasing the stirring speed in a stirred cell ultrafiltration unit is the simplest way
to enhance mixing. Slater et al. [39] found that in the fractionation of bovine alkaline
phosphate (MW 140,000) and BSA in a stirred-cell UF module with 100 kDa MWCO
regenerated cellulose membrane, the efficiency of fractionation was higher at higher
speeds. In the purification of lysozyme from chicken egg whte (CEW), the permeate
flux was found to increase with increase in stirring speed; while the transmission of
lysozyme was found to decrease with the increase in stimng speed, and to be highest
when the stirring speed was totally stopped [23]. Thls apparently goes against the rule
of thumb, but the rule is applicable only when there is incomplete rejection of the
preferentially retained species. In the system discussed by Ghosh and Cui [23], the
preferentially retained species was almost totally rejected and hence selectivity could
not really be defined. The reduced transmission of lysozyme was explained in terms
of the concentration polarization effect for this protein.
Balakrishnan and Agarwal [20] adopted a vortex flow filtration device fitted with
100 kDa MWCO polyacrylonitrile membrane to investigate the effects of system
hydrodynamics on permeate flux and protein transmission in single component
ultrafiltration systems. Taylor vortices were generated in their experiments to affect
the system hydrodynamics, and lower protein transmission was reported at higher
rotation speeds and higher axial velocities. The effect of permeate flux on
transmission was also found to be significant. It was explained by a combined
concentration polarisation-irreversible thermodynamics model. In a subsequent paper
[2 11, they investigated the fractionation of simulated mixtures of lysozyme/ovalbumin
and lysozyme/myoglobin utilizing Taylor vortices and found that the ultrafiltration
Fractionation of Proteins Using Ultrafiltration: Developments and Challenges
characteristics of dilute protein mixtures were virtually identical to those of their
individual components at low transmembrane pressures and high membrane rotation
speeds. The selectivity of separation was controlled primarily by the extent of
polarisation of the smaller, preferentially transported species (lysozyme). Kaur and
Agarwal [40] studied the effect of Dean vortices on protein transmission in a thin
channel flow module. The model protein used was lysozyme and the membrane was
hydrophilic (cellulose acetate-type with molecular weight cut-off of 30 m a ) .
Experimental results indicated that, as in the case of Taylor vortices, Dean vortices
could be effective in minimizing concentration polarization and enhancing mass
Bellhouse and co-workers [30, 32, 411 carried out a detailed study of the
performance of helical screw thread inserts in tubular membranes. The design
combined predominantly helical flow in which Dean vortices were generated.
Experimental results showed that a dramatic increase in permeate flux (by factors of
6- 10) could be obtained under typical microfiltration, ultrafiltration and nanofiltration
conditions. Helical screw thread flow promoters were first used for the ultrafiltration
of BSA in tubular membrane geometry [41]. The screw thread inserts are simple to
construct and operate well below laminar quasi-steady flow conditions, and this
makes the system ideal for scale-up and handling of shear-sensitive fluids. Very
significant permeate flux enhancement was observed with filtration of 60 g/l BSA
solution. Millward et al. [42] also investigated the enhancement of plasma filtration
using oscillatory flow and flow deflectors, which were used to generate vortex waves.
The vortex waves were found to be effective in improving permeate flux, and a flux
enhancement factor of 3.5 relative to a flat unobstructed channel was reported.
Najarian and Bellhouse [31] examined the effects of pressure pulsation and flow
deflector on fractionation of bovine plasma proteins using a flat-plate ultrafilter
equipped with a ladder-like flow deflector. Application of transmembrane pressure
pulsation improved the selectivity of albumin-globulins by a factor of 3 with GR40PP
(100 D a MWCO polysulfone membrane).
Rodgers and Sparks [43, 441 studied the effects of transmembrane pressure
pulsing on protein ultrafiltration. They found that transmembrane pressure pulsing
was very effective in reducing both the fouling and concentration polarisation in
protein ultrafiltration either using hlly or partially retentive membranes. More
recently, Wilharm and Rodgers [45] studied the effects of varying the pulse amplitude
and duration in pressure pulsation ultrafiltration of BSA and IgG. It was found that
the variation in pulse duration did not significantly affect permeate flux but the mass
flux of BSA could be increased by a factor of three. No data was presented for the
transmission of IgG. It was suggested that transmembrane pressure pulsing might
enhance solute flux by removing lodged solute molecules from the membrane pores.
The creation of gas-liquid two-phase flow in a membrane module by sparging
compressed air has been shown to be an effective way for concentration polarisation
control [34-361. Cui [34] reported an increase in rejection coefficient from 0.80 to
0.93 for 87 kDa dextran when using gas sparging during ultrafiltration with a 100 kDa
MWCO membrane. Bellara et al. [46] employed gas sparging in cross-flow hollow
fibre ultrafiltration of dextran and albumin solutions. Flux enhancements of 20-50%
for dextran and 10-60% for albumin were obtained. While in a downward flow of
feed and gas bubbles with a tubular ultrafiltration membrane, flux increases of up to
Y. Wan, R. Ghosh and 2. Cui
320% were observed [35]. Li et al. [47] studied the effects of gas sparging on
permeate flux and selectivity on the fractionation of HSA and IgG, using 100 kDa
MWCO polyethersulfone membrane in a tangential-flow flat-sheet device and found
that gas sparging could enhance both permeate flux and selectivity. In a subsequent
paper, Li et al. [13] studied the fractionation of HSA and IgG by gas-sparged
ultrafiltration in a tubular membrane module (polyvinylidene fluoride membrane,
100 m a ) . Introduction of gas bubbles greatly increased the selectivity of the
fractionation. Typical values of the reversed selectivity were around 60, about a sixfold increase compared with the un-sparged operation. Under optimal conditions
almost complete separation of the two proteins was obtained. Ghosh et al. [38]
discussed the fractionation of BSA and lysozyme using gas-sparged ultrafiltration,
with a 100 kDa MWCO polysulfone membrane. The effects of gas flow rate, liquid
flow rate and feed concentration on the selectivity of fractionation were examined.
Gas sparging was found to enhance protein fractionation and under suitable solution
conditions, nearly complete separation of BSA and lysozyme was obtained with gassparged ultrafiltration. The permeate flux was also increased by gas sparging. The
mechanism of enhancement was explained in terms of disruption of the concentration
polarisation layer and enhanced mass transfer due to bubble-induced secondary flow.
(iv) Membrane Fouling Control
The effect of membrane fouling on the permeate flux and solute sieving can often be
similar to that associated with concentration polarisation. In spite of having been
extensively studied, membrane fouling remains a poorly understood phenomenon.
One of the major problems involved in developing a fundamental understanding of
membrane fouling is the difficulty in identifying the actual foulant, and in
distinguishing between the symptoms of fouling and the effects of concentration
polarisation and membrane compaction [48]. Fouling control in protein ultrafiltration
processes, as in other fields of application of ultrafiltration, is a key issue for both
research and practical usage.
The rate and extent of membrane fouling depend upon the membrane material
(porosity, hydrophobicity), feed protein concentration, protein type, solution
environment (pH, ionic strength), and operating parameters such as transmembrane
pressure, crossflow velocity, temperature, etc. [49-531. Electrokinetic effects
(membrane and solute charge, pH and ionic strength), in particular, have been shown
to have a significant effect on both fouling and retention [54-571.
Fouling is strongly dependent on protein-protein and protein-membrane
interactions. The severest fouling in protein filtration usually occurs at the isoelectric
point of the protein being filtered and results in pronounced flux decline. At the same
time, many authors found that the protein transmission was hghest when the solution
pH was equal to its isoelectric point [17, 56, 581. Fane et al. [54] studied the
ultrafiltration of protein (BSA) solutions with retentive membranes over a range of
pH values (2-10) and salt concentrations. A flux minimum occurred at pH 5 in the
absence of salts, but in the presence of 0.2 M NaCl the flux increased monotonically
with pH. Maximum protein adsorption also occurred at the isoionic/isoelectric points,
and it was greater in the presence of salts. Conformational changes and charge
properties of the BSA appeared to be the dominant factors determining the permeate
Fractionation of Proteins Using Ultrafiltration: Developments and Challenges
flux. Oppenheim et al. [55] examined the fouling of 100 kDa MWCO polysulfone
membrane by BSA. The BSA solutions were at pH 5 and 7 with 0.05 and 0.15 M
NaCl concentration. It was found that solution properties did not affect the amount of
adsorbed materials, but they did affect the resistance per mass of adsorbed species,
with the maximum being at the isoelectric point of the protein. Huisman et al. [57]
studied the effect of pH on permeate flux, streaming potential and protein
transmission during filtration of BSA solutions using various membranes with
different MWCO. They also found that the lowest flux occurred at the isoelectric
point of BSA. The protein transmission during a fouling experiment changed slowly
from a value related to the ratio of protein size/pore size to the structure of the fouling
Ricq et al. [56] examined fouling and protein transmission using mineral
membranes (Carbosep M1, MWCO 150 kDa, pore radius 10 nm). The model proteins
were P-lactoglobulin and lysozyme. In single protein ultrafiltration, minimum
permeate flux and maximum protein transmission were observed at the isoelectric
point of the protein. In binary protein ultrafiltration, maximum transmission of
P-lactoglobulin was found not at its isoelectric point (pH 5), but at pH values between
7 and 8 when the salt concentration was 0.001 M NaCl; the transmission of lysozyme
was undetectable under those operating conditions, indicating reversed selectivity. At
pH 3.5 and 7.0, selectivity of separation decreased with the increase in ionic strength
when NaCl concentration was more than 0.01 M. Membrane fouling either remained
constant or decreased with increase in ionic strength.
Rogers and Sparks [43] studied the effects of negative transmembrane pressure
pulsing on fouling and selectivity of separation in binary protein ultrafiltration. The
feed solution used in their experiments was 1% albumin (BSA, 69 m a ) and 0.3%
y-globulin (159 kDa) in 0.15 M NaCl, pH 7.4. They found that transmembrane
pressure pulsing could be an effective method to overcome membrane fouling.
However, no obvious increase in selectivity of separation was observed since the
transmission of both albumin and y-globulin increased proportionally.
New Techniques and Their Applications
The literature reviewed in pervious sections reveals that high-resolution fractionation
of proteins using ultrafiltration is feasible only at hlghly fine-tuned conditions [ 1, 4,
19, 22, 38, 58-60]. Parameters that need to be optimised include pH, ionic strength,
permeate flux and system hydrodynamics. Process optimization is time-consuming
and expensive since a large number of experiments need to be carried out, and thls
requires relatively large amounts of proteins. Moreover, conventional modes of
carrying out ultrafiltration, i.e. dead-end and cross-flow, are not ideally suited for
protein fractionation processes [60]. Therefore, the development of efficient process
optimisation techniques must be backed up by efficient modes of operation. Two new
techniques have recently been developed at the University of Oxford, namely pulsed
sample injection ultrafiltration [61] allowing rapid optimisation of operating
conditions, and carrier phase ultrafiltration (CPUF) [60] which facilitates highresolution fractionation of high-value biological macromolecules. Both of these
techniques are based on constant permeate flux ultrafiltration.
Y. Wan,R. Ghosh and Z. Cui
High-performance tangential flow filtration (HPTFF) is a novel techruque for
protein purification [58, 62, 631. In HPTFF, high selectivity is obtained by careful
choice of both membrane and buffer conditions. The latter (i.e. pH and ionic strength)
is optimised to maximise the difference in hydrodynamic volume for proteins to be
separated. Concentration polarisation is exploited to enhance, rather than to limit,
protein separation. Using appropriate dynamic and mass transfer conditions can
control membrane fouling. van Reis et al. [58] and Christy et al. [64] have shown that
it is feasible to separate proteins with similar molecular weights using HPTFF.
(i) Pulsed Sample Injection Technique and Carrier Phase Ultrafiltration (CPUF)
Pulsed sample injection technique facilitates rapid and efficient optimisation of
ultrafiltration [60, 611, and is based on constant flux operation in an appropriate
membrane module which is integrated with a standard liquid chromatography system.
A small volume of protein solution is injected into the system and the instantaneous
protein concentrations in the permeate, bulk feed and on the membrane surface are
determined using a “quasi-steady state” approximation. This techmque can also be
used to provide useful information on protein adsorption on membranes, unsteadystate protein transport through membranes, fouling and concentration polarization
phenomena [65]. The scope of this technique has been further extended by ‘parameter
scanning ultrafiltration’ [66]. The parameter scanning ultrafiltration involves
continuous change of a particular process parameter (e.g. pH, salt concentration)
during pulsed sample ultrafiltration. The time taken and the quantity of materials
required for process optimisation are further reduced, and information about the
effects of a wide range of parameters on protein ultrafiltration can be rapidly
obtained. The aim of parameter scanning ultrafiltration is to rapidly identify and
narrow down operating conditions likely to be suitable for fractionation. Fractionation
of human serum albumin (HSA) and human immunoglobulins (HIgG) with Omega
polyethersulfone membranes was investigated using parameter scanning
ultrafiltration, and the likely optimal conditions were predicted [66, 671. These
experimental results were verified in subsequent fractionation experiments.
CPUF is based on a modification of dead-end UF [60]. A solution corresponding
to the optimised physicochemical conditions (i.e. pH, ionic strength) is employed as a
carrier phase and this is passed through an appropriate ultrafiltration module in whch
the permeate flux and system hydrodynamics can be adjusted independently. The
hgh-resolution separation can thus be carried out under optimised conditions. The
experimental results obtained in the fractionation of lysozyme and myoglobin using
CPUF with a 25 kDa MWCO polysulfone membrane demonstrated that CPUF could
indeed maintain very stable operation over time, thereby resulting in consistently high
selectivity. The usefulness and effectiveness of the combination of pulsed sample
injection technique with CPUF has been demonstrated in the fractionation of human
serum albumin (HSA) and human immunoglobulins (HIgG) [66, 671, bovine serum
albumin (BSA) and monoclonal antibody Campath-1H [66].
(ii) High Performance Tangential Flow Filtration (HPTFF)
High performance tangential flow filtration (HPTFF) is based on conventional
tangential flow filtration (TFF) (also referred to as a cross-flow filtration). However,
Fractionation of Proteins Using UltrajZtration:Developments and Challenges
unlike conventional TFF, which typically operates in the pressure-independent part of
the permeate flux curve, HPTFF is canied out in the pressure-dependent flux regime.
In a HPTFF system, the formation of a pressure gradient along the length of the feed
channel is overcome by operating a co-current permeate stream that maintains
constant transmembrane pressure along the length of the TFF module. Thus both high
selectivity and high mass throughput can be obtained, since the permeate flux and the
local transmembrane pressure can be carehlly controlled at optimum levels. By
exploiting both size and charge mechanisms, HPTFF can be used to separate
monomers from dimers based solely on size differences and proteins of equal size on
the basis of charge differences [68]. van Reis et al. [63] studied the separation of BSA
monomer and oligomers, BSA and IgG using HPTFF. The combination of size and
charges mechanisms can be maximized on an optimization diagram [62]. A limitation
of HPTFF is the possibility of fouling caused by feed streams containing precipitates
Thus far, most of research work in the area of protein fractionation using
ultrafiltration has been done with simulated protein mixtures, with the focus on
understanding basic mechanisms involved. There are relatively fewer papers dealing
with fractionation of real protein mixtures. The wider acceptance of ultrafiltration as a
promising technology in the ever-growing field of biotechnology requires much more
high-quality experimental work in this area.
In protein ultrafiltration, in addition to protein size difference, factors such as
protein-protein and protein-membranes interactions, system hydrodynamics, and mass
transfer play a significant role in determining selectivity. The first step towards
obtaining high selectivity and permeate flux is the selection of a suitable membrane.
If necessary a membrane can be modified using appropriate methods in order to
M h e r enhance the possibility of obtaining h g h selectivity. Environmental conditions
(such as pH and salt concentration) and operating conditions (such as system
hydrodynamics and transmembrane pressure or permeate flux) offer convenient
handles by which to fine-tune a fractionation process. It is possible to significantly
enhance the selectivity for protein fractionation by manipulating electrostatic
interactions through the control of solution pH and salt concentration. Controlling the
salt concentration can modulate the protein-protein and protein-membrane
interactions. In some cases, “reverse selectivity” could be obtained by simply
adjusting the solution environment such that the larger molecule can preferentially
pass through the membrane. Another effective way to improve selectivity is by
controlling the permeate flux and system hydrodynamics to minimize fouling and
exploit the effects of concentration polarisation. It is also important to note that, in
protein fractionation using ultrafiltration, the transmission behavior of proteins may
be altered by the presence of other proteins in solution. Conformational changes and
complex formation can influence protein transmission and must also be taken into
Development of new techniques and technological innovation will lead to wider
acceptance of ultrafiltration as method for high-resolution protein fractionation. It is
Y. Wan, R. Ghosh and Z. Cui
expected that the use of the pulsed sample injection techmque will enable rapid and
inexpensive process optimization. Novel modes of operation such as CPUF and
HPTFF will facilitate high-resolution fractionation of target protein under optimized
conditions. The successful commercialization of these techniques will provide a
competitive purification tool to complement chromatographic processing of proteins.
The authors acknowledge the Engineering and Physical Sciences Research Council
(EPSRC), UK for financial support (GRRO1729).
1. Saksena, S., and Zydney, A.L. 1994. Effect of solution pH and ionic strength on the separation of
albumin from immunoglobulins (IgG) by selective filtration, Biotechnol. Bioeng., 44,960-968.
2. Nakatsuka, S . , and Michaels, A S . 1992. Transport and separation of proteins by ultrafiltration through
sorptive and non-sorptive membranes, J. Membrane Sci., 69,189-21 1.
3. Ohno, S.; Koyama, K., and Fukada, M. 1981. US Patent 4,347,138.
4. Higuchi, A.; Mishima, S., and Nakagawa, T. 1991. Separation of proteins by surface modified
polysulfone membranes, J. Membrane Sci., 57, 175-185.
5. Zhang, L., and Spencer, H.G. 1993. Selective separation of proteins by microfiltration with formed-inplace membranes, Desalinaiion, 90, 137-146.
6. Millesime, L.; Amiel, C., and Chaufer, B. 1994. Ultrafiltration of lysozyme and bovine serum albumin
with polysulfone membranes modified with quatemized polyvinyl imidazole, J. Membrane Sci., 89,
7. Millesime, L.; Dulieu, J., and Chaufer, B. 1996. Fractionation of proteins with modified membranes,
Bioseparation, 6(3), 135-145.
8. Rabiller-Baudry. M.; Chaufer, 8.; Lucas, D., and Michel, F. 2001. Ultratiltration of mixed protein
solutions of lysozyme and lactofemn: role of modified inorganic membranes and ionic strength on the
selectivity,J. Membrane Sci., 184, 137-148.
9. Ehsani, N., and Nystrom, M. 1994. Fractionation of model proteins with modified and unmodified
ultrafiltration membranes, Engineering of Membrane Processes Ii: Environmental Applications,
Ciocco, Italy.
10. Ghosh, R., and Cui, Z.F. 1998. Fractionation of BSA and lysozyme using ultrafiltration: Effect of pH
and membrane surface pretreatment.J. Membrane Sci., 139, 17-28.
11. Pujar, N.S., and Zydney, A.L. 1994. Electrostatic and electrokinetic interactions during protein
transport through narrow pore membranes, Ind. Eng. Chem. Res., 33,2473-2482.
12. Nel, R.G.; Oppennheim, S.F., and Rodgers, V.G.J. 1994. Effects of solution properties on solute and
permeate flux in bovine serum albumin - IgG ultrafiltration,Biotechnol. Prog., 10,539-542.
13. Li, Q.Y.;Cui, Z.F., and Pepper, D.S. 1997b. Fractionation of HSA and IgG by gas sparged
ultrafi I tration, J. Membrane Sci., 136, 18 1-190.
14. Ingham, K.C.; Busby, T.F.; Sahlestrom, Y.,and Castino, F. 1980. In: A.R. Cooper (ed.), Polymer
Science and Technology, Vo1.13 (Ultrafiltration Membranes and Applications), pp. 141-1 58. Plenum
Press, New York.
15. Iritani, E.; Mukai, Y., and Murase, T. 1995. Upward dead-end ultrafiltration ofbinary protein mixtures,
Sep. Sci. Technol.,30(3), 369-382.
16. Nakao, S.; Osada, H.; Kurata,H.; Tsuru, T., and Kimura, S. 1988. Separation of proteins by charged
ultrafiltration membranes, Desalination, 70, 191-205.
17. Nystrom, M.; Aimar, P.; Luque, S.; Kulovaara, M., and Metsamuuronen, S. 1998. Fractionation of
model proteins using their physiochemicalproperties, Colloids Surj: A, 138, 185-205.
18. Sudareva, N.N.; Kurenbin, 0.1.. and Belenkii, B.G. 1992. Increase in efficiency of membrane
fractionation, J. Membrane Sci., 68,263-270.
19. van Eijndhoven, R.H.C.M.; Saksena, S., and Zydney, A.L. 1995. Protein fractionation using
electrostatic interactions in membrane filtration, Biotechnol. Bioeng., 48,406-414.
Fractionation of Proteins Using Ultrafiltration:Developments and Challenges
20. Balakrishnan, M., and Aganval, G.P. 1996a. Protein fractionation in a vortex flow filter. I: Effect of
system hydrodynamics and solution environment on single protein transmission, J. Membrane Sci.,
21. Balakrishnan, M., and Aganval, G.P. 1996b. Protein fractionation in a vortex flow filter. 11: Separation
of simulated mixtures, J. Membrane Sci., 112, 75-84.
22. Ehsani, N.; Parkkinen, S., and Nystrom, M. 1997. Fractionation ofnatural and model egg-white protein
solutions with modified and unmodified polysulfone UF membranes, J. Membrane Sci., 123, 105-119.
23. Ghosh, R., and Cui, Z.F. 2000. Purification of lysozyme using ultrafiltration, Biotechnol. Bioeng., 68,
24. Mehra, R.K., and Donnelly, W.J. 1993. Fractionation of whey protein components through a large
pore-size hydrophilic cellulosic membrane, J. Daity Res., 60,89-97.
25. Kawasaki, Y.; Hawakami, H.; Tanimoto, M.; Dosako, S.; Tomizawa, A.; Kotake, M., and Nakajma, 1.
1993. pH dependent molecular weight changes in K-casein glycomacropeptide and its preparation by
ultrafiltration, Milchwissenschaji - Milk Science International, 48(4), 194-196.
26. Nau, F.; Kerherve, F.L.; Leonil, J., and Daufin, G. 1995. Selective separation of tryptic B-casein
peptides through ultrafiltration membranes - Influence of ionic interactions, Biotechnol. Bioeng., 46,
27. Zydney, A.L. 1998. Protein separation using membrane filtration: New opportunities for whey
fractionation, Inr. Dairy Journal, 8.243-250.
28. Belfort, G.; Pimbley, J.M.; Greiner, A., and Chung, K.Y. 1993. Diagnosis of membrane fouling using a
rotating annular filter: 1. Cell culture media, J. Membrane Sci., 77, 1-22.
29. Bellhouse, B.J.; Sobey, I.J.; Alani, S., and DeBlois, B.M. 1994. Enhanced filtration using flat
membranes and standing vortex waves, Bioseparation, 4, 127-138.
30. Najarian, S.,and Bellhouse, B.J. 1996a. Enhanced microfiltration of bovine blood using a tubular
membrane with a screw-threaded insert and oscillatory flow, J. Membrane Sci., 112,249-261.
31. Najarian, S., and Bellhouse, B.J. 1996b. Effect of liquid pulsation on protein fractionation using
ultrafiltration processes, J. Membrane Sci., 114,245-253.
32. Bellhouse, B.J.; Costigan, G.; Abhinava, K., and Merry, A. 2001. The performance of helical screwthread inserts in tubular membranes, Sep. furif: Technol., 22-23: 89-1 13.
33. Wang, Y.Y.; Howell, J.A.; Field, R.W., and Wu, D.X. 1994. Simulation of cross-flow filtration for
baffled tubular channels and pulsatile flow, J. Membrane Sci., 95,243-258.
34. Cui, Z.F. 1993. In: R. Patterson (ed.), Mechanical Engineering, p.237. London.
35. Cui, Z.F., and Wright, K.I.T.1994. Gas-liquid two-phase flow ultrafiltration of BSA and dextran
solution,J. Membrane Sci., 90, 183-189.
36. Cui, Z.F., and Wright, K.I.T. 1996. Flux enhancement with gas sparging in downwards cross-flow
ultrafiltration: Performance and mechanism, J. Membrane Sci., 123, 109-1 16.
37. Mercier, M.; Fonade, C., and Lafforgue-Delonne, C. 1997. How slug flow can enhance ultrafiltration
flux in mineral tubular membranes, J. Membrane Sci., 128, 103-1 13.
38. Ghosh, R.; Li, Q.Y., and Cui, Z.F. 1998. Fractionation of BSA and lysozyrne using ultrafiltration:
Effect of gas sparging. AlChE J.,44,61-67.
39. Slater, C.S.; Huggins Jr., T.C.; Brookes 111, C.A., and Hollein, H.C. 1986. Purification of alkaline
phosphatse by ultrafiltration in a stirred batch cell, Sep. Sci. Technol., 21,575-590.
40. b u r , J., and Aganval, G.P. 2002. Studies on protein transmission in thin channel flow module: the
role of Dean vortices for improving mass transfer, J. Membrane Sci., 196: 1-1 1 .
41, Millward, H.R.; Bellhouse. B.J., and Walker, G . 1995a. Screw-thread flow promoter - An experimental
study of ultrafiltration and microfiltration performance, J. Membrane Sci., 106,269-279.
42. Millward, H.R.; Bellhouse, B.J.; Sobey, 1.J.. and Lewis, RW.H. 1995b. Enhancement of plasma
filtration using the concept of vortex wave, J. MembraneSci., 100, 121-129.
43. Rodgers, V.G.J., and Sparks, R.E. 1991. Reduction of membrane fouling in the ultrafiltration of binary
protein mixtures, AIChEJ., 37, 1517-1528.
44. Rodgers, V.G.J., and Sparks, R.E. 1992. Effect of transmembrane pressure pulsing on concentration
polarization, J. Membrane Sci.,68, 149-168.
45. Wilhann, C., and Rodgers, V.G.J. 1996. Significance of duration and amplitude in transmembrane
pressure pulsed ultrafiltration of binary protein mixtures, J. Membrane Sci., 121,217-228.
46. Bellara, S.R.; Cui, Z.F., and Pepper, D.S. 1996. Gas sparging to enhance permeate flux in ultrafiltration
using hollow fiber membrane,J. Membrane Sci., 121, 175-184.
47. Li, Q.Y.; Ghosh, R.; Cui, Z.F., and Pepper, D.S. 1997. Enhancement of ultrafiltration with flat sheet
membrane modules by gas sparging, Euromembrane 97, Twente, The Netherlands.
Y. Wan, R. Ghosh and 2. Cui
48. Zeman, L.J., and Zydney, A.L. 1996. Microfiltration and Ultrafiltration: Principles and Applications.
Marcel Dekker, Inc. New York.
49. Matthiasson, E. 1983. The role of macromolecular adsorption in fouling of ultrafiltration membranes,
J. Membrane Sci., 16,23-36.
50. KO, M.K.; Pellegrino, J.J.; Nassimbene, R., and Marko, P. 1993. Characterization of the adsorptionfouling layer using globular proteins on ultrafiltration membranes, J. Membrane Sci., 76, 101-120.
51. Crozes, G.F.; Jacangelo, J.G.; Anselme, C., and k i n e , J.M. 1997. Impact of ultrafiltration operating
conditions on membrane irreversible fouling, J. Membrane Sci., 124,63-76.
52. Babu, P.R., and Gaikar, V.G. 2001. Membrane characteristics as determinant in fouling of UF
membranes, Sep. Purg Technol.,24, 23-34.
53. Jones, K.L., and O’Melia, C.R. 2001. Ultrafiltration of protein and humic substances: effect of solution
chemistry on fouling and flux decline, J. Membrane Sci., 193, 163-173.
54. Fane, A.G.; Fell, C.J.D., and Suki, A. 1983. The effect of pH and ionic environment on the
ultrafiltration of protein solutions with retentive membranes, J. Membrane Sci., 16, 195-210.
55. Oppenheim, S.F.;Philips, C.B., and Rodgers, V.G.J. 1996. Analysis of initial protein surface coverage
on fouled ultrafiltration membranes, J. Colloid Inteface Sci., 184,639-651.
56. Ricq, L.; Narcon, S.; Reggiani, J.C., and Pagetti, J. 1999. Streaming potential and protein
transmission ultrafiltration of single proteins and proteins in mixture: P-lactoglobulin and lysozyme,
J. Membrane Sci., 156,s1-96
57. Huisman, I.H.; Pradanos, P., and Henandez, A. 2000. The effect of protein-protein and proteinmembrane interactions on membrane fouling in ultrafiltration, J. Membrane Sci., 179,79-90.
Charkoudian, J.; Bums, D.B., and Zydney, A.L. 1999. High performance
58. van Reis, R.; Brake, J.M.;
tangential flow filtration using charged membranes. J. Membrane Sci., 159, 133-142.
59. Ghosh, R., and Cui, Z.F. 2000a. Protein purification by ultrafiltration with pretreated membrane, J.
Membrane Sci., 167,47-53.
60. Ghosh, R. 2001. Fractionation of biological macromolecules using camer phase ultrafiltration.
Biotechnol. Bioeng., 74, 1-1 1.
61. Ghosh, R.; Wan, Y.H.; Cui, Z.F., and Hale, G. 2003. Parameter scanning ultrafiltration: Rapid
optirnisation of protein separation, Biotechnol. Bioeng., 81,673-682.
62. van Reis, R., and Saksena, S. 1997. Optimization diagram for membrane separation, J. Membrane Sci.,
129, 19-29.
63. van Reis, R.; Gadam, S.; Frautschy, L.N.; Orlando, S.; Goodrich, E.M.; Saksena, S.; Kuriyel, R.;
Sirnpson, C.M.; Pearl, S., and Zydney, A.L. 1997. High performance tangential flow filtration,
Biotechnol. Bioeng., 56,71-82.
64. Christy, C.; Adam, G.; Kuriyel, R.; Bolton, G., and Seilly, A. 2002. High-performance tangential flow
filtration: a highly selective membrane separation process, Desalination, 144, 133-136.
65. Ghosh, R. 2002. Study of membrane fouling by BSA using pulsed injection technique, J. Membrane
Sci., 195, 115-123.
66. Ghosh, R., and Cui, Z.F. 2000b. Analysis of protein transport and polarisation through membrane
using pulsed sample injection technique, J. Membrane Sci., 175, 75-84.
67. Wan, Y.H.; Ghosh, R., and Cui, Z.F. 2002. High-resolution plasma protein fractionation using
ultrafiltration, Desalination, 144, 301-306.
68. van Reis, R., and Zydney, A.L. 2001. Membrane separations in biotechnology, Current Opinions
Biotechnof., 12,208-21 1.
Без категории
Размер файла
1 139 Кб
development, ultrafiltration, using, challenge, protein, fractionation
Пожаловаться на содержимое документа