FUNCTIONALIZED POROUS MEMBRANES FOR RAPID
PROTEIN PURIFICATION AND CONTROLLED
PROTEIN DIGESTION PRIOR TO
MASS SPECTROMETRY ANALYSIS
By
Jinlan Dong

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Chemistry-Doctor of Philosophy
2015

i

ABSTRACT
FUNCTIONALIZED POROUS MEMBRANES FOR RAPID PROTEIN
PURIFICATION AND CONTROLLED PROTEIN DIGESTION PRIOR
TO MASS SPECTROMETRY ANALYSIS
By
Jinlan Dong
With the increasing demand for proteins in biotechnological research and
applications, high-throughput protein purification is crucial. Convective mass transport
and short radial diffusion distances in membrane pores make membrane absorbers
attractive for rapid protein capture. However, the binding capacity of membranes is
modest compared to bead-based techniques. Layer-by-layer polyelectrolyte adsorption
should provide a simple way to functionalize membranes for extensive protein capture.
Preliminary studies showed that poly(allylamine) (PAH)/poly(acrylic acid) (PAA) films
adsorbed at pH 3 on Au-coated wafers swell up to 500% in pH 7 buffer. At neutral pH,
(PAH/PAA)n films are highly ionized and capture the equivalent of many monolayers of
lysozyme without causing changes in film roughness.
Nylon membranes containing PAA/polyethyleneimine (PEI)/PAA films adsorbed at
pH 3 provide a porous platform with a high density of –COOH groups that are available
for

further

derivatization

with

NTA-Ni2+ complexes.

When

modified

with

PAA/PEI/PAA-NTA-Ni2+, membranes enable rapid isolation of His-tagged proteins
from cell extracts with > 95% purity and recovery. Amazingly, the whole protein
purification takes less than 30 min.

A new membrane device developed by our

commercial partners should require only 5 min to purify His-tagged protein from cell
extracts. Finally, a polymer containing NTA as part of its side chain will avoid
derivatization steps and greatly simplify the membrane-modification process.
ii

Membrane adsorbers can also include immobilized proteins as their affinity agents.
However, a membrane containing Concanavalin A (Con A, a protein that binds specific
glycoproteins) did not provide a high glycoprotein binding capacity.

Nevertheless, an

antibody-containing membrane has the potential for rapid immunoaffinity purification
of specific antigens with a binding capacity significantly higher than that of commercial
resins.
Thin membranes also present an excellent platform for creating enzyme reactors.
Covalent binding to PAA-modified membranes immobilizes about 35 mg of trypsin
per cm3 of membrane pore, which is a 2800-fold increase in enzyme concentration
relative to in-solution digestion.

This high enzyme concentration also yields a 75-fold

increase in digestion rate compared to typical in-solution methods. In addition, the
covalent immobilization successfully prevents enzyme leaching and stabilizes the
enzyme in 6 M urea or at pH 5, which aids digestion of proteins that require
denaturation. Even after 6 months of storage at room temperature, the membrane can
still digest protein in ms residence times. Moreover, thin (110 μm) membrane reactors
afford fine control over proteolysis time and, hence, peptide lengths to enhance protein
analysis. For β-casein digestion, msec digestion gives a large peptide that carries all five
phosphorylation sites. Thus, such large peptides should enable correlation of
post-translational protein modification events such as phosphorylation.

iii

I dedicate this dissertation to my parents, Xinmin Dong and Junzhi
Wu, and my brother, Yueqiang Dong, for their love and support.

iv

ACKNOWLEDGEMENTS

I would first like to express my sincerest appreciation to my advisor, Dr. Merlin
Bruening, for his continuous guidance and support. His patience and trust encouraged me
to overcome many challenges and pursue my research. In addition to what I learned from
him on how to be a better scientist, he also taught me how to present my science and
myself to the audience in different circumstances. It was an honor to do my PhD research
with him and I am proud to be an alumnus of the Bruening group.
I also want to thank my second reader, Dr. Gavin Reid, for his support in my mass
spectrometry studies. I learned mass spectrometry not only from his classes, but also
from his suggestions on my research.
I am very grateful for my committee members: Dr. Dana Spence, Dr. James Geiger
and Dr. Gregory Baker. I am indebted to Dr. Geiger and his student Stacy Hovde for
their generous support in my His-tagged protein purification studies and to Dr. Spence
for his insightful guidance in my research. Last but not least, I wish to thank Dr. Baker
for the valuable training I received on how to grow polymer brushes.
I also would like to give my deepest thanks to my fellow group members, including
those who have graduated from the lab before me and those who are still working, for
their support and assistance with my research and studies, especially the tough times. I
will always remember all the work and fun times we had together. Similarly, thanks to
my friends all over the world for your continuing friendship and everything we have
experienced together.
I gratefully acknowledge Dr. Kathy Severin for her guidance in metal-ion analysis
v

using AA and ICP-OES, and during my teaching of CEM435, Dr. Ruth Smith for the
training on ATR-IR, and Dr. Stanley Flegler, Carol Flegler and Abigail Vanderberg for
their help with my SEM studies.
Above all, it would be impossible for me to become who I am today without the
unconditional love and support of my family. I am fortunate to have my wonderful
parents and brother in my life. Even though I spent most of the past several years on the
other side of the Earth, love always keeps us together!

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... …..x
LIST OF FIGURES ........................................................................................ xi
KEY TO ABBREVIATIONS ....................................................................... xv
Chapter 1. Introduction and recent progress ............................................. 1
1.1.

Membrane modification with polymeric thin films ........................................ 1
1.1.1. Advantages of membranes ....................................................................... 1
1.1.2. Modification of membranes with functional polymeric thin films .............. 4
1.2. Protein purification using affinity membranes ....................................................... 6
1.2.1. His-tagged protein purification .................................................................... 6
1.2.2. Lectin purification using glycosylated membranes ...................................... 8
1.2.3. Molecularly imprinted polymer membranes .............................................. 11
1.3. Protein purification via ion-exchange and hydrophobic interactions .................. 12
1.3.1. Ion-exchange membranes for protein capture ............................................ 13
1.3.2. Hydrophobic and multimodal protein purification membranes ................. 15
1.4. Phosphopeptide enrichment with TiO2-containing membranes ........................... 16
1.5. Protease-containing membranes for controlled protein digestion prior to MS
analysis....................................................................................................................... 19
1.6. Outline of the dissertation .................................................................................... 22
REFERENCES............................................................................................................ 25

Chapter 2. Characterization of polyelectrolyte films used for protein
capture .......................................................................................................... 34
2.1. Introduction .......................................................................................................... 34
2.2. Experimental section ............................................................................................ 36
2.2.1. Materials ..................................................................................................... 36
2.2.2. Preparation of PEMs................................................................................... 37
2.2.3. Characterization of PEMs........................................................................... 37
2.2.4. Determination of the average pKa for PAA ............................................... 38
2.3. Results and discussion ......................................................................................... 38
2.3.1. Formation of (PAH/PAA)n films ................................................................ 38
2.3.2. PAA ionization in (PAH/PAA)n films........................................................ 40
2.3.3. Lysozyme binding to (PAH/PAA)5 films ................................................... 43
2.3.4. Film swelling and morphology changes upon protein sorption ................. 49
2.3.5. Film stability and repetitive binding experiments ...................................... 54
2.4. Conclusions .......................................................................................................... 59
REFERENCES............................................................................................................ 61

Chapter 3. Purification of His-tagged protein using functionalized
vii

membranes ................................................................................................... 66
3.1. Introduction ...............................................................................................................................67
3.2. Experimental section ............................................................................................ 68
3.2.1. Materials ..................................................................................................... 68
3.2.2. Membrane modification ............................................................................. 69
3.2.3 Metal-ion binding and leaching ................................................................... 71
3.2.4 Protein biding .............................................................................................. 72
3.2.5. Purification of HisU from a model protein mixture ................................... 72
3.2.6. Protein isolation from a cell extract............................................................ 73
3.3. Results and discussion ......................................................................................... 74
3.3.1. Permeability of membranes with PAA/PEI/PAA films ............................. 74
3.3.2. Functionalization of membranes with NTA and studies of metal-ion
binding .................................................................................................................. 76
3.3.3. Purification of HisU from model protein mixtures .................................... 80
3.3.4. Purification of His-tagged CSN 8 from cell extracts.................................. 82
3.3.5. Modification of electrospun fibers ............................................................. 83
3.3.5.1. Lysozyme binding .............................................................................. 84
3.3.5.2. Con A binding .................................................................................... 85
3.3.5.3 Challenges with electrospun membranes ............................................ 86
3.3.6. Membrane modification with PNTA .......................................................... 88
3.4. Conclusions .......................................................................................................... 94
REFERENCES............................................................................................................ 95

Chapter 4. Enrichment of N-Glycoproteins Using Membranes Modified
with Lectins .................................................................................................. 99
4.1. Introduction .......................................................................................................... 99
4.2. Experimental section .......................................................................................... 101
4.2.1. Materials ................................................................................................... 101
4.2.2. Initiator adsorption ................................................................................... 101
4.2.2.1. Trichlorosilane initiator .................................................................... 101
4.2.2.2. Macroinitiator ................................................................................... 102
4.2.3. .Preparation of catalyst and monomer solutions....................................... 102
4.2.4. Membrane modification with polymeric films ......................................... 103
4.2.5. Derivatization of polymeric films with Con A ......................................... 103
4.2.5.1. Cu2+-affinity binding ........................................................................ 103
4.2.5.2. Covalent derivatization..................................................................... 104
4.2.5.3. Immobilization of protected Con A .................................................. 105
4.2.6. Measurement of protein binding capacities .............................................. 105
4.3. Results and Discussion....................................................................................... 106
4.3.1. Membrane modification with poly(MES) films ....................................... 106
4.3.2. Con A immobilization via Cu2+-affinity and subsequent glycoprotein
binding ................................................................................................................ 110
4.3.3 Con A immobilization via covalent interactions and subsequent
glycoprotein binding ........................................................................................... 116
4.4. Conclusion ......................................................................................................... 118
viii

REFERENCES.......................................................................................................... 119

Chapter 5. Rapid and Controlled Protein Digestion in Porous
Membrane Reactors Containing Covalently Immobilized Trypsin ..... 124
5.1. Introduction ........................................................................................................ 124
5.2. Experimental section .......................................................................................... 126
5.2.1. Materials ................................................................................................... 126
5.2.2. Covalent immobilization of trypsin in membranes .................................. 127
5.2.3. Quantitation of trypsin immobilization .................................................... 127
5.2.4. Quantification of enzyme activity ............................................................ 128
5.2.4.1. In-solution enzyme activity ............................................................. 128
5.2.4.2. In-membrane enzyme activity ......................................................... 128
5.2.5. Protein digestion ....................................................................................... 129
5.2.5.1. In-solution digestion ......................................................................... 129
5.2.5.2. Membrane digestion ......................................................................... 129
5.2.6. Desalting using a C18 Sep-Pak cartridge ................................................. 130
5.2.7. Mass spectrometry and data analysis ....................................................... 131
5.3 Results and discussion ........................................................................................ 131
5.3.1. Covalent immobilization of trypsin in membranes .................................. 132
5.3.2. Proteolytic activity of Trypsin Membranes .............................................. 133
5.3.3. Controlled β-casein digestion as a function of residence time ................. 136
5.3.4. Apomyoglobin digestion with high concentrations of urea and variable
pH ....................................................................................................................... 140
5.3.5. Stability of the trypsin membrane reactor ................................................ 143
5.4. Conclusion ......................................................................................................... 144
REFERENCES.......................................................................................................... 145

Chapter 6. Summary and future work .................................................... 151
6.1. Research summary ............................................................................................. 151
6.2. Future work ........................................................................................................ 153
6.2.1. His-tagged protein purification using PNTA-containing membranes ...... 153
6.2.2. Immunoaffinity purification with antibody-containing membranes ........ 153
6.2.3. Probing protein structures using limited protein digestion with the trypsin
membrane reactor ............................................................................................... 157
6.3. Potential impact.................................................................................................. 157
REFERENCES.......................................................................................................... 159

ix

LIST OF TABLES

Table 2.1. Efficiency of lysozyme elution from (PAH/PAA)5 films as a function of the
eluent KSCN concentration and elution times. .................................................................. 47
Table 3.1. Water or buffer permeabilities of bare and PAA/PEI/PAA-modified nylon
membranes in sequential measurements. ........................................................................... 76
Table 3.2. Cu2+ or Ni2+ leaching from membranes modified with PAA/PEI/PAA-NTA
and PAA/PEI/PAA. ........................................................................................................... 80
Table 3.3. Lysozyme binding capacities of electrospun nylon fiber membranes with
and without PAA/PEI/PAA films and of a PAA/PEI/PAA-modified PET support that
was detached from the nylon membrane. .......................................................................... 84
Table 3.4. Water permeabilities of an unmodified electrospun nylon fiber membrane and
of similar membranes after modification with PAA/PEI/PAA and
PAA/PEI/PAA-NTA-Cu2+films......................................................................................... 87
Table 3.5. Cu2+ binding capacities of nylon membranes (1.2 µm pores) modified with
PNTA-containing films deposited under different conditions. .......................................... 90

x

LIST OF FIGURES

Figure 1.1. Transport phenomena involved in bead-based chromatography and
membraneadsorbers. ............................................................................................................ 2
Figure 1.2. Schematic diagram of growth of polymer brushes from initiators attached to a
membrane surface. ............................................................................................................... 4
Figure 1.3. Schematic presentation for lectin-binding membranes modified with (a)
poly(AG), (b) poly(LAMA) and comblike poly(LAMA) and (c)poly(TMSPA)-DGT
using different pathways…………………………………………………………... ......... 10
Figure 1.4. Protein imprinting via two step grafting using SI-ATRP and UV-initiated
grafting/crosslinking copolymerization on the membrane surfaces. ................................. 12
Figure 1.5. Cartoon of selective phosphopeptide capture in a membrane containing TiO2
nanoparticles... ................................................................................................................... 17
Figure 1.6. Peptide size distributions in ESI-Orbitrap mass spectra of BSA digested using
a pepsin-containing membrane and residence times of 2 (blue) and 0.05 (red) s.............. 21
Figure 2.1. Ellipsometric thicknesses of (PAH/PAA)n films deposited from 20 mM
polyelectrolyte solutions containing 0.5 M NaCl at pH 3 (red squares) or pH 9 (blue
triangles).. .......................................................................................................................... 39
Figure 2.2. Titration of deprotonated PAA with 0.1 M HCl. ............................................ 41
Figure 2.3. Reflectance FTIR spectra of (PAH/PAA)5 films grown from pH 3 and later
immersed in 20 mM phosphate solutions adjusted to different pH values. ....................... 43
Figure 2.4. FTIR spectra of (PAH/PAA)5 (top) and (PAH/PAA)5PAH (bottom) films
before and after sorption of lysozyme.. ............................................................................. 44
Figure 2.5. Lysozyme sorption in (PAH/PAA)5 films as a function of the pH of the 1
mg/mL lysozyme sorption solution ................................................................................... 46
Figure 2.6. Lysozyme sorption in (PAH/PAA)5 multilayers as a function of the
concentration of NaCl in the 1 mg/mL lysozyme sorption solution. ................................. 48
Figure 2.7. Time-evolution of lysozyme sorption in (PAH/PAA)5 multilayers
immersed in a 1 mg/mL lysozyme solution in 20 mM phosphate buffer (pH 7.4). .......... 49

xi

Figure 2.8. AFM images of (PAH/PAA)5 films (a) before and (b) after sorption of
lysozyme. ........................................................................................................................... 52
Figure 2.9. AFM images of (PAH/PAA)5 films (a) before and (b) after adsorption of
lysozyme. ........................................................................................................................... 53
Figure 2.10. Reflectance FTIR spectra of (PAH/PAA)5 films before and after immersion
in a 20 mM phosphate solution adjusted to different pH values.. ..................................... 55
Figure 2.11. Repetitive lysozyme sorption in (PAH/PAA)2 films from a 1 mg/mL
lysozyme sorption solution. ............................................................................................... 57
Figure 2.12. FTIR spectra of (PAH/PAA)2 films before and after immersion in a 20 mM
phosphate solution (pH 7.4). .............................................................................................. 58
Figure 2.13. FTIR spectra of (PAH/PAA)5 films before and after immersion in a 20 mM
phosphate solution adjusted to different pH values. .......................................................... 59
Figure 3.1. Apparatus for membrane modification and protein-binding studies. .............. 70
Figure 3.2. Schematic illustration of PAA/PEI/PAA derivatization using NTA-Ni2+. ..... 78
Figure 3.3. SDS-PAGE analysis (Coomassie blue staining) of HisU purification from a
mixture of BSA, Ovalbumin, Con A, β-Lactoglobulin B, and HisU. ................................ 81
Figure 3.4. SDS-PAGE analysis (Coomassie blue staining) of His-tagged CSN 8 purified
from a cell extract……………………………………………………………………… .. 83
Figure 3.5. Lysozyme breakthrough curves during passage of 0.3 mg/mL of lysozyme
through PET-supported nylon fiber membranes with and without modification by a
PAA/PEI/PAA film and through a PAA/PEI/PAA-modified PET support. ..................... 85
Figure 3.6. Breakthrough curve during passage of 0.3 mg/mL Con A through an
electrospun nylon membrane modified with PAA/PEI/PAA-NTA-Cu2+. ......................... 86
Figure 3.7. SEM images of electrospun nylon fiber membranes: (a) an unmodified
membrane, (b) an unmodified membrane after water flux measurements, (c) a membrane
modified with a PAA/PEI/PAA film and (d) a membrane modified with
PAA/PEI/PAA-NTA-Cu2+. ................................................................................................ 88
Figure 3.8. Membrane modification with PNTA-Cu2+-containing films. ......................... 89
Figure 3.9. Percentage of the bound Cu2+ leached from PSS(PAH/PNTA)2-,
PAA(PAH/PNTA)2- and PAA/PEI/PAA-NTA-modified membranes when passing 5 mL
xii

of buffers sequentially through the membranes. ................................................................ 93
Figure 4.1. Growth of poly(MES) brushes from nylon membranes using (a) a
trichlorosilane initiator and (b) a macroinitiator. ............................................................. 108
Figure 4.2. Lysozyme breakthrough curves during passage of a 0.3 mg/mL lysozyme
solution through nylon membranes modified with poly(MES) brushes grown from
trichlorosilane initiator for 30 or 15 min or macroinitiator for 5 min.............................. 110
Figure 4.3. Conformation of Con A (left) and illustration of its glycoprotein binding sites
…………………. ............................................................................................................ 111
Figure 4.4. ATR-IR spectra of nylon membranes before and after steps for membrane
functionalization. (a) Nylon membrane; (b) Nylon membrane-poly(MES); (c) Nylon
membrane-poly(MES)-NHS; (d) Nylon membrane-poly(MES)-NTA; (e) Nylon
membrane -poly(MES)-NTA-Cu2+. ................................................................................. 113
Figure 4.5. Breakthrough curve during passage of a 0.3 mg/mL Con A solution through a
nylon membrane modified with poly(MES)-NTA-Cu2+. ................................................ 114
Figure 4.6. Breakthrough curves during the passage of 0.3 mg/mL asialofetuin through
membranes modified with membranes with poly(MES)-NTA-Cu2+,
poly(MES)-NTA-Cu2+-Con A or poly(MES)-NTA-Cu2+-Con A/BSA........................... 115
Figure 4.7. Breakthrough curve of asialofetuin for membranes with covalently
immobilized Con A.......................................................................................................... 117
Figure 5.1. Apparatus for in-membrane protein digestion............................................... 130
Figure 5.2. Covalent immobilization of trypsin to PAA adsorbed in a membrane pore. 132
Figure 5.3. Residence time as a function of the BAEE concentration in the permeate for
BAEE digestion during passage through a trypsin-modified membrane. ....................... 135
Figure 5.4. In-solution digestion time as a function of the evolving BAEE concentration
……………… ................................................................................................................. 136
Figure 5.5. SDS-PAGE of intact β-casein (lane 2) and membrane digests of β-casein
obtained with different residence times (lanes 3-5).. ....................................................... 137
Figure 5.6. Manually deconvoluted ESI-Orbitrap mass spectra of 0.1 mg/mL β-casein
digested with residence times of (a) 33 ms and (b) 3.3 s in a trypsin-modified
membrane......................................................................................................................... 139

xiii

Figure 5.7. ESI-Orbitrap mass spectra of 0.1 mg/mL apomyoglobin digested (a)
in-solution in 10 mM, pH 7.6 NH4HCO3 for 30 s; (b) in a trypsin-modified membrane for
3.3 s using 10 mM, NH4HCO3, pH 7.6, no urea; and (c) in a trypsin-modified membrane
for 3.3 s using 10 mM, NH4HCO3, pH 7.6 with 6 M urea. For (c), urea was removed
using a C18 Sep-Pak cartridge ......................................................................................... 141
Figure 5.8. ESI-Orbitrap mass spectra of 0.1 mg/mL apomyoglobin digested (a) in
solution for 30 s at pH 5 and (b) in a trypsin-modified membrane for 3.3 s at pH 5. ..... 143
Figure 5.9. Conversion of BAEE in a continuous flow mode using a membrane with an
exposed 2-cm diameter. ................................................................................................... 144
Figure 6.1. Work flow of immunoaffinity purification of target proteins using either (a)
pre-immobilized antibody or (b) free antibodies that are eventually captured by
antibody-binding resins.................................................................................................... 154
Figure 6.2. Two-step immobilization of IgG to a PAA/PEI/PAA-modified membrane. 156
Figure 6.3. Limited digestion at the most exposed site of a protein during rapid passage
through a protease-containing membrane ........................................................................ 157

xiv

KEY TO ABBREVIATIONS

AAS

Atomic absorption spectroscopy

CAN

Acetonitrile

AEX

Anion-exchange

AG

α-D-allyl glucoside

Aminobutyl NTA Nα,Nα-Bis(carboxymethyl)-L-lysine
Asn

Asparagine

ATR-FTIR

Attenuated total reflectance fourier transform infrared

BA

Na-benzoyl-L-arginine

BAEE

.Nα-benzoyl-L-arginine ethyl ester hydrochloride

BSA

Bovine serum albumin

CSN8

COP 9 signalosome complex subunit 8

Con A

.Concanavalin A

DEAE

.Diethylaminoethanol

DMF

.N,N-dimethylformamide

EDC

1-(3-(Dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride

EDTA

Ethylenediaminetetraacetic acid

FA

Formic acid

FPT

Freeze-pump-thaw

HEMA

.2-hydroxyethyl methacrylate

His-tag

.Histidine-tag

xv

HisU ..
IgG

.Histidine6-tagged Ubiquitin
.Immunoglobulin G

LAMA

2-lactobionamidoethyl methacrylate

LbL

.Layer-by-layer

LCST

Lower critical solution temperature

Me6TREN

Tris(2-(dimethylamino)ethyl)amine

MPA

3-mercaptopropionic acid

MS

Mass spectrometry

nESI

Nanoelectrospary ionization

NHS

N-Hydroxy succinimide

NTA

Nitrilotriacetate

PAA

.Poly(acrylic acid)

PAH

.Poly(allylamine)

PEI

.Polyethyleneimine

PEMs

.Polyelectrolyte multilayers

PET

Polyethylene terephthalate

PNTA

.Poly(2,2-(5-acrylamido-1-carboxypentylazanediyl) diacetic acid)

Poly(MES)

.Poly(2-(methacryloyloxy)ethyl succinate)

PSS

.Poly(sodium styrene sulfonate)

PTMs

Post-translational modifications

PVAm

Poly(vinyl amine)

PVCL

Poly(N-vinylcaprolactam)

xvi

RC

.Regenerated cellulose

RHD3

Root hair defective protein

RT

Room temperature

PAGE

.Polyacrylamide gel electrophoresis

SEM

.Scanning electron microscope

SI-ATRP

.Surface initiated atom-transfer radical polymerization

TMSPA

.Poly(3-(trimethyl) propargyl methacrylate

xvii

This chapter is adapted from an invited review paper that will be published in Annual
Review of Analytical Chemistry (Dong, J. and Bruening, M. L. Annual Review of
Analytical Chemistry 2015, 8, xxx-xxx).

Chapter 1. Introduction and recent progress

This dissertation discusses development of functionalized porous membranes for
protein purification and protein digestion prior to mass spectrometry (MS) analysis.
When modified with polymeric films that swell in water, these membranes provide rapid
Histidine-tagged (His-tagged) protein purification and immunopurification with binding
capacities higher than those of commercial beads. The thin membrane also serves as a
unique

platform

for

enzyme

reactors.

Compared

to

in-solution

digestion,

trypsin-containing membrane reactors increase the rate of protein digestion and afford
control over the length of proteolytic peptides. To put this work in perspective, I first
introduce the advantages and development of membranes modified with polymeric thin
films (section 1.1). Subsequent sections review recent progress in membrane-based
protein purification via affinity interactions (section 1.2) and ion-exchange and
hydrophobic interactions (section 1.3), phosphopeptide enrichment (section 1.4) and
protease membrane reactors (section 1.5). A final section outlines this dissertation.

1.1.
1.1.1.

Membrane modification with polymeric thin films
Advantages of membranes

1

The current therapeutic protein market is nearly 100 billion U.S. dollars per year and
growing rapidly (2). Because downstream purification and recovery account for almost
half of the manufacturing costs for cell-derived pharmaceuticals (3-5), high-throughput
and high-recovery separation techniques are essential in therapeutic protein production.
Convenient, high-throughput purification is also important for fundamental protein
research, especially with proteins that degrade rapidly, e.g. those in protease-containing
solutions (6, 7).

Figure 1.1. Transport phenomena involved in bead-based chromatography and
membrane adsorbers. Blue arrows: convective flow; Red arrows: diffusion. The
figures neglects film diffusion at pore and bead surfaces. For interpretation of the
references to color in this and all other figures, the reader is referred to the electronic

version of this thesis (or dissertation).
Bead-based chromatography is the dominant technique in protein purification.
Despite the success of this technique, slow diffusion of biomacromolecules into porous
beads restricts throughput (8-11). Membrane-based methods should overcome diffusion

2

limitations to enhance rates of protein purification because convective flow passes
through membrane pores and radial diffusion distances are small (Figure 1.1).
Moreover, pressure drops across membranes with thicknesses around 100 μm are low, so
high flow rates are easy to achieve (3). Thus, membrane adsorbers are a promising
alternative to packed columns for rapid protein isolation (10, 12, 13).
Nevertheless, membranes historically suffered from modest binding capacities due
to low specific surface areas relative to porous beads. This chapter highlights recent
membrane functionalization strategies that achieve selective protein binding with
capacities equal to or greater than those of porous beads. Moreover, protein
immobilization techniques enable fabrication of enzyme-containing membranes that
serve as proteolytic reactors with residence times as low as a few milliseconds (ms). In
membrane-based protein digestion, control over residence time enables the formation of
large proteolytic peptides that often facilitate protein characterization by MS. This
review first describes membrane modification by coating pores with polymer brushes or
polyelectrolyte multilayers (PEMs). Subsequent sections explore protein purification
using metal-affinity adsorption, ion-exchange and multimodal interactions. Modification
of membranes with functional polymeric films leads to highly selective protein capture
with high binding capacities. Later sections show that membranes containing TiO2
nanoparticles or proteases enhance MS studies of protein structure and phosphorylation.
Finally, I summarize recent advances and highlight a few challenges for membrane
applications.

3

1.1.2. Modification of membranes with functional polymeric thin films

A wide range of porous materials including alumina (14), silica (15) and many
polymers (16) have served as membranes for protein purification. With all materials,
however, selective binding relies on functionalization of the porous support with affinity
groups. Moreover, high-capacity protein capture requires a 3-dimensional binding
scaffold, i.e. a highly swollen polymer film, on the surface of membrane pores.

Figure 1.2. Schematic diagram of growth of polymer brushes from initiators
attached to a membrane surface.
Polymer brushes, assemblies of polymer chains with one end attached to a substrate
and one end extended from the surface (17), are attractive for binding multilayers of
protein in membrane pores. Brush formation in membranes usually employs
polymerization from surfaces to achieve high polymer-chain areal densities (10, 18, 19).
Thus, the brush synthesis typically includes initiator attachment to the membrane and
polymer growth from these immobilized initiators (Figure 1.2) (10, 19). Among many
techniques for brush growth, surface initiated atom-transfer radical polymerization
(SI-ATRP) is particularly useful. ATRP offers controlled polymerization of a wide range
of monomers under mild conditions and uses readily available catalysts and initiators
(20-23). Several groups modified a variety of membranes using ATRP from immobilized
4

initiators, and the binding capacities of such membranes often exceed 100 mg of protein
per mL of membrane (24-28).
The amount of protein binding in polymer brushes varies with the polymer-chain
areal density. Low chain densities give a small number of binding sites and minimal
protein capture, whereas high densities may result in steric hindrance to protein entry
into the brush. Hence, an intermediate chain areal density will likely lead to the most
protein binding. Chain density depends in part on the density of initiation sites anchored
to membrane surfaces, and anchoring typically occurs through surface functionalities
such as hydroxyl groups (29) and carboxylic acids (30). However, some membranes have
low densities of such surface functional groups. Jain and coworkers overcame this
obstacle using LbL adsorption of a macroinitiator onto a polyethersulfone membrane and
successfully modified the membrane using ATRP from the macroinitiator (24).
Several studies demonstrated control of chain densities in polymer brushes through
dilution of initiators during their immobilization (27, 30-32). Reduced polymer-chain
density decreases film thickness but increases the degree of swelling to allow protein
penetration to binding sites (30, 31). Recently, Bhut and Husson controlled the grafting
density of poly(2-(methacryloyloxy)ethyl)trimethylammonium chloride) in regenerated
cellulose (RC) membranes by varying the ratios of ATRP-initiator and a
non-ATRP-active diluent molecule during initiator immobilization (28). Membranes
containing the highest density of polymer chains gave the highest IgG binding capacity,
indicating minimally restricted IgG access to binding sites in this case. However, the
protein-binding capacity did decrease at high polymer densities in commercial strong

5

cation-exchange resins (33).

1.2. Protein purification using affinity membranes

Among the strategies for protein purification, binding to immobilized affinity
ligands is probably the most powerful (16). Ligands exploited for affinity capture include
antibodies, protein A and protein G (IgG-binding proteins), lectins (a group of
carbohydrate-binding proteins), carbohydrates (for lectin capture), and metal-ion
complexes that capture His-tagged proteins or phosphopeptides (10). This section
focuses on recent development of membranes that capture His-tagged proteins and
lectins and then briefly examines the fabrication of affinity membranes through
molecular imprinting

1.2.1. His-tagged protein purification

His-tags are oligo-histidine (commonly 6 residues) clusters fused to the N- or Ctermini of recombinant proteins. These clusters form chelating complexes with
immobilized Cu(II), Ni(II), Zn(II) or Co(II) during immobilized metal-ion affinity
chromatography (34) and are the most common affinity tags for protein purification (35).
Two reviews summarized early work on creating membranes for His-tagged protein
purification (10, 19), so this section describes recent development of simpler but
effective membrane-modification techniques.
Using SI-ATRP, Jain and coworkers grew poly(2-(methacryloyloxy)ethyl succinate)
(poly(MES)) from nylon membranes previously modified by reaction with a
trichlorosilane initiator (35). The aqueous MES polymerization is compatible with a wide
6

range of polymeric membrane supports. After derivatization of the –COOH groups on
poly(MES) with aminobutyl NTA, formation of Cu2+ and Ni2+ complexes allows protein
adsorption through metal-ion affinity. Modified membranes containing Cu2+ have a
bovine serum albumin (BSA)-binding capacity of 80 mg/mL, whereas membranes with
Ni2+ complexes capture 85 mg of His-tagged ubiquitin per mL of membrane.
Remarkably, these binding capacities are about twice those of commercial metal-ion
affinity resins (36). Moreover, the membrane modified with poly(MES)-NTA-Ni2+
isolated His-tagged cellular retinaldehyde-binding protein from E. coli cell lysates in
only 10 min.
In the above-mentioned procedure, MES polymerization occurs in water but reaction
of the membrane with the trichlorosilane initiator takes place in tetrahydrofuran, which is
incompatible with some polymeric substrates. Anuraj and coworkers developed a
completely aqueous membrane modification pathway that uses a macroinitiator (25).
During 10-min of circulation through the membrane, the aqueous macroinitiator adsorbs
to membrane pores via hydrophobic interactions. Remarkably, after only 5 min of
subsequent MES polymerization from the macroinitiator, modified membranes attain
protein-binding capacities as high as those after a 1-hour polymerization from
membranes modified using a trichlorosilane initiator. Polymerization from the
macroinitiator is a robust modification, as no significant drop in protein-binding capacity
occurred after 5 cycles of protein binding and elution.
To clearly demonstrate the rapid protein binding that takes place in membranes,
Anuraj and coworkers showed that the dynamic lysozyme-binding capacities of

7

poly(MES)-modified membranes do not decrease when increasing the flow through the
membrane 30-fold, and binding occurs in residence times of only 35 ms. After reaction
with aminobutyl NTA and formation of Ni2+ complexes, these membranes selectively
capture His-tagged myo-inositol-1-phosphate synthase from a crude E.coli cell lysate.
This aqueous modification procedure should apply to a wide range of porous membrane
materials.

1.2.2. Lectin purification using glycosylated membranes

Carbohydrate-lectin interactions play important roles in biological processes such as
cell–cell adhesion, signal transduction, tumor progression and host-pathogen recognition
(37-42). Despite the importance and specificity of carbohydrate-protein interactions,
however, the binding affinity is low (Ka=103-104 M-1) (43). Nevertheless, multiple
interactions between lectins and dense sugar ligands overcome low affinity (44). The
mechanism of multivalent interactions is not elucidated completely, but the so called
“cluster glycoside effect” improved binding affinity in several studies (45-47). Thus,
dense carbohydrates in membranes should bind lectins with high affinity, and convective
mass transport will yield rapid capture.
Yang and coworkers recently grafted a sugar-containing monomer, α-D-allyl
glucoside, onto a polypropylene membrane using UV-induced polymerization (Figure
1.3a). The membrane showed the “cluster glycoside effect”, as increasing the ligand
density enhanced the binding capacity for a common lectin, Con A, with selectivity over
contaminating BSA (44). Subsequently, the same group found that comblike

8

glycopolymers in membranes exhibit higher lectin binding than linear polymers. The
comblike structure likely leads to both thicker films and higher densities of
carbohydrates (Figure 1.3b). Membranes with a comblike glycopolymer containing
galactose end groups captured 24 mg/mL of the lectin peanut agglutinin. Considering the
~ 10% porosity of the track-etched polyethylene terephthalate (PET) in that study,
membranes with 50% porosity might capture 100 mg of lectin per mL of membrane (48).
Additionally, the high density of sugar-containing chains provides resistance to
nonspecific adsorption, and BSA did not bind significantly to the membrane.
Using a two-step grafting strategy, Wang and coworkers established a second route
to modify polypropylene membranes with carbohydrates. After pretreatment, they
grafted poly(3-(trimethyl) propargyl methacrylate) (TMSPA) to the membrane using
UV-initiated graft polymerization. Subsequent “Click” coupling by reaction of
2,3,4,6-tetra-o-acetyl-β-D-glucopyranoside

thiol

with

the

alkyne

groups

on

poly(3-(trimethyl) propargyl methacrylate) gave a high density of carbohydrate sites
(Figure 1.3c). This sugar-containing polymer selectively binds Con A and resists capture
of peanut agglutinin (49). These studies clearly show the promise of membrane adsorbers
for lectin purification.

9

Figure 1.3. Schematic presentation for lectin-binding membranes modified with
(a)

poly(AG),

(b)

poly(LAMA)

and

comblike

poly(LAMA)

and

(c)poly(TMSPA)-DGT using different pathways. AG: α-D-allyl glucoside; HEMA:
2-hydroxyethyl methacrylate; LAMA: 2-lactobionamidoethyl methacrylate; and
TMSPA: 3-(trimethylsilyl) propargyl methacrylate.

10

1.2.3. Molecularly imprinted polymer membranes

Molecular imprinting usually creates recognition sites through templating of a
polymer scaffold (50). However, imprinting is challenging with macromolecules because
of their flexible conformation (51). To overcome this limitation, Ulbricht and coworkers
first created a poly(methacrylic acid) scaffold that adsorbs template proteins in
track-etched PET membranes. After protein adsorption, UV-initiated cross-linking using
polyacrylamide led to a cross-linked hydrogel, and the “cavities” left by template elution
served as recognition sites for target proteins (Figure 1.4). Tailoring the polymerization
and crosslinking eventually yielded a lysozyme:cytochrome C selectivity as high as 18
from a 1:1 mixture of these proteins, which have similar sizes and isoelectric points (52).
They also developed an IgG-imprinted membrane to separate IgG from human serum
albumin, a major contaminating protein in antibody downstream processing. The binding
capacity for IgG was 8 mg/mL and the IgG/ human serum albumin selectivity was 5 (53).
Further improvements in binding capacity and selectivity are needed to make separations
with imprinted membranes competitive with current techniques.

11

Figure 1.4. Protein imprinting via two step grafting using SI-ATRP and
UV-initiated grafting/crosslinking copolymerization on the membrane surfaces.

1.3. Protein purification via ion-exchange and hydrophobic interactions
In addition to affinity methods, protein isolation with membrane adsorbers can occur
through less specific ion-exchange and hydrophobic interactions or their combination.
The subsections below describe recent advances in these areas.

12

1.3.1. Ion-exchange membranes for protein capture

Ion-exchange protein separations rely on electrostatic interactions between charged
moieties on membranes, e.g. carboxylate (54), sulfonate (55) and ammonium (5) groups,
and surface charges on biomolecules. Although several companies provide ion-exchange
membranes (16), the materials have relatively low binding capacities (<50 mg of
lysozyme/mL of membrane) (56). Thus, several research groups aimed to improve the
protein-binding capacity of ion-exchange membranes while maintaining reasonable
membrane permeability.
Wang and coworkers prepared a cation-exchange membrane through UV-initiated
grafting of poly (acrylic acid (PAA) and PAA co-polymers on macroporous RC
membranes. These membranes showed dynamic binding capacities (protein adsorbed at
the point of 10% break through) of 90 mg of lysozyme per mL of membrane and 66 mg
of IgG per mL of membrane (54). Subsequently, Chenette and coworkers developed
strong cation-exchange RC membranes via organic solvent-based SI-ATRP of
3-sulfopropyl methacrylate (57). Protein purification with sulfonates avoids potential
challenges with incomplete dissociation of carboxylate groups. The poly(3-sulfopropyl
methacrylate)-modified membranes exhibited a dynamic lysozyme binding capacity as
high as 70 mg/mL at a throughput of 100 column volumes/min and a productivity of 300
mg/mL/min whereas the productivity of a commercial Capto S resin column was only 4
mg/mL/min. Wei and coworkers fabricated a high-capacity strong cation-exchange
membrane using SI-ATRP of poly(sodium 4-styrenesulfonate) on RC membranes (55).
The controlled polymerization reduced the membrane roughness and provided a
13

lysozyme static binding capacity of 138 mg/mL and a binding capacity of 93 mg/mL at
50% break-through.
A number of studies reported new methods for fabrication of anion-exchange (AEX)
membranes (5, 13, 27, 28). Bhut and coworkers used SI-ATRP to create poly
(2-dimethylaminoethyl methacrylate) nanolayers on RC membranes. The static BSA
binding capacity was 66 mg/mL (5). They dramatically improved the performance of
their membranes by optimizing the initiator density and poly (2-dimethylaminoethyl
methacrylate) molecular weight to achieve a BSA binding capacity of 130 mg/mL at a
linear flow velocity >350 cm/h (27). The binding capacity was >35 mg/mL even with
100 mM ionic strength, and BSA isolation from a protein mixture occurred with >98%
recovery and >97 % purity.
Compared to commercial AEX Sartobind○R D membranes and fast flow columns
prepacked with diethylaminoethanol (DEAE) Sepharose beads (HiTrapTM DEAE FF
resins), Bhut’s AEX brush membranes showed superior performance. The breakthrough
curves of the HiTrapTM DEAE FF resin columns changed at different flow rates,
presumably because diffusion of BSA to the capture sites is not fast enough for binding
to keep up with convection. Even worse, binding capacities decreased as flow rate
increased. In contrast, for the Sartobind○R D membrane and the new AEX membranes, the
breakthrough curves and binding capacity were independent of flow rate because
convective flow dominates the mass transport and the kinetics of binding are rapid (13).
Notably, the AEX brush-modified membrane shows a binding capacity 3 times that for
the Sartobind ○R D membrane. Bhut and coworkers also developed a strong AEX

14

membrane containing quaternary amines (Q-type AEX membrane) (28). By controlling
the graft density, the Q-type AEX membrane showed a flow-rate independent dynamic
IgG binding capacity of greater than 130 mg/mL, enabling high throughput protein
purification. Clearly, brush-modified membranes can greatly increase the rate of
ion-exchange protein separations, while enhancing binding capacity 2- to 3-fold
compared to commercial ion-exchange systems.

1.3.2. Hydrophobic and multimodal protein purification membranes

Most protein-binding membranes have hydrophilic surfaces to increase pore wetting
with water and minimize non-specific binding and membrane fouling (58, 59). However,
interactions between hydrophobic surface patches on proteins and hydrophobic ligands
on a solid support can also effect protein separations (16, 60). Hydrophobic
chromatography usually relies on protein loading from solutions with high
concentrations of salts and subsequent elution with solutions of decreasing ionic strength
(61, 62). Recently, Himstedt and Wickramasinghe used SI-ATRP to graft
poly(N-vinylcaprolactam) (PVCL) from RC and create membranes for hydrophobic
chromatography (63). PVCL is a thermo-responsive polymer. At temperatures above the
lower critical solution temperature (LCST) the polymer loses water and collapses, and
the LCST of PVCL decreases with increasing ionic strength and moves below room
temperature in buffers containing 1.8 M (NH4)2SO4. Thus, in solutions with >1.8 M
(NH4)2SO4, PVCL assumes a collapsed conformation, which is favorable for
hydrophobic protein binding. Subsequent elution can occur at lower ionic strength where

15

the LCST is greater than room temperature. PVCL-modified membranes achieved a BSA
binding capacity of 5 mg/mL, which is higher than that with non-modified PVDF
membranes and comparable to Sartorius hydrophobic membranes. The IgG binding
capacity is 21 mg/mL or about twice that of Sartorius membranes (64, 65).
To improve membrane performance, Wang and Husson developed a multimodal
cation-exchange membrane that employs both Coulombic and hydrophobic interactions
(66). They grafted poly(glycidyl methacrylate) from RC membranes and then coupled
4-mercaptobenzoic acid to the polymer side chains. At pH 6.5 without salt, these
membranes bind 150 mg IgG /mL of membrane in static measurements. Remarkably, the
membranes retain 90% of their static binding capacity in the presence of 1 M sodium
citrate, a kosmotropic salt found in elution buffers for protein A chromatography.
However, effective use of these multimodal membranes will require development of an
efficient elution strategy. In addition, protein binding and selectivity in dynamic modes
need more study.

1.4. Phosphopeptide enrichment with TiO2-containing membranes

Adsorption of nanoparticles provides another option for introducing selective binding
sites into membranes. Specifically, Tan and coworkers modified membranes with TiO2
nanoparticles to capture the phosphopeptides in protein digests (Figure 1.5). Protein
phosphorylation on serine, threonine or tyrosine amino acid residues is one of the most
important mechanisms for creating protein diversity in eukaryotic cells, and regulation of
protein phosphorylation affects nearly all aspects of cell life, including signal

16

transduction and metabolism. Abnormal phosphorylation can cause diseases including
cancer (67-69), so successful analysis of phosphorylated proteins is vital for
understanding disease stages and discovering potential pharmaceutical treatment targets
(70, 71).

Figure 1.5. Cartoon of selective phosphopeptide capture in a membrane containing
TiO2 nanoparticles. A small holder attached to a syringe pump enables phosphopeptide
elution in as little as 10 μL of solution. Adapted from reference 76 with permission of the
American Chemical Society.

Because of the low relative abundance of phosphorylated proteins, the detection and
identification of phosphorylation sites is challenging even with recent advances in MS
(72). However, a number of selective phosphopeptide enrichment methods help to
overcome this challenge (19, 73, 74). In particular, phosphopeptide capture can occur

17

through selective adsorption on TiO2 or ZrO2 columns or on MALDI plates containing
TiO2 nanoparticles (75, 76). Membranes are attractive for immobilizing TiO2
nanoparticles that are too small for column formats, and such particles have high surface
areas for phosphopeptide capture and may show different selectivities than larger
particles.
Tan and coworkers modified nylon membranes by sequentially adsorbing
poly(sodium styrene sulfonate) (PSS) and TiO2 nanoparticles in membrane pores (77).
Negatively charged sulfonate groups bind the positively charged nanoparticles, and
energy-dispersive X-ray spectroscopy shows a uniform Ti coverage over the whole
cross-section of the membrane. With 22-mm-diameter membranes adsorption of two
PSS/TiO2 bilayers results in 55 μmol of Ti per membrane, and the immobilized TiO2
binds 540 nmol (12 mg/mL) of phosphoangiotensin when saturated. This capacity is
significantly lower than the protein-binding capacities of polymer brush-modified
membranes because PSS/TiO2 nanoparticle adsorption does not give thick, highly
swollen films. Notably, however, the TiO2-containing membranes show 90%
phosphoangiotensin recoveries even after 10 cycles of loading, washing and elution. In
competition experiments, the phosphopeptide recovery is >70% in a 15-fold excess of
BSA.
Membranes with working areas of only 0.02 cm2 can isolate phosphopeptides from
solution volumes ranging from a few mL to 10’s of μL. For example, Tan and coworkers
selectively captured phosphopeptides from 5 mL of a 1 fmol/μL protein digest and
concentrated these peptides into elution volumes of 10’s of μL. Additionally, they

18

enriched phosphopeptides from a digest of hyperphosphorylated tau (p-tau), a protein
whose abundance correlates well with the severity of neurodegeneration in Alzheimer’s
disease. Before enrichment, MALDI-MS analysis of the digest gave 63 p-tau peptide
signals with intensities >10% of the highest signal, but none of these peptides were
phosphorylated. In contrast, after enrichment using the TiO2-containing membrane,
signals from four phosphopeptides appeared, one of which showed the highest intensity
in the MALDI mass spectrum. The enrichment led to identification of 7 phosphorylation
sites in MS/MS analysis.

1.5. Protease-containing membranes for controlled protein digestion prior to MS
analysis

With methods in hand to immobilize proteins in membranes (see sections 1.2 and
1.3), the Bruening group began employing enzyme-modified membranes as controlled
reactors for protein digestion prior to MS analysis. MS is the most common and powerful
technique for identifying proteins and their posttranslational modifications (78).
Nevertheless,

because

peptides

are

more

amenable

to

MS

and

Liquid

Chromatography-MS analysis than proteins, proteolytic digestion is usually a vital initial
step for MS protein analysis. Conventionally, digestion occurs after mixing a protease
such as trypsin with substrate proteins in solution. However, this strategy requires low
enzyme concentrations to restrict self-digestion of the protease, so digestion times are
generally 1 hour or more (79). To overcome this drawback and facilitate online MS
analysis, several research groups developed reactors with proteases immobilized on solid
supports including monoliths (80, 81), membranes (82, 83), polymeric microfluidic
19

channels (82, 84) and resins (85, 86). With thicknesses of only 10-200 μm, membranes
provide a unique platform for controlling protein digestion.
Xu and coworkers formed membrane reactors by sequential adsorption of PSS and
trypsin in nylon membranes (87). Adsorption occurred at pH 3, where trypsin is
positively charged and inactive, whereas digestion took place in 10 mM NH4HCO3 to
optimize trypsin activity (88). This simple adsorption procedure immobilizes 10 mg of
trypsin per mL of membrane pores, which is 450 times the enzyme concentration during
typical digestion in solution. Moreover, the immobilized trypsin rapidly digests protein.
After digestion of α-casein, gel electrophoresis showed no bands for intact protein even
at membranes residence times as short as 0.79 s, indicating digestion of at least 93% of
the protein. With a 3.2-s residence time, the MALDI-MS sequence coverages of the α-S1
and α-S2 chain were 53 % and 43%, respectively. In contrast after a 16-hour in-solution
digestion, the sequence coverages were only 46% and 26%, respectively. Thus, in much
shorter time, the trypsin membrane reactor provided equal or better digestion than the
conventional in-solution digestion. Membranes were active during continual digestion
for 33 h, and their shelf life is several months.

20

Figure 1.6. Peptide size distributions in ESI-Orbitrap mass spectra of BSA
digested using a pepsin-containing membrane and residence times of 2 (blue) and
0.05 (red) s. The numbers above the bars are the number of recognized peptides in that
mass range. Adapted from reference 88 with permission of the American Chemical
Society.
Perhaps the biggest asset of membrane digestion is the control of peptide size
afforded by variation of residence times down to the ms level. Short residence times
should give large peptides due to missed cleavage sites. Tan and coworkers (89) reported
limited proteolysis using membranes modified with PSS/pepsin coatings. Apomyoglobin
digestion at pH 2 showed increasing peptide sizes with decreasing residence times. The
controlled digestion improved protein sequence coverage in ESI-Orbitrap MS analysis.
In limited digestion of denatured BSA, which contains 583 amino acid residues and 17
disulfide bonds, a 0.05-s resident time in a pepsin-containing membrane yielded 82%
sequence coverage in ESI-Orbitrap MS analysis, whereas a 2-s residence time (a

21

complete digestion) gave only 53% sequence coverage. Complete digestion likely gives
small peptides that are sometimes difficult to detect. Figure 1.6 shows the peptide size
distributions and clearly demonstrates that the 0.05-s digestion leads to larger peptides
than digestion for 2 s. Larger peptides should enhance detection of posttranslational
modifications because of greater sequence coverage.
Tan and coworkers employed limited digestion in PSS/trypsin-containing
membranes to examine the structure of Root Hair Defective (RHD3) protein (89). RHD3
participates in fusion of the endoplasmic reticulum membrane and is essential in the
growth of Arabidopsis thaliana (90), but with a complex structure including a GTPase
domain, a helix bundle, two trans-membrane helices and one cytosolic tail, the biological
roles and structure of RHD3 are difficult to interrogate. Tan and coworkers employed
low membrane residence times to determine which sites are most accessible to digestion.
Non-denatured RHD 3 was passed through a PSS/trypsin-modified membrane with
residence times of 0.045, 0.2 and 1 s. Analysis of the digests by MS revealed the
strongest signal for the peptide from amino acid residues 673-691 (RHD 3 has 691 total
amino acid residues) for all residence times. Signals for other peptides showed a relative
increase in intensity as digestion time increased. These results suggest that R672 is in the
most flexible region of RHD 3, and truncating the protein at R672 may lead to
crystallization.

1.6. Outline of the dissertation

The literature reviewed above clearly shows the potential of membranes to simplify

22

and enhance the rate of protein purification and also to control protein digestion. This
dissertation first aims to develop new methods to more simply and inexpensively modify
porous membranes for protein purification. In a second aim, the research examines
covalent trypsin immobilization to stabilize membranes for protein digestion. To
simplify membrane modification, we employ layer-by-layer polyelectrolyte deposition
(LbL), which simply involves passing polyelectrolyte solutions through a membrane.
Chapter 2 of this dissertation describes fundamental studies of LbL polyelectrolyte film
properties on flat surfaces mainly using ellipsometry and reflectance FT-IR. With control
over the growth conditions, these films swell extensively in water and capture
multiple-layer of protein to provide high protein binding capacity. However, protein
binding depends on pH, ionic strength and binding kinetics. This chapter examines the
factors that affect protein binding LbL films.
In chapter 3, I adsorb PAA/polyethyleneimine (PEI)/PAA films in porous
membranes and subsequently derivatize these films to capture His-tagged protein.
His-tagged proteins bind to Ni2+ that is bound to NTA groups appended to the
PAA/PEI/PAA films. This chapter investigates both metal-ion binding and protein
binding and compares the recovery and specificity of protein capture to purification with
commercial resins. To simplify the membrane fabrication process and decrease cost,
Salinda Wijeratne of our group developed new polymers that I used to modify the
membranes for his-tagged protein purification.
Membrane-based protein purification is not limited to His-tagged proteins. In
chapter 4, I describe efforts to create glycoprotein purification membranes. Purification

23

of glycoprotein relies on interactions with lectins such as Con A. Thus, I immobilized
Con A on the membranes using covalent or metal-affinity methods. The immobilization
employed membranes containing polymer brushes or LbL films, but unfortunately
glycoprotein binding was not extensive.
As reviewed above, thin membranes are unique supports for enzyme reactors
because they afford control over reaction times down to the ms level. Chapter 5 describes
the development and application of a reaction containing covalently immobilized trypsin.
The covalent linkage prevents enzyme leaching and provides a high enzyme
concentration that leads to digestion that is much faster than conventional in-solution
methods.

Moreover, these membranes afford control of peptide lengths by variation of

the flow rate through the membrane. These reactors are more stable than membrane
containing electrostatically adsorbed trypsin, and this stability enables digestion in harsh
conditions such as 6 M urea. Digestion under such heavily denaturing conditions is not
possible with in-solution methods.
Finally, chapter 6 summarizes the major findings in this work and highlight future
applications.

24

REFERENCES

25

REFERENCES

1.

GBI research.Therapeutic proteins market to 2017 - high demand for monoclonal
antibodies
will
drive
the
market.
2011.
Https://www.Marketresearch.Com/GBI-research-v3759/therapeutic-proteins-highdemand-monoclonal-6511062/ (accessed on 09/04/2014).

2.

Ghosh R. Protein separation using membrane chromatography: Opportunities
and challenges. J. Chromatogr. A 2002, 952, 13-27.

3.

Low D, O'Leary R, Pujar NS.
2007, 848, 48-63.

4.

Bhut BV, Wickramasinghe SR, Husson SM.
Preparation of high-capacity,
weak anion-exchange membranes for protein separations using surface-initiated
atom transfer radical polymerization. J. Membr. Sci. 2008, 325, 176-183.

5.

Conley AJ, Joensuu JJ, Richman A, Menassa R.
Protein body-inducing fusions
for high-level production and purification of recombinant proteins in plants. Plant
Biotechnol. J. 2011, 9, 419-433.

6.

Idiris A, Tohda H, Kumagai H, Takegawa K.
Engineering of protein secretion
in yeast: Strategies and impact on protein production. Appl. Microbiol. Biotechnol.
2010, 86, 403-417.

7.

Boi C.
Membrane adsorbers as purification tools for monoclonal antibody
purification. J. Chromatogr. B 2007, 848, 19-27.

8.

Saxena A, Tripathi BP, Kumar M, Shahi VK.
Membrane-based techniques for
the separation and purification of proteins: An overview. Adv. Colloid Interface
Sci. 2009, 145, 1-22.

9.

Yang Q, Adrus N, Tomicki F, Ulbricht M.
Composites of functional polymeric
hydrogels and porous membranes. J. Mater. Chem. 2011, 21, 2783-2811.

10.

Weaver J, Husson SM, Murphy L, Wickramasinghe SR.
Anion exchange
membrane adsorbers for flow-through polishing steps: Part I. Clearance of minute
virus of mice. Biotechnol. Bioeng. 2013, 110, 491-499.

11.

Monster A, Villain L, Scheper T, Beutel S.
One-step-purification of penicillin
G amidase from cell lysate using ion-exchange membrane adsorbers. J. Membr.
Sci. 2013, 444, 359-364.

Future of antibody purification. J. Chromatogr. B

26

12.

Bhut BV, Christensen KA, Husson SM.
Membrane chromatography: Protein
purification from E. coli lysate using newly designed and commercial
anion-exchange stationary phases. J. Chromatogr. A 2010, 1217, 4946-4957.

13.

Sun L, Dai JH, Baker GL, Bruening ML.
High-capacity, protein-binding
membranes based on polymer brushes grown in porous substrates. Chem. Mat.
2006, 18, 4033-4039.

14.

Kawakita H, Masunaga H, Nomura K, Uezu K, Akiba I, Tsuneda S.
Adsorption of bovine serum albumin to a polymer brush prepared by
atom-transfer radical polymerization in a porous inorganic membrane. J. Porous
Mat. 2007, 14, 387-391.

15.

Orr V, Zhong LY, Moo-Young M, Chou CP.
Recent advances in
bioprocessing application of membrane chromatography. Biotechnol. Adv. 2013,
31, 450-465.

16.

Milner ST.

17.

Huang WX, Skanth G, Baker GL, Bruening ML.
Surface-initiated thermal
radical polymerization on gold. Langmuir 2001, 17, 1731-1736.

18.

Jain P, Baker GL, Bruening ML.
Applications of polymer brushes in protein
analysis and purification. Annu. Rev. Anal. Chem. 2009, 2, 387-408.

19.

Chung ID, Britt P, Xie D, Harth E, Mays J.
Synthesis of amino acid-based
polymers via atom transfer radical polymerization in aqueous media at ambient
temperature. Chem. Commun. 2005, 1046-1048.

20.

Tsarevsky NV, Pintauer T, Matyjaszewski K.
Deactivation efficiency and
degree of control over polymerization in ATRP in protic solvents.
Macromolecules 2004, 37, 9768-9778.

21.

Kim JB, Huang WX, Miller MD, Baker GL, Bruening ML.
Kinetics of
surface-initiated atom transfer radical polymerization. J. Polym. Sci. Pol. Chem.
2003, 41, 386-394.

22.

Matyjaszewski K, Xia JH.
2001, 101, 2921-2990.

23.

Jain P, Dai J, Grajales S, Saha S, Baker GL, Bruening ML.
Completely
aqueous procedure for the growth of polymer brushes on polymeric substrates.
Langmuir 2007, 23, 11360-11365.

24.

Anuraj N, Bhattacharjee S, Geiger JH, Baker GL, Bruening ML.
An
all-aqueous route to polymer brush-modified membranes with remarkable
permeabilites and protein capture rates. J. Membr. Sci. 2012, 389, 117-125.

Polymer brushes. Science 1991, 251, 905-914.

Atom transfer radical polymerization. Chem. Rev.

27

25.

Jain P, Sun L, Dai JH, Baker GL, Bruening ML.
High-capacity purification of
his-tagged proteins by affinity membranes containing functionalized polymer
brushes. Biomacromolecules 2007, 8, 3102-3107.

26.

Bhut BV, Husson SM.
Dramatic performance improvement of weak
anion-exchange membranes for chromatographic bioseparations. J. Membr. Sci.
2009, 337, 215-223.

27.

Bhut BV, Weaver J, Carter AR, Wickramasinghe SR, Husson SM.
The role of
polymer nanolayer architecture on the separation performance of anion-exchange
membrane adsorbers: I. Protein separations. Biotechnol. Bioeng. 2011, 108,
2645-2653.

28.

Schwark S, Ulbricht M.
Toward protein-selective membrane adsorbers: A
novel surface-selective photo-grafting method. Eur. Polym. J. 2012, 48,
1914-1922.

29.

Tomicki F, Krix D, Nienhaus H, Ulbricht M.
Stimuli-responsive track-etched
membranes via surface-initiated controlled radical polymerization: Influence of
grafting density and pore size. J. Membr. Sci. 2011, 377, 124-133.

30.

Bao ZY, Bruening ML, Baker GL.
Control of the density of polymer brushes
prepared by surface-initiated atom transfer radical polymerization.
Macromolecules 2006, 39, 5251-5258.

31.

Yang Q, Kaul C, Ulbricht M.
Anti-nonspecific protein adsorption properties of
biomimetic glycocalyx-like glycopolymer layers: Effects of glycopolymer chain
density and protein size. Langmuir 2010, 26, 5746-5752.

32.

Franke A, Forrer N, Butte A, Cvijetic B, Morbidelli M, Johnck M, Schulte M.
Role of the ligand density in cation exchange materials for the purification of
proteins. J. Chromatogr. A 2010, 1217, 2216-2225.

33.

Gaberc-Porekar V, Menart V.
Potential for using histidine tags in purification
of proteins at large scale. Chem. Eng. Technol. 2005, 28, 1306-1314.

34.

Jain P, Vyas MK, Geiger JH, Baker GL, Bruening ML.
Protein purification
with polymeric affinity membranes containing functionalized poly(acid) brushes.
Biomacromolecules 2010, 11, 1019-1026.

35.

QIAGEN sample & assay technologies. Ni-NTA superflow cartridges.
Http://www.Qiagen.Com/search/ni-nta-superflow-cartridges#technicalspecificatio
n
(accessed on 09/04/2014).

36.

Bertozzi CR, Kiessling LL.
2357-2364.

Chemical glycobiology. Science 2001, 291,

28

37.

Sacchettini JC, Baum LG, Brewer CF.
Multivalent protein-carbohydrate
interactions. A new paradigm for supermolecular assembly and signal
transduction. Biochemistry 2001, 40, 3009-3015.

38.

Kiessling LL, Gestwicki JE, Strong LE.
Synthetic multivalent ligands as
probes of signal transduction. Angew. Chem.-Int. Edit. 2006, 45, 2348-2368.

39.

Holgersson J, Gustafsson A, Breimer ME.
Characteristics of
protein-carbohydrate interactions as a basis for developing novel
carbohydrate-based antirejection therapies. Immunol. Cell Biol. 2005, 83,
694-708.

40.

Mammen M, Choi SK, Whitesides GM.
Polyvalent interactions in biological
systems: Implications for design and use of multivalent ligands and inhibitors.
Angew. Chem.-Int. Edit. 1998, 37, 2755-2794.

41.

Fernandez-Alonso MD, Diaz D, Berbis MA, Marcelo F, Canada J,
Jimenez-Barbero J.
Protein-carbohydrate interactions studied by NMR: From
molecular recognition to drug design. Curr. Protein Pept. Sci. 2012, 13, 816-830.

42.

Nagahori N, Nishimura SI.
Tailored glycopolymers: Controlling the
carbohydrate-protein interaction based on template effect. Biomacromolecules
2001, 2, 22-24.

43.

Ragoussi ME, Casado S, Ribeiro-Viana R, de la Torre G, Rojo J, Torres T.
Selective carbohydrate-lectin interactions in covalent graphene- and
SWCNT-based molecular recognition systems. Chem. Sci. 2013, 4, 4035-4041.

44.

Yang Q, Hu MX, Dai ZW, Tian J, Xu ZK.
Fabrication of glycosylated surface
on polymer membrane by UV-induced graft polymerization for lectin recognition.
Langmuir 2006, 22, 9345-9349.

45.

Hu MX, Fang Y, Xu ZK.
Glycosylated membranes: A promising biomimetic
material. J. Appl. Polym. Sci. 2014, 131, 39658.

46.

Wang ZH, Chen GJ, Lu JW, Hong LZ, Ngai T.
Investigation of the factors
affecting the carbohydrate-lectin interaction by ITC and QCM-D. Colloid Polym.
Sci. 2014, 292, 391-398.

47.

Yang Q, Ulbricht M.
Cylindrical membrane pores with well-defined grafted
linear and comblike glycopolymer layers for lectin binding. Macromolecules 2011,
44, 1303-1310.

48.

Wang C, Fan Y, Hu MX, Xu W, Wu J, Ren PF, Xu ZK.
Glycosylation of the
polypropylene membrane surface via thiol-yne click chemistry for lectin
adsorption. Colloids Surf., B 2013, 110, 105-112.

29

49.

Vasapollo G, Del Sole R, Mergola L, Lazzoi MR, Scardino A, Scorrano S, Mele
G.
Molecularly imprinted polymers: Present and future prospective. Int. J. Mol.
Sci. 2011, 12, 5908-5945.

50.

Chen LX, Xu SF, Li JH.
Recent advances in molecular imprinting technology:
Current status, challenges and highlighted applications. Chem. Soc. Rev. 2011, 40,
2922-2942.

51.

Yin DX, Ulbricht M.
Protein-selective adsorbers by molecular imprinting via a
novel two-step surface grafting method. J. Mat. Chem. B 2013, 1, 3209-3219.

52.

Yin DX, Ulbricht M.
Antibody-imprinted membrane adsorber via two-step
surface grafting. Biomacromolecules 2013, 14, 4489-4496.

53.

Wang J, Faber R, Ulbricht M.
Influence of pore structure and architecture of
photo-grafted functional layers on separation performance of cellulose-based
macroporous membrane adsorbers. J. Chromatogr. A 2009, 1216, 6490-6501.

54.

Wei YM, Ma JJ, Wang CZ.
Preparation of high-capacity strong cation
exchange membrane for protein adsorption via surface-initiated atom transfer
radical polymerization. J. Membr. Sci. 2013, 427, 197-206.

55.

Pall corporation. Acrodisc® units with Mustang® Q and S membranes.
Http://www.Pall.Com/main/laboratory/product.Page?Lid=gri78l6d (accessed on
09/04/2014).

56.

Chenette HCS, Robinson JR, Hobley E, Husson SM.
Development of
high-productivity, strong cation-exchange adsorbers for protein capture by graft
polymerization from membranes with different pore sizes. J. Membr. Sci. 2012,
423, 43-52.

57.

Yusof AHM, Ulbricht M.
Polypropylene-based membrane adsorbers via
photo-initiated graft copolymerization: Optimizing separation performance by
preparation conditions. J. Membr. Sci. 2008, 311, 294-305.

58.

Susanto H, Roihatin A, Aryanti N, Anggoro DD, Ulbricht M.
Effect of
membrane hydrophilization on ultrafiltration performance for biomolecules
separation. Mater. Sci. Eng. C-Mater. Biol. Appl. 2012, 32, 1759-1766.

59.

Roper DK, Lightfoot EN.
Separation of biomolecules using adsorptive
membranes. J. Chromatogr. A 1995, 702, 3-26.

60.

Chen J, Luo Q, Breneman CM, Cramer SM.
Classification of protein
adsorption and recovery at low salt conditions in hydrophobic interaction
chromatographic systems. J. Chromatogr. A 2007, 1139, 236-246.

61.

Shang XJ, Wittbold W, Ghosh R.
30

Purification and analysis of

mono-PEGylated HSA by hydrophobic interaction membrane chromatography. J.
Sep. Sci. 2013, 36, 3673-3681.
62.

Charlet G, Delmas G.
Thermodynamic properties of polyolefin solutions at
high-temperature.1. Lower critical solubility temperatures of polyethylene,
polypropylene and ethylene-propylene co-polymers in hydrocarbon solvents.
Polymer 1981, 22, 1181-1189.

63.

Himstedt HH, Qian XH, Weaver JR, Wickramasinghe SR.
Responsive
membranes for hydrophobic interaction chromatography. J. Membr. Sci. 2013,
447, 335-344.

64.

Ghosh R, Wang L.
Purification of humanized monoclonal antibody by
hydrophobic interaction membrane chromatography. J. Chromatogr. A 2006, 1107,
104-109.

65.

Wang J, Sproul RT, Anderson LS, Husson SM.
Development of multimodal
membrane adsorbers for antibody purification using atom transfer radical
polymerization. Polymer 2014, 55, 1404-1411.

66.

Graves JD, Krebs EG.
Protein phosphorylation and signal transduction.
Pharmacol. Ther. 1999, 82, 111-121.

67.

Hunter T.

68.

Schettini G, Govoni S, Racchi M, Rodriguez G.
Phosphorylation of
APP-CTF-AICD domains and interaction with adaptor proteins: Signal
transduction and/or transcriptional role - relevance for alzheimer pathology. J.
Neurochem. 2010, 115, 1299-1308.

69.

Cohen P.
The role of protein phosphorylation in human health and disease. Eur.
J. Biochem. 2001, 268, 5001-5010.

70.

Huang DM, Li YJ, Liu N, Zhang ZL, Peng ZK, Duan CJ, Tang XW, Tan GJ, Yan
GR, Tang FQ.
Identification of novel signaling components in
N,N'-Dinitrosopiperazine-mediated metastasis of nasopharyngeal carcinoma by
quantitative phosphoproteomics. Bmc Cancer 2014, 14, 243.

71.

Witze ES, Old WM, Resing KA, Ahn NG.
Mapping protein post-translational
modifications with mass spectrometry. Nat. Methods 2007, 4, 798-806.

72.

Dunn JD, Reid GE, Bruening ML.
Techniques for phosphopeptide enrichment
prior to analysis by mass spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.

73.

Cheng G, Wang ZG, Liu YL, Zhang JL, Sun DH, Ni JZ.
Phosphoprotein/phosphopeptide enrichment and analysis based on nanostructured
materials. Prog. Chem. 2013, 25, 620-632.

Signaling - 2000 and beyond. Cell 2000, 100, 113-127.

31

74.

Wang WH, Palumbo AM, Tan YJ, Reid GE, Tepe JJ, Bruening ML.
Identification of p65-associated phosphoproteins by mass spectrometry after
on-plate phosphopeptide enrichment using polymer-oxotitanium films. J.
Proteome Res. 2010, 9, 3005-3015.

75.

Wang WH, Dong JL, Baker GL, Bruening ML.
Bifunctional polymer brushes
for low-bias enrichment of mono- and multi-phosphorylated peptides prior to
mass spectrometry analysis. Analyst 2011, 136, 3595-3598.

76.

Tan YJ, Sui DX, Wang WH, Kuo MH, Reid GE, Bruening ML.
Phosphopeptide enrichment with TiO2-modified membranes and investigation of
tau protein phosphorylation. Anal. Chem. 2013, 85, 5699-5706.

77.

Han XM, Aslanian A, Yates JR.
Chem. Biol. 2008, 12, 483-490.

78.

Ji J, Zhang Y, Zhou X, Kong J, Tang Y, Liu B.
Enhanced protein digestion
through the confinement of nanozeolite-assembled microchip reactors. Anal.
Chem. 2008, 80, 2457-2463.

79.

Krenkova J, Lacher NA, Svec F.
Highly efficient enzyme reactors containing
trypsin and endoproteinase LysC immobilized on porous polymer monolith
coupled to MS suitable for analysis of antibodies. Anal. Chem. 2009, 81,
2004-2012.

80.

Spross J, Sinz A.
A capillary monolithic trypsin reactor for efficient protein
digestion in online and offline coupling to ESI and MALDI mass spectrometry.
Anal. Chem. 2010, 82, 1434-1443.

81.

Gao J, Xu JD, Locascio LE, Lee CS.
Integrated microfluidic system enabling
protein digestion, peptide separation, and protein identification. Anal. Chem. 2001,
73, 2648-2655.

82.

Cooper JW, Chen JZ, Li Y, Lee CS.
Membrane-based nanoscale proteolytic
reactor enabling protein digestion, peptide separation, and protein identification
using mass spectrometry. Anal. Chem. 2003, 75, 1067-1074.

83.

Liuni P, Rob T, Wilson DJ.
A microfluidic reactor for rapid, low-pressure
proteolysis with on-chip electrospray ionization. Rapid Commun. Mass Spectrom.
2010, 24, 315-320.

84.

Slysz GW, Schriemer DC.
Blending protein separation and peptide analysis
through real-time proteolytic digestion. Anal. Chem. 2005, 77, 1572-1579.

85.

Slysz GW, Lewis DF, Schriemer DC.
Detection and identification of
sub-nanogram levels of protein in a nanoLC-trypsin-MS system. J. Proteome Res.
2006, 5, 1959-1966.

Mass spectrometry for proteomics. Curr. Opin.

32

86.

Xu F, Wang WH, Tan YJ, Bruening ML.
Facile trypsin immobilization in
polymeric membranes for rapid, efficient protein digestion. Anal. Chem. 2010, 82,
10045-10051.

87.

Buck FF, Vithayathil AJ, Nord FF, Bier M.
On the mechanism of enzyme
action. LXXIII. Studies on trypsins from beef, sheep and pig pancreas. Arch.
Biochem. Biophys. 1962, 97, 417-424.

88.

Tan YJ, Wang WH, Zheng Y, Dong JL, Stefano G, Brandizzi F, Garavito RM,
Reid GE, Bruening ML.
Limited proteolysis via millisecond digestions in
protease-modified membranes. Anal. Chem. 2012, 84, 8357-8363.

89.

Saltzman MR, Young SA, Kump LR, Gill BC, Lyons TW, Runnegar B.
Pulse
of atmospheric oxygen during the late cambrian. Proc. Natl. Acad. Sci. U. S. A.
2011, 108, 3876-3881.

33

Part of this chapter is adapted from our published paper in Langmuir with permission of
the American Chemical Society (Ma, Y., Dong, J., Bhattacharjee, S., Wijeratne, S.,
Bruening, M. L. and Baker, G. L. Langmuir 2013, 29, 2946-2954.).

Chapter 2. Characterization of polyelectrolyte films used for protein
capture

The long-range goal of this research is the development of coatings that
conveniently capture proteins in porous membranes. Layer-by-layer (LbL) adsorption of
polyelectrolytes is a particularly convenient method for preparing such coatings. This
chapter describes fundamental studies of the growth, swelling, ionization, stability, and
lysozyme binding of protonated poly(allylamine) (PAH)/poly(acrylic acid) (PAA) films
on Au-coated Si wafers. Deposition of these films on flat wafers enables characterization
using ellipsometry and reflectance FT-IR spectroscopy, which is not possible with films
in membranes pores. With control over adsorption conditions, (PAH/PAA)n films formed
at low pH swell extensively in pH 7.4 buffers and capture multiple layers of lysozyme.
However, lysozyme sorption occurs via ion exchange and thus depends on pH and ionic
strength.

2.1. Introduction

Despite the successful application of polymer brushes to membrane functionalization,
brush synthesis is complicated because it relies on initiator attachment and
polymerization under anaerobic conditions (10). LbL adsorption of functional
34

polyelectrolytes, exposing a surface to alternating aqueous polyanion and polycation
solutions with a water rinse between each exposure, provides a much simpler method to
coat substrates (91, 92). The LbL method is essentially a multi-point “grafting to”
technique, but sequential adsorption steps, judicious polyelectrolyte selection, and
control over deposition conditions can tune polyelectrolyte multilayers (PEMs) for
different purposes (93-95).
Since its introduction by Decher and Hong, LbL polyelectrolyte adsorption has
emerged as one of the simplest and most versatile coating methods as it applies to a
range of materials and geometries (96-98), including flat surfaces (91, 96), nanoparticles
(99) and membranes (100, 101). Over the last 20 years, research groups have applied
PEMs in sensors (102-105), biomimetic material (106-108), tissue engineering (109, 110)
and drug delivery (111-113). Building blocks of PEMs can include polymers (91,
114-117), viruses (118, 119), enzymes (120, 121) and inorganic particles (96, 122, 123).
The PEMs can also serve as a “reservoir” for biomolecules and preserve their bioactivity
(92). For example, Jewell and coworkers incorporated DNA, RNA and nucleic
acid-based agents in PEMs for drug delivery purposes (124).
My research employs PEMs for protein capture, and extensive protein sorption will
likely require films that swell extensively in water. Variation of deposition parameters
such as ionic strength and pH allows tuning of PEM ionization and swelling to control
protein sorption (125). For example, Dragan and coworkers reported sorption of the
equivalent of 30 nm of human serum albumin in cross-linked PEMs (126). They
generated the PEMs by selectively cross-linking the poly(vinyl amine) (PVAm) layers in

35

(PVAm/ PAA)n thin films and subsequently removing the PAA layer. This chapter aims
at developing PEMs that adsorb the equivalent of many monolayers of protein without
post-construction treatment, with an eventual goal of providing high protein binding
capacity in modified membranes. Specifically, I focus on (PAH/PAA)n films because
under physiological condition PAA-containing PEMs show minimal nonspecific binding
and their –COOH groups are readily available for derivatization with protein binding
ligands (93, 127). Control over deposition conditions also leads to (PAA/PAH)n films
that capture large amounts of protein. This chapter characterizes the growth of
(PAA/PAH)n coatings on flat substrates as well as their swelling, stability, ionization,
and lysozyme sorption. Later chapters investigate derivatization of these films to
specifically capture His-tagged proteins, glycoproteins, and antigens both on flat surfaces
and in membranes.

2.2. Experimental section
2.2.1. Materials

Poly(allylamine hydrochloride) (PAH, molecular weight 120,000 - 200,000 Da) was
purchased from Alfa Aesar, and poly(acrylic acid) was obtained from Polysciences
(molecular weight 90000 Da, 25 wt% solution in water) or Sigma-Aldrich (molecular
weight 100000 Da, 35 wt% solution in water). Aqueous solutions of 0.02 M PAH and
0.01 M PAA were prepared in deionized water (18.2 MΩcm, Milli-Q) with 0.5 M
aqueous NaCl, and solution pH values were adjusted by dropwise addition of 0.1 M
NaOH or HCl. (Polymer concentrations are given with respect to the repeating unit.)

36

Lysozyme from egg white, sodium phosphate and sodium phosphate dibasic were
purchased from Sigma-Aldrich and used without further purification. Au-coated silicon
wafers (200 nm of sputtered Au on 20 nm of Cr on Si (100) wafers) were cleaned in a
UV/O3 chamber for 15 min prior to use.

2.2.2. Preparation of PEMs
Au-coated Si substrates (24 mm × 11 mm) were immersed in 5 mM
3-mercaptopropionic acid (MPA) in ethanol for 2 h, rinsed with ethanol and dried with
N2 to form a monolayer of MPA and create a negatively charged surface. These
substrates were immersed in 0.02 M aqueous PAH (with 0.5 M NaCl, adjusted to the
desired pH) for 5 min and subsequently rinsed with 10 mL of deionized water and blown
dry with N2. Substrates were then immersed in a polyanion-containing solution (0.02 M
PAA with 0.5 M NaCl, adjusted to the desired pH value) for 5 min followed by the same
rinsing and drying procedures. The process was repeated to form multilayer films.

2.2.3. Characterization of PEMs
“Dry” film thicknesses were determined using a spectroscopic ellipsometer (J.A.
Woollam M-44) assuming a film refractive index of 1.5. Fitting the refractive index did
not significantly alter the calculated thickness. In situ ellipsometry in aqueous solutions
was performed using a home-built cell described previously (128, 129). In this case, both
refractive index and thickness were fitting parameters. Reflectance FTIR spectra were
obtained with a Thermo Nicolet 6700 FTIR spectrometer using a Pike grazing angle (80º)
apparatus. A UV/ozone-cleaned gold wafer served as a background. Atomic force
37

microscope (AFM) images were acquired using a NanoScope IIIa scanning probe
microscope (Veeco Instruments, CA) operating in the contact mode. Pyramidal-shaped
Si3N4 tips, mounted on gold cantilevers (100 µm legs, 0.38 N/m spring constant) were
used in air.

2.2.4. Determination of the average pKa for PAA

The average pKa value for PAA in aqueous solution was calculated from titrations
with 0.1 M HCl. The polyelectrolyte was dissolved in aqueous NaOH (pH ~12), and the
pH was adjusted to 10 before titration. For films, pKa values for PAA were estimated
based on the ratio of –COOH and –COO- absorbances (peak heights) in FTIR spectra (1).

2.3. Results and discussion

Protein binding likely varies with film thickness and ionization. Thus, the following
two subsections present ellipsometric studies of film thickness as a function the number
of adsorbed layers and reflectance FTIR spectra that investigate the extent of ionization.
Subsequent sections examine lysozyme binding and film stability.

2.3.1. Formation of (PAH/PAA)n films

The driving force for PEM formation includes electrostatic interactions between
polycations and polyanions and more importantly release of counterions to the solution
to increase entropy (42, 43). With strong polyelectrolytes, addition of supporting
electrolyte to deposition solutions (44) and temperature affect polyelectrolyte
conformations and film thickness (45). In the case of weak polyelectrolytes such as PAH

38

and PAA, the pH of the deposition solutions also influences polyelectrolyte conformation
and film growth (7, 46). Typically, adsorption pH values that give a low degree of
ionization of a weak polyelectrolyte lead to thick films (7, 47). Because thick films are
attractive for high protein sorption, I studied the adsorption of (PAH/PAA)n films from
polyelectrolyte solutions at pH 3, where PAA is only partially deprotonated, and at pH 9,
where PAH is only partly protonated. Shiratori and Rubner performed a similar study,
but they employed polyelectrolyte solutions with no added salt (5). Because Ma and
other researchers found that the addition of 0.5 M NaCl to PAA and PAH deposition
solutions greatly increases film thickness, this work examines adsorption from solutions
containing 0.5 M NaCl (46, 48, 49).
70
pH 3

Thickness (nm)

60

pH 9

50

40
30
20
10
0
0

1

2
3
4
Number of bilayers, n

5

Figure 2.1. Ellipsometric thicknesses of (PAH/PAA)n films deposited from 20
mM polyelectrolyte solutions containing 0.5 M NaCl at pH 3 (red squares) or
pH 9 (blue triangles). When n is an integer, film deposition ends after PAA
adsorption, whereas noninteger values indicate termination with PAH.
39

Figure 2.1 shows the growth of (PAH/PAA)n films with the number of adsorption
steps at pH 3 and pH 9. The total film thickness is similar for the two deposition pH
values over the first 3 bilayers. However, for additional bilayers, films deposited at pH 9
show higher thicknesses. At pH 3, PAA is mostly protonated and contains a high density
of –COOH in its structure. As mentioned above, the low charge density on the polymer
at low pH likely gives a compact, rather than extended, polymer chain and this results in
thicker films compared to adsorption of highly ionized PAA. Similarly, at pH 9
deprotonation of PAH leads to a low charge density and compact conformation of this
polyelectrolyte. As a result, deposition at either pH 3 or 9 leads to relatively thick films
compared to adsorption at neutral pH where film thickness is less than 10 nm for 5
bilayers (130).

2.3.2. PAA ionization in (PAH/PAA)n films

Adsorption of PAA in (PAA/PAH)n films changes its pKa value. Based on the titration
curve in Figure 2.2, the average pKa of free PAA in solution is ~ 6.3. For this titration,
the PAA solution was adjusted to pH 10 where we assumed all the –COOH groups were
deprotonated, and titration with 0.1 M HCl followed. Equation (2.1) describes the degree
of ionization, where COO- and COOH represent the moles of deprotonated and
protonated acid groups, respectively.
𝐶𝑂𝑂 −

𝛼 = 𝐶𝑂𝑂−+𝐶𝑂𝑂𝐻 x 100%

(2.1)

40

degree of ionization(%)

100
90
80
70
60
50
40
30
20
10
0

PAA

2

3

4

5 pH 6

7

8

9

10

Figure 2.2. Titration of deprotonated PAA with 0.1 M HCl. The number of –COOH
groups at a given pH value was determined with the following equation, assuming that all
the –COOH groups were deprotonated at pH 10.
𝐻+ 𝑎𝑑𝑑𝑒𝑑 = 𝐶𝑂𝑂𝐻 + 𝑂𝐻− 𝑛𝑒𝑢𝑡𝑟𝑎𝑙𝑖𝑧𝑒𝑑 + 𝑖𝑛𝑐𝑟𝑒𝑎𝑠𝑒 𝑖𝑛 𝑓𝑟𝑒𝑒 𝐻+ 𝑖𝑛 𝑠𝑜𝑙𝑢𝑡𝑖𝑜𝑛.
The titration curve for PAA is similar to that in the literature (1). The text defines the
degree of ionization.

To estimate the pKa of the PAA in (PAA/PAH)n films, we employed reflectance
FT-IR spectroscopy (Figure 2.3). With the assumption of the same extinction
coefficients for acid carbonyl and carboxylate absorption bands (1), equation (2.2) gives
the degree of ionization in the film.
𝛼=𝐴

𝐴𝐶𝑂𝑂−

𝐶𝑂𝑂− +𝐴𝐶𝑂𝑂𝐻

x 100% 





















In this equation, 𝐴𝐶𝑂𝑂− is the maximum absorbance of the carboxylate band (~1580
cm-1) and 𝐴𝐶𝑂𝑂𝐻 is the maximum absorbance of the acid carbonyl band (~1720 cm-1).
We obtained values for the degree of ionization by immersing (PAH/PAA)5 films in

41

different buffers and subsequently acquiring their reflectance FTIR spectra. Based on
equation (2.2) and the data in Figure 2.3, the estimated pKa value of the -COOH groups
in PAA is 5.3 (the pH at which 𝐴𝐶𝑂𝑂− = 𝐴𝐶𝑂𝑂𝐻 ), which is about 1-unit lower than the
pKa value of the polymer in solution. Note that the pKa value for PAA is an average
value, and that removal of protons becomes more difficult as the polymer reaches an
increasing degree of ionization. Thus the average pKa value for PAA in solution is about
2 units higher than the pKa value of acetic acid. In the film, electrostatic interactions with
neighboring ammonium groups likely decrease the energy of required to form –COOgroups (130).

42

Figure 2.3. Reflectance FTIR spectra of (PAH/PAA)5 films grown from pH 3 and
later immersed in 20 mM phosphate solutions adjusted to different pH values. Prior
to obtaining the IR spectra the films were rinsed only in deionized water, which based on
the difference in the spectra does not alter the degree of ionization.

2.3.3. Lysozyme binding to (PAH/PAA)5 films

We examined lysozyme sorption in (PAH/PAA)5 films because this protein is
positively charged and adsorbs through ion-exchange reactions to native (PAH/PAA)5
coatings. These studies employed (PAH/PAA)5 multilayers deposited from 20 mM
polyelectrolyte solutions containing 0.5 M NaCl at pH 3, and lysozyme sorption occurred

43

from 20 mM phosphate solutions at various pH values. We used reflectance FTIR spectra
(e.g. Figure 2.4) to determine the amide absorbance of the captured lysozyme, and prior
spectra of spin-coated protein films with a variety of thicknesses allowed us to convert
the amide absorbance to an equivalent thicknesses of sorbed protein (131).

Figure 2.4. FTIR spectra of (PAH/PAA)5 (top) and (PAH/PAA)5PAH (bottom)
films before and after sorption of lysozyme. Polyelectrolyte films were deposited from
solutions containing 0.5 M NaCl at pH 3.0. Sorption of lysozyme occurred for 16 h from
20 mM phosphate buffer (pH 7.4) containing 1 mg/mL of lysozyme, and films were
rinsed with water prior to obtaining the spectra.
Lysozyme sorption from pH 3 and pH 5 solutions yields lysozyme equivalent
thicknesses of 7.1 ± 1.7 nm and 18.9 ± 0.2 nm, respectively. In contrast, sorption at pH
44

7.4 and pH 10 leads to lysozyme equivalent thicknesses around 150 nm (Figure 2.5). At
low pH, protons effectively compete with lysozyme for –COO- ion-exchange sites to
greatly decrease binding. Notably, terminating films with PAH rather than PAA does not
lead to a significant difference in lysozyme adsorption (Figure 2.4). Thus, the lysozyme
penetrates the film during sorption. Effective lysozyme elution with 1 M KSCN is also
consistent with binding by ion-exchange. Exposure to the KSCN solution for only 1 min
removes 96% of the bound lysozyme. In contrast, elution with 0.1 M KSCN removes
only 85% of the lysozyme in 1 h (Table 2.1). Binding also varies with the ionic strength
of the lysozyme solution. In 20 mM phosphate buffer at pH 7.4, addition of 0.1 M NaCl
to the adsorption solution decreases the effective thickness of bound lysozyme
marginally from 147 ± 29 nm to 133 ± 1 nm. However, further increasing the NaCl
concentration to 0.3 or 0.5 M decreases the effective lysozyme thickness to essentially
zero (Figure 2.6). The Na+ effectively competes with lysozyme for the ion-exchange
sites.

45

Lysozyme Equivalent Thickness (nm)

180
160
140
120
100
80
60
40
20
0
0

2

4
6
8
pH of binding buffer

10

Figure 2.5. Lysozyme sorption in (PAH/PAA)5 films as a function of the pH of the
1 mg/mL lysozyme sorption solution.

(PAH/PAA)5 coatings were deposited from

solutions containing 0.5 M NaCl at pH 3. Lysozyme sorption solutions contained 20
mM phosphate adjusted to different pH values and no supporting salt. The equivalent
thickness is the thickness of spin-coated lysozyme that would give an FTIR
absorbance equivalent to that of the sorbed lysozyme.

46

Table 2.1. Efficiency of lysozyme elution from (PAH/PAA)5 films as a function
of the eluent KSCN concentration and elution times.

Sorption solutions

contained 1 mg/mL of lysozyme in 20 mM phosphate buffer (pH 7.4). The eluent
was prepared in the same buffer. The elution efficiency was calculated from the
amount of lysozyme in the film before and after elution as determined from FTIR
spectra. The (PAH/PAA)5 film was adsorbed from a pH 3 solution containing 0.5
M NaCl.

Ratio of Eluted - to - Bound lysozyme
Elution time

0.10 M KSCN

0.25 M KSCN

1.00 M KSCN

1 min

0.56

0.88

0.96

10 min

0.81

0.93

-

30 min

0.84

0.93

-

1 hour

0.85

0.94

-

47

Lysozyme Equivalent Thickness (nm)

200
180
160
140
120
100
80
60
40
20
0
-20
0

0.1
0.2
0.3
0.4
0.5
Concentration of NaCl in binding buffer (M)

Figure 2.6. Lysozyme sorption in (PAH/PAA)5 multilayers as a function of the
concentration of NaCl in the 1 mg/mL lysozyme sorption solution. (PAH/PAA)5
films were deposited from solutions containing 0.5 M NaCl at pH 3. The sorption
solution also contained 20 mM phosphate buffer at pH 7.4. The equivalent thickness is
the thickness of spin-coated lysozyme that would give a reflectance FTIR absorbance
equivalent to that of the sorbed lysozyme.

The above results employed a 16-h adsorption time for lysozyme sorption to ensure
equilibrium, but binding actually occurs in a few minutes. Figure 2.7 shows how the
binding capacity of a (PAH/PAA)5 film (deposited from 0.5 M NaCl solutions at pH 3)
varies with sorption time. Even after a 4-min exposure to lysozyme, the binding reaches
85% of the value at 16 h, and the difference in the amount of lysozyme sorption during 1

48

h and 16 h is negligible. The rapid binding suggests a highly swollen film that affords
access to interior binding sites.

Lysozyme Equivalent Thickness
(nm)

160
140
120
100
80
60
40
20
0
0

5

10
Time (hour)

15

20

Figure 2.7. Time-evolution of lysozyme sorption in (PAH/PAA)5 multilayers
immersed in a 1 mg/mL lysozyme solution in 20 mM phosphate buffer (pH 7.4). The
(PAH/PAA)5 films were deposited from pH 3 solutions containing 0.5 M NaCl. The
equivalent thickness is the thickness of spin-coated lysozyme that gives an FTIR
absorbance equivalent to that of the sorbed lysozyme. The error bars represent standard
deviations of measurements on 3 different films.
2.3.4. Film swelling and morphology changes upon protein sorption

To examine the extent of film swelling, we performed in situ ellipsometry with

49

(PAA/PAH)5 films immersed in deionized water. The films were initially deposited at
pH 3 from a 0.5 M NaCl solution. After an overnight immersion in deionized water,
swelling increased the film thickness 5-fold relative to the thickness in ambient air.
(Most of the swelling occurs in the first two hours of immersion). Consistent with ~80%
water in the immersed coating, the film refractive index decreases from 1.50 to 1.36 after
swelling. (The refractive index of water at the wavelengths of the spectroscopic
ellipsometer is about 1.333.) After lysozyme binding, the dry film thickness is essentially
the thickness of the water-swollen film. The coating containing bound protein swells by
only 30% in water and the swollen film refractive index is around 1.50, reflective of the
large amount of sorbed protein. Thus, lysozyme likely replaces water in the film.
AFM images of (PAA/PAH)5 multilayers deposited from 0.5 M NaCl at pH 3
suggest that film morphology does not change greatly upon lysozyme sorption (Figures
2.8 a and b). The root-mean square roughness values for (PAA/PAH)5 films increase
from 2.0 to 3.4 nm after sorption. Given the large amount of protein binding (150 nm of
lysozyme in films with an initial dry thickness of 36 nm), these images provide another
evidence for diffusion of lysozyme into the interior of the film. Adsorption of multilayers
of lysozyme at the surface would likely increase roughness.
Interestingly, although (PAA/PAH)5 films deposited at pH 3 and pH 9 (from
solutions containing 0.5 M NaCl) sorb similar amounts of lysozyme (around 150 nm in
both cases), only the latter coating shows a large morphology change upon lysozyme
sorption. With (PAA/PAH)5 deposited at pH 9, the rms roughness increases from 2.7 to
14.7 nm upon lysozyme sorption (Figures 2.9 a and b). For these particular films, the

50

large positive charge due to both protonation of amine groups and sorption of lysozyme
may cause film reconstruction (132).

51

Figure 2.8. AFM images of (PAH/PAA)5 films (a) before and (b) after sorption

of lysozyme. The films were adsorbed from pH 3 solutions containing 0.5 M NaCl.
Lysozyme sorption occurred from a 1 mg/mL lysozyme solution in phosphate buffer
(7.4) for 16 h, and films were rinsed with deionized water prior to imaging.

52

Figure 2.9. AFM images of (PAH/PAA)5 films (a) before and (b) after adsorption

of lysozyme. The films were adsorbed from pH 9 solutions containing 0.5 M NaCl.
Lysozyme adsorption occurred from a 1 mg/mL lysozyme solution in phosphate buffer
(7.4) for 16 h, and films were rinsed with deionized water prior to imaging.

53

2.3.5. Film stability and repetitive binding experiments

The high swelling of (PAA/PAH)5 films deposited from 0.5 M NaCl at pH 3 is
attractive for protein binding, but these films are not stable in neutral pH buffers in the
absence of lysozyme. (They are stable in deionized water, the matrix for swelling studies.)
Figure 2.10 shows FTIR spectra of (PAA/PAH)5 films before and after an overnight
immersion in 20 mM phosphate buffers adjusted to pH 3, 5, 7.4, and 10 (no added NaCl).
At pH 3, the overnight immersion does not affect the IR spectrum of the film, but after
immersion in pH 7.4 or pH 10 phosphate buffer, the absorbance due to PAA decreases
by about 80%. Even after immersion in pH 5 buffer, the absorbance due to PAA may
decrease by 20%. However, this 20% decrease is not definitive because although the acid
carbonyl absorbance near 1720 cm-1 decreases, the carboxylate absorbance increases
slightly. (Even though we immersed films in a pH 3 phosphate solution prior to obtaining
the spectra, the degree of protonation may change slightly.). After immersion in a neutral
pH 7.4 or pH 10 buffer, deprotonation of -COOH groups creates electrostatic repulsion
as well as increased solubility of PAA, and this apparently leads to dissolution of much
of the film. The IR spectra in Figure 2.3 confirm partial deprotonation of –COOH
groups as the ambient pH increases from 3.0 to 6.0.

54

Figure 2.10. Reflectance FTIR spectra of (PAH/PAA)5 films before and after
immersion in a 20 mM phosphate solution adjusted to different pH values.
(PAH/PAA)5 films were deposited from 20 mM polyelectrolyte solutions containing 0.5
M NaCl at pH 3.0. These coatings were immersed in the buffers overnight and washed
sequentially with pH 3.0 phosphate and water before obtaining FTIR spectra.
Lysozyme binding stabilizes the adsorbed PAA in (PAA/PAH)5 films. Although
binding occurs at pH 7.4 where the film is unstable, the thickness of films containing
sorbed lysozyme is 180 ± 23 nm. The thickness prior to lysozyme sorption is 37 ± 6 nm.
However, after elution of lysozyme with 1 M KSCN in 20 mM phosphate buffer (pH

55

7.4), the spectrum of the film is similar to that after immersion of the film overnight in
pH 7.4 buffer, showing that the (PAA/PAH)5 is unstable in 1 M KSCN. The protein
binding could involve some replacement of PAH by lysozyme (133, 134), but given the
thickness increase upon binding, there is much more sorbed lysozyme than initially
adsorbed PAH.
Interestingly, after immersion in pH 7.4 phosphate buffer, the spectrum of a
(PAH/PAA)5 film is similar to that of a (PAH/PAA)2 film (also deposited from a pH 3,
0.5 M NaCl solution). Thus, we investigated repetitive lysozyme binding to (PAA/PAH)2
multilayers deposited at pH 3.0 and pH 9.0. Figure 2.11 shows that 5 cycles of binding
and elution on the same (PAA/PAH)2 film deposited from a pH 3, 0.5 M NaCl solution
lead to only a ~25% decrease in the binding capacity. For a (PAA/PAH)2 film deposited
from a pH 9 solution, the binding capacity decreases by 50% in the same experiment.
Nevertheless, immersion of a (PAH/PAA)2 film in pH 7.4 buffers overnight results in a
40% decrease in the carbonyl absorbance of the film (Figure 2.12). In contrast to films
deposited from pH 3, 0.5 M NaCl solutions, (PAH/PAA)5 films deposited at pH 9 are
relatively stable at neutral pH (compare Figure 2.10 and Figure 2.13), as these films
contain a small fraction of –COOH groups after deposition. However, these films show a
decreased protein binding capacity with repetitive use suggesting that lysozyme binding
and reuse cause some instability.

56

45
pH 3 2 bilayers

Lysozyme Equivalent Thickness (nm)

40

pH 9 2 bilayers
35
30
25
20
15
10
5
0
0

1

2

3
Repeat Times

4

5

Figure 2.11. Repetitive lysozyme sorption in (PAH/PAA)2 films from a 1 mg/mL
lysozyme sorption solution. (PAH/PAA)2 films were deposited from polyelectrolyte
solutions containing 0.5 M NaCl at pH 3.0. Between each binding, the sorbed
lysozyme was eluted with 20 mM phosphate buffer containing 1 M KSCN. The
sorption solution also contained 20 mM phosphate buffer at pH 7.4. The equivalent
thickness is the thickness of spin-coated lysozyme that would give an FTIR
absorbance equivalent to that of the sorbed lysozyme.

57

Figure 2.12. FTIR spectra of (PAH/PAA)2 films before and after immersion in a
20 mM phosphate solution (pH 7.4). (PAH/PAA)2 films were deposited from pH 3.0
polyelectrolyte solutions containing 0.5 M NaCl. The film was immersed in the buffer
overnight and washed with pH 3.0 phosphate buffer and water before FT-IR
measurements.

58

Figure 2.13. FTIR spectra of (PAH/PAA)5 films before and after immersion in a 20
mM phosphate solution adjusted to different pH values. (PAH/PAA)5 films were
deposited from pH 9.0 polyelectrolyte solutions containing 0.5 M NaCl. The films were
immersed in the buffers overnight and washed with pH 9.0 buffer and water before
FT-IR measurements.

2.4. Conclusions

LbL adsorption of PAA and PAH gives PEMs whose properties depend on the
deposition pH. Films grown from pH 3 swell by 500% in pH 7.4 buffer and capture
multilayers of lysozyme throughout the film structure. Lysozyme binding depends on the

59

pH and ionic strength of the binding buffers because of the ion-exchange mechanism.
At neutral pH and above, deprotonation of the –COOH groups in the film increases
swelling and creates ion-exchange sites for lysozyme binding. In contrast, electrolytes in
the elution solution can compete with the protein for ion-exchange sites to enable
efficient protein elution from the films. Films adsorbed on Au-coated wafers at pH 3
bind large amounts of protein, but they are not stable at neutral pH in the absence of
protein.

However, film stability may be greater on other substrates such as nylon

membranes where hydrophobic interactions between polyelectrolytes and initial layers
occur.

60

REFERENCES

61

REFERENCES

1.

Yang Q, Adrus N, Tomicki F, Ulbricht M.
Composites of functional polymeric
hydrogels and porous membranes. J. Mater. Chem. 2011, 21, 2783-2811.

2.

Bruening ML, Dotzauer DM, Jain P, Ouyang L, Baker GL.
Creation of
functional membranes using polyelectrolyte multilayers and polymer brushes.
Langmuir 2008, 24, 7663-7673.

3.

Gribova V, Auzely-Velty R, Picart C.
Polyelectrolyte multilayer assemblies on
materials surfaces: From cell adhesion to tissue engineering. Chem. Mat. 2012, 24,
854-869.

4.

Salloum DS, Schlenoff JB.
Protein adsorption modalities on polyelectrolyte
multilayers. Biomacromolecules 2004, 5, 1089-1096.

5.

Shiratori SS, Rubner MF.
pH-dependent thickness behavior of sequentially
adsorbed layers of weak polyelectrolytes. Macromolecules 2000, 33, 4213-4219.

6.

von Klitzing R.
Internal structure of polyelectrolyte multilayer assemblies.
Phys. Chem. Chem. Phys. 2006, 8, 5012-5033.

7.

Choi J, Rubner MF.
Influence of the degree of ionization on weak
polyelectrolyte multilayer assembly. Macromolecules 2005, 38, 116-124.

8.

Decher G.
Fuzzy nanoassemblies: Toward layered polymeric multicomposites.
Science 1997, 277, 1232-1237.

9.

Boudou T, Crouzier T, Ren KF, Blin G, Picart C.
Multiple functionalities of
polyelectrolyte multilayer films: New biomedical applications. Adv. Mater. 2010,
22, 441-467.

10.

Decher G, Hong JD.
Buildup of ultrathin multilayer films by a self-assembly
process .1. Consecutive adsorption of anionic and cationic bipolar amphiphiles on
charged surfaces. Makromol. Chem. Macromol. Symp. 1991, 46, 321-327.

11.

Johnston APR, Cortez C, Angelatos AS, Caruso F.
Layer-by-layer engineered
capsules and their applications. Curr. Opin. Colloid Interface Sci. 2006, 11,
203-209.

12.

Liu GQ, Dotzauer DM, Bruening ML.
Ion-exchange membranes prepared
using layer-by-layer polyelectrolyte deposition. J. Membr. Sci. 2010, 354,
198-205.

13.

Dai JH, Baker GL, Bruening ML.

Use of porous membranes modified with
62

polyelectrolyte multilayers as substrates for protein arrays with low nonspecific
adsorption. Anal. Chem. 2006, 78, 135-140.
14.

Borges J, Rodrigues LC, Reis RL, Mano JF.
Layer-by-Layer assembly of
light-responsive polymeric multilayer systems. Adv. Funct. Mater. 2014, 24,
5624-5648.

15.

Raoufi N, Surre F, Sun T, Rajarajan M, Grattan KTV.
Wavelength dependent
pH optical sensor using the layer-by-layer technique. Sens. Actuator B-Chem.
2012, 169, 374-381.

16.

Zhang B, Cui TH.
An ultrasensitive and low-cost graphene sensor based on
layer-by-layer nano self-assembly. Appl. Phys. Lett. 2011, 98, 073116.

17.

Zhang H, Xie AJ, Shen YH, Qiu LG, Tian XY.
Layer-by-layer inkjet printing
of fabricating reduced graphene-polyoxometalate composite film for chemical
sensors. Phys. Chem. Chem. Phys. 2012, 14, 12757-12763.

18.

Tang ZY, Wang Y, Podsiadlo P, Kotov NA.
Biomedical applications of
layer-by-layer assembly: From biomimetics to tissue engineering. Adv. Mater.
2006, 18, 3203-3224.

19.

Lord MS, Foss M, Besenbacher F.
Influence of nanoscale surface topography
on protein adsorption and cellular response. Nano Today 2010, 5, 66-78.

20.

Tang YW, Zhao Y, Wang XG, Lin T.
Layer-by-layer assembly of silica
nanoparticles on 3D fibrous scaffolds: Enhancement of osteoblast cell adhesion,
proliferation, and differentiation. J. Biomed. Mater. Res. Part A 2014, 102,
3803-3812.

21.

Shukla A, Almeida B.
Advances in cellular and tissue engineering using
layer-by-layer assembly. Wiley Interdiscip. Rev.-Nanomed. Nanobiotechnol.
2014, 6, 411-421.

22.

Xia Y, Cheng C, Wang R, Qin H, Zhang Y, Ma L, Tan H, Gu ZW, Zhao CS.
Surface-engineered nanogel assemblies with integrated blood compatibility, cell
proliferation and antibacterial property: Towards multifunctional biomedical
membranes. Polym. Chem. 2014, 5, 5906-5919.

23.

Zelikin AN.
Drug releasing polymer thin films: New era of surface-mediated
drug delivery. ACS Nano 2010, 4, 2494-2509.

24.

Pavlukhina S, Sukhishvili S.
Polymer assemblies for controlled delivery of
bioactive molecules from surfaces. Adv. Drug Deliv. Rev. 2011, 63, 822-836.

25.

Lynn DM.
Peeling back the layers: Controlled erosion and triggered
disassembly of multilayered polyelectrolyte thin films. Adv. Mater. 2007, 19,
63

4118-4130.
26.

Gomes AP, Mano JF, Queiroz JA, Gouveia IC.
Layer-by-layer deposition of
antimicrobial polymers on cellulosic fibers: A new strategy to develop bioactive
textiles. Polym. Adv. Technol. 2013, 24, 1005-1010.

27.

Lee SW, Kim BS, Chen S, Shao-Horn Y, Hammond PT.
Layer-by-Layer
assembly of all carbon nanotube ultrathin films for electrochemical applications. J.
Am. Chem. Soc. 2009, 131, 671-679.

28.

Leite AJ, Sher P, Mano JF.
Chitosan/chondroitin sulfate multilayers as
supports for calcium phosphate biomineralization. Mater. Lett. 2014, 121, 62-65.

29.

Martins GV, Mano JF, Alves NM.
Nanostructured self-assembled films
containing chitosan fabricated at neutral pH. Carbohydr. Polym. 2010, 80,
570-573.

30.

Liu AH, Abbineni G, Mao CB.
Nanocomposite films assembled from
genetically engineered filamentous viruses and gold nanoparticles:
Nanoarchitecture- and humidity-tunable surface plasmon resonance spectra. Adv.
Mater. 2009, 21, 1001-1005.

31.

Nam KT, Kim DW, Yoo PJ, Chiang CY, Meethong N, Hammond PT, Chiang
YM, Belcher AM.
Virus-enabled synthesis and assembly of nanowires for
lithium ion battery electrodes. Science 2006, 312, 885-888.

32.

Sakr OS, Borchard G.
Encapsulation of enzymes in Layer-by-Layer (LbL)
structures: Latest advances and applications. Biomacromolecules 2013, 14,
2117-2135.

33.

Baek H, Lee C, Lim KI, Cho J.
Resistive switching memory properties of
layer-by-layer assembled enzyme multilayers. Nanotechnology 2012, 23, 155604.

34.

Kim Y, Kook K, Hwang SK, Park C, Cho J.
Polymer/perovskite-type
nanoparticle multilayers with multielectric properties prepared from ligand
addition-induced Layer-by-Layer assembly. ACS Nano 2014, 8, 2419-2430.

35.

Xi Q, Chen X, Evans DG, Yang WS.
Gold nanoparticle-embedded porous
graphene thin films fabricated via Layer-by-Layer self-assembly and subsequent
thermal annealing for electrochemical sensing. Langmuir 2012, 28, 9885-9892.

36.

Jewell CM, Lynn DM.
Multilayered polyelectrolyte assemblies as platforms
for the delivery of DNA and other nucleic acid-based therapeutics. Adv. Drug
Deliv. Rev. 2008, 60, 979-999.

37.

Lichter JA, Van Vliet KJ, Rubner MF.
Design of antibacterial surfaces and
interfaces: Polyelectrolyte multilayers as a multifunctional platform.
64

Macromolecules 2009, 42, 8573-8586.
38.

Dragan ES, Bucatariu F, Hitruc G.
Sorption of proteins onto porous
single-component poly(vinyl amine) multilayer thin films. Biomacromolecules
2010, 11, 787-796.

39.

Mendelsohn JD, Yang SY, Hiller J, Hochbaum AI, Rubner MF.
Rational
design of cytophilic and cytophobic polyelectrolyte multilayer thin films.
Biomacromolecules 2003, 4, 96-106.

40.

Harris JJ, Bruening ML.
Electrochemical and in situ ellipsometric investigation
of the permeability and stability of layered polyelectrolyte films. Langmuir 2000,
16, 2006-2013.

41.

Miller MD, Bruening ML.
Correlation of the swelling and permeability of
polyelectrolyte multilayer films. Chem. Mat. 2005, 17, 5375-5381.

42.

Ma YD, Dong JL, Bhattacharjee S, Wijeratne S, Bruening ML, Baker GL.
Increased protein sorption in poly(acrylic acid)-containing films through
incorporation of comb-like polymers and film adsorption at low pH and high ionic
strength. Langmuir 2013, 29, 2946-2954.

43.

Dai JH, Bao ZY, Sun L, Hong SU, Baker GL, Bruening ML.
High-capacity
binding of proteins by poly(acrylic acid) brushes and their derivatives. Langmuir
2006, 22, 4274-4281.

44.

Mendelsohn JD, Barrett CJ, Chan VV, Pal AJ, Mayes AM, Rubner MF.
Fabrication of microporous thin films from polyelectrolyte multilayers. Langmuir
2000, 16, 5017-5023.

45.

Ball V, Hubsch E, Schweiss R, Voegel JC, Schaaf P, Knoll W.
Interactions
between multivalent ions and exponentially growing multilayers: Dissolution and
exchange processes. Langmuir 2005, 21, 8526-8531.

46.

Jomaa HW, Schlenoff JB.
Accelerated exchange in polyelectrolyte multilayers
by "catalytic" polyvalent ion pairing. Langmuir 2005, 21, 8081-8084.

65

Part of this chapter is adapted from our published paper in Langmuir with permission of
the American Chemical Society (Bhattacharjee, S., Dong, J., Ma, Y., Hovde, S., Geiger,
J., Baker, G., and Bruening, M. L. Langmuir, 2012, 28, 6885-6892.). Professor Jeffery
McCutcheon at the University of Connecticut provided the electrospun nylon membranes,
and Salinda Wijeratne prepared the newly designed polyelectrolyte (manuscript is in
preparation).

Chapter 3. Purification of His-tagged protein using functionalized
membranes

Chapter 2 examined protein adsorption in model films on flat substrates. This
chapter describes the adsorption of poly(acrylic acid) (PAA)/polyethyleneimine
(PEI)/PAA films in porous membranes and subsequent derivatization of these films to
capture His-tagged protein. His-tagged proteins bind to Ni2+-nitrilotriacetate (NTA)
complexes appended to the PAA/PEI/PAA films. (Note that a previous study showed
slightly higher lysozyme binding when polyelectrolyte films contain PEI rather than
PAH (1), so this work employs PEI as a polycation.) We investigate both metal-ion
binding and protein binding and compare the recovery and specificity of protein capture
to purification with commercial resins. To simplify the membrane modification process
and decrease cost, Salinda Wijeratne of our group developed a new polymer that I use to
modify the membranes for His-tagged protein purification.

66

3.1. Introduction

Since the approval of recombinant insulin as a therapeutic agent in 1982 (2), the
global therapeutic protein market has expanded rapidly with the continuous growth of
biotechnologies (3, 4). However, purification of proteins remains a bottleneck in their
production because of the complexity of protein mixtures. Affinity adsorption of target
proteins to ligands immobilized in packed columns is probably the most powerful step in
protein purification (5, 6), and this method is also vital to obtaining purified protein for
fundamental research.
To facilitate isolation of target proteins, affinity tags (usually peptides or fusion
proteins) are attached to these proteins through gene engineering (7). Such tags include
polyhistidine (His) (8), the octapeptide FLAG (9), glutathione S-transferase (10) and
maltose binding protein (11). Among these tags, His-tag is the most common because it
often does not alter protein properties, and purification is relatively inexpensive (12). The
His-tag, usually 6 histidine residues, chelates to immobilized metal ions complexed on
resins (8). However, resin-based protein purification requires relatively long separation
times due to the slow diffusion of macromolecules to binding sites in nanopores, and
slow purification is especially troublesome for unstable proteins (5, 13-15).
As summarized in Chapter 1, porous membranes modified with affinity ligands
overcome some of the limitations of resin-based affinity techniques. For His-tagged
protein purification, the Bruening group successfully developed affinity membranes
through growth of polymer brushes in membrane pores (16-21). This chapter aims to
create His-tagged protein-purification membranes using layer-by-layer adsorption of
67

PEMs, which is much easier than surface-initiated polymerization. Based on our
previous studies, membranes containing PAA/PEI/PAA films deposited at pH 3 provide
high lysozyme capture even when proteins spend only 35 ms in membranes, and the
membranes are stable for at least 6 repetitive purifications (1, 22). Thus, below I examine
metal-ion

binding

and

His-tagged

protein

purification

using

Ni2+-

or

Cu2+-NTA-functionalized PAA/PEI/PAA films. The work also examines adsorption of
PAA/PEI/PAA-NTA-Cu2+ on mats of electrospun nylon fibers. Finally I investigate
protein binding to PEMs containing poly(2,2-(5-acrylamido-1-carboxypentylazanediyl)
diacetic acid) (PNTA), a new polymer synthesized by Salinda Wijeratne, to simplify the
membrane fabrication.

3.2. Experimental section
3.2.1. Materials
Hydroxylated nylon (LoProdyne® LP, Pall, 1.2 μm pore size, 110 µm thick) and
electrospun nylon fibers membranes (obtained from Professor Jeffery McCutcheon at the
University of Connecticut) were cut into 25 mm-diameter discs prior to use. Unless
specified, all proteins and chemicals were obtained from Sigma-Aldrich. Coomassie
protein assay reagent (Thermo Scientific), Histidine6-tagged Ubiquitin (HisU, human
recombinant, Enzo Life Sciences), Concanavalin A from Canavaliaensiformis (Con A,
Jack bean), albumin from chicken egg white (Ovalbumin); lysozyme from chicken egg
white; bovine serum albumin (BSA), and β-Lactoglobulin B from Bovine milk were used
as received. His-tagged COP9 signalosome complex subunit 8 (CSN 8) was

68

overexpressed in BL21DE3 cells by Stacy Hovde in Professor James Geiger’s group.
Buffers were prepared using analytical grade chemicals and deionized water (Milli-Q,
18.2 MΩ cm). Polyethyleneimine (PEI, branched, Mw = 25,000), poly(sodium
4-styrenesufonate) (PSS, Mw = 70 000), poly(acrylic acid) (PAA, Mw = 90,000, 25%
aqueous

solution,

Polysciences),

TWEEN-20

N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide
N-Hydroxysuccinimide

(NHS),

and

hydrochloride

Nα,Nα-Bis(carboxymethyl)-L-lysine

surfactant,
(EDC),
Hydrate

(Aminobutyl NTA) were used without further purification. PNTA was synthesized by
Salinda Wijeratne and will be described in his dissertation.

3.2.2. Membrane modification

Membrane discs were cleaned for 10 min with UV/ozone and placed in a homemade
Teflon holder (similar to an Amicon cell) that exposed 3.1 cm2 of external membrane
surface area. Subsequently, a 5-mL solution containing PAA (0.02 M with 0.5 M NaCl)
was circulated through the membrane for 20 min at a flow rate of 1 mL/min using a
peristaltic pump (Figure 3.1). Following a literature protocol, additional polycation (PEI,
2 mg/mL without NaCl in the deposition solution) and polyanion (PAA, with 0.5 M
NaCl in the deposition solution) layers were deposited similarly. After adsorption of each
polyelectrolyte, 20 mL of water was passed through the membrane at the same flow rate.
The pH values of PAA, PAH, and PEI deposition solutions were adjusted to 3 with 1 M
NaOH or 1M HCl for PAA/PEI/PAA films. For the PNTA-containing PEMs, the
concentration of polyelectrolytes in deposition solutions was 1 mg/mL. To optimize

69

metal-ion binding, we investigated adsorption with and without 0.5 M NaCl in the
polyelectrolyte solution. We also employed PNTA, PSS and PAA as the first layer
attached to the membrane to initiate the film growth. In some cases, to grow thicker films
in the membrane, the pH of polyanion solutions, including PNTA, PSS and PAA, was set
to 3, whereas the PAH solution was adjusted to pH 3 or pH 9. The adsorption and rinsing
process was the same as for PAA/PEI/PAA films. Membrane hydraulic permeabilities
were determined as described previously by measuring the mass flux through the
membrane at fixed pressures (16).

Figure 3.1. Apparatus for membrane modification and protein-binding studies.
Drawn by Zhefei Yang.
To derivatize PAA side chains in adsorbed films, 10 mL of 0.1 M NHS, 0.1 M EDC
in water was circulated through the membrane for 1 h prior to rinsing with 20 mL of
70

deionized water and 20 mL of ethanol. Subsequently, 10 mL of aqueous aminobutyl
NTA (0.1 M, adjusted to pH 10.2 with NaOH) was circulated through the NHS-modified
substrate for 1 h followed by rinsing with 20 mL of water.

Finally, the NTA-Cu2+ (or

Ni2+) complex was formed by circulating 10 mL of aqueous 0.1 M CuSO4 (or NiSO4)
through the membrane for 1 h followed by rinsing with 20 mL of water. The substrate
was dried with N2 prior to protein binding.

3.2.3 Metal-ion binding and leaching
After derivatization of PEMs with Ni2+ or Cu2+, bound metal ions were eluted with
two 5-mL aliquots of 0.1 M Ethylenediaminetetraacetic acid (EDTA) (pH 7.6). The
concentrations of metal ions were determined with atomic absorption spectroscopy (AAS)
using calibration curves with metal-ion concentrations ranging from 0 ppm to 10 ppm in
0.1 M EDTA. The metal-ion binding capacity was calculated by dividing the mass of
eluted metal ions by the volume of membrane, which is about 0.035 cm3 (the membranes
are 110 μm thick and have an exposed area of 3.1 cm2). To study the metal-ion leaching,
the membrane was washed sequentially with 5 mL of four different buffers and 0.1 M
EDTA (pH 7.4). The buffer compositions were: binding buffer- 20 mM phosphate buffer
with 0.3 M NaCl and 10 mM imidazole; washing buffer I- 20 mM phosphate buffer with
0.15 M NaCl and 0.1% Tween-20; washing buffer II- 20 mM phosphate buffer with 0.15
M NaCl and 45 mM imidazole; and elution buffer- 20 mM phosphate buffer with 0.5 M
NaCl and 0.3 M imidazole. The pH of all buffers was 7.4. The samples were diluted
5-fold with 0.1 M EDTA before the metal-ion concentration was determined with AAS.

71

3.2.4 Protein biding

A solution of lysozyme (0.3 mg/mL) in 20 mM phosphate buffer (pH 7.4) was
pumped through the modified membrane at a flow rate of 1 mL/min, and the permeate
was collected for analysis at specific time intervals. Subsequently, the membrane was
rinsed with 20 mL of washing buffer I followed by 20 mL of phosphate buffer. The
protein was then eluted using 5-10 mL of 20 mM phosphate buffer (pH 7.4) containing 1
M KSCN. Unless specified, flow rates were 1 mL/min.
Con A solutions (0.3 mg/mL) were prepared in 20 mM phosphate buffer at pH
6because this protein is not stable at pH 7.4. For Con A, washing buffer I and elution
buffer were also adjusted to pH 6. The membranes were loaded with buffered protein
solution, rinsed with 20 mL of washing buffer I followed by 20 mL of phosphate buffer
at pH 6, and eluted with 8-9 mL of imidazole elution buffer (see above). The
concentrations of protein in loading, rinsing, and eluate solutions were determined using
a Bradford assay (16).

3.2.5. Purification of HisU from a model protein mixture

To test protein binding specificity and recovery, we prepared a solution containing
0.05 mg/mL (each) of His-U, Con A, BSA, Ovalbumin, and β-Lactoglobulin B in 20 mM
phosphate buffer (pH 7.4). Ten mL of this protein solution was passed through a
PAA/PEI/PAA-NTA-Ni2+-modified membrane at 1.5 mL/min after the membrane was
equilibrated with 20 mL of phosphate buffer. Subsequently, the membrane was washed
with 20 mL of washing buffer I and 20 mL of phosphate buffer. The bound protein was

72

initially eluted with 10 mL (2 mL for each aliquot) of elution buffer to study the elution
efficiency, but the first aliquot contained more than 95% of the bound protein. The purity
of eluted protein was examined by Polyacrylamide gel electrophoresis (SDS-PAGE, 4-20%
gradient gel from Bio-Rad with standard Coomassie blue staining protocols).

3.2.6. Protein isolation from a cell extract

At a flow rate of 1.5 mL/min, 3 mL of lysate supernatant diluted 1:2 using binding
buffer was passed through a PAA/PEI/PAA-NTA-Ni2+-modified membrane that was
previously equilibrated with lysate buffer (20 mM pH 8 phosphate buffer containing 10
mM imidazole and 300 mM NaCl). To be consistent with the lysis buffer, all the binding,
washing and elution buffers were adjusted to pH 8.0. After washing with 20 mL of
washing buffer I, 20 mL of washing buffer II and 20 mL of phosphate buffer at a flow
rate of 5 mL/min, protein elution and gel electrophoresis followed the same procedure
for HisU purification.
To compare the membrane to commercial products, CSN 8 was also purified with a
spin column packed with HisPur Ni-NTA resins from Thermo Scientific (0.5 mL resins
in each 1.5 mL column). Before use, the column was equilibrated with 1 mL of binding
buffer for 10 min andcentrifuged for 2 min to remove the binding buffer. Subsequently, 3
mL of 3-fold diluted cell extract was loaded on the column in two steps (1.5 mL for each
loading). In each loading step, the column was incubated for 10 min before centrifuging
for 2 min. The column was then sequentially washed with 2 aliquots of washing buffer I,
washing buffer II and binding buffer before the protein was eluted with 2 aliquots of

73

elution buffer. As with the loading, 10 min of incubation and 2 min of centrifugation
were applied for each washing and elution step. The whole process took about 3 hours
including the protein loading, washing and elution.

3.3. Results and discussion

Membranes are attractive for rapid protein purification, but they often have a low
binding capacity. Prior work showed that adsorption of PEMs in membrane pores greatly
increases lysozyme-binding capacities (1). However, the PEM must not plug membrane
pores.

The

following

four

subsections

discuss

the

permeability

of

PAA/PEI/PAA-modified membranes and derivatization of these membranes for capture
of His-tagged protein from cell extracts. Subsequent subsections examine modification
of electrospun nylon fibers as well as adsorption of a NTA-containing polymer to
simplify membrane modification.

3.3.1. Permeability of membranes with PAA/PEI/PAA films

Chapter 2 showed that (PAH/PAA)n films can serve as coatings with high
lysozyme-binding capacities. Bhattacharjee and coworkers used PAA/PAH/PAA and
PAA/PEI/PAA films to modify hydroxylated nylon membranes and found that the
lysozyme binding capacity was higher with PEI as the polycation (1). As with films on
flat surfaces, adsorption pH plays an important role in the properties of
PAA/PEI/PAA films (23). With an adsorption pH of 3, membranes modified with
PAA/PEI/PAA captured over 100 mg of lysozyme per cm3 of membrane, and the
lysozyme breakthrough curve showed little dependence on flow rate (1). Based on those
74

promising results, much of the research in this chapter examines membranes modified
with PAA/PEI/PAA films deposited at pH 3.
Plugging of pores is always a potential problem when modifying membranes by
adsorption. Moreover, directly quantifying the amount of thin film adsorption in highly
porous substrates is difficult. Nevertheless, permeabilities give an indication of changes
in membrane pore size distribution due to modification (Table 3.1). The permeability of
membranes to pH 7.4 phosphate buffer (20 mM) decreased from 78 to 20 mL/(cm2 min
atm) when comparing a bare membrane and a membrane containing a PAA/PEI/PAA
film.

Although this is a significant decline in permeability, rapid flow through the

membrane can still occur using a simple peristaltic pump. In contrast, the permeability of
PAA/PEI/PAA-modified membranes to deionized water after treatment with buffer is <1
mL/(cm2 min atm). This suggests that at low ionic strength (deionized water), unscreened
electrostatic repulsions lead to high film swelling. Rinsing membranes with 2.7 mM HCl
protonates –COOH groups, and the presumed resulting collapse of polymers restores
water permeability to 96 mL/(cm2 min atm). The lower permeability in pH 7.4 buffer
than in deionized water after HCl treatment is consistent with deprotonation of the
polymer at neutral pH, and these results are similar to prior studies of pH-responsive
membranes (24-26).
We roughly estimated the pore size under each condition using the Hagen-Poiseuille
equation, which states that the volumetric flux is proportional to d4, where d is the
diameter of the open fraction of the pore (27). Assuming that the bare membrane has a
pore size of 1.2 μm, the pore diameter decreased to ~1 μm after PAA/PEI/PAA

75

adsorption and then declined to 0.21 μm after exposure to buffer and then water. As
discussed above, contact with 2.7 mM HCl opens the pore, presumably due to minimal
film swelling in the protonated state. Water flux also dropped (relative to the membrane
treated with HCl) after exposure to 1 M KSCN in phosphate buffer. This demonstrates
that the film is stable in high salt concentrations. A subsequent exposure to 2.7 mM HCl
again restores flux.
Table 3.1. Water or buffer permeabilities of bare and PAA/PEI/PAA-modified
nylon membranes in sequential measurements. Buffer: 20 mM phosphate buffer
(pH 7.4); KSCN: 1M KSCN in 20 mM phosphate buffer (pH 7.4) and HCl: 2.7 mM
HCl. The pressure was maintained at 0.68 atm and the unit for permeability is
mL/(cm2 min atm). The drop in peremeability is calculated with respect to the water
permeability of the bare membrane.
Bare

PAA/PEI/PAA modified membrane

membrane

Water

water

Buffer

Water

Water

Water

after

after

afer

after

buffer

HCl

KSCN

HCl

rinse

rinse

rinse

rinse

average

115

77.8

58.9

20.0

0.1

95.8

0.1

87.6

drop (%)

0.0

27.7

48.6

82.5

99.9

16.4

99.9

23.5

1.20

1.09

1.02

0.78

0.21

1.15

0.23

1.12

pore
size/μm
a

Buffer

Water

The table shows data for one bare and one modified membrane. Corresponding

experiments on other membranes showed less than 20% deviation from these values.

3.3.2. Functionalization of membranes with NTA and studies of metal-ion binding

During purification, His-tags chelate metal ions in complexes immobilized on a solid

76

support (8). Although PAA contains –COOH groups, isolated carboxylates do not bind
metal ions strongly. Thus, to effectively create immobilized metal-ion complexes, we
anchored aminobutyl NTA to PAA/PEI/PAA films using the strategy in Figure 3.2.
Films deposited at pH 3 contain a high density of free (not paired with ammonium
groups) –COOH groups that are available for reaction with aminobutyl NTA through
NHS/EDC coupling (1, 16, 18). The immobilized NTA strongly complexes ions such as
Cu2+ or Ni2+.

77

Figure 3.2. Schematic illustration of PAA/PEI/PAA derivatization using NTA-Ni2+.
RT: room temperature. Reproduced from reference 1 with permission of the American
Chemical Society.
PAA/PEI/PAA-NTA-modified membranes have binding capacities of 18 mg/cm3
and 9.5 mg/cm3 for Cu2+ and Ni2+, respectively. However, in addition to the formation of
NTA-metal ion complexes, weak Cu2+ and Ni2+ binding to underivatized –COOH groups
may complicate the interpretation of metal-ion binding data. Thus, we compared Cu2+ or
Ni2+ binding to PAA/PEI/PAA and PAA/PEI/PAA-NTA membranes. For both Cu2+ and
Ni2+, the binding capacities for the NTA-derivatized membranes were ~30% higher than

78

those for underivatized membranes, but significant binding occurs to the PAA/PEI/PAA
film. Derivatization will likely decrease binding to free –COO- groups, but it still occurs.
To distinguish between metal-ion binding to –COO- and NTA groups, I tried to
selectively elute the metal ions from the membranes using buffers that contain different
concentrations of imidazole. Table 3.2 shows the percentage of Cu2+ or Ni2+ (out of the
total immobilized Cu2+ or Ni2+) leached in different solutions. In the case of Cu2+, about
10% more Cu2+ was eluted from membrane without NTA by the 10 mM imidazole in
binding buffer but the opposite trend occurred in the 45 mM imidazole in washing buffer
II. As a result, the percentage of Cu2+ eluted in EDTA was almost the same for both
membranes. For Ni2+, the high percentage of ions leached by the binding buffer from the
PAA/PEI/PAA-NTA membrane clearly demonstrates that Ni2+ also binds to –COOH
groups. We tried unsuccessfully to selectively elute Ni2+ bound to underivatized –COOH
groups, but the percentages of leached Ni2+ ions were similar for PAA/PEI/PAA-NTA
and PAA/PEI/PAA modified membranes. Low metal-ion leaching is important to avoid
contaminating protein solutions, so we designed a new polyelectrolyte, PNTA, discussed
in section 3.3.6

79

Table 3.2. Cu2+ or Ni2+ leaching from membranes

modified with

PAA/PEI/PAA-NTA and PAA/PEI/PAA. The numbers in the table represent the
percentage of metal-ion loss (relative to the total loaded metal) in each solution.
Five mL of each solution was passed through the membrane during the leaching.
Membranes
PAA/PEI/PAA
-NTA (%)
PAA/PEI/PAA
(%)

Metal
ions
Cu2+
Ni2+
Cu2+
Ni2+

Binding
buffer
17.1
55.9
26.8
56.4

Washing
buffer I
6.8
2.3
5.5
3.4

Washing
buffer II
26.6
9.6
14.9
10.1

Elution
buffer
31.8
8.9
35.0
14.1

0.1 M
EDTA
17.7
23.2
13.6
16.0

3.3.3. Purification of HisU from model protein mixtures
To demonstrate the selectivity of PAA/PEI/PAA-Ni2+-modified membranes for
His-tagged protein capture, we first separated HisU from a mixture of HisU, Con A,
BSA, Ovalbumin, and β-Lactoglobulin B. These model proteins, except HisU, do not
have His-tags and serve as contaminating proteins in this experiment. SDS-PAGE
analysis of the mixed-protein solution exiting the membrane suggests successful removal
of HisU (Figure 3.3 lane 3), whereas the eluate shows only bands due to HisU (Figure
3.3 lane 4, note that even the as-received HisU shows two bands). Thus, the membranes
are highly selective for capturing HisU. Five aliquots of 2 mL elution were used to elute
the bound protein, but the elution mainly occurred during the first 2 mL and no bands
appeared for the second eluate as shown in Figure 3.3 lane 4 and 5.

80

Figure 3.3. SDS-PAGE analysis (Coomassie blue staining) of HisU purification
from a mixture of BSA, Ovalbumin, Con A, β-Lactoglobulin B, and HisU: lane 1- a
protein ladder; lane 2- the protein solution; lane 3- the protein solution after passage
through a PAA/PEI/PAA-NTA-Ni2+-modified membrane; Lane 4- the 1st eluate from
the membrane, lane 5 –the 2nd eluate from the membrane; Lanes 6 to lane 10 are the
pure model proteins. Two μg of each protein was loaded in each lane. The eluate should
contain 2 μg of HisU if all the protein was captured and eluted.

The purity of the eluted HisU allowed us to use Bradford assays to quantify the
recovery from the membrane purification. (These assays examine changes in Coomassie
blue absorbance due to interactions with the protein.) Based on the absorbance of a series
of standard solutions of HisU in elution buffer and the absorbance of the eluate solution,
the recovery was 95 ± 3%. Thus, less than 10% of the protein was lost in the purification
process.

81

3.3.4. Purification of His-tagged CSN 8 from cell extracts

To further demonstrate the high selectivity and potential biological applications of
PAA/PEI/PAA-NTA-Ni2+-modified membranes, we purified His-tagged CSN 8 protein
from a whole-cell extract. Figure 3.4 shows the SDS-PAGE analysis of the cell extract
(lane 2) and purification of His-tagged CSN8 using a spin column packed with
Ni2+-resin or a PAA/PEI/PAA-NTA-Ni2+-modified membrane. Remarkably, the
membrane eluate contains only one strong band with a purity of above 95% based on the
darkness of the band. The elution (lanes 7 and 8) from the resin-containing spin column
is less efficient. The first eluate from the spin column shows clear bands for
contaminating proteins, and the elution was not complete. Additionally, compared to the
commercial Ni2+ spin columns, the membrane purification process takes less time. A
complete membrane purification requires less than 20 min, including passing cell lysate
through the membrane, 3 washes using different buffers and elution, while for the spin
column, the whole process takes about 3 hours. This membrane is in the final stage of
commercialization, and we expect products to appear in 2015. With the commercial
membrane holder, the whole process only requires 5 min.

82

Figure 3.4. SDS-PAGE analysis (Coomassie blue staining) of His-tagged CSN8
purified from a cell extract: lane 1- a protein ladder; lane 2- a cell extract form
BL21DE3 cells with overexpressed His-tagged CSN8 protein; lane 3- the cell extract after
passing through a spin column packed with Ni2+-resins; lane 4- the cell extract after
passing through a PAA/PEI/PAA-NTA-Ni2+-modified membrane; Lane 5-column wash;
Lane 6-membrane wash; Lane 7- the 1st eluate from the spin column; lane 8- the 2nd
eluate from spin column, lane 9- the 1st eluate from the membrane and lane 10-the 2nd
eluate from the membrane.
3.3.5. Modification of electrospun fibers
Mats of electrospun nylon fibers have a high porosity and could potentially serve as
membrane supports with high permeability and protein-binding capacities. Thus, we
experimented with layer-by-layer modification of nylon fibers electrospun on a
poly(ethylene terephthalate) (PET) support. Provided by professor Jeffery McCutcheon,
the mats contain 30 μm of spun fiber on a 30μm-thick PET support. The modification
procedure was the same as that for the nylon membranes described above. Lysozyme

83

binding

to

PAA/PEI/PAA-modified

fibers

and

Con

A

binding

to

PAA/PEI/PAA-NTA-Cu2+-modified fibers served as metrics for the performance of the
modified fibrous mats.

3.3.5.1. Lysozyme binding

PAA/PEI/PAA multilayers deposited at pH 3 on the electrospun nylon membrane
contain a high density of –COOH groups that bind lysozyme through ion-exchange.
Table 3.3 and Figure 3.5 show results for lysozyme binding to a control electrospun
membrane, the modified electrospun membrane, and a modified PET membrane support.
The breakthrough curves in Figure 3.4 clearly show the most extensive binding to the
PAA/PEI/PAA-modified nylon fiber membrane, and the lysozyme binding capacity in
this membrane is 115 mg/cm3 of membrane (the binding capacity was calculated only
using the thickness of the nylon fibers which is 30μm), similar to values for traditional
nylon membranes with 1.2 μm pores (1). Moreover, the low binding capacities for the
unmodified membrane and the modified PET support indicate that the binding mainly
occurs to the PAA/PEI/PAA films on the electrospun fibers.
Table 3.3.

Lysozyme binding capacities of electrospun nylon fiber membranes with

and without PAA/PEI/PAA films and of a PAA/PEI/PAA-modified PET support that
was detached from the nylon membrane.
unmodified
membrane

PAA/PEI/PAA- PAA/PEI/PAAmodified
modified PET
membrane
layer

binding capacity
4
115±5*
7
(mg/mL)
*The ± value represents the standard deviation of binding capacities for three different
membranes.
84

Concentration (mg/mL)

0.35
0.3
0.25
0.2
0.15

umodified membrane
PET support

0.1

modified membrne

0.05
0
0

10
20
Permeat Volume(mL)

30

Figure 3.5. Lysozyme breakthrough curves during passage of 0.3 mg/mL of
lysozyme through PET-supported nylon fiber membranes with and without
modification by a PAA/PEI/PAA film and through a PAA/PEI/PAA-modified PET
support. The PAA/PEI/PAA films were deposited at pH 3 from 0.02 M PAA solutions
containing 0.5 M NaCl and 2 mg/mL PEI solutions. The lysozyme feed solution
contained 20 mM phosphate buffer at pH 7.4.

3.3.5.2. Con A binding
Derivatization of the electrospun fibers with PAA/PEI/PAA-NTA-Cu2+ allows
capture of Con A, a protein that has 4 subunits, each of which contains 6 histidine
residues. These residues interact with Cu2+ ions loaded in the membrane. The nylon fiber
membrane gave a Con A binding capacity of 86±12 mg/cm3 of membrane (Figure 3.6).
Again this is similar to the binding capacity for traditional spongy nylon membranes
(pore diameter of 1.2 μm) modified with PAA/PEI/PAA-NTA-Cu2+ (1).
85

Concentration (mg/mL)

0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0

5
10
15
20
Permeate Volume (mL)

25

Figure 3.6. Breakthrough curve during passage of 0.3 mg/mL Con A through an
electrospun nylon membrane modified with PAA/PEI/PAA-NTA-Cu2+. The binding
solution contained 20 mM phosphate buffer at pH 6.

3.3.5.3 Challenges with electrospun membranes

Unfortunately, large flux drops occurred after modification of the electrospun fibers
(Table 3.4). Compared to a bare membrane, the adsorption of polyelectrolytes decreased
the water flux 80%. This flux drop mainly stems from modification of the membrane
fibers rather than modification of the PET support layer, as the flux through the modified
PET support still has high flux after modification. Scanning electron microscope (SEM)
images (Figure 3.7) suggest the high flux drop in modified membranes results from
PEMs that span gaps between fibers. Even worse, membranes modified with
PAA/PEI/PAA-NTA-Cu2+ are almost impermeable. The flow rate dropped to about 0.1
mL/min when passing a Con A solution through such membranes using a peristaltic
pump.

86

Table 3.4. Water permeabilities of an unmodified electrospun nylon fiber
membrane and of similar membranes after modification with PAA/PEI/PAA and
PAA/PEI/PAA-NTA-Cu2+films. (Buffer was not passed through the substrates before
flux measurements). The table also shows the permeability of a PAA/PEI/PAA-modified
PET electrospinning support. The pressure was 0.68 atm and the unit for permeability is
mL/(cm2 min atm). The ± value represents the standard deviation of values for three
different membranes.

flux

Unmodified
electrospun
membrane

PAA/PEI/PAA
-modified PET
support

63.7±1.4

56.2±0.9

PAA/PEI/PAAPAA/PEI/PAA-NTA-Cu2+modified
modified electrospun
electrospun
membrane
membrane
14.1±0.5

87

0.35±0.09

a

b

1 μm

1 μm
d

c

1 μm

1 μm

7 SEM images of electrospun nylon fiber membranes: (a) an unmodified
Figure 3.7.
membrane, (b) an unmodified membrane after water flux measurements, (c) a
membrane modified with a PAA/PEI/PAA film and (d) a membrane modified with
PAA/PEI/PAA-NTA-Cu2+.

3.3.6. Membrane modification with PNTA
Modification of nylon membranes with PAA/PEI/PAA-NTA-Ni2+ provides
membranes with both high capacity and high selectivity for binding His-tagged proteins
(see above). However, the derivatization of these films with aminobutyl NTA is both
inefficient and expensive. The reagents for this derivatization represent more than 95%
of the membrane cost, and most of the aminobutyl NTA does not bind to the membrane.
To simply the membrane fabrication, Salinda Wijeratne synthesized PNTA, which
contains the NTA functionality and can also directly participate in layer-by-layer

88

adsorption to avoid the derivatization step (Figure 3.8).

Figure 3.8. Membrane modification with PNTA-Cu2+-containing films.
Because PNTA is a new polymer for LbL deposition, we examined a variety of
conditions for membrane modification. Based on our previous work (1), we chose PNTA,
PSS, and PAA as initial layers for subsequent growth of PAH/PNTA films. Both PSS
and PAA adhere well to nylon, probably through hydrophobic interactions, and are
effective priming layers. However, an initial layer of PNTA may enhance metal-ion and
protein binding. Table 3.5 shows the Cu2+-binding capacities in membranes with
PNTA-containing films deposited under different conditions. The quantity of Cu2+ bound
by the membranes serves as an indicator of the amount of adsorbed PNTA. We first
directly deposited PNTA-containing films in the membrane using 1 mg/mL PNTA and
PAH solutions at pH 3 with no NaCl in the deposition solution. PEI was not chosen as

89

the polycation because it also binds Cu2+ to interfere Cu2+ binding on PNTA (28, 29).
Cu2+-binding

The

capacity

for

these

PNTA/PAH/PNTA-

and

PNTA(PAH/PNTA)2-modified membranes is <2 mg/mL (Table 3.5, entries 1 and 2).
Moreover, addition of 0.5 M NaCl to polyelectrolyte adsorption solutions did not
increase

the

Cu2+

PNTA(PAH/PNTA)2

binding
films

capacities

even

with

(Table

3.5,

entry

PNTA/PAH/PNTA
3).

In

and

contrast,

PAA/PAH/PAA-NTA-modified membranes (films grown with 0.5 M NaCl ) capture 14
mg of Cu2+/cm3 of membrane (1). The low Cu2+-binding capacity for films containing
PNTA suggests limited polyelectrolyte adsorption in membrane pores. Thus, we
developed strategies to increase the film growth.
Table 3.5. Cu2+ binding capacities of nylon membranes (1.2 µm pores) modified
with PNTA-containing films deposited under different conditions.
Entry
1
2
3
4
5
6

Deposition
conditions
pH 3.0 for PNTA
and PAH,
no NaCl
pH 3.0 for PNTA
and PAH,
0.5 M NaCl
pH 3.0 for PNTA
and PSS,
pH 9 for PAH,
no NaCl

7
8
9
10
11
12

pH 3 for PNTA,
PSS and/or
PAA,
pH 9 for PAH,
0.5 M NaCl

Film

Cu2+ binding
capacity (mg/cm3)

PNTA/PAH/PNTA

1.2

PNTA(PAH/PNTA)2

1.5

PNTA(PAH/PNTA)2

1.4

PNTA/PAH/PNTA)

1.5

PNTA(PAH/PNTA)2

1.8

PSS/PAH/PNTA

1.0

PSS(PAH/PNTA)2

3.8

PSS/PAH/PNTA

1.9

PSS(PAH/PNTA)2

9.4

PAA/PAH/PNTA

8.4

PAA(PAH/PNTA)2

19.0

PAA/PAH

2.1

90

As discussed in chapter 2, pH greatly affects the adsorption of weak polyelectrolytes
such as PNTA and PAH. Adsorption of PAH at pH 9 rather than 3 increased the Cu 2+
binding to PNTA/PAH/PNTA- and PNTA(PAH/PNTA)2-modified membranes (Table
3.5, entries 4 and 5) about 30%. The high deposition pH increases PAH adsorption,
either by decreasing the charge on PAH or increasing the charge on already adsorbed
PNTA.

Either mechanism will require adsorption of more PAH to compensate the

surface charge. Thus all subsequent experiments employed pH 3 for PNTA adsorption
and pH 9 for PAH adsorption.
Our previous work suggests that PSS adsorbs strongly to nylon membranes (30).
Thus, in some experiments we employed PSS as an adhesion layer for subsequent
adsorption of PAH/PNTA. For deposition in the absence of salt, PSS/PAH/PNTA films
give lower Cu2+-binding capacities than PNTA/PAH/PNTA films (compare entries 4
and 6 in Table 3.5), likely because the inner PNTA layer binds Cu2+ and PSS does not.
However, addition of a second PAH/NTA bilayer leads to twice as much binding for the
PSS-initiated film relative to the PNTA(PAH/PNTA)2 film (compare entries 5 and 7 in
Table 3.5). The initial PSS layer probably templates more polyelectrolyte adsorption
than a first layer of PNTA.
Salt added to deposition solutions screens the charges on polyelectrolytes to enhance
their deposition.

Addition of 0.5 M NaCl to all polyelecrolyte deposition solutions

increases Cu2+ binding by a factor of 1.9 (compare entries 6 and 8 in Table 3.5) and 2.5
(compare entries 7 and 9 in Table 3.5) for PSS(PAH/PNTA) and PSS(PAH/PNTA)2
films, respectively. Finally, we examined PAA as the initial layer using deposition from

91

salt-containing solutions. This clearly leads to the highest Cu2+ binding with values
reaching 19.4 mg/cm3 for PAA/(PAH/PNTA)2 films. Because free –COOH groups in
PAA might be responsible for some of this binding, we also determined the Cu 2+ capture
by membranes modified with a PAA/PAH film. Such membranes bind only 2.1 mg
Cu2+/cm3 of membrane (compare entries 11 and 12 in Table 3.5). Even if we take this
adsorption into account, the use of PAA as an initial layer in nylon membranes clearly
increases Cu2+ capture.
For His-tagged protein capture both the quantity of metal-ion binding and the
strength of binding are important, as weakly bound ions may simply leach from the
membrane. Figure 3.9, shows the Cu2+ leaching in the buffers we employ for protein
binding, washing, elution, and metal-ion stripping.

These studies examined

membranes containing PSS(PAH/PNTA)2 and PAA(PAH/PNTA)2 deposited from
solutions containing 0.5 M NaCl because they have high metal-ion binding capacities. As
Figure 3.9 shows, in binding and washing buffers, the PNTA-containing membranes
exhibits much less Cu2+ leaching than PAA/PEI/PAA-NTA-modified membranes. The
PAA/PEI/PAA-NTA membranes lose nearly 20% of their Cu2+ in a binding buffer
containing 10 mM imidazole. Much of this leaching likely includes Cu2+ bound to
underivatized free –COO- groups of PAA. After subsequent washes with washing buffers
I and II, only half of the Cu2+ remains in the PAA/PEI/PAA-NTA-modified membrane.
The PNTA-containing films are more resistant to leaching, even when the films have
PAA as the first layer. For these films, during adsorption of PAH at pH 9, -COOH
groups in PAA will deprotonate and form ion pairs with ammonium groups of PAA.

92

Such groups likely do not bind significant amounts of metal ions, and Cu2+ capture
occurs primarily through NTA groups.
60

Percentage of the total binding (%)

PSS(PAH/PNTA)2
50

PAA/PEI/PAA-NTA

PAA(PAH/PNTA)2

40

30

20

10

0
binding
buffer

washing I washing II
elution
buffer
buffer
buffer
Leaching buffers

0.1 M EDTA

Figure 3.9. Percentage of the bound Cu2+ leached from PSS(PAH/PNTA)2-,
PAA(PAH/PNTA)2- and PAA/PEI/PAA-NTA-modified membranes when passing 5
mL of buffers sequentially through the membranes. Deposition conditions for
PSS(PAH/PNTA)2, PAA(PAH/PNTA)2 are those in footnote c in Table 3.4.
PAA/PEI/PAA-NTA was deposited as described in section 3.2.2.
For all films, extensive leaching occurs in the 0.5 M imidazole solution that elutes
proteins. At high concentration, imidazole effectively competes with NTA for Cu2+ ions
and leaching occurs. Still, the PNTA-containing membranes retain half of their Cu2+
even after passing 0.5 M imidazole through the membrane. Importantly, His-tagged

93

protein purification will employ Ni2+ rather than Cu2+ because Ni2+ has a lower affinity
for imidazole, which should lead to less non-specific adsorption of proteins that contain
histidine residues and not polyhistidine. The relatively low affinity of imidazole for Ni2+
should also decrease leaching in the elution buffer.

3.4. Conclusions

PAA/PEI/PAA multilayers provide a platform for creating derivatizable membranes.
When deposited from pH 3 solutions, PAA/PEI/PAA films swell extensively at neutral
pH because of a high density of free –COO- groups. These groups are also available for
further modification with aminobutyl NTA, which chelates Cu2+ or Ni2+ ions.
Membranes modified with PAA/PEI/PAA-NTA-Ni2+ enable rapid isolation of
His-tagged proteins from cell extracts with > 95% purity and recovery. Spongy nylon
membranes are excellent substrates for PEM modification, so we also examined whether
modification of highly porous mats of electrospun nylon fibers could create highly
permeable membranes with high protein-binding capacities. However, the flux through
the electrospun membranes declined dramatically after modification.
Membranes with PAA/PEI/PAA-NTA-Ni2+ films successfully isolate His-tagged
protein, but metal-ion leaching is extensive. A newly designed polymer that contains
NTA groups, PNTA, greatly simplifies formation of membranes with chelating groups,
and importantly the membranes show less Cu2+ leaching than PAA/PEI/PAA-NTA. In
the future, we will examine the PNTA-modified membranes for Ni2+ binding and
His-tagged protein purification.

94

REFERENCES

95

REFERENCES

1.

Bhattacharjee S, Dong JL, Ma YD, Hovde S, Geiger JH, Baker GL, Bruening ML.
Formation of high-capacity protein-adsorbing membranes through simple
adsorption of poly(acrylic acid)-containing films at low pH. Langmuir 2012, 28,
6885-6892.

2.

Ladisch MR, Kohlmann KL.
1992, 8, 469-478.

3.

Pavlou AK, Reichert JM.
Recombinant protein therapeutics - success rates,
market trends and values to 2010. Nat. Biotechnol. 2004, 22, 1513-1519.

4.

Wurm FM.
Production of recombinant protein therapeutics in cultivated
mammalian cells. Nat. Biotechnol. 2004, 22, 1393-1398.

5.

Kawai T, Saito K, Lee W.
Protein binding to polymer brush, based on
ion-exchange, hydrophobic, and affinity interactions. J. Chromatogr. B 2003, 790,
131-142.

6.

Clarke W, Hage DS.
Clinical applications of affinity chromatography. Sep.
Purif. Rev. 2003, 32, 19-60.

7.

Terpe K. Overview of tag protein fusions: From molecular and biochemical
fundamentals to commercial systems. Appl. Microbiol. Biotechnol. 2003, 60,
523-533.

8.

Hochuli E, Dobeli H, Schacher A.
New metal chelate adsorbent selective for
proteins and peptides containing neighboring histidine-residues. Journal of
Chromatography 1987, 411, 177-184.

9.

Einhauer A, Jungbauer A.
The FLAG (tm) peptide, a versatile fusion tag for
the purification of recombinant proteins. J. Biochem. Biophys. Methods 2001, 49,
455-465.

10.

Smith DB, Johnson KS.
Single-step purification of polypeptides expressed in
Escherichia-Coli as fusions with glutathione s-transferase. Gene 1988, 67, 31-40.

11.

Diguan C, Li P, Riggs PD, Inouye H.
Vectors that facilitate the expression and
purification of foreign peptides in Escherichia-Coli by fusion to maltose-binding
protein. Gene 1988, 67, 21-30.

12.

Lichty JJ, Malecki JL, Agnew HD, Michelson-Horowitz DJ, Tan S.
Comparison of affinity tags for protein purification. Protein Expr. Purif. 2005, 41,
98-105.

Recombinant human insulin. Biotechnol. Prog.

96

13.

Datta S, Bhattacharyya D, Ray PD, Nath A, Toborek M.
Effect of pre-filtration
on selective isolation of Tat protein by affinity membrane separation: Analysis of
flux, separation efficiency, and processing time. Sep. Sci. Technol. 2007, 42,
2451-2471.

14.

Ghosh R.
Protein separation using membrane chromatography: Opportunities
and challenges. J. Chromatogr. A 2002, 952, 13-27.

15.

Bhut BV, Husson SM.
Dramatic performance improvement of weak
anion-exchange membranes for chromatographic bioseparations. J. Membr. Sci.
2009, 337, 215-223.

16.

Anuraj N, Bhattacharjee S, Geiger JH, Baker GL, Bruening ML.
An
all-aqueous route to polymer brush-modified membranes with remarkable
permeabilites and protein capture rates. J. Membr. Sci. 2012, 389, 117-125.

17.

Jain P, Baker GL, Bruening ML.
Applications of polymer brushes in protein
analysis and purification. Annu. Rev. Anal. Chem. 2009, 2, 387-408.

18.

Jain P, Dai J, Grajales S, Saha S, Baker GL, Bruening ML.
Completely
aqueous procedure for the growth of polymer brushes on polymeric substrates.
Langmuir 2007, 23, 11360-11365.

19.

Jain P, Dai JH, Baker GL, Bruening ML.
Rapid synthesis of functional
polymer brushes by surface-initiated atom transfer radical polymerization of an
acidic monomer. Macromolecules 2008, 41, 8413-8417.

20.

Jain P, Sun L, Dai JH, Baker GL, Bruening ML.
High-capacity purification of
his-tagged proteins by affinity membranes containing functionalized polymer
brushes. Biomacromolecules 2007, 8, 3102-3107.

21.

Jain P, Vyas MK, Geiger JH, Baker GL, Bruening ML.
Protein purification
with polymeric affinity membranes containing functionalized poly(acid) brushes.
Biomacromolecules 2010, 11, 1019-1026.

22.

Ma YD, Dong JL, Bhattacharjee S, Wijeratne S, Bruening ML, Baker GL.
Increased protein sorption in poly(acrylic acid)-containing films through
incorporation of comb-like polymers and film adsorption at low pH and high ionic
strength. Langmuir 2013, 29, 2946-2954.

23.

Choi J, Rubner MF.
Influence of the degree of ionization on weak
polyelectrolyte multilayer assembly. Macromolecules 2005, 38, 116-124.

24.

Ito Y, Inaba M, Chung DJ, Imanishi Y.
Control of water permeation by pH and
ionic-strength through a porous membrane haveing poly(carboxylic acid)
surface-grafted. Macromolecules 1992, 25, 7313-7316.

97

25.

Ito Y, Park YS, Imanishi Y.
Visualization of critical pH-controlled gating of a
porous membrane grafted with polyelectrolyte brushes. J. Am. Chem. Soc. 1997,
119, 2739-2740.

26.

Lee D, Nolte AJ, Kunz AL, Rubner MF, Cohen RE.
pH-induced hysteretic
gating of track-etched polycarbonate membranes: Swelling/deswelling behavior
of polyelectrolyte multilayers in confined geometry. J. Am. Chem. Soc. 2006, 128,
8521-8529.

27.

Sutera SP, Skalak R. 1993. The history of poiseuille's law. In Annual review of
fluid mechanics, ed. JL Lumley, M Van Dyke, pp. 1-19.

28.

Kobayashi S, Hiroishi K, Tokunoh M, Saegusa T.
Chelating properties of
linear and branched poly(ethylenimines). Macromolecules 1987, 20, 1496-1500.

29.

Linden JB, Larsson M, Coad BR, Skinner WM, Nyden M.
Polyethyleneimine
for copper absorption: Kinetics, selectivity and efficiency in artificial seawater.
RSC Adv. 2014, 4, 25063-25066.

30.

Liu GQ, Dotzauer DM, Bruening ML.
Ion-exchange membranes prepared
using layer-by-layer polyelectrolyte deposition. J. Membr. Sci. 2010, 354,
198-205.

98

Chapter 4. Enrichment of N-Glycoproteins Using Membranes
Modified with Lectins

Chapter 3 described our development of membranes for His-tagged protein
purification. This chapter presents my efforts to create membranes that capture
N-glycoproteins via binding of glycans to an immobilized lectin, Concanavalin A (Con
A). Thus, I immobilized Con A in membranes using covalent or metal-affinity methods
with either polymer brushes or LbL films.

4.1. Introduction

Post-translational modifications (PTMs) of proteins are chemical modifications that
occur on specific amino acids after protein expression (2, 3). These PTMs can affect
protein structures and regulate catalytic activities, localization, and cell signaling (2, 4-7).
In the human genome, DNA only encodes the 20 essential amino acids (8), but PTMs
enhance the range of protein functions, particularly in response to cellular development
requirements or changes in the physiological environment (2, 8). There are more than
200 types of PTMs and over 80000 specific PTMs have been characterized
experimentally (9-11).
Glycosylation, the attachment of sugar moieties (glycans) to proteins, is one of the
most common PTMs in eukaryotes (12). Protein glycosylation can be divided into two
types: N-glycosylation and O-glycosylation, based on the glycosylation sites and the type
of linkage between the protein and the glycans (13-15). In N-glycosylated proteins,
amide linkages connect N-acetylglucosamines to Asparagine (Asn) residues that must
99

reside in consensus sequences containing Asn-any amino acid-Serine/Threonine (15).
With O-glycosylated proteins, ether linkages attach glycans to Serine or Threonine (16).
A complex series of enzymes catalyze glycan synthesis and their attachment to proteins
(17). In the endoplasmatic reticulum, N-glycosylation on Asn employs more than 23
enzymatic steps in assembly, trimming, and maturation of the carbohydrate groups (18).
Many factors, especially the cell stage affect this complicated process and increase the
heterogeneity of the glycans. For example, at least 5 different glycan structures appear at
60

Asn in bovine Ribonuclease B (19).
Several studies report that N-glycans play a role in physiological and pathological

activities including cell adhesion, migration, signaling, differentiation, and tumor
invasion (20-22). Some aberrant glycans reflect the development and progress of cancers.
For example, E-cadherin, a cell surface protein primarily found in neural tissues and
fibroblasts, is the main epithelial cell-cell adhesion molecule involved in Ca2+-dependent
cell-cell adhesion associated with cancer metastasis. Attachment of N-acetylglucosamine
to a branching position on the N-glycans connected to E-cadherin promotes cancer
progression and metastasis. In contrast, N-glycans with a bisecting conformation may
suppress the cancer metastasis (23).

Therefore, glycoproteins are a promising source of

biomarkers for cancers, including CA125 for ovarian cancer (24) and Her2/neu for breast
cancer (25).
Despite advances in glycoprotein research, exploration of glycosylation on a
proteome scale is challenging due to the low abundance and heterogeneities of the
glycans. In vivo, less than 1% of total proteins are glycosylated (2), so purification of

100

glycoproteins is vital to their identification. Currently, lectin affinity chromatography is
the most common method for enriching glycoproteins (26). Lectin capture in membranes
may simplify glycoprotein isolation due to the advantage of convective transport to
binding sites. Thus in this work, I aimed to develop a Con A-containing membrane for
rapid glycoprotein purification. This chapter investigates different strategies for Con A
immobilization and subsequent glycoprotein binding to immobilized Con A.

4.2. Experimental section
4.2.1. Materials

Hydroxylated nylon (LoProdyne® LP, Pall, 1.2 µm pore size, 110 µm thick) was cut
into 25 mm-diameter discs prior to use. Coomassie protein assay reagent (Thermo
Scientific), Tris(2-(dimethylamino)ethyl)amine (Me6TREN, ATRP Solutions) and other
chemicals from Sigma-Aldrich were used as received. The trichlorosilane initiator
(11-(2-bromo-2-methyl)propionyloxy)-undecyltrichlorosilane)
((poly(2-(trimethylammonium

and

macroinitiator
iodide)ethyl

methacrylate-co-2-(2-bromoisobutyryloxy)ethyl acrylate)) were synthesized according to
literature procedures (27-29). Buffers were prepared using analytical grade chemicals
and deionized (Milli-Q, 18.2 MΩ cm) water.

4.2.2. Initiator adsorption
4.2.2.1. Trichlorosilane initiator

Before the attachment of initiator, the hydroxylated nylon membrane was cleaned

101

with UV/ozone for 15 min. Initiator attachment occurred by circulating 1 mM
trichlorosilane initiator in anhydrous tetrahydrofuran through the membrane for 2 h at a
flow rate of approximately 3 mL/min using a peristaltic pump. This was followed by
sequentially passing 20 mL of tetrahydrofuran, 20 mL of ethanol, and 20 mL of water
through the membrane before drying with a N2 stream.

4.2.2.2. Macroinitiator

The UV/ozone-cleaned membrane was modified by flowing 10 mL of 2 mg/mL
aqueous macroinitiator through the membrane for 10 min at a flow rate of 1 mL/min
using a peristaltic pump (30). Subsequently, 20 mL of water was passed through the
membrane, which was then dried and stored for polymerization.

4.2.3. .Preparation of catalyst and monomer solutions

Atom transfer radical polymerization (ATRP) was performed with a Cu (I)/Cu(II)
catalyst (31). For a 20-mL catalyst solution, 0.0447 g of CuBr2 was dissolved in 20 mL
of

N,N-dimethylformamide (DMF) in a round-bottom flask and the solution was

degassed using 2 freeze-pump-thaw (FPT) cycles on a Schlenk line. Subsequently, 0.184
mL of Me6TREN was added to the mixture of CuBr2 and DMF using a syringe. Two
more FPT cycles were applied immediately to degass the solution. Subsequently, 0.0574
g CuBr was added, and the solution was degassed using 3 FPT cycles. In a glove bag, the
catalyst solution was transferred into 20 mL vials and stored there at room temperature
for

no

more

than

a

week.

The

monomer

solution,

which

contained

2-(methacryloyloxy)ethyl succinate (MES) mixed with 1 M aqueous NaOH (1:1, v/v),
102

was degassed using FPT cycles before polymerization just prior to polymerization (30).

4.2.4. Membrane modification with polymeric films

To prepare polymerization solutions, 10 mL of freshly degassed monomer solution
was mixed with 1 mL of catalyst solution in a glove bag. The catalyst concentration in
the mixture was CuBr2 (1 mM), CuBr (2 mM) and Me6TREN (6 mM). Still in the glove
bag, the monomer/catalyst solution was circulated through an initiator-modified
membrane using a peristaltic pump at a flow rate of 1 mL/min. After the desired
polymerization time, 20 mL of ethanol and 20 mL of water were passed sequentially
through the membrane at the same flow rate. PAA/PEI/PAA-modified membranes were
also employed for Con A immobilization, and these membranes were prepared as
described in section 3.2.2 of chapter 3 and treated the same as membranes containing
poly(MES) for derivatization. Attenuated total reflectance Fourier transform infrared
(ATR-FTIR) spectroscopy (Perkin Elmer Spectrum One Instrument, air background) was
employed to verify the modifications.

4.2.5. Derivatization of polymeric films with Con A
4.2.5.1. Cu2+-affinity binding

The poly(MES) in the membrane was activated by passing an aqueous solution of
0.05

M

N-hydroxysuccinimide

(NHS)/0.05

M

1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride (EDC) through the
membrane for 1 h at 1 mL/min. The membrane was then washed with 20 mL of water

103

and 20 mL of ethanol. After activation, an aqueous solution of 0.1 M
Nα,Nα-bis(carboxymethyl)-L-Lysine hydrate (aminobutyl NTA, pH 10) was circulated
through NHS-modified membrane for 1 h at a flow rate of 1 mL/min followed by a
20-mL water wash. Then 20 mL of 0.1 M aqueous CuSO4 was circulated through the
membrane for 2 hours at the same flow rate followed by washing with 20 mL water. Con
A was immobilized to the poly(MES)-NTA-Cu2+-modified membrane by circulating a
0.3 mg/mL Con A solution in 20 mM phosphate buffer (pH 6) through the membrane for
1 h. Excess Cu2+ sites were blocked with bovine serum albumin (BSA, which does not
bind to Con A) by subsequently circulating a 0.3 mg/mL BSA solution in phosphate
buffer (20 mM, pH 7.4) through the membrane for 15 min (32). The amount of Con A
immobilized on polymer brushes was estimated using a protein Bradford assay to
determine the amount of Con A in the loading solution before and after circulating
through the membrane. The membranes were dried and stored for glycoprotein binding
tests.

4.2.5.2. Covalent derivatization
Similarly to Cu2+-affinity immobilization, the first step is to activate the carboxylic
acid groups on the polymer brushes with 0.05 M NHS/0.05 M EDC for 1 hour. After
rinsing as described above, 3 mL of 0.3 mg/mL Con A solution in 20 mM phosphate
buffer (pH 6) was circulated through the membrane for 4 hours at a flow rate of 1
mL/min followed by 40 mL of phosphate buffer (pH 6). The amount of Con A
immobilized on the polymer brushes was calculated as described in section 4.2.5.1. The

104

membranes were dried with N2 and stored for further analysis.

4.2.5.3. Immobilization of protected Con A

To

preserve

the

glycan

binding

sites,

Con

A

was

protected

with

methyl-α-D-mannopyranoside by incubating 0.3 mg/mL Con A in 20 mM Tris buffer
(pH 6) containing 500 mM NaCl, 5mM MgCl2, 5 mM MnCl2, 5 mM CaCl2 and 200 mM
methyl-α-D-mannopyranoside for 30 min. After membrane activation as described above,
this Con A solution (containing all the reagents including the protecting agent) was
employed to anchor Con A to the NHS-activated membrane and the Cu2+-membrane in
the same way as described in sections 4.2.5.1 and 4.2.5.2. After immobilization, the Con
A was regenerated by washing the membrane with 3 mL of 100 mM Tris buffer
supplemented with 500 mM NaCl (pH 8.5) followed by 3 mL of 100 mM sodium acetate
with 1.0 M NaCl (pH 4.5). This process was performed three times. Prior to tests of
glycoprotein binding, the membranes were equilibrated with 20 mM phosphate buffer at
pH 7.4 (33).

4.2.6. Measurement of protein binding capacities

Bradford Assays were also used to determine protein-binding capacities from
breakthrough curves and protein elutions. In this study, lysozyme (in 20 M phosphate
buffer, pH 7.4), Con A (in 20 mM phosphate buffer, pH 6) and Asialofetuin (in 20 mM
Tris, 0.15 M NaCl, 1 mM MnCl2, and 1 mM CaCl2, pH 7.4) were pumped through
membranes at a flow rate of about 1 mL/min. The permeate solution was collected at
specific time intervals, and 40 mL of phosphate or Tris buffers were used to rinse the
105

membrane before elution. Then, lysozyme was eluted with 20 mM phosphate buffer
containing 1 M KSCN (pH 7.4), Cu2+-chelated Con A was eluted with 20 mM phosphate
buffer containing 0.1 M ethylenediamine tetraacetic acid (EDTA, pH 6) and asialofetuin
was eluted with 20 mM Tris buffer containing 0.5 M methyl-a-D-mannopyranoside (pH
7.4). To determine effluent and eluate concentrations, a series of standard solutions with
concentrations ranging from 0 to the feed concentration for each protein were prepared
and 30 L of protein solution including the standards and samples was diluted into 1.5
mL Coomassie Brilliant Blue solution (used as received) at room temperature (34). The
absorbance of these solutions was determined using a Perkin-Elmer UV/vis (model
Lambda 40) spectrophotometer at 595 nm.

4.3. Results and Discussion
4.3.1. Membrane modification with poly(MES) films

The first step in creating lectin-based glycoprotein-binding membranes is lectin
immobilization. Anchoring through poly(MES) brushes may yield multilayers of
immobilized lectin (27, 30, 34, 35) to give a high capacity in subsequent capture of
glycoproteins. Growth of poly(MES) in membranes begins with initiator attachment, and
Figure 4.1 shows the strategies we employ to introduce the initiating groups. The
trichlorosilane initiator attaches to the hydroxylated nylon membrane via Si-O bonds
(Figure 4.1a). We polymerized MES from the immobilized trichlorosilane initiator for 1
hour, 30 min and 15 min to see if more polymerization would increase protein binding.
However, after 1 hour of polymerization, the polymer clogged the membrane pores. In

106

initial studies of protein capture, we examined lysozyme binding to poly(MES) via ion
exchange. Based on breakthrough curves (Figure 4.2), the 30-min polymerization and
15-min polymerization give similar binding amounts of 85-90 mg/cm3. Note that this is
not the saturation binding capacity, as the effluent lysozyme concentration is still less
than that in the feed at the end of loading. Unfortunately, the flux through the membranes
modified with 30 min of polymerization membrane declined severely during the
lysozyme binding studies. Even with the 15-min polymerization the permeability
declined from 115 mL/(cm2 min atm) (bare membrane) to 2.9 mL/(cm2 min atm) after
lysozyme binding. In our previous studies, we also noticed flux drop when using the
trichlorosilane initiator (35).

107

Figure 4.1. Growth of poly(MES) brushes from nylon membranes using (a) a
trichlorosilane initiator and (b) a macroinitiator. MES was polymerized using a

Cu(I)/Cu(II)/Me6TREN/DMF catalyst at room temperature.

To increase permeability, we employed a macroinitiator to initiate polymerization in
the membranes. The macroinitiator attaches to the pore surface via hydrophobic
interactions (Figure 4.1b) and likely gives less dense polymer brushes than initiation

108

from the trichlorosilane initiator. Using the macroinitiator, Anuraj and coworkers found
that lysozyme binding capacity increased as the following MES polymerization time
increased from 2 to 5 min but decreased for longer polymerization, with a corresponding
flux decreased as a function of time (30). Thus, we use only the 5-min polymerization
time for the macroinitator in this chapter, and under those conditions, the poly(MES)
binds around 107 mg lysozyme/cm3 of membrane (Figure 4.2), although again this is not
the saturation binding capacity. Importantly the flux was still above 35 mL/(cm2 min atm)
after protein sorption. In addition, the macroinitiator interacts adsorbs to the membrane
from aqueous solution in only 10 min and thus is easier to apply than the trichlorosilane
initiator.

Thus,

the

following

discussion

macroinitiator-modified nylon membranes.

109

focuses

on

brushes

grown

from

Concentration (mg/mL)

0.2
macroinitiator 5 min
0.16

trichlorosilane 15 min
trichlorosilan 30 min

0.12

0.08

0.04

0
0

5

10

15

Permeate volume (mL)
Figure 4.2. Lysozyme breakthrough curves during passage of a 0.3 mg/mL
lysozyme solution through nylon membranes modified with poly(MES) brushes
grown from trichlorosilane initiator for 30 or 15 min or macroinitiator for 5 min.
All lysozyme feed solutions contained 20 mM phosphate buffer (pH 7.4). Flow rates
were around 1 mL/min through the membranes (area of 3.1 cm2) modified using the
macroinitiator or the 15-min polymerization from trichlorosilane initiator, but much
smaller with the membrane modified with 30-min of polymerization from the
trichlorosilane initiator because the peristaltic pump could not maintain the flow rate.

4.3.2. Con A immobilization via Cu2+-affinity and subsequent glycoprotein binding

Con A contains four identical subunits, each of which has a binding site for

110

glycoprotein, and is the most common lectin for N-glycoprotein binding (Figure 4.3) (1,
36-38). The presence of Mn2+ and Ca2+ in the N-glycoprotein binding buffer is crucial to
maintain Con A’s binding activity. With regard to immobilization, each Con A subunit
contains six histidine residues that could interact with Cu2+ on a solid support (39).

Figure 4.3. Conformation of Con A (left) and illustration of its glycoprotein
binding sites.

Each subunit contains Mn2+ and Ca2+ binding sites that help maintain

the active conformation. A series of hydrogen bonding on
(circled in green) and hydrophobic interactions with

12

208

Asp,

228

Arg and

14

Asn

Try (circled in brown) lead to

the glycan (1). Redrawn from reference 1 and 38 with permission of Elsevier and the
American Chemical Society, respectively.
In one method to immobilize Con A, we formed Cu2+-NTA complexes in the
membrane by first attaching aminobutyl NTA to the polymer brushes with NHS/EDC
coupling as described in chapter 3. We monitored the reaction with aminobutyl NTA
using ATR-IR spectroscopy (Figure 4.4). Bare nylon membranes have amide I and II
absorbance peaks at 1630 and 1533 cm-1, respectively (Figure 4.4, spectrum a). After

111

polymerization, the new peak at 1720 cm-1 (Figure 4.4, spectrum b) arises from the acid
carbonyl group on poly(MES). To functionalize the polymer brushes with NTA, the –
COOH groups of poly(MES) were firstly activated by NHS. The peaks at 1810 and 1770
cm-1 due to newly formed succininidyl ester groups verify the activation (Figure 4.4,
spectrum c). After reaction with NTA, the absorbances from the succinimidyl ester
disappeared. However, the ATR-IR spectra of the NTA- and NTA-Cu2+-functionalized
membrane (Figure 4.4, spectra d and e), do not show any readily apparent new
absorbances, presumably because they are masked by the strong amide absorbances of
nylon. These amide absorbances also preclude detection of protein binding through
ATR-IR spectroscopy. Therefore, we determined the Con A binding using a
breakthrough curve (Figure 4.5), which indicated a Con A binding capacity of about 69
mg/cm3 of membrane.

112

Figure 4.4. ATR-IR spectra of nylon membranes before and after steps for membrane
functionalization. (a) Nylon membrane; (b) Nylon membrane-poly(MES); (c) Nylon
membrane-poly(MES)-NHS; (d) Nylon membrane-poly(MES)-NTA; (e) Nylon
membrane -poly(MES)-NTA-Cu2+. The membranes were washed with 20 mM phosphate
buffer (pH 7.4) followed by deionized water prior to obtaining the spectra.

113

0.3

Concentration (mg/mL)

0.25
0.2
0.15
0.1
0.05
0
0

5
10
Permeate volume (mL)

15

20

Figure 4.5. Breakthrough curve during passage of a 0.3 mg/mL Con A solution
through a nylon membrane modified with poly(MES)-NTA-Cu2+. The Con A
solution in 20 mM phosphate buffer (pH 6) was passed through the membrane at a flow
rate of 1 mL/min. The membrane had an exposed area of 3.1 cm2.

To examine glycoprotein capture by Con A immobilized via Cu2+ affinity, we
determined the asialofetuin (a glycoprotein binding to Con A) binding capacity for nylon
membranes containing poly(MES)-NTA-Cu2+, poly(MES)-NTA-Cu2+-Con A and
poly(MES)-NTA-Cu2+-Con A/BSA (Figure 4.6) (40-42). For the membrane with
poly(MES)-NTA-Cu2+-Con A, the asialofetuin binding capacity was 40 mg/cm3 of
membrane. Unfortunately, asialofetuin binds to Cu2+ as well as to Con A, and the
binding capacity of the membrane with only poly(MES)-NTA-Cu2+ (no Con A) was 55
mg Asialofetuin /cm3 of membrane. After trying to block excess Cu2+ sites and other
114

non-specific binding sites with BSA, the asialofetuin binding capacity decreased to about
8 mg/cm3 of membrane. The low capacity might stem from steric hindrance to binding
sites. The molecular mass of asialofetuin is 48.4 kDa, whereas that of Con A is 104 kDa.
If each of the four subunits of Con A captured one asialofetuin molecule, the binding
capacity would be about 120 mg asialofetuin/cm3 of membrane based on the
immobilization of 69 mg Con A/cm3 of membrane.
0.3

Concentration (mg/mL)

0.25

0.2

0.15

0.1
membrane-poly(MES)-NTA-Cu-Con A/BSA

0.05

membrane-poly(MES)-NTA-Cu
membrane-poly(MES)-NTA-Cu-Con A

0
0

5

10

15

20

25

Permeate volume (mL)
Figure 4.6. Breakthrough curves during the passage of 0.3 mg/mL asialofetuin
through membranes modified with

membranes with poly(MES)-NTA-Cu2+,

poly(MES)-NTA-Cu2+-Con A or poly(MES)-NTA-Cu2+-Con A/BSA. Asialofetuin
was dissolved in pH 7.4 Tris buffer and passed through the membrane (3.1 cm 2) at 1
mL/min.

115

4.3.3 Con A immobilization via covalent interactions and subsequent glycoprotein
binding

We also immobilized Con A via reaction of some of its lysine residues with the
activated –COOH groups of poly(MES). In this anchoring strategy, the pH of the buffer
plays an important role. Normally, the rate of amide formation would increase at high pH
due to deprotonation of amino groups. However, Con A aggregates at pH values > 7 (43,
44). Thus we attempted to immobilize Con A at pH 6 using a relatively long reaction
time of 4 h. This strategy immobilizes only 10 mg of Con A/cm3, as determined from
decrease in the concentration of Con A in the immobilization solution after
immobilization. Considering Con A’s dimensions of 4.2 ×4×3.9 nm, the binding capacity
indicates a single layer protein binding (45). Moreover, the Asialofetuin binding capacity
is only 5 mg/cm3 of membrane (Figure 4.7). This Asialofetuin binding capacity suggests
that only 25% of the binding sites on Con A are accessible, although this percentage
should be viewed with caution due to uncertainties in determining low binding capacities.
The random covalent linkages may disrupt the protein structure or limit access to the
binding site (46).

116

Concentration (mg/mL)

0.3

0.2

0.1

0
0

4

8
12
16
Permeate volume (mL)

20

Figure 4.7. Breakthrough curve of asialofetuin for membranes with covalently
immobilized Con A. Asialofetuin was dissolved in pH 7.4 Tris buffer at a concentration
of 0.3 mg/mL, and this solution was passed through the membrane at 1 mL/min. The
membrane area was 3.1 cm2.

In an effort to protect the binding site, we modified Con A with
methyl-α-D-mannopyranoside during immobilization. Binding of this sugar to Con A
should prevent formation of covalent bonds to the binding site (33). The presence of
methyl-α-D-mannopyranoside does not decrease the amount of Con A immobilization.
However, after removal of methyl-α-D-mannopyranoside, the asialofetuin binding
capacity

was

the

same

as

that

for

Con

A

immobilized

without

methyl-α-D-mannopyranoside.
All the experiments were repeated using PAA/PEI/PAA-modified membranes with
similar methods to immobilize Con A via Cu2+ affinity or covalent linkages. However,
117

the binding capacity for glycoprotein was again not extensive.

4.4. Conclusion

Using macroinitiator adsorption, a 5-min polymerization of MES within membrane
pores successfully created a platform that binds at least 100 mg of lysozyme/cm3 of
membrane and still maintains sufficient hydraulic permeability for flow with a peristaltic
pump. After functionalization with NTA-Cu2+ complexes, these membranes can capture
69 mg Con A/cm3. However, after blocking nonspecific binding sites with BSA, this
high Con A binding capacity gave an asialofetuin binding capacity of only 10 mg/cm 3 of
membrane. Even when Con A was immobilized with its binding sites protected,
asialofetuin binding was not extensive. There are two possible reasons for the low
glycoprotein capture: 1) The immobilization process might damage Con A, which is a
pH-sensitive protein (43), or more likely 2) The immobilization of Con A leads to
significant steric hindrance to binding.

118

REFERENCES

119

REFERENCES

1.

Loris R, Hamelryck T, Bouckaert J, Wyns L. Legume lectin structure. Biochim.
Biophys. Acta-Protein Struct. Molec. Enzym. 1998, 1383, 9-36.

2.

Zavala-Cerna MG, Martinez-Garcia EA, Torres-Bugarin O, Rubio-Jurado B,
Riebeling C, Nava A. The clinical significance of posttranslational modification
of autoantigens. Clin. Rev. Allergy Immunol. 2014, 47, 73-90.

3.

Seo J, Lee KJ. Post-translational modifications and their biological functions:
Proteomic analysis and systematic approaches. J. Biochem. Mol. Biol. 2004, 37,
35-44.

4.

Kang T, Kim JH, Hong I, Park N, Heinsen H, Lee JY, Ravid R, Ferrer I, Yoo JS,
Kwon KH, et al. Large-scale analysis of posttranslational modifications in the
hippocampus of patients with Alzheimer's disease using pI shift and label-free
quantification without enrichment. Anal. Bioanal. Chem. 2014, 406, 5433-5446.

5.

Baenziger JU.
A major step on the road to understanding a unique
posttranslational modification and its role in a genetic disease. Cell 2003, 113,
421-422.

6.

Mann M, Jensen ON. Proteomic analysis of post-translational modifications.
Nat. Biotechnol. 2003, 21, 255-261.

7.

Prabakaran S, Lippens G, Steen H, Gunawardena J.
Post-translational
modification: Nature's escape from genetic imprisonment and the basis for
dynamic information encoding. Wiley Interdiscip. Rev.-Syst. Biol 2012, 4,
565-583.

8.

Tia SQ, Brown K, Chen D, Herr AE. Protein post-translational modification
analyses using on-chip immunoprobed isoelectric focusing. Anal. Chem. 2013, 85,
2882-2890.

9.

Barrero-Gil J, Salinas J. Post-translational regulation of cold acclimation
response. Plant Sci. 2013, 205, 48-54.

10.

Posttranslational
Http://en.Wikipedia.Org/wiki/posttranslational_modification
11/18/2014).

11.

Khoury GA, Baliban RC, Floudas CA. Proteome-wide post-translational
modification statistics: Frequency analysis and curation of the swiss-prot database.
Sci Rep 2011, 1, 90.

120

modification.
(accessed
on

12.

Kaji H, Yamauchi Y, Takahashi N, Isobe T. Mass spectrometric identification
of N-linked glycopeptides using lectin-mediated affinity capture and glycosylation
site-specific stable isotope tagging. Nat. Protoc. 2006, 1, 3019-3027.

13.

Varki A. Biological roles of oligosaccharides - all of the theories are correct.
Glycobiology 1993, 3, 97-130.

14.

Temporini C, Callerli E, Massolini G, Caccialanza G. Integrated analytical
strategies for the study of phosphorylation and glycosylation in proteins. Mass
Spectrom. Rev. 2008, 27, 207-236.

15.

Harvey DJ. Proteomic analysis of glycosylation: Structural determination of Nand O-linked glycans by mass spectrometry. Expert Rev. Proteomics 2005, 2,
87-101.

16.

Van den Steen P, Rudd PM, Dwek RA, Opdenakker G. Concepts and principles
of O-linked glycosylation. Crit. Rev. Biochem. Mol. Biol. 1998, 33, 151-208.

17.

Hagglund P, Bunkenborg J, Elortza F, Jensen ON, Roepstorff P. A new strategy
for identification of N-glycosylated proteins and unambiguous assignment of their
glycosylation sites using hilic enrichment and partial deglycosylation. J. Proteome
Res. 2004, 3, 556-566.

18.

Walsh CT, Garneau-Tsodikova S, Gatto GJ.
Protein posttranslational
modifications: The chemistry of proteome diversifications. Angew. Chem.-Int.
Edit. 2005, 44, 7342-7372.

19.

Dalpathado DS, Desaire H.
Analyst 2008, 133, 731-738.

20.

Zhao YY, Takahashi M, Gu JG, Miyoshi E, Matsumoto A, Kitazume S,
Taniguchi N. Functional roles of N-glycans in cell signaling and cell adhesion in
cancer. Cancer Sci. 2008, 99, 1304-1310.

21.

Bones J, Mittermayr S, O'Donoghue N, Guttman A, Rudd PM. Ultra
performance liquid chromatographic profiling of serum N-glycans for fast and
efficient identification of cancer associated alterations in glycosylation. Anal.
Chem. 2010, 82, 10208-10215.

22.

Lau KS, Dennis JW. N-glycans in cancer progression. Glycobiology 2008, 18,
750-760.

23.

Gu J, Sato Y, Kariya Y, Isaji T, Taniguchi N, Fukuda T. A mutual regulation
between cell-cell adhesion and N-glycosylation: Implication of the bisecting
glcnac for biological functions. J. Proteome Res. 2009, 8, 431-435.

24.

O'Brien TJ, Beard JB, Underwood LJ, Dennis RA, Santin AD, York L.

Glycopeptide analysis by mass spectrometry.

121

The CA

125 gene: An extracellular superstructure dominated by repeat sequences. Tumor
Biol. 2001, 22, 348-366.
25.

Ross JS, Fletcher JA, Linette GP, Stec J, Clark E, Ayers M, Symmans WF,
Pusztai L, Bloom KJ. The HER-2/neu gene and protein in breast cancer 2003:
Biomarker and target of therapy. Oncologist 2003, 8, 307-325.

26.

A new monolithic ConA affinity column for purification and analysis of glycans,
glycopeptides,
and
glycoproteins.
Http://www.Dionex.Com/en-us/webdocs/87395-po-lc-ConA-glycoprotein-analysi
s-30mar2011-lpn2751-01.Pdf (accessed on 11/18/2014).

27.

Matyjaszewski K, Miller PJ, Shukla N, Immaraporn B, Gelman A, Luokala BB,
Siclovan TM, Kickelbick G, Vallant T, Hoffmann H, et al. Polymers at
interfaces: Using atom transfer radical polymerization in the controlled growth of
homopolymers and block copolymers from silicon surfaces in the absence of
untethered sacrificial initiator. Macromolecules 1999, 32, 8716-8724.

28.

Jain P, Dai J, Grajales S, Saha S, Baker GL, Bruening ML. Completely aqueous
procedure for the growth of polymer brushes on polymeric substrates. Langmuir
2007, 23, 11360-11365.

29.

Chen XY, Armes SP. Surface polymerization of hydrophilic methacrylates from
ultrafine silica sols in protic media at ambient temperature: A novel approach to
surface functionalization using a polyelectrolytic macroinitiator. Adv. Mater.
2003, 15, 1558-1562.

30.

Anuraj N, Bhattacharjee S, Geiger JH, Baker GL, Bruening ML. An all-aqueous
route to polymer brush-modified membranes with remarkable permeabilites and
protein capture rates. J. Membr. Sci. 2012, 389, 117-125.

31.

Jain P, Dai JH, Baker GL, Bruening ML. Rapid synthesis of functional polymer
brushes by surface-initiated atom transfer radical polymerization of an acidic
monomer. Macromolecules 2008, 41, 8413-8417.

32.

He JT, Liu YS, Xie XL, Zhu T, Soules M, DiMeco F, Vescovi AL, Fan X,
Lubman DM.
Identification of cell surface glycoprotein markers for
glioblastoma-derived stem-like cells using a lectin microarray and LC-MS/MS
approach. J. Proteome Res. 2010, 9, 2565-2572.

33.

Ferreira JA, Daniel-da-Silva AL, Alves RMP, Duarte D, Vieira I, Santos LL,
Vitorino R, Amado F. Synthesis and optimization of lectin functionalized
nanoprobes for the selective recovery of glycoproteins from human body fluids.
Anal. Chem. 2011, 83, 7035-7043.

34.

Li YS, Ge YS, Zhang Y, Zhang AQ, Sun SF, Jiang FL, Liu Y. Interaction of
coomassie brilliant blue G250 with human serum albumin: Probing of the binding
122

mechanism and binding site by spectroscopic and molecular modeling methods. J.
Mol. Struct. 2010, 968, 24-31.
35.

Jain P, Vyas MK, Geiger JH, Baker GL, Bruening ML. Protein purification with
polymeric affinity membranes containing functionalized poly(acid) brushes.
Biomacromolecules 2010, 11, 1019-1026.

36.

Concanavalin
A
(Con
A)
Http://www.Gbiosciences.Com/researchproducts/ConA_agarose.Aspx
on 11/18/2014).

37.

Wang JL, Edelman GM. Binding and functional properties of concanavalin-a
and its derivatives.1. Mono-valent, divalent, and tetravalent derivatives stable at
physiological pH. J. Biol. Chem. 1978, 253, 3000-3007.

38.

Hardman KD, Ainswort.Cf. Structure of Concanavalin-A at 2.4-A resolution.
Biochemistry 1972, 11, 4910-4919.

39.

Concanavalin-A.
11/18/2014).

40.

Spiro RG. Studies on fetuin, a glycoprotein of fetal serum : II. Nature of
carbohydrate units. J. Biol. Chem. 1962, 237, 382-388.

41.

Mori K, Emoto M, Inaba M. Fetuin-A: A multifunctional protein. Recent Pat.
Endocr. Meta.b Immune Drug. Discov. 2011, 5, 124-146.

42.

Motoyama K, Nakashima Y, Aramaki Y, Hirayama F, Uekama K, Arima H. In
Vitro gene delivery mediated by asialofetuin-appended cationic liposomes
associated with gamma-cyclodextrin into hepatocytes. J. drug delivery 2011, 2011,
476137.

43.

Zand R, Agrawal BBL, Goldstei.Ij. pH-dependent conformational changes of
Concanavalin-A. Proc. Natl. Acad. Sci. U. S. A. 1971, 68, 2173-2176.

44.

Fischer MJE. 2010. Amine coupling through EDC/NHS: A practical approach. In
Surface plasmon resonance: Methods and protocols, ed. NJ DeMol, MJE Fischer,
pp. 55-73.

45.

Reeke GN, Becker JW, Edelman GM. Covalent and 3-dimensional structure of
Concanavalin-A .4. Atomic coordinates, hydrogen-bonding, and quaternary
structure. J. Biol. Chem. 1975, 250, 1525-1547.

46.

Brady D, Jordaan J. Advances in enzyme immobilisation. Biotechnol. Lett. 2009,
31, 1639-1650.

Http://www.Uniprot.Org/uniprot/p81461

123

agarose.
(accessed

(accessed

on

Chapter 5. Rapid and Controlled Protein Digestion in Porous
Membrane Reactors Containing Covalently Immobilized Trypsin

The previous chapters described the development of a membrane platform for
purification of proteins, including His-tagged protein and glycoproteins. Convective
mass transport and minimal radial diffusion distances allow rapid protein purification,
and modification of membrane micropores with PEMs often gives a high protein binding
capacity. Thin (~100 µm) microporous membranes also provide a unique platform for
creating enzyme reactors. This chapter describes covalent immobilization of trypsin in a
PAA-modified membrane and examines the enzymatic activity of the immobilized
enzyme. The thin membrane reactor affords fine control over digestion time to control
peptide lengths and enhance protein analysis with MS.

5.1. Introduction

Since the sequencing of the human genome, research interests have shifted to
uncovering the secrets of the human proteome (1, 2). An understanding of protein
structures, PTMs and protein-protein interactions sometimes provides valuable
information about the development of diseases such as cancer (3). MS is the most
comprehensive and versatile analytical technique to identify proteins and their PTMs
(4-6), and driven by the need for large-scale protein analysis, MS techniques continue
evolving to give higher resolution, faster analysis, and higher mass accuracy (5-7).
However, even with recent advances in MS, analysis of intact proteins with molecular
weights greater than 50 kDa is challenging (8-10). As a result, proteolytic digestion to
124

convert proteins to peptides is a vital step in sample preparation for successful MS
analysis (4, 11). Conventionally, digestion occurs after directly mixing the proteolytic
enzyme with proteins solutions. Several proteases including chymotrypsin (12), trypsin
(13), Lys-C (14) and Asp-N (15) are effective, but trypsin, which cleaves at the carboxyl
side of Lys or Arg when neither is followed by Pro, is the most popular enzyme due to its
high cleavage specificity and low cost (16). The enzyme to substrate ratio is usually low
to avoid protease self-digestion, but this low ratio leads to long digestion times for
complete digestion (17).
To enhance the protein-digestion rate, several research groups developed reactors
containing immobilized trypsin (18-21). Substrates for trypsin immobilization include
polymer plates (22), membranes (11, 23, 24), monoliths (25), and microfluidic channels
(26, 27), and enzyme anchoring can occur via hydrophobic (23, 27), covalent (28) or
electrostatic interactions (11, 24). Membranes present a unique platform for enzyme
reactors because their small thickness causes low transmembrane pressure drops. As a
result rapid flow rates through the membrane can yield digestion times as low as 1 msec.
(11, 24). Xu and Tan from the Bruening group already developed trypsin and pepsin
membrane reactors using electrostatic interactions to immobilize the enzyme (11, 24).
Millisecond residence times in these membranes lead to large peptides that facilitate
protein analysis. Unfortunately, at high ionic strengths, the electrostatic immobilization
is insufficient to prevent enzyme leaching. Thus, this chapter focuses on immobilizing
trypsin on the membrane via covalent bonds to enable digestion under harsh conditions.
Similar to previous chapters, adsorption of a layer of PAA in the membrane provides

125

free –COOH groups for extensive enzyme immobilization using NHS/EDC coupling.
Importantly, PAA also serves as a shielding layer on the porous surface to minimize
nonspecific binding and enhance the recovery of digested peptides (29, 30). This chapter
describes enzyme immobilization, quantification of the enzymatic activity, and
examination of controlled digestion of model proteins in the trypsin-containing
membranes.

5.2. Experimental section
5.2.1. Materials

Hydroxylated nylon (LoProdyne® LP, Pall, 1.2 µm pore size, 110 µm thick, 50 %
porosity) was cut into 2.5-cm disks. Trypsin from bovine pancreas (type I, 12200
units/mg solid), benzamidine hydrochloride,

Nα-benzoyl-L-arginine ethyl ester

hydrochloride (BAEE), poly(acrylic acid) (PAA, average molecular weight ~100,000, 35%
aqueous

solution),

N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide

hydrochloride

(EDC), N-hydroxysuccinimide (NHS), bovine -casein and apomyoglobin (protein
sequencing grade from horse skeletal muscle, salt-free lyophilized powder) were
purchased from Sigma-Aldrich. Urea (J.T.Baker), Sodium phosphate (dibasic and
monobasic) and Ammonium bicarbonate were obtained from Columbus Chemical.
Acetonitrile (ACN) was obtained from Mallinckrodt Baker. Buffers were prepared using
analytical grade chemicals and deionized (Milli-Q, 18.2 MΩ cm) water, and Sep-Pak®
C18 Cartridges (100 mg, 55-92 μm beads, Part No.WAT023590) from Waters were
employed for desalting.

126

5.2.2. Covalent immobilization of trypsin in membranes

Membrane modification employed an Amicon cell (Model 8010, 10 mL, Millipore)
connected to a peristaltic pump. Hydroxylated nylon 6,6 membranes (2.5 cm diameter,
1.2 µm pore size) were exposed to UV/ozone for 15 min, placed in the Amicon cell, and
5 mL of a solution containing 20 mM PAA (pH 3) was passed through the membrane
using a peristaltic pump. Subsequently, 20 mL of deionized water was pumped through
the membrane at 1 mL/min. The adsorbed PAA was activated by circulating an aqueous
solution containing 20 mM NHS and 20 mM EDC through the membrane for 2 hours
followed by passage of 20 mL of water and 20 mL of ethanol through the membrane.
Subsequently, 5 mL of 0.4 mg/mL trypsin solution in 20 mM phosphate buffer (pH 7.6)
supplemented with 20 mM benzamidine was circulated through the membrane for 2
hours to complete the covalent immobilization. These membranes were washed by
passing 20 mL of phosphate buffer containing 20 mM benzamidine and 20 mL of 1 mM
HCl through the membrane before drying with N2 and storage at room temperature. In
some cases, we examined membrane activity after 6 months of storage.

5.2.3. Quantitation of trypsin immobilization

The amount of trypsin bound to the membrane was determined by measuring the
absorbance at 280 nm of the loading solution before and after circulating through the
membrane. These studies employed a PerkinElmer Lamda 25 spectrophotometer and
calibration solutions containing 20 mM benzamidine and trypsin at a concentration range
of 0-0.3 mg/mL. Control experiments employed membranes with PAA but no NHS/EDC

127

activation and membranes without any modification. Because NHS absorbs at 280 nm
under basic conditions, 2.5 mL of standard or sample solution was mixed with 0.5 mL of
1 M HCl before absorbance measurements (31).

5.2.4. Quantification of enzyme activity

We determined the proteolytic activity of trypsin both in solution and in membranes
using the trypsin substrate BAEE dissolved in 10 mM NH4HCO3 at pH 7.6.

5.2.4.1. In-solution enzyme activity
Ten μL of 1 mg/mL trypsin in 2.7 mM HCl was added to 3 mL of 1 mM BAEE in 10
mM NH4HCO3 (pH 7.6) buffer and stirred with a pipet for ~ 5 s. The UV absorbance at
253 nm was monitored every 10 s with 1 mM BAEE as background until the absorbance
plateaued.

5.2.4.2. In-membrane enzyme activity

Twenty mM BAEE was passed though the membrane (2-cm exposed diameter) at
flow rates of 60 mL/hour, 40 mL/hour, 30 mL/hour, 20 mL/hour, 15 mL/hour, 10
mL/hour, 8 mL/hour and 6 mL/hour using a 10-mL syringe pump. With a membrane
thickness of 110 μm and 50% porosity (values provided by Pall), these flow rates give
residence time of 0.017 to 0.17 min. Because BAEE has a very strong absorbance at 253
nm, all the samples were diluted 20-fold with phosphate buffer before measurement
against a 1 mM BAEE background.

128

5.2.5. Protein digestion
Proteins were dissolved in water at 10 mg/mL, separated into 100 μL aliquots, dried
using a SpeedVac and stored at -20 oC until use.

5.2.5.1. In-solution digestion
For in-solution digestion without denaturation, 10 μL of freshly made 0.5 mg/mL
trypsin solution in 1 mM HCl was added to 0.1 mg/ml protein solutions in 10 mM
NH4HCO3 to achieve a 1:20 ratio of trypsin to substrate protein. After the desired
digestion time, the reaction was quenched by addition of 11μL acetic acid.

5.2.5.2. Membrane digestion

For protein digestion, the membrane was cut to fit into a miniature membrane holder
(the holder exposes 0.02 cm2 of external membrane area, which was connected to a
syringe pump (KDS210, KD scientific) as shown in Figure 5.1 (11). Before digestion,
pH 7.6 buffer was passed through the membrane for 30 min to activate the immobilized
trypsin. (Prior to storage the membrane had been rinsed with 1 mM HCl). For digestion
at different pH values, the membrane was activated with phosphate buffers adjusted to
the pH of the digestion. Solutions containing 0.1 mg/mL protein in 10 mM NH4HCO3
buffer were sequentially passed through the trypsin containing membrane at flow rates of
0.12 mL/hour, 1.2 mL/hour, 12 mL/hour and 120 mL/hour. The corresponding residence
times are 3.3 s, 0.33 s, 33 ms and 3.3 ms. For membrane digestions without salts and
urea, 50μL of permeate was collected and dried with a Speed-Vac, and then reconstituted

129

in MS buffer (50 μL of 1% acetic acid/49% H2O/50% methanol). For digests with urea,
desalting was performed using a C18 Sep-Pak cartridge.

Figure 5.1. Apparatus for in-membrane protein digestion. A syringe pump forces a
protein solution through a miniature trypsin-containing membrane. Drawn by Zhefei Yang.

5.2.6. Desalting using a C18 Sep-Pak cartridge

Digests containing urea were acidified with formic acid (FA) to get a 0.5% final FA
concentration and desalted using a C18 cartridge (syringe cartridge with 1 mL volume).
The C18 cartridges were first washed and equilibrated with acidic solutions by
sequentially passing 3 mL of 0.05% FA in 99.95% ACN and 3 mL of aqueous 0.05% FA
through the cartridge. 150 μL of acidified protein solution was then forced through the
cartridge followed by 0.5 mL of 0.05% FA for washing. Protein bound to the solid
phase was recovered by sequentially passing 0.2 mL of 0.05% FA buffers containing
40%, 60%, 80% and 99.95% ACN through the cartridge. All the above procedures were
performed manually with a ~0.7 mL/min flow rate. The combined elution fractions were
finally dried with a SpeedVac and reconstituted in MS buffer before analysis.

130

5.2.7. Mass spectrometry and data analysis
All the digests were analyzed with ESI-MS. Forty μL of reconstituted protein digest
in MS buffer was loaded into a Whatman multichem 96-well plate and sealed with
Teflon Ultrathin Sealing Tape. The samples were introduced into the high-resolution
accurate mass Thermo LTQ Orbitrap Velos mass spectrometer using an Advion Triversa
Nanomate nanoelectrospary ionization (nESI) source. The spray voltage was 1.4 kV, and
the gas pressure was 1.0 psi. High-resolution mass spectra were acquired in positive
ionization mode using the FT analyzer operating at 100,000 resolving power with
relative intensity as the Y-axis. Peptides were identified manually by comparing the
experimental data to the theoretical peptides generated from the ProteinProspector
MS-Product program (32). For the theoretical peptide generation, the data base was user
protein, the enzyme was set to trypsin, the maximum missed cleavage was 99, the
peptide mass was 200 to 30000 Da, the minimum peptide length was 2 amino acids and
the digestion was performed with the protein sequence without the signal peptides.

5.3 Results and discussion

Membranes are attractive solid supports for enzyme reactors because they allow fine
control over residence time through variation of flow rate. The first two subsections
below describe trypsin immobilization in the membrane and studies of the immobilized
enzyme’s activity. Subsequent subsections discuss in-membrane protein digestion and
membrane stability.

131

5.3.1. Covalent immobilization of trypsin in membranes

Figure 5.2 shows our strategy for enzyme immobilization, which includes
adsorption of PAA, activation of –COOH groups with NHS/EDC, and covalent coupling
via amide linkages (33). Several previous studied reported covalent immobilization using
accessible primary amine groups on trypsin (25, 28, 34-37), mainly through reactions
with epoxy or aldehyde groups on the solid support (38). Amide coupling using
NHS/EDC is more rapid than reaction of amine groups with epoxides or aldehydes, and
the amide bond is more stable than the imine formed by reaction with aldehydes. At pH
7.6, NHS esters that do not react with amine groups undergo hydrolysis to reform –
COOH groups and avoid possible covalent capture of other proteins during digestion
with trypsin-containing membranes. To avoid trypsin self-digestion, we add benzamidine
as a competitive inhibitor during the immobilization step (34, 35). The benzamidine may
also minimize the formation of covalent bonds with amino acids near the active site of
trypsin and stabilize the protein tertiary structure (35, 39).

Figure 5.2. Covalent immobilization of trypsin to PAA adsorbed in a membrane
pore.
132

Based on the decrease in the concentration of trypsin in the solution circulated through
the membrane, the modified membrane contains ~35 mg of trypsin per cm3 of membrane
or ~70 mg trypsin per cm3 of pores, assuming 50% porosity. This binding capacity is
higher than the 11 mg per cm3 of membrane that we achieved previously through
adsorption of poly(styrene sulfonate) in the membrane and subsequent electrostatic
capture of trypsin (11, 24). Moreover, the trypsin “concentration” in membrane pores is
2800 times the 0.025 mg/cm3 trypsin concentration in a typical in-solution digestion (11),
and this high concentration enables rapid digestion. Control experiments on bare nylon
membranes showed <5 mg trypsin per cm3 of membrane.
Hydrolysis of active esters during trypsin binding could lead to some immobilization
of trypsin through electrostatic interactions with –COO- groups. However, washing of
trypsin-modified membranes with buffers containing as much as 1 M NaCl, led to no
detectable absorbance at 280 nm, indicating minimal protein elution under conditions
that should disrupt electrostatic interactions. In contrast, electrostatic binding of trypsin
to PAA adsorbed in nylon membranes (no NHS/EDC activation) captured ~ 40 mg
trypsin per cm3 of membrane, but elution with 1 M NaCl removed all the bound trypsin.
These data imply that trypsin capture after NHS/EDC activation occurs through covalent
bonds.

5.3.2. Proteolytic activity of Trypsin Membranes

After creating trypsin-modified membranes, we initially examined their enzyme
activity spectrophotometrically by monitoring the cleavage of the ester group of BAEE, a

133

synthetic trypsin substrate (28, 35, 36, 40). The UV absorbance at 253 nm due to the
digestion product, Na-benzoyl-L-arginine (BA), is proportional to the rate of ester
cleavage (41). Trypsin activity typically follows Michaelis-Menten kinetics, where
equation (5.1) describes
[𝑆]

𝑉 = 𝑉𝑚𝑎𝑥 [𝑠]+𝐾

(5.1)

𝑀

the reaction rate, V or more explicity −

𝑑[𝑆]
𝑑𝑡

, in terms of the rate when the enzyme is

saturated with substrate, Vmax, the substrate concentration, [S], and the Michaelis constant,
Km. Usually, a plot of

1
𝑉

versus

1
[𝑆]

yields values for 𝑉𝑚𝑎𝑥 and 𝐾𝑚 . However, for the

membrane reactor, the value of [S] varies significantly from the feed to the permeate side
of the membrane. At low BAEE feed concentrations, the high trypsin loading in the
pores leads to essentially complete BAEE cleavage. Thus, we employ equation (5.2), the
integrated Michaelis-Menten equation, to determine kinetic parameters. In this
expression
𝐾𝑀

𝑡=𝑉

𝑚𝑎𝑥

[𝑆]

𝑙𝑛 ( [𝑆]0 ) +

[𝑆]0 −[𝑆]

(5.2)

𝑉𝑚𝑎𝑥

t is the digestion time and [𝑆]0 is the initial substrate concentration. Specifically for the
case where the membrane behaves as a plug-flow reactor, t is the residence time, tres, of
solution in the membrane, [𝑆]0 is the substrate concentration in the feed, and [S] is the
substrate concentration in the membrane permeate. We employ equation (5.3) to
calculate the residence time, assuming a porosity, ɛ, of 0.5; a membrane thickness, l, of
0.011 cm; and a membrane area, A, of 3.1 cm2.
𝐴𝑙

𝑡𝑟𝑒𝑠 = 𝑄𝜀

(5.3)

Variation of the volumetric flow rate, Q, using a syringe pump affords different

134

residence times, and Figure 5.3 shows a plot of the BAEE concentration in the
membrane permeate as a function of the residence time. Fitting of the data in Figure 5.3
with equation (5.2) yields Vmax =1 mM/s and Km ≅ 4 mM. However, given that the data
in Figure 5.3 are essentially linear, the fit does not depend significantly on Km, so that
value

is

highly

uncertain.

Figure 5.3. Residence time as a function of the BAEE concentration in the permeate
for BAEE digestion during passage through a trypsin-modified membrane. The feed
solution contained 20 mM BAEE. The curve is the fit to the data using equation (5.2).

We also determined kinetic parameters using in-solution digestion and with the same
fitting method. In this procedure, we collected the UV absorbance every 10 s until the
reading plateaued. From the fitting shown in figure 5.4, we get Vmax = 0.013 mM/s and
Km = 0.35 mM in a typical solution digestion, and the Vmax is consistent with reported

135

values of Vmax = 0.0134 mM/s determined using equation (5.1) (35). However, the Vmax
value in the membrane is about 75 times that in solution because of the high trypsin
“concentration” in the membrane pores. Nevertheless, immobilization does decrease the
enzyme activity, as the enzyme concentration in the membrane pores is 2800 times that
in
solution.

Figure 5.4. In-solution digestion time as a function of the evolving BAEE
concentration. The initial BAEE concentration in the solution was 1 mM. The curve is
the fit to the data using equation 5.2.

5.3.3. Controlled β-casein digestion as a function of residence time

In addition to rapid digestion, protease-containing membranes provide a platform for
controlling the extent of protein cleavage by varying residence times down to the ms
136

level.

β-casein served as our initial model protein for examining the extent of protein

digestion as a function of flow rate because digestion can occur in 10 mM NH4HCO3
without denaturation. The electrophoretic gel in Figure 5.5 shows the molecular weights
of digestion products obtained at flow rates of 120 mL/hour, 12 mL/hour and 0.12
mL/hour, which correspond to residence times of 3.3 ms, 33 ms and 3.3 s, respectively.
The gel shows a clear trend toward increased protein digestion at longer residence times.
The 3.3-ms residence times leads to a significant amount of intact protein along with
several proteolytic peptides with masses between 10 and 25 kDa, whereas a 3.3-s
residence times yields no visible peptides, indicating complete digestion into peptides
with masses <10 kDa. The intermediate residence time of 33 ms gives minimal intact
protein along with peptides with masses between 10 and 25 kDa.

Figure 5.5. SDS-PAGE of intact β-casein (lane 2) and membrane digests of β-casein
obtained with different residence times (lanes 3-5). Five-μg protein samples were
loaded for each lane, and lane 1 shows a protein ladder. Digestion occurred using 0.1
mg/mL β-casein in 10 mM NH4HCO3.

137

We analyzed the digests obtained at different residence times with ESI-Orbitrap mass
spectrometry using direct injection. Figure 5.6 compares the deconvoluted mass spectra
for digestion with 33-ms and 3.3-s residence times. For the 33-ms digestion (Figure
5.6a), as few as two peptides (e.g. 14 and 17) can cover the entire 209 amino acid
sequence. In contrast, at the long residence time (3.3 s, Figure 5.6b), all detected
peptides have masses less than 8 kDa, which is consistent with the gel electrophoresis
(Figure 5.5, lane 5). The long residence times also lead to relatively high intensities of
peptides with masses less than 1000. More missed cleavage occurs at short residence
times to decrease the intensities of small peptides, e.g., with the 33-ms residence time,
the signal for peptide 8 (amino acid (AA) 170-209) largely replaces signals for peptides 1
(AA 170-176), 2 (AA 177-183), and 4 (AA 184-209). Sometimes, researchers employ
highly selective enzymes to obtain large peptides to achieve high sequence coverage (42,
43), but the peptide size depends on the protein sequence. The trypsin enzyme reactor
may enable production of large peptides independent of sequence.
In addition to improving sequence coverage, controlled limited digestion may
facilitate simultaneous analysis of multiple post translational modifications in single
peptides (“combinatorial” PTM characterization) (44). Such studies can show whether
PTMs occur simultaneously or randomly. β-casein is phosphorylated at Ser15, Ser17,
Ser18, Ser19 and Ser35. Previous analysis of β-casein phosphorylation based on small
tryptic phosphopeptides inevitably required phosphopeptide enrichment from the
complete β-casein digest (45). The existence of one Arg and three Lys residues between
Ser35 and the other four phosphorylation sites makes the simultaneous characterization

138

of the five phosphorylation sites on a single peptide almost impossible after traditional
in-solution tryptic digestion. However, peptide 17 (Figure 5.6a, AA 1-107) contains all
five phosphorylation sites on β-casein. The signal of this peptide is likely low in
positive-ion mode because of the 5 acidic phosphate groups, but it clearly shows that all
five phosphorylation events occur together on this peptide.

Figure 5.6. Manually deconvoluted ESI-Orbitrap mass spectra of 0.1 mg/mL β-casein
digested with residence times of (a) 33 ms and (b) 3.3 s in a trypsin-modified
membrane. Signals above 1% of the highest signal were assigned to specific peptides by
comparison to theoretical m/z values. The normalized intensities are the sum of signal
intensities for all detected charge states of a given peptide, and the spectra show the
peptides only at the +1 charge state for the monoisotopic mass.

139

5.3.4. Apomyoglobin digestion with high concentrations of urea and variable pH
After successful β-casein digestion under native conditions (10 mM NH4CO3),
apomyoglobin served as a model protein for membrane digestion under denaturing
conditions because its tryptic map is well documented (35). Apomyoglobin is a
monomeric protein with 153 amino acid residues and no disulfide bonds, but it
undergoes slow tryptic proteolysis at pH 7-8 because of its compact hydrophobic core at
neutral pH (11, 46, 47). Figure 7a shows the ESI-Orbitrap mass spectrum of
apomyoglobin after in-solution digestion at pH 7.6 for 30 s. The appearance of dominant
signals from intact protein (marked with *) confirms the slow proteolysis under this
condition. Apomyoglobin digestion in the trypsin-containing membrane (pH 7.6 without
urea) for 3.3 s (Figure 7b) provides 100% sequence coverage from proteolytic peptides,
and the intact protein peaks are less intense than for the 30-s in-solution digestion.
Consistent with the Michaelis-Menten study in section 4.3.2, this confirms that the
membrane reactor has a faster digestion rate. After addition of 6 M urea to the
apomyoglobin solution, a 3.3-s membrane digestion shows 100% sequence coverage and
no detected intact protein peaks in the spectrum (Figure 7c), indicating that the
membrane reactor still works in the presence of a high concentration of urea. This is
especially important for protein samples with low solubility. In-solution digestion in 6 M
urea is typically not possible due to loss of enzymatic activity (48).

140

Figure 5.7. ESI-Orbitrap mass spectra of 0.1 mg/mL apomyoglobin digested (a)
in-solution in 10 mM, pH 7.6 NH4HCO3 for 30 s; (b) in a trypsin-modified membrane
for 3.3 s using 10 mM, NH4HCO3, pH 7.6, no urea; and (c) in a trypsin-modified
membrane for 3.3 s using 10 mM, NH4HCO3, pH 7.6 with 6 M urea. For (c), urea
was removed using a C18 Sep-Pak cartridge (peaks labeled with

correspond to

intact apomyoglobin).
The extent of apomyoglobin digestion also depends on pH (46, 47, 49). Our
previous studies using membranes with electrostatically immobilized pepsin showed that
digestion of apomyoglobin at low pH occurs in residence times as short as 0.02 s (11).
This study explores the effect of pH on apomyoglobin digestion in membranes with
covalently immobilized trypsin. Figures 8a and 8b show mass spectra of apomyoglobin
141

digests performed in solution and in trypsin-containing membranes, respectively, at pH 5.
No urea was added to these solutions. With membrane digestion at pH 5, no intact
protein signals appear, which indicates a more complete digestion compared to
proteolysis at pH 7.6 (compare Figures 8b and 7b). Presumably, partial apomyoglobin
denaturation at pH 5 enhances digestion (45, 46, 48). Previous studies suggest that
trypsin is inactive at pH below 6 (37), but these results imply an extended working pH
range from 5 to 8.5 for trypsin immobilized in membranes. The relatively wide working
range with the trypsin-containing membrane will allow digestion under weakly acidic
conditions that may denature many proteins. Although digestion at even lower pH can
occur with pepsin (11), trypsin is attractive because of its higher specificity. For all the
protein-digestion experiments, we observed no trypsin self-digestion peaks under all
digestion conditions, which implies that the covalently immobilization gives a stable
membrane reactor.

142

Figure 5.8. ESI-Orbitrap mass spectra of 0.1 mg/mL apomyoglobin digested (a) in
solution for 30 s at pH 5 and (b) in a trypsin-modified membrane for 3.3 s at pH 5.

Apomyoglobin was dissolved in 10 mM NH4HCO3 (peaks labeled with

correspond to

intact apomyoglobin).
5.3.5. Stability of the trypsin membrane reactor

In addition to their rapid and controlled protein digestion, membrane reactors are
also durable. Eeven with storage in water at room temperature, the membrane does not
lose enzymatic activity after more than 24 hours when we examine BAEE digestion
every hour in two time intervals (Figure 5.9). In contrast, in solution (pH 7.5) the
enzyme completely loses its activity after this period of time (28). The hydrophilic PAA
surface may reduce the denaturation of trypsin after immobilization (22, 28, 50).
Typically, after fabrication, we dry the membrane and store it in a dessicator at room
143

temperature. After storage for at least 6 months, the membrane provides the same
β-casein digestion results as shown in Figure 5.6.
Day 2

Day 1
Conversion of BAEE
(%)

100
left
wet for
12
hours
in

80
60
40
20
0
0h1h2h3h4h5h6h

0 h1 h 2 h 3 h 4 h 5 h 6 h 7 h8 h
time
Figure 5.9. Conversion of BAEE in a continuous flow mode using a membrane with
an exposed 2-cm diameter. The low conversion in the first tests on both days is due to
sample dilution by the system dead volume. The feed solution contains 20 mM BAEE in
10 mM NH4HCO3, and the flow rate was 10 mL/hour. h: hour.
5.4. Conclusion
Porous membranes modified with one layer of PAA bind 35 mg of trypsin/cm 3 of
membrane, and this high enzyme concentration leads to protein digestion that is 75 times
faster than conventional in-solution digestion. In addition, the covalent immobilization
stabilizes the enzyme and prevents leaching. The trypsin membrane reactor digests
protein even in the presence of 6 M urea or at pH 5, which enables denaturation of
proteins that typically resist digestion. When dry, membranes stored for over 6 months at
room temperature are still active. A miniature membrane holder and syringe pump afford
control over digestion times and, hence, peptide lengths. For β-casein, we found a large
peptide (AA 1-107) that carries all the phosphorylation sites.
144

REFERENCES

145

REFERENCES

1.

Capelo JL, Carreira R, Diniz M, Fernandes L, Galesio M, Lodeiro C, Santos HM,
Vale G. Overview on modern approaches to speed up protein identification
workflows relying on enzymatic cleavage and mass spectrometry-based
techniques. Anal. Chim. Acta 2009, 650, 151-159.

2.

Mann M, Kelleher NL. Precision proteomics: The case for high resolution and
high mass accuracy. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 18132-18138.

3.

Pastwa E, Somiari SB, Czyz M, Somiari RI.
research. Proteom. Clin. Appl. 2007, 1, 4-17.

4.

Somiari RI, Renganathan K, Russell S, Wolfe S, Mayko F, Somiari SB. A
colorimetric method for monitoring tryptic digestion prior to shotgun proteomics.
International journal of proteomics 2014, 2014, 125482.

5.

Domon B, Aebersold R. Review - mass spectrometry and protein analysis.
Science 2006, 312, 212-217.

6.

Yates JR, Ruse CI, Nakorchevsky A. 2009. Proteomics by mass spectrometry:
Approaches, advances, and applications. In Annual review of biomedical
engineering, pp. 49-79. Palo Alto: Annual Reviews.

7.

Xian F, Hendrickson CL, Marshall AG. High resolution mass spectrometry.
Anal. Chem. 2012, 84, 708-719.

8.

Han XM, Aslanian A, Yates JR. Mass spectrometry for proteomics. Curr. Opin.
Chem. Biol. 2008, 12, 483-490.

9.

McLafferty FW, Breuker K, Jin M, Han XM, Infusini G, Jiang H, Kong XL,
Begley TP. Top-down MS, a powerful complement to the high capabilities of
proteolysis proteomics. Febs J. 2007, 274, 6256-6268.

10.

Kelleher NL, Lin HY, Valaskovic GA, Aaserud DJ, Fridriksson EK, McLafferty
FW.
Top down versus bottom up protein characterization by tandem
high-resolution mass spectrometry. J. Am. Chem. Soc. 1999, 121, 806-812.

11.

Tan YJ, Wang WH, Zheng Y, Dong JL, Stefano G, Brandizzi F, Garavito RM,
Reid GE, Bruening ML. Limited proteolysis via millisecond digestions in
protease-modified membranes. Anal. Chem. 2012, 84, 8357-8363.

12.

Appel W. Chymotrypsin - molecular and catalytic properties. Clin. Biochem.
1986, 19, 317-322.

146

Proteomics in human cancer

13.

Lee TD, Shively JE. Enzymatic and chemical digestion of proteins for
mass-spectrometry. Method Enzymol. 1990, 193, 361-374.

14.

Raijmakers R, Neerincx P, Mohammed S, Heck AJR. Cleavage specificities of
the brother and sister proteases lys-c and lys-n. Chem. Commun. 2010, 46,
8827-8829.

15.

Drapeau GR. Substrate-specificity of a proteolytic-enzyme isolated from a
mutant of pseudomonas-fragi. J. Biol. Chem. 1980, 255, 839-840.

16.

Olsen JV, Ong SE, Mann M. Trypsin cleaves exclusively C-terminal to arginine
and lysine residues. Mol. Cell. Proteomics 2004, 3, 608-614.

17.

Ma J, Zhang L, Liang Z, Zhang W, Zhang Y. Monolith-based immobilized
enzyme reactors: Recent developments and applications for proteome analysis. J.
Sep. Sci. 2007, 30, 3050-3059.

18.

Ma JF, Hou CY, Sun LL, Tao DY, Zhang YY, Shan YC, Liang Z, Zhang LH,
Yang L, Zhang YK. Coupling formic acid assisted solubilization and online
immobilized pepsin digestion with strong cation exchange and microflow
reversed-phase liquid chromatography with electrospray ionization tandem mass
spectrometry for integral membrane proteome analysis. Anal. Chem. 2010, 82,
9622-9625.

19.

Ma JF, Zhang LH, Liang Z, Shan YC, Zhang YK. Immobilized enzyme reactors
in proteomics. Trac-Trends Anal. Chem. 2011, 30, 691-702.

20.

Rey M, Man P, Brandolin G, Forest E, Pelosi L. Recombinant immobilized
rhizopuspepsin as a new tool for protein digestion in hydrogen/deuterium
exchange mass spectrometry. Rapid Commun. Mass Spectrom. 2009, 23,
3431-3438.

21.

Kim J, Kim BC, Lopez-Ferrer D, Petritis K, Smith RD. Nanobiocatalysis for
protein digestion in proteomic analysis. Proteomics 2010, 10, 687-699.

22.

Yamada K, Nakasone T, Nagano R, Hirata M. Retention and reusability of
trypsin activity by covalent immobilization onto grafted polyethylene plates. J.
Appl. Polym. Sci. 2003, 89, 3574-3581.

23.

Cooper JW, Chen JZ, Li Y, Lee CS. Membrane-based nanoscale proteolytic
reactor enabling protein digestion, peptide separation, and protein identification
using mass spectrometry. Anal. Chem. 2003, 75, 1067-1074.

24.

Xu F, Wang WH, Tan YJ, Bruening ML. Facile trypsin immobilization in
polymeric membranes for rapid, efficient protein digestion. Anal. Chem. 2010, 82,
10045-10051.

147

25.

Krenkova J, Lacher NA, Svec F. Highly efficient enzyme reactors containing
trypsin and endoproteinase LysC immobilized on porous polymer monolith
coupled to MS suitable for analysis of antibodies. Anal. Chem. 2009, 81,
2004-2012.

26.

Liuni P, Rob T, Wilson DJ. A microfluidic reactor for rapid, low-pressure
proteolysis with on-chip electrospray ionization. Rapid Commun. Mass Spectrom.
2010, 24, 315-320.

27.

Gao J, Xu JD, Locascio LE, Lee CS. Integrated microfluidic system enabling
protein digestion, peptide separation, and protein identification. Anal. Chem. 2001,
73, 2648-2655.

28.

Dulay MT, Baca QJ, Zare RN. Enhanced proteolytic activity enzymes in
photopolymerized of covalently bound sol gel. Anal. Chem. 2005, 77, 4604-4610.

29.

Dai JH, Baker GL, Bruening ML. Use of porous membranes modified with
polyelectrolyte multilayers as substrates for protein arrays with low nonspecific
adsorption. Anal. Chem. 2006, 78, 135-140.

30.

Dainiak MB, Galaev IY, Mattiasson B. Polyelectrolyte-coated ion exchangers
for cell-resistant expanded bed adsorption. Biotechnol. Prog. 2002, 18, 815-820.

31.

Miron T, Wilchek M. A spectrophotometric assay for soluble and immobilized
N-hydroxysuccinimide esters. Anal. Biochem. 1982, 126, 433-435.

32.

Mammen M, Choi SK, Whitesides GM. Polyvalent interactions in biological
systems: Implications for design and use of multivalent ligands and inhibitors.
Angew. Chem.-Int. Edit. 1998, 37, 2755-2794.

33.

Bhattacharjee S, Dong JL, Ma YD, Hovde S, Geiger JH, Baker GL, Bruening ML.
Formation of high-capacity protein-adsorbing membranes through simple
adsorption of poly(acrylic acid)-containing films at low pH. Langmuir 2012, 28,
6885-6892.

34.

Bencina K, Podgornik A, Strancar A, Bencina M. Enzyme immobilization on
epoxy- and 1,1 '-carbonyldiimidazole-activated methacrylate-based monoliths. J.
Sep. Sci. 2004, 27, 811-818.

35.

Yao CH, Qi L, Hu WB, Wang FY, Yang GL. Immobilization of trypsin on
sub-micron skeletal polymer monolith. Anal. Chim. Acta 2011, 692, 131-137.

36.

Nicoli R, Gaud N, Stella C, Rudaz S, Veuthey JL. Trypsin immobilization on
three monolithic disks for on-line protein digestion. J. Pharm. Biomed. Anal. 2008,
48, 398-407.

37.

Nicoli R, Rudaz S, Stella C, Veuthey JL.
148

Trypsin immobilization on an

ethylenediamine-based monolithic minidisk for rapid on-line peptide mass
fingerprinting studies. J. Chromatogr. A 2009, 1216, 2695-2699.
38.

Svec F. Less common applications of monoliths: I. Microscale protein mapping
with proteolytic enzymes immobilized on monolithic supports. Electrophoresis
2006, 27, 947-961.

39.

Buch I, Giorgino T, De Fabritiis G.
Complete reconstruction of an
enzyme-inhibitor binding process by molecular dynamics simulations. Proc. Natl.
Acad. Sci. U. S. A. 2011, 108, 10184-10189.

40.

Spross J, Sinz A. A capillary monolithic trypsin reactor for efficient protein
digestion in online and offline coupling to ESI and MALDI mass spectrometry.
Anal. Chem. 2010, 82, 1434-1443.

41.

Schwert GW, Takenaka Y. A spectrophotometric determination of trypsin and
chymotrypsin. Biochimica Et Biophysica Acta 1955, 16, 570-575.

42.

Laskay UA, Lobas AA, Srzentic K, Gorshkov MV, Tsybin YO. Proteome
digestion specificity analysis for rational design of extended bottom-up and
middle-down proteomics experiments. J. Proteome Res. 2013, 12, 5558-5569.

43.

Wu C, Tran JC, Zamdborg L, Durbin KR, Li MX, Ahlf DR, Early BP, Thomas
PM, Sweedler JV, Kelleher NL. A protease for 'middle-down' proteomics. Nat.
Methods 2012, 9, 822-824.

44.

Young NL, Plazas-Mayorca MD, Garcia BA.
Systems-wide proteomic
characterization of combinatorial post-translational modification patterns. Expert
Rev. Proteomics 2010, 7, 79-92.

45.

Dunn JD, Reid GE, Bruening ML. Techniques for phosphopeptide enrichment
prior to analysis by mass spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.

46.

Fontana A, Zambonin M, deLaureto PP, DeFilippis V, Clementi A, Scaramella E.
Probing the conformational state of apomyoglobin by limited proteolysis. J. Mol.
Biol. 1997, 266, 223-230.

47.

Fontana A, Spolaore B, Mero A, Veronese FM. Site-specific modification and
pegylation of pharmaceutical proteins mediated by transglutaminase. Adv. Drug
Deliv. Rev. 2008, 60, 13-28.

48.

Harris JI. Effect of urea on trypsin and alpha-chymotrypsin. Nature 1956, 177,
471-473.

49.

Fontana A, de Laureto PP, Spolaore B, Frare E, Picotti P, Zambonin M. Probing
protein structure by limited proteolysis. Acta Biochim. Pol. 2004, 51, 299-321.

149

50.

Qu HY, Wang HT, Huang Y, Zhong W, Lu HJ, Kong JL, Yang PY, Liu BH.
Stable microstructured network for protein patterning on a plastic microfluidic
channel: Strategy and characterization of on-chip enzyme microreactors. Anal.
Chem. 2004, 76, 6426-6433.

150

Chapter 6. Summary and future work

6.1. Research summary

This dissertation describes my efforts to develop functionalized membrane platforms
for protein purification and protein digestion prior to MS characterization. The
functionalization starts with membrane modification using polymeric thin films. Chapter
2 describes my fundamental study of PEMs and shows that the properties of
(PAH/PAA)n films depend on deposition conditions. For films grown at pH 3,
deprotonation of –COOH groups at neutral pH leads to a 500% increase in thickness
upon swelling in water and creates binding sites throughout the film structure to capture
multilayers of lysozyme through ion-exchange. Consistent with the ion-exchange
mechanism, lysozyme binding to these PEMs depends on the pH and ionic strength of
the binding buffers, and elution solution occurs in 1 M KSCN. (PAH/PAA)n films
adsorbed on Au-coated wafers at pH 3 bind large amounts of lysozyme at neutral pH
without increasing surface roughness, presumably because lysozyme replaces water in
the film. However, in the absence of lysozyme (PAA/PAH)n films are not stable in a
neutral pH buffer. Fortunately, film stability is greater in nylon membranes than on
Au-coated wafers, which allows the development of stable, functional membranes.
Chapter 3 describes the development of Ni2+-containing membranes for isolation of
His-tagged proteins. When deposited from pH 3 solutions, PAA/PEI/PAA films have a
high density of free -COO- groups which are available for further modification with
aminobutyl NTA to chelate Cu2+ or Ni2+ ions. Spongy nylon membranes modified with

151

PAA/PEI/PAA-NTA-Ni2+ enable rapid isolation of His-tagged CSN8 from cell extracts
with > 95% purity and recovery within 30 min. With a commercial membrane holder, the
whole process only requires 5 min.
Membranes with PAA/PEI/PAA-NTA-Ni2+ films successfully capture His-tagged
proteins, but derivatization with aminobutyl NTA is inefficient and expensive.
Additionally, weak metal-ion binding to unreacted –COO- groups of PAA leads to high
metal-ion leaching.

A newly designed polymer, PNTA, contains NTA groups and no

isolated –COO- groups.

Direct adsorption of PNTA simplifies formation of membranes

with chelating groups, and importantly these modified membranes show less Cu2+
leaching than membranes with PAA/PEI/PAA-NTA. In the future, we will use the
PNTA-modified membranes for Ni2+ binding and His-tagged protein purification.
Chapter 4 presents my efforts to develop a lectin-containing membrane to capture
glycoproteins. After adsorption of a macroinitiator, a 5-min polymerization of MES
gives a membrane that binds at least 100 mg of lysozyme/cm3 of membrane. However,
the binding of a glycoprotein was not extensive when Con A was immobilized in
membranes using covalent or metal-affinity methods with either polymer brushes or LbL
films. The low binding capacity may stem from damage to Con A, which is a
pH-sensitive protein, or more likely steric hindrance to the binding sites of Con A.
Thin (~100 µm) microporous membranes also provide a unique platform for creating
enzyme reactors. Chapter 5 describes development of membrane reactor containing
covalently anchored trypsin. The porous membrane modified with one layer of PAA
captures 35 mg of trypsin/cm3 of membrane and affords digestion rates 75 times those

152

for a typical in-solution digestion. In addition, the covalent immobilization stabilizes the
enzyme in 6 M urea or at pH 5 to enable denaturation of proteins that typically resist
digestion in their native state. These trypsin-containing membranes have a
room-temperature shelf life of over 6 months. A miniature membrane holder and syringe
pump afford control over digestion times and, hence, peptide lengths.

6.2. Future work
6.2.1. His-tagged protein purification using PNTA-containing membranes
Chapter 3 describes the Cu2+-binding and leaching studies for membranes modified
with PNTA-containing films. The data showed much less Cu2+ leaching compared to a
PAA/PEI/PAA-NTA-modified membrane. In the next step, we will investigate Ni2+
binding and leaching on the PNTA-contain membranes. Further we will examine
His-tagged protein binding specificity and recovery during isolation of His-tagged
proteins from cell lysates or protein mixtures.
6.2.2. Immunoaffinity purification with antibody-containing membranes
Chapter 4 describes my work to develop a lectin-containing membrane to isolate
specific glycoproteins. Unfortunately, these membranes have a low glycoprotein binding
capacity. However, we still can expand the membrane-based technique for
immunoaffinity purification using more specific antibody-antigen interactions.
Figure 6.1 shows typical bead-based immunopurification strategies. Incubation of
either a free or immobilized antibody with the target lysate leads to specific binding of
antigen, and subsequent capture of antibody-antigen complexes on beads followed by

153

elution and analysis allows detection of the antigen and its associated proteins. For
example, Nicol and coworkers quantified several lung-cancer biomarker candidates with
low ng/mL concentrations using enrichment with hydrazide resins modified with
multiple antibodies (1). However, these bead-based pathways require long incubation
times that lead to sacrifice for throughput (2). In addition, they suffer from nonspecific
binding due to the low binding capacity (< 1 mg/mL) (3).

Figure 6.1. Work flow of immunoaffinity purification of target proteins using either

(a) pre-immobilized antibody or (b) free antibodies that are eventually captured by
antibody-binding resins. Both the pathways require hours of incubation time.
Similar to the binding of His-tagged protein, membranes should provide an attractive
alternative to beads for rapid immunoaffinity capture. In preliminary studies, I

154

immobilized human IgG to a membrane through electrostatic interactions, but the
antibody leached from the membrane. Moreover, direct covalent immobilization to a
PAA-containing membrane immobilized < 5 mg of antibody/cm3 of membrane. Most
recently, I employed a two-step immobilization strategy, electrostatic adsorption
followed by covalent crosslinking, to immobilize IgG to the PAA/PEI/PAA modified
membrane (Figure 6.2) (4). With this strategy, I immobilized ~40 mg of antibody/cm3 of
membrane and also prevented antibody leaching. The IgG-containing membrane captures
7 mg of anti-IgG/cm3 of membrane, which is an order of magnitude higher than typical
capacities of commercial resins (3).

155

Figure 6.2. Two-step immobilization of IgG to a PAA/PEI/PAA-modified membrane.
At pH 4.5, IgG is positively charged and binds to –COO- groups in the PAA/PEI/PAA
film. IgG and EDC were dissolved in separate 20 mM phosphate solution at pH 4.5 at
concentrations of 1 mg/mL and 5 mM, respectively. The solutions were circulated through
the membrane (area of 3.1 cm2) at 1 mL/min.
This high level of IgG immobilization may introduce steric hindrance, so future
studies should optimize the amount of IgG bound to the membrane to achieve maximum
anti-IgG binding capacity. Immobilization of multiple antibodies is possible for multiple
target protein capture.

156

6.2.3. Probing protein structures using limited protein digestion with the trypsin
membrane reactor

Chapter 5 presents the development of a trypsin membrane reactor. With a syringe
pump and a miniature membrane holder, we can control the digestion time and, hence,
peptide lengths. Through a statistical analysis of proteolytic events and protein 3D
structures, Kazanov and coworkers ranked the importance of the factors affecting
proteolysis as exposure > flexibility > local interactions (5). As a result, if a protein
passes through a membrane at a high flow rate, exposed, flexible sites have the highest
probability of cleavage (Figure 6.3). The exposed sites depend on the protein
conformation, so in principle we can use limited digestion coupled with MS to probe
protein conformational changes that may occur when protein binds a ligand, such as a
drug (6-8).

Figure 6.3. Limited digestion at the most exposed site of a protein during rapid

passage through a protease-containing membrane. (The protein is not drawn to scale, as
it is much smaller than the membrane pores.)

6.3. Potential impact

The membrane-modification techniques described in this dissertation provide new
157

devices for protein purification and digestion. The PAA/PEI/PAA-NTA-Ni2+ modified
membrane technology is in the final stage of licensing, and new membrane devices will
require only 5 min for the entire His-tagged protein purification process. Adsorption of
PNTA in membranes should simplify the membrane modification to further decrease the
cost of membrane-based purification of His-tagged proteins. Rapid protein associations
with

newly

developed

antibody-containing

membranes

should

enhance

immunopurification and may improve applications such as detection of biomarkers.
The protein-digestion membranes discussed in chapter 5 contain high concentrations
of trypsin within membrane pores and can digest proteins in ms. Using the trypsin
membrane reactor, we can control protein digestion to improve sequence coverage and
PTM analysis by MS. Furthermore, we may use the limited digestion to probe changes in
protein structure due to events such as drug binding.

Such a technique may help

elucidate drug binding sites. Overall, new methods for protein capture and
immobilization in membranes should fill an important niche for rapid purification and
sample preparation in biotechnology.

158

REFERENCES

159

REFERENCES

1.

Nicol GR, Han M, Kim J, Birse CE, Brand E, Nguyen A, Mesri M, FitzHugh W,
Kaminker P, Moore PA, et al.
Use of an immunoaffinity-mass
spectrometry-based approach for the quantification of protein biomarkers from
serum samples of lung cancer patients. Mol. Cell. Proteomics 2008, 7, 1974-1982.

2.

Shi TJ, Su D, Liu T, Tang KQ, Camp DG, Qian WJ, Smith RD. Advancing the
sensitivity of selected reaction monitoring-based targeted quantitative proteomics.
Proteomics 2012, 12, 1074-1092.

3.

Anti-FLAG®
M2
magnetic
beads.
Http://www.Sigmaaldrich.Com/catalog/product/sigma/m8823?Lang=en&region=
us (accessed 11/18/2014).

4.

Qu ZY, Chen KM, Gu HC, Xu H. Covalent immobilization of proteins on 3D
poly(acrylic acid) brushes: Mechanism study and a more effective and
controllable process. Bioconjugate Chem. 2014, 25, 370-378.

5.

Kazanov MD, Igarashi Y, Eroshkin AM, Cieplak P, Ratnikov B, Zhang Y, Li ZW,
Godzik A, Osterman AL, Smith JW. Structural determinants of limited
proteolysis. J. Proteome Res. 2011, 10, 3642-3651.

6.

Jiang FG, Ramanathan A, Miller MT, Tang GQ, Gale M, Patel SS, Marcotrigiano
J. Structural basis of RNA recognition and activation by innate immune receptor
RIG-I. Nature 2011, 479, 423-428.

7.

Prota AE, Danel F, Bachmann F, Bargsten K, Buey RM, Pohlmann J, Reinelt S,
Lane H, Steinmetz MO. The novel microtubule-destabilizing drug BAL27862
binds to the colchicine site of tubulin with distinct effects on microtubule
organization. J. Mol. Biol. 2014, 426, 1848-1860.

8.

Fontana A, de Laureto PP, Spolaore B, Frare E, Picotti P, Zambonin M. Probing
protein structure by limited proteolysis. Acta Biochim. Pol. 2004, 51, 299-321.

160