w. . .
.7... m
It 1 .CH.
f 3.1

.
g
, )

 

I wilt
clay
Eliot

or ..u

A ~ xi}.
WWI-1‘ .39,
nil”

L.
,.a. a; r”. .
r11. a: #IPHJ9I

.u
5.... 19.1 .
:1. h .‘tvnli

- I251}...
. ,

IAIv-XG.‘ .i
0

1‘...

~ (Vii-1'. 15" ‘1‘: I
..-§ . I
.NWWA.’L

9.42,... mm»... Lug. .
«it!
Il‘l'l ‘
1"

t0: nv!’
II£.'€‘K
fill-0‘2!

.‘|h‘.l“l’ - l
\nn.Ll.x.IuJIOI¥
A\ 3.11.4313.
5.3!». .96), K...» .

... .,.;, .........,

 

 

LIF‘ ‘ \RY

 

 

“#:an ' “rah:
IIUI "you I vsusv
Universrty

 

This is to certify that the
dissertation entitled

CREATION OF AFFINITY MEMBRANES CONTAINING
FUNCTIONALIZED POLYMER BRUSHES FOR HIGH CAPACITY
PURIFICATION OF HISTIDINE-TAGGED PROTEINS

presented by

Parul Jain

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Chemisty

 

 

' ' Major Professor’s SW
041L111 0 ‘21

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

 

 

 

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Acc&Pres/ClRC/DateDueindd

 

 

CREATION OF AFFINITY MEMBRANES CONTAINING
FUNCTIONALIZED POLYMER BRUSHES FOR
HIGH-CAPACITY PURIFICATION OF HISTIDINE-TAGGED
PROTEINS

By

Parul Jain

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Chemistry

2009

ABSTRACT
CREATION OF AFFINITY MEMBRANES CONTAINING FUNCTIONALIZED
POLYMER BRUSHES FOR HIGH-CAPACITY PURIFICATION OF
HISTIDINE-TAGGED PROTEINS
By

Parul Jain

Porous membrane absorbers are attractive for rapid protein purification, but their
binding capacity is low relative to nanoporous beads. Modification of membranes with
functionalized polymer brushes, however, can greatly enhance capacity. Porous alumina
membranes containing poly(2-hydroxyethyl methacrylate) (poly(HEMA)) brushes
derivatized with nitrilotriacetate-Ni2+ (NTA-Ni2+) complexes facilitate purification of
polyhistidine-tagged ubiquitin (HisU) in less than 30 min. These materials have a
binding capacity of 120 mg HisU/cm3 of membrane, and the purity of the eluted HisU is
>99%, even when the feed contains 10% bovine serum.

Unfortunately, the submicron pore size in commercial alumina membranes limits
their use to simple solutions because complex mixtures such as cell extracts often plug
the small pores. Polymeric membranes, on the other hand, can have larger pore
diameters (1-10 pm) that allow lower pressure drops and purification of more complex
solutions. Nylon membranes containing poly(HEMA)-NTA-Ni2+ brushes can isolate
polyhistidine tagged (His-tagged) cellular retinaldehyde binding protein (CRALBP) from
a cell extract with purities at least as good as those obtained with commercial Ni2+

columns. Unfortunately, these membranes have a low protein binding capacity (25 mg

protein/cm3 of membrane), perhaps because the organic solvents employed in brush
synthesis and derivatization partially damage the membrane structure.

A newly created aqueous procedure for growth of polymer brushes inside
polymeric membranes avoids contact of polymer membranes with organic solvents. This
method includes layer-by-layer adsorption of macroinitiators and subsequent aqueous
ATRP from these immobilized initiators. However, the formation of protein-binding
brushes still requires conversion of poly(HEMA) hydroxyl groups to carboxylic acid
moieties, which involves the use of organic solvents that may damage the membrane. A
new surface-initiated aqueous ATRP of an acidic monomer, 2-(methacryloyloxy)ethyl
succinate (MES), overcomes this problem and provides a rapid, one-step route to
polyacid brushes. ATRP from initiators immobilized on Au—coated Si wafers yields
poly(MES) films with an ellipsometric thickness of 120 nm in less than 15 min. FTIR
spectroscopy and ellipsometry studies Show that poly(MES) brushes and their derivatives
bind the equivalent of many monolayers of BSA as well as lysozyme.

Finally, modification of nylon membranes with poly(MES)-NTA—Ni2+ allows for
high-capacity purification of His-tagged proteins directly from a cell lysate. These
membranes show remarkable HisU, BSA, and lysozyme binding capacities of 85, 80 i 2
and 118 d: 8 mg protein per cm3 of membrane, respectively. Most importantly, the
poly(MES)-NTA-Ni2+-modified membranes allow isolation of His-tagged CRALBP
directly from a cell extract in less than 15 min with purities comparable to commercial
affinity columns. Thus, porous nylon membranes modified by poly(MES)-NTA-Ni2+
brushes are attractive candidates for rapid, high-capacity purification of His-tagged

proteins.

I dedicate this dissertation to my parents, Ashok Jain and Sangeeta Jain, my
husband, Sameer Patel and my siblings, Mayura Jain and Kamesh Jain, for their

love and support.

iv

ACKNOWLEDGMENTS

I would like to express my sincere thanks to my dissertation advisor, Dr. Merlin
Bruening, for his insightful guidance and continuous support. He is a great mentor and
an excellent scientist. He gave me the freedom to explore on my own and at the same
time provided the guidance to recover when my steps faltered. I have learned a lot of
useful presentation and scientific Skills from him. I could not thank him enough for his
patience and understanding.

I am deeply grateful to my second supervisor, Dr. Gregory Baker, for the discussions
that helped me sort out the technical details of my work. I am also thankful to him for
carefully reading and commenting on countless reviSions of the manuscripts.

I also thank my committee members, Dr. James Geiger and Dr. Gary Blanchard for
their kind help and great suggestions. Thanks to Dr. Geiger and his students, Dr. Xiaofei
Jia and Remie Touma, for kindly providing the cell extracts needed for this study.

My deepest thanks to Dr. J inhua Dai for being a great mentor to me. I would also like
to thank current and past Bruening research group members. I

I acknowledge the help of several people for training me on different instruments.
Thanks to Dr. Kathy Severin for the guidance on transmission F TIR and ICP, Dr. Ruth
Smith for training on ATR-IR, Dr. Dan Holmes for help with NMR and Dr. Baokang Bi
for training on SEM. Thanks to Per Askeland for help with XPS studies.

Dr. S.K. Jain, my mentor in undergraduate studies, is one of the best teachers I had. I
am indebted to him for his continuous encouragement, guidance and for always believing

in me.

Most importantly, none of this would have been possible without the love and patience
of my family. I am fortunate to have a family who has been a constant source of love,
concern, support and strength all these years. I owe the maximum gratitude to my
parents, Ashok Jain and Sangeeta Jain, for their unwavering faith and confidence in me
and for making me the person I am today. Thanks to my husband, Sameer, for his love,
support, understanding and constant encouragement. He believed in me more than I did
and never let me settle down for less. Without him, I could not have done what I was
able to do. Thanks to my sister, Mayura Jain and my brother, Kamesh Jain, for their

immense love and concern. I owe my success to my family.

vi

TABLE OF CONTENTS

Page
List of Figures ....................................................................................... xi
List of Schemes ...................................................................................... xv
List of Abbreviations .............................................................................. xvii
Chapter 1. Introduction and background ...................................................... 2
1.1. Polymer brushes ........................................................................... 2
1.1.1. Definition of polymer brushes ........................................................... 2
1.1.2. Synthesis of polymer brushes ......................................................... 3
1.2. Applications of polymer brushes .............. - .......................................... 10
1.2.1. Protein immobilization and purification with polymer
brush-modified substrates ............................................................ 10
1.2.1-a. Advantages of using polymer brushes for protein
immobilization and purification .............................................. 10
1.2.1-b. Protein immobilization on brush-modified flat surfaces .................. 11
1.2. l-c. Protein immobilization and purification using
brush-modified beads .......................................................... l3
1.2.1-d. Protein immobilization and purification using
brush-modified monoliths ..................................................... 15
1.2.1-e. Polymer brush modified substrates as protein microarrays..... . . . . . . . ...l6
1.2.1-f. Polymer brushes for enzyme immobilization .............................. 26
1.2.2. Polymer brush-based capture of proteins for analysis by
mass spectrometry ..................................................................... 27
1.2.3. Polymer brushes for capillary chromatography and
electrochromatography ............................................................... 30

1.2.4. Selective protein purification in polymer brush-modified membranes... . . . . .32
124-3. Membrane absorbers for protein purification via

affinity interactions ............................................................ 32
1.2.4-b. Membrane absorbers for protein purification via

ion-exchange interactions ..................................................... 35

1.3. Outline of the dissertation ............................................................... 37
1.4. References ................................................................................. 40

Chapter 2. High-capacity purification of His-tagged proteins by alumina

membranes containing functionalized poly(HEMA) brushes ................ 47

2.1. Introduction ................................................................................ 47
2.2. Experimental section ...................................................................... 50

vii

2.2.1. Materials ................................................................................ 50

2.2.2. Polymerization of HEMA in porous alumina membranes ....................... 51
2.2.3. Polymerization of HEMA on Au substrates ........................................ 51
2.2.4. Poly(HEMA) derivatization and protein immobilization. . . . . . . . 5.1
2.2.5. Film characterization methods. ...52
2.2.6. Determination of the amount of cOordinated Ni” in ”the membrane .............. 53
2.2.7. Protein quantification... ...53
2.2.8. Determination of protein purity by” sodium dOdecyl sulfate-polyacrylamrde
gelelectrophoresis (SDS— PAGE)... 5.4
2.3. Results and discussion ..................................................................... 54
2.3.1. Characterization of poly(HEMA)- derivatized membranes ...................... 54
2.3.2. HisU binding to poly(HEMA)-NTA-Ni2+ brushes on gold-coated silicon
substrates ............................................................................... 58
2.3.3. HisU binding to poly(HEMA)—NTA-Ni2+ brushes In membranes ............. 64
2.3.4. Separation of protein mixtures using poly(HEMA)-NTA-Ni2+ brushes
on Au substrates” ...................................................................... 67
2.3.5. Separation of protein mixtures using poly(HEMA)-NTA-Ni2+ brushes
in membranes" . .................... ‘ ...................................................... 68
2.4. Conclusions ............................................................................... 71
2.5. References ................................................................................. 73

Chapter 3. Purification of His-tagged proteins by polymer membranes

containing functionalized poly(HEMA) brushes ........................... 76
3. 1 . Introduction ................................................................................ 76
3.2. Experimental Section ..................................................................... 77
3.2.1. Materials .................................................................................... 77
3.2.2. Polymerization of HEMA in porous polymer membranes... . . . . . . . . . . . . .......77
3.2.3. Poly(HEMA) derivatization and protein immobilization... ..................... 78
3.2.4. Characterization methods ............................................................. 80
3.2.5. Protein quantification ................................................................. 80
3.2.6. Determination of protein purity by sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE) ................................................... 81
3.3. Results and Discussion ................................................................... 81
3.3.1. Characterization of poly(HEMA)- derivatized membranes... ........81
3. 3. 2. BSA binding to poly(HEMA)-NTA- Cu2+ brushes In membranes” .....84
3. 3. 3. Purification of His-Tagged CRALBP using nylon membranes modified
with poly(HEMA)-NTA-Ni2+ brushes .............................................. 86
3.4. Conclusions .................................................................................... 92

3.5. References93

Chapter 4. Completely aqueous procedure for growth of polymer brushes on
polymeric substrates ............................................................... 95

4.1. Introduction .................................................................................... 95

viii

4.2. Experimental Section ...................................................................... 97

4.2.1. Materials .................................................................................... 97
4.2.2. Characterization methods ............................................................ 98
4.2.3. Determination of the internal surface area of a PBS membrane ............... 98

4.2.4. Synthesis of 2-(2-bromoisobutyryloxy)ethy1 acrylate..... . . . . . . . . . ..............99

4.2.5. Synthesis of poly(2-(dimethylamino)ethyl methacrylate-co-Z-(Z-
bromoisobutyryloxy)ethyl acrylate) (poly(DMAEMA-co—BIEA), the
precursor of the polycationic initiator .............................................. 100

4.2.6. Synthesis of poly(2-(trimethylammonium iodide) ethyl methacrylate-co-
BIEA) (poly(TMAEMA-co-BIEA)), the polycationic

 

initiator ................................................................................ 101
4.2.7. Polymer brush synthesis ............................................................. 102
4.3. Results and Discussion .................................................................. 103
4.3.1. Synthesis ofpolycationic macroinitiator ........................................... 103
4.3.2. Polymer brush growth from surfaces modified with
poly(TMAEMA-co-BIEA). . . . . . .. ................................................................ 107
4.3.3. Polymer brush growth on PBS ....................................................... 112
4.4. Conclusions........ ............................................................................................ 118
4.5. References........ .............................................................................................. 119
Chapter 5. Rapid synthesis of functional polymer brushes by surface-initiated
atom transfer radical polymerization of an acidic monomer ............................. 122
5. 1 . Introduction .................................................................................. 122
5.2. Experimental Section .................................................................... 126
5.2.1. Materials .................................................................................. 126
5.2.2. Polymerization of MES on Au substrates ....................................... 127
5.2.3. Derivatization of carboxylic acid groups and protein immobilization. 128
5.2.4. Quantification of protein binding .................................................. 129
5.2.5. Kinetics of solution polymerization of MES, MAA and HEMA..... . . . .......129
5.2.6. Instrumentation ....................................................................... l 30
5.3. Results and Discussion .................................................................. 130
5.3.1. Kinetics of surface-initiated MES polymerization. ..130
5. 3. 2. Comparison of ATRP of MES, MAA and AA .................................. 134
5. 3. 3. Effect ofNaOH on MES polymerization... 140
5.3 .4 Protein binding to poly(MES) brushes and their derrvatrves .................... 141
5 .4 Conclusions" . ............................................................................................ 144
5.5. References....”m.... .............................................................................................. 145
Chapter 6. His-tagged protein purification with high capacity affinity membranes
containing functionalized poly(MES) brushes , 149
6. 1 . Introduction .................................................................................. 1 49
6.2. Experimental Section .................................................................... 150
6.2.1. Materials .................................................................................. 150
6.2.2. Initiator attachment in porous polymer membranes ............................ 151

ix

6.2.2-a. Initiator attachment in hydroxylated (LoProdyne® LP) nylon

membranes ..................................................................... 151

6.2.2-b. Initiator attachment in non-hydroxylated nylon membranes ........... 151
6.2.3. Polymerization of MES in porous nylon membranes ........................... 153
6.2.4. Poly(MES) derivatization and protein immobilization ......................... 153
6.2.5. Characterization ofbrush growth and derivatization.... . 155
6.2.6. Protein quantification ............................................................... 156
6.2.7. Determination of protein purity by SDS-PAGE .................................. 156
6.3. Results and Discussion .................................................................. 156
6.3.1. Characterization of poly(MES)-derivatized membranes. . . . . . . . . . .........156
6.3.2. Protein binding to polymer brush-modified membranes .............................. 159

6.3.2—a. Protein binding capacity in two membranes with different

pore sizes ...................................................................... 159
6.3.2-b. Protein binding capacity as a function of polymerization time ......... 164
6.3.2-c. Protein binding capacity as a function of flow rate ....................... 165
6.3.2-d. Protein binding in membranes with different compositions ............ 166

6.3.3. HisU binding to poly(MES)-NTA-Ni2+ brushes in membranes. . ..........167
6.3.4. Purification of His-Tagged CRALBP from cell extracts.............. . . . . ......168

 

6.4. Conclusions......... ........................................................................................... 170
6.5. References. . . . . . .. .............................................................................................. 172
Chapter 7. Summary and future work ..................... 174
7.1. Research summary174
7.2. Future work ............................................................................... 176
7.3. Potential Impact ........................................................................... 180
7.4. References ................................................................................. l 82

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 1.4.

Figure 1.5.
Figure 1.6.

Figure 2.1.

Figure 2.2.

Figure 2.3.

Figure 2.4.

Figure 2.5.

Figure 2.6.

Figure 2.7.

Figure 2.8.

LIST OF FIGURES

Structures of the polymers included in chapter 1 ....................................... 9

Schematic representation of protein immobilization on polymer brush-
modified magnetic beads ........................................................................ 14

Schematic representation of lipase immobilized in a porous hollow-fiber
polyethylene membrane .................................................................... 27

Protocol for phosphopeptide enrichment using poly(HEMA)— NTA-Fe(III)

films on Au MALDI plates ........................................................................ 29
His-tagged protein binding to a NTA-Ni2+ modified surface .................... 33
Protein binding to a polymer brush inside a membrane pore ................... 34

Binding of His-tagged protein to a NTA-Ni2+ derivatized poly(HEMA)
brush inside a membrane pore. . . . . . . . . .. .................................................... 49

Transmission FTIR spectra of an alumina membrane modified with
poly(HEMA) brushes before and after reaction with succinic anhydride,
activation with EDC/N HS, reaction with aminobutyl NTA, and exposure
to 0.1 M NiSO4 .................................................................................... 56

Subtracted reflectance FTIR spectra of protein immobilized on
poly(HEMA)-NTA-Ni2+ brushes after exposure of the films to
O. 01 mg/mL HisU, 0. 1 mg/mL BSA, or 0.1 mg/mL ubiquitin” .............59

Reflectance FTIR spectra of spin-coated BSA, myoglobin, lysozyme,
and ubiquitin films on Au .............................................................. 61

Amount of HisU bound to poly(HEMA)-NTA—NI2+ films as a function of
concentration and time. .... .63

HisU binding capacity versus thickness of poly(HEMA) films that were
subsequently derivatized with NTA-Niy ................................................ 64

Breakthrough curves for absorption of 0. 3 mg/mL HisU, 0.3 mg/mL BSA,
and 0.35 mg/mL myoglobin In membranes modified with poly(HEMA)-
NTA Ni2+ or poly(HEMA)-NTA Cu2 .... .............65

SDS- PAGE analysis of protein solutions and eluents from Au-
poly(HEMA)-NTA-Ni2+ films and poly(HEMA)-NTA-Ni2+ modified
alumina membranes ................................................................................ 68

xi

Figure 2.9.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 3.4.

Figure 3.5.

Figure 3.6.

Figure 4.1.

Figure 4.2.

Figure 4.3.
Figure 4.4.

Figure 4.5.

Figure 4.6.

Figure 4.7.

SDS- PAGE analysis of 10% bovine serum spiked with 0.3 mg/mL HisU
and the imidazole eluent from an alumina-poly(HEMA)-NTA-Ni2+
membrane loaded with this solution ............................................. 70

ATR-FTIR spectra of a hydroxyl functionalized nylon membrane before
and after formation of poly(HEMA) brushes inside the membrane;
reaction with succinic anhydride; activation with EDC/N HS; and reaction
with aminobutyl NTA ................................................................................ 82

SEM images of a bare nylon membrane with a 1.2 pm nominal
filtration cutoff, a Similar membrane modified with poly(HEMA)-NTA-
Cu2+ ,and a cross-sectional Image of the membrane” ... ......84

Breakthrough curve for absorption of 1 mg/mL BSA In a nylon membrane
modified with poly(HEMA)-NTA- Cu"2+ ............................................ 85

SDS-PAGE analysis of an extract from E. coli containing over—expressed
His-tagged CRALBP; and CRALBP purified from these extracts using
flow-through and immersion methods ........................................... 88
SDS-PAGE analysis of an extract from E. coli containing over-expressed
His-tagged CRALBP and CRALBP purified from cell extracts containing
10 mM and 5 mM imidazole .................................................................. 89
SDS- PAGE analysis of an extract from E. coli containing over-expressed
His-tagged CRALBP and CRALBP purified from the cell extracts using
poly(HEMA)—NTA-Ni2+-modified membranes and spin-trap columns... 9.1
1H NMR of 2-(2-bromoisobutyryloxy)ethyl acrylate .............................. 100

1H NMR Spectra of poly(DMAEMA-co-BIEA) and its quatemized
product poly(TMAEMA-co-BIEA) ................................................. 101

GPC trace ofpoly(TMAEMA-co-BIEA)...........................................104
XPS elemental analysis of quatemized poly(DMAEMA-co-BIEA) ....... 106

1H NMR spectrum of a lO-month old sample of
poly(TMAEMA-co-BIEA) ...................................................... 107

Reflectance FTIR spectra of poly(TMAEMA-co-BIEA)/
[PSS/poly(TMAEMA -co-BIEA)]n films deposited on
Au—MPA substrates (n=0-4). . . . . . . .. ......................................................... 108

Reflectance FTIR spectra poly(HEMA) films grown from

xii

Figure 4.8.

Figure 4.9.

Figure 4.10.

Figure 4.11.

Figure 4.12.

Figure 4.13.

Figure 4.14.

Figure 5.1.

Figure 5.2.

Figure 5.3.

Figure 5.4.

Figure 5.5.

Figure 5.6.

poly(TMAEMA-co-BIEA)/ [PSS/poly(TMAEMA -c0-BIEA)]n films
deposited on Au-MPA substrates (n=0-4). ..................................... 109

Poly(HEMA) brush ellipsometric thickness versus number of bilayers for
grth from poly(TMAEMA-co-BIEA)/[PSS/poly(TMAEMA -co-
BIEA)]n films deposited on Au—MPA substrates (n=0—4) ................... 110

Poly(HEMA) brush thickness versus polymerization time for growth from
Au-MPA-[PDADMAC/PSS]2-poly(TMAEMA-co-BIEA) films
and from Au-MUD-BribBr films ................................................ 111

AF M Images of (a): a Au-MPA-(PDADMAC/PSS)2-poly(TMAEMA
-co-B1EA) film; (b): (a) + a poly(HEMA) brush .................................... 1 13

Reflectance F TIR of (a): spin-coated PES on Au;
(b): (a) + PSS(PDADMAC/PSS)4; (c): (b) + poly(TMAEMA-co-BIEA);
(d): (c) + poly(DMAEMA) brush; and (e): (c) + poly(HEMA) brush....114

Transmission FTIR Spectra of a bare PES membrane and a PBS membrane
modified with a PSS(PDADMAC/PSS)4/poly(TMAEMA—co-BIEA)
film and 1 h polymerization of poly(HEMA) .......................................... 116

Calibration curve of the absorbance at 1730 cm'1 versus the areal
concentration of poly(HEMA) in KBr pellets ......................................... 1 17

SEM image of a bare PES membrane and 3 PES membrane dissolved in
dichloromethane afier derivatization with poly(HEMA) ......................... 1 18

Reflectance F TIR spectra of a poly(MES) brush on a Au-coated Si wafer
before and after activation with EDC/NHS; and reaction with

aminobutyl NTA ...................................................................................... 132
Evolution of ellipsometric thickness with time for surface-initiated
polymerization of MES using HMTETA; Me4Cyclam/anbpy;
MeéTREN; and bpy catalyst systems ................................................... 133

Evolution of ellipsometric film thickness with time for surface-initiated
polymerization of MES, HEMA, and MAA ............................................ 135

Percent monomer conversion as a function of time for solution
polymerization of MES and MAA ........................................................... 137

1H NMR Spectra of a MES polymerization solution 8 min and
8 h afier the addition of initiator ............................................................. 138

1H NMR spectra of a MAA polymerization solution 8 min and

xiii

Figure 5.7.

Figure 5.8.

Figure 5.9.

Figure 6.1.

Figure 6.2.

Figure 6.3.

Figure 6.4.

Figure 6.5.

Figure 6.6.

Figure 6.7.

Figure 6.8.

Figure 6.9.

Figure 7.1.

Figure 7.2.

8 h after the addition of initiator ............................................................. 139

Thickness of poly(MES) brushes as a function of the molar ratio of
NaOH to MES added to the polymerization solution .............................. 140

Reflectance FTIR spectra of a poly(MES) film and
a poly(MES)-lysozyme film .................................................................... 142

BSA binding capacity as a function of poly(MES) film thickness ........ 143

ATR-FTIR spectra of a hydroxyl functionalized nylon membrane before
and after formation of poly(MES) brushes inside the membrane; activation
with EDC/NHS; and reaction with aminobutyl NTA .............................. 157

SEM images of a bare nylon membrane with a 1.2 pm
nominal filtration cutoff and a similar membrane modified with
apoly(MES)-NTA-Cu2+ film ................................................................... 159

Breakthrough curve for absorption of 1 mg/mL lysozyme in poly(MES)-
modified nylon membranes with 1.2 pm and 5.0 pm nominal

filtration cutoffs ....................................................................................... 160

Breakthrough curve for absorption of BSA in nominal 1.2 pm and 5.0 pm

nylon membranes modified with poly(MES)-NTA-Cu2+ brushes ........... 161
SEM images of pristine nylon membranes with 1.2 pm and 5.0 pm

nominal filtration cutoffs ......................................................................... 163
Lysozyme binding capacity as a function of polymerization time. 164

Breakthrough curves for absorption of 1 mg/mL BSA in
poly(MES)-NTA-Cu2+-modifed nylon membranes at different flow ...... 166

Breakthrough curve for absorption of 0.3 mg/mL HisU in
poly(MES)-NTA-Ni2+-modifed nylon membranes .................................. 168

SDS-PAGE analysis of an extract from E. coli containing over-expressed
His-tagged CRALBP, and CRALBP purified from the cell extracts using a
poly(MES)-NTA-Ni2+-modified membrane, and a spin-trap column ..... 169

Picture of syringe filter with a disposable syringe and a
membrane holder ................................................................. 179

Purification of His-tagged protein from a cell extract using a stack of
poly(MES)-NTA-Ni2+ membrane3180

xiv

Scheme 1.1.

Scheme 1.2.

Scheme 1.3.

Scheme 1.4.

Scheme 1.5 .

Scheme 1.6.

Scheme 1.7.

Scheme 1.8.

Scheme 1.9.

Scheme 1.10.

Scheme 1.11.

Scheme 2.1.

Scheme 3.1.

Scheme 4.1.

LIST OF SCHEMES

Schematic representation of different approaches to polymer brush
synthesis .................................................................................................... 4

Possible methods for attaching initiators to hydroxylated surfaces, Au, and
a1um1na5

Growth of polymer brushes via ATRP from macroinitiator deposited on a
surface using layer-by-layer adsorption ........................................... 6

Mechanism of ATRP ................................................................................ 8

Protein immobilization on poly(AA) brush-modified films via
ion-exchange interactions and metal-ion affinity interactions ................... 12

A schematic illustration of anti-hCG immobilization on NHS-modified

poly(CBMA) brushes .............................................................................. 19
Fabrication of protein-functionalized poly(oligo(ethy1ene glycol)

methacrylate) brushes ............................................................. 20
Pattering of PEG and poly(AA) brushes on a silicon surface .................... 22

Immobilization of Fab’ fiagments in defined orientation on poly(MPC)-b-
poly(GMA) brushes. . . . . . . .. ...................................................................... 24

Attachment of trichlorosilane initiator to a silica capillary surface, ATRP
of HEMA from the immobilized initiator, activation of poly(HEMA) by
1,1'-carbonyldiimidazole, and derivatization with ethylenediamine ....... 31

Modification of a porous high-density polyethylene hollow fiber
membrane with poly(GMA) brushes and subsequent reaction of the
brushes with 3-aminopropionic acid, (2-aminoethy1)phosphonic acid,

or 2-aminoethane-1-sulfonic ac1d36

Derivatization of poly(HEMA) with NTA-Ni2+ prior to protein
adsorption ........................................................................... 55

Schematic illustration of the growth and derivatization of poly(HEMA)
brushes inside polymer membranes ............................................. 79

Growth of polymer brushes by ATRP from macroinitiators adsorbed in a
membrane pore ..................................................................... 96

XV

Scheme 5.1. Polymerization of HEMA or MES from initiator immobilized on a
gold substrate ........................................................................................... 123

Scheme 5.2. Synthesis of protein-binding poly(MES) brushes on Au surfaces.......... 1 31

Scheme 6.1. Schematic illustration of (a) activation of the amide groups in a nylon
membrane to introduce hydroxyl functionalities (b) trichlorosilane
initiator immobilization on the hydroxylated membrane and (c)

polymerization of MES from initiator-modified membrane... ................ 152

Scheme 7.1. Schematic illustration of grth of poly(MES) brushes by ATRP from
macroinitiators adsorbed in a membrane pore ........................................ 176

xvi

AA

AGT

AIBN

LIST OF ABBREVIATIONS

Acrylic acid
O6—alkylguanine-DNA-alkyltransferase

2,2’-Azobis(2-methylpropionitri1e)

Aminobutyl NTA Na, Na—bis(carboxymethyl)-L-lysine hydrate

ATR
ATRP
BIEA
pr
BSA
CD1
CRALBP
DMAEMA
DMAP
DMF
anbe
EDC
EDTA

en
FESEM
FITC
FTIR

GPC

Attenuated total reflectance

Atom-transfer radical polymerization
2-(2-Bromoisobutyryloxy)ethyl acrylate
2,2’-Bipyridyl

Bovine serum albumin

1 ,1 '-carbony1diimidazole

Cellular retinaldehyde binding protein
2-(Dimethylamino)ethyl methacrylate
4-Dimethylaminopyridine
Dimethylformamide
4,4'-Dinony1-2,2'-bipyridyl
l-[3-(Dimethylamino)propyl]-3-ethy1carbodiimide hydrochloride
Ethylenediamine tetraacetic acid
Ethylenediamine

Field-emission scanning electron microscopy
Fluorescein isothiocyanate

Fourier transform infrared

Gel permeation chromatography

xvii

HEMA
His-tag
HisU
HMTETA
IMAC
MAA
MALDI
MALDI-MS
Me4Cyclam
MEHQ
MES
Me6TREN
MPA

MUD

NHS

NTA
OT-CEC
PDADMAC
PDMS
PEG

PES

PGA

PMI

2—Hydroxyethy1 methacrylate

Polyhistidine tag

His-tagged ubiquitin

1 , 1 ,4,7,10,10-Hexamethyltriethylenetetramine
Immobilized metal-affinity chromatography
Methacrylic acid

Matrix-assisted laser desorption/ionization
Matrix-assisted laser desorption/ionization mass spectrometry
1 ,4,8, 1 1 -Tetraaza- 1 ,4,8, 1 1 -tetramethy1cyclotetradecane
Monomethyl ether hydroquinone I
2-(Methacryloyloxy)ethy1 succinate
TriS[2-(dimethylamino)ethy1]amine
3-Mercaptopropionic acid

1 l-Mercapto- 1 -undecanol
N-Hydroxysuccinimide

Nitrilotriacetate

Open-tubular capillary electrochromatography
Poly(diallyldimethylammonium chloride)
Poly(dimethyl siloxane)

Polyethylene glycol

Polyethersulfone

Penicillin G acylase

Polycationic macroinitiator

xviii

Poly(CBMA) Poly(carboxybetaine methacrylate)

Poly(GMA) Poly(glycidyl methacrylate)

Poly(MAA) Poly(methacrylic acid)

Poly(MPC) Poly(2-methacry1oyloxyethylphosphorylcholine)

Poly(TMAEMA) Poly(2-(trimethylammonium iodide)ethy1 methacrylate)

PSS Poly(styrene sulfonate)
PTFE Polytetrafluoroethylene
PVDF Polyvinylidine fluoride
SAM Self-assembled monolayer

SDS-PAGE Sodium dodecyl sulfate-polyacrylamide gel electrophoresis

THF Tetrahydrofuran

xix

This chapter is adapted from our recently published work in
Annual Review of Analytical Chemistry (Jain, R, Baker, G.L.,
Bruening, M.L. Annual Review of Analytical Chemistry 2009, 2,

387-408).

Chapter 1

Introduction and background

This dissertation describes the grth of polymer brushes on and in porous
supports to form high-capacity membrane absorbers capable of purifying proteins
modified with polyhistidine tags (His-tags). The research builds on previous studies of
the synthesis and application of polymer brushes towards protein immobilization, so to
put this work in perspective, I first define polymer brushes and then describe different
approaches for the synthesis of polymer brushes (Section 1.1). Subsequent sections
discuss the applications of polymer brush-modified surfaces for protein immobilization
and isolation. Specifically, I describe the use of polymer brush-modified flat surfaces,
beads, and monoliths towards protein immobilization and purification followed by the
application of polymer brush-modified substrates as protein microarrays (Section 1.2.1.).
Sections 1.2.2. and 1.2.3. present polymer-brush based capture of proteins for analysis by
mass spectrometry, capillary chromatography, and electrochromatography. Subsequently
I describe selective protein purification in polymer brush-modified membranes via
affinity and ion-exchange interactions (Section 1.2.4.). Finally, I present an outline of the
dissertation.

1.1. Polymer brushes
1.1.1. Definition of polymer brushes

Polymer brushes are assemblies of polymeric molecules tethered to a substrate
such that the graft density is high enough to force the polymer chains to extend away
from the surface.I The end of the polymer chain is usually held on the surface by

physisorption or covalent bonding, whereas the bulk of the chain extends into the solution

or air interface as shown in Scheme 1.].2 In an appropriate solvent, brushes can be
swollen and highly extended for rapid capture and purification of proteins or other
analytesf"5 Highly swollen, hydrophilic brushes are also useful for minimizing non-
specific adsorption of proteins,6'8 and the functional groups in such brushes can be readily
tailored for specific separations.
1.1.2. Synthesis of polymer brushes

Initially, polymer brushes were formed by physical adsorption of block
copolymersf”10 In this case, one block has affinity for the surface, while the other block
extends into the solvent. However, because such systems are often unstable, recently
developed synthetic techniques use covalent attachment of polymers to substrates to

provide more robust brushes. Covalent grafting generally occurs using either “grafting

9911 9912-14

to or “grafting from techniques. In the “grafting to” method, end-functionalized
polymer chains bind to a substrate via chemical reaction between active groups on the
surface and active end groups in the polymer chains (Scheme 1.1.A.).ll This method
results in relatively low grafting densities because steric hindrance prevents incoming
polymer chains from diffusing through previously deposited chains to reactive surface
sites. In contrast, in the “grafting fiom” strategy, the polymer chains grow directly from
surface-tethered initiators (Scheme 1.1.8.)”14 The “grafting from” approach yields a
high density of chains because small monomers can readily reach growing chains or
initiators on the surface. Initiator immobilization on the surface is a vital step in the

15.16

“grafting from” method and affords some control over brush density. Typical

initiator-attachment strategies include reaction of surface hydroxyl/amino groups with

17,18

acid chlorides or acid bromides (Scheme 1.2.(a)), modification of Au surfaces using

A. “Grafting to” B. “Grafting from”

mm m

\ktéff {mi

1 __ ,

     

 

. Reactive end group in m Monomer
polymer chain

Y Reactive group on the ? Initiation site
surface

Scheme 1.1. Schematic representation of different approaches to polymer brush
synthesis. (A) “Grafting to” method in which active groups on the surface and
reactive end groups in the polymer chains react to form a covalent bond. (B)
“Grafting from” method in which the polymer chains grow fiom surface-tethered

initiators.

thiols or disulfides (Scheme 1.2.(b)),'8’l9 reaction of alumina or silica with silanes
(Scheme 1.2.(c)),'8'20 and adsorption of polyelectrolyte macroinitiators (Scheme
1.3.).2"22 A number of recent review articles provide an extensive discussion of the

synthesis of polymer brushes.23'25

T—OH Et N —0
(a) + Brfir —’3 \n)(Br

 

 

 

 

 

 

 

 

0 Dry 0
DMF

__ L

S (CH2)11 Br 0

/
_ /O O

S (CH2)11 Br
__1 O _.
T T
L “’OH Cl L

\
(C) I "—OH +Cl/S\I'(CH2)11 Br I _0\S'i/(CH2)11 31'

N —OH CI 0 N (I) O
..6. A \

 

 

 

 

Scheme 1.2. Possible methods for attaching initiators to (a) hydroxylated surfaces,

(b) An, and (c) alumina.

Polymer brush—>

  
   

MacrornItIator\

 
 

  
 
    

(1 $101!! '1'}: "s (911'! 3'9.\
Polyelectrolyte { , ' ' ' ‘
multilayers

(5 hbfc'i '9?) 9'! V9131 6'93

 

Polymerization
————p

..t‘; ' ' A“ - ..4

- Suabstrte -

0.. .' A.

‘ Suabstrte -

Example of macroinitiator:

W
o *o o’C“o

Poly (2-(trimethylamino)ethyl methacrylate-co-
2-(2-bromoisobutyryloxy)ethy| acrylate)

Scheme 1.3. Growth of polymer brushes via ATRP from macroinitiator deposited on

a surface using layer-by-layer adsorption.

The “grafting from” approach has been used to modify various surfaces with

26

almost all of the known polymerization techniques including cationic, anionic,27

29.30 I 31.32 133
’

radical,28 ring-opening metathesis, photochemica and electrochemica

polymerization. However, controlled radical polymerization techniques such as atom-
transfer radical polymerization (ATRP),34'36 reversible addition-fragmentation

37,38

transfer, and nitroxide-mediated polymerization39 have emerged as some of the most

powerful synthetic methods for brush formation because they afford controlled polymer

growth and relatively low polydispersity. Additionally, surface-initiated polymerization
with these techniques frequently results in minimal polymerization in solution.

ATRP is especially attractive for controlled polymerizations due to its mild
reaction conditions (room temperature in many cases), use of readily available catalysts,
initiators and monomers, and tolerance to impurities.”42 This polymerization technique
proceeds as described in Scheme 1.4. Radical generation in ATRP involves an initiator

(an organic halide) undergoing a reversible redox process catalyzed by a transition metal
complex,43 and R' is the reactive radical that initiates the polymerization. The controlled

nature of ATRP is due to the reversible activation-deactivation reaction between the
growing polymer chains and a copper halide—ligand species (km and kdeact are the rate
constants for activation and deactivation reactions respectively). The ligand forms a
complex with the cuprous and cupric salts and helps to solubilize them in the organic
reaction system. Fast deactivation by reaction with the CuBr2(ligand) complex leads to a
low concentration of propagating radicals, thereby minimizing chain termination and
radical transfer reactions. A successful ATRP requires (a) fast and quantitative activation
of initiator so that all the propagating species begin grth at the same time and (b) rapid
reversible deactivation of growing radicals to minimize termination of living polymers.
The combination of these steps ensures a narrow molecular weight distribution because
all the propagating chains grow at the same time and for the same duration of time.43

The use of controlled polymerization is particularly important for creating films in
complex geometries such as the pores of membranes, where the brushes can serve as
high-capacity, selective adsorbents. In the synthesis of such materials, uncontrolled

polymerization or formation of polymer in solution can rapidly plug pores to prevent

flow. Surface-initiated ATRP in membrane pores minimizes polymerization in solution
and allows for control over the polymer brush thickness through variation of
polymerization time. Husson and coworkers used ATRP to grow poly(poly(ethylene
g1ycol)methacrylate) brushes (Figure 1.1.(i)) inside regenerated-cellulose ultrafiltration
membranes.44 At a constant pressure, the water flux through the membrane decreased
monotonically with increasing polymerization time because growth of the brushes
decreased the pore diameters. This study also showed that the molecular weight cutoff of
the membrane decreases with increasing polymerization time, further confirming that
ATRP provides control over the diameter of the pores in the membrane. Yusof and
Ulbricht studied the effects of photo-grafting conditions and monomer concentration on
polymer brush grth inside membrane pores.45 They found that the density of polymer

chains in the membrane correlates with the density of the entrapped photo-initiator,

 

benzophenone.
kact
R— Br + CuBr (ligand) ‘——‘ R0 + CuBr2 (ligand)
kdeact

l Monomer (M)
RMn -
CuBr2 (ligand)

RMn— Br + CuBr(ligand)

Scheme 1.4. Mechanism of ATRP. (Redrawn from Odian, G. Principles of

Polymerization; 2004, Fourth ed.; Wiley-Interscience, pp 316.).

CH3 CH3

—l—CH2-C|3-}-n— —{-CH2-CH]Ln— ——[-CH2-C+—

C C
o/ l\ /C\ci0H 0/ /\0H COH\
H2 0/
(i) Poly(poly(ethylene (ii) Poly(acrylic acid) (iii) Poly(methacrylic acid)
glycol)methacrylate)
CH3 CH3
__[_CH2-C_}n_H ‘lCHZ 0+}. H —['CH2‘CH]l—
/C H2 2
H c /C
o/ C/\ 0’C\7 C/ \ O/\C\C/ OH 0/ NH2
0 I
OH
(iv) Poly(glycidyl (v) Poly(glycerol (vi) Poly(acrylamide)
methacrylate) monomethacrylate)
CH3
_{_CH2_C_]n_H —‘['CH2‘C‘l'n—H g2 H2
0 0 $1032in Ho‘cf \ 43% /°“
0/ \O Cic/OH /CO\ C: \C/ \FI/ O\fi/ \IE'CH3 H2 0 n 32
H2 H2 0- 2 CH3

(vii) Poly(hydroxyethyl (viii) Poly(Z-methacryloyloxyethyl (ix) Poly(ethylene glycol)
methacrylate) phosphorylcholine)

—+CH2-CHl—n—

CH3
—[—CH2-C-ln—H —i'CH2‘CHl“
/C CH3 H2 0 n
0/ 0:1 C / / N
\o c— —N+—C/ \C/ CH2 \l
H2 I H2 (IT N'+C|
CH3 H3C/C | N\CH3
CH3
(x) Poly(carboxybetaine methacrylate) (xi) Poly(vinylbenzyl (xii) Poly(Z-vinylpyridine)
trimethylammonium
chloride)

Figure 1.1. Structures of the polymers included in this chapter.

1.2. Applications of polymer brushes
In recent years, polymer brush-modified surfaces were examined for

applications in various fields of science and technology. This section describes recent
developments in the use of polymer brush-modified flat surfaces, beads and monoliths for
protein immobilization and purification. Subsequently, I discuss the applications of
polymer brushes in protein microarrays, for enzyme immobilization, in mass
spectrometry, and in chromatography and electrochromatography.
1.2.1. Protein immobilization and purification with polymer brush-modified
substrates
1.2.1.-a. Advantages of using polymer brushes for protein immobilization and
purification

Polymer brushes are attractive for protein immobilization and purification
because swollen brushes can potentially bind the equivalent of many monolayers of
protein. Increases in binding capacity may enhance the efficiency or sensitivity of
analytical devices such as membrane absorbers, protein microarrays, and modified
matrix-assisted laser desorption/ionization (MALDI) plates used for protein capture prior
to analysis. Several methods for immobilizing proteins in brushes have been reported,
including covalent binding, electrostatic adsorption (ion exchange), and binding to metal-

20.46.47

ion complexes. Brushes containing carboxylic acid and epoxide groups are

(1.48’49 Poly(acrylic acid)

particularly common because they can be readily derivatize
(Poly(AA) Figure 1.1.(ii)) brushes are especially attractive for protein immobilization
because in aqueous solution these films swell to four times their initial thickness to

facilitate binding of large biomolec:ules.‘“"49'50

10

1 .2.1-b. Protein immobilization on brush-modified flat surfaces

Due to the ease in handling and characterization, polymer brush-modified flat
surfaces such as silicon and gold-coated silicon wafers have been extensively used for
studying protein immobilization. Dai and coworkers modified Au-coated Si with
pol y(AA) brushes and their derivatives and immobilized lysozyme in these films via ion-
exchange (Scheme 1.5.A.) and metal-ion affinity interactions (Scheme 1.5.B.).50 Both
methods give high protein-binding capacities. Remarkably, about 80 monolayers (16.2
ug/cmz) of lysozyme adsorb on a 55 nm thick poly(AA) film on Au via electrostatic
adsorption. Functionalization of the poly(AA) brushes with nitrilotriacetate (NTA)-Cu2+
complexes yields films capable of adsorbing large amounts of protein via metal-ion
affinity interactions. A 55 nm poly(AA) film modified with NTA-Cu2+ binds 5.8 ).lg/cm2
of bovine serum albumin (BSA), 7.7 (lg/cm2 of myoglobin and 9.6 (lg/cm2 of anti-
Immunoglobulin G. This corresponds to around 20 monolayers of protein in these films.
Recently, Cullen, et al. used poly(AA)-NTA-Cu2+ brushes to immobilize Ribonuclease A
at a capacity of 11 pg Ribonuclease A per cm2 of film, which is 30 monolayers of the

immobi l ized enzyme.49

11

l N l /

 

l /
S‘(-CH2)O-C—CH o o StCHziO-C—CH
11 [H\C—E-l—Br Au 11 [H\C—E—l—Br
2 I n (NHS) 2 l n
c c
// \ ——O // \
OH 0
O EDC O |
Poly(AA) brush oflfo
(A) Protein
1. Aminobutyl NTA
0 CH3 (3) 2. Cu2+
ll /
. S(—CH27\1(1)-C—C\H H 3. Protein
[HZC—CIHEBr
c
/ +
o/
+ + +

 

 

Scheme 1.5. Protein immobilization on poly(AA) brush-modified films via (A) ion-
exchange interactions and (B) metal-ion affinity interactions. For metal-ion affinity
binding, polymer brushes were activated with N-hydroxysuccinimide (NHS) in the
presence of 1-[3—(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC).
The NH S-activated poly(AA) brushes were derivatized with NTA-Cu2+ complexes.

RedYaWn from Dai, J. H.; Bao, Z. Y.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M.

L, Langmuir 2006, 22, 4274-4281.

1.2.l-c. Protein immobilization and purification using brush-modified beads

Beads are attractive substrates for protein purification because of their high
surface area. Several groups used beads modified with polymer brushes to purify
proteins directly from egg white via ion-exchange.”53 As an example, Bayramoglu and
coworkers used poly(methacrylic acid) (poly(MAA), Figure l.1.(iii))-grafied chitosan
beads for purification of lysozyme from 50% diluted egg white at pH 6.0.52 The
chitosan-g-poly(MAA) beads showed a lysozyme binding capacity of ~66 mg
lysozyme/g of beads. Single-step purification of lysozyme from egg white with these
beads resulted in 94% pure lysozyme as determined by high-performance liquid
chromatography. The high purity is achieved because lysozyme has a net positive charge
at pH 6.0, whereas most other proteins in egg white are negatively charged at this pH.

Recently, a number of groups demonstrated the immobilization of proteins on
magnetic beads modified with polymer brushes (Figure 1.2.).54‘55 Because these beads
can be collected or focused using a modest magnetic field, they are attractive for use in
drug delivery, immunoassays, protein and enzyme immobilization, and in the separation,
isolation, and analysis of biomolecules. Huang and coworkers immobilized BSA in
poly(glycidyl methacrylate-co-glycerol monomethacrylate) brushes grafted to magnetic
microspheres.54 The epoxide groups in poly(glycidyl methacrylate) (poly(GMA), Figure
l.1.(iv)), units can covalently bind proteins, whereas the hydrophilic poly(glycerol
monomethacrylate) (Figure 1.1.(v)) units enable the microspheres to disperse efficiently
in aqueous solution. These materials have a binding capacity of ~27 mg BSA per g of
beads. The poly(glycidyl methacrylate-co-glycerol monomethacrylate) brushes can also

immobilize penicillin G acylase (PGA), and the activity of this enzyme depends on the

13

(A)

ATRP % Protein.
—-> —-—9

Magnetic bead Polymer brush-modified Protein immobilization
bead

magnetic
particles"—>

Figure 1.2. (A) Schematic representation of protein immobilization on polymer

 

(B)

     
 
 

brush-modified magnetic beads. (B) Image of lOO-nm diameter silica-coated

magnetic beads with (right) and without (lefi) collection by a magnet.

ratio of the constituent monomers.56 The maximum PGA activity (753 U per g of beads)

occurs when the weight ratio of GMA to glycerol monomethacrylate used to form the

brushes is 60/40. Enzyme activity decreases when more of the hydrophilic monomer,

glycerol monomethacrylate, is present because the enzyme substrate, penicillin, must

diffuse through the hydrophilic polymer to the enzyme. With >60% of the hydrophobic

polymer, GMA, the activity of the beads also decreases because the brushes collapse in

water and bind little enzyme. Compared to its free form, microsphere-immobilized PGA

l4

is less sensitive to changes in temperature and pH. The activity of the immobilized
enzyme decreased by 8.5% when the temperature was changed from 45 °C to 55 °C,
while for the free enzyme under the same conditions, the enzyme activity decreased by
80%. Approximately 64% of the enzyme activity was retained after ten cycles of
repeated use.

1.2.l-d. Protein immobilization and purification using brush-modified monoliths

Polymer brush-modified silica monoliths have also been used for purification of

57,58 28,59,60

proteins using ion-exchange, size exclusion, and hydrophobic-interaction

chromatography.6l'63

Kikuchi et al. reviewed the use of polymer brush-modified
stationary phases for applications in different areas of chromatography.62 lmportantly,
controlled, surface-initiated radical polymerization enables fine control over the thickness
of polymer brushes so modification of porous stationary phases with brushes affords

. 9
control over pore srze.28’5

Huang and Wirth modified nanoporous silica gel with
poly(acrylamide) (Figure l.1.(vi)) brushes and used this gel for a size-exclusion-based
chromatographic separation of proteins.28 The thickness of the poly(acrylamide) brushes
was 10 nm, which was much smaller than the average pore size of the silica gel (86 nm).
Thus the polymer film reduced the pore size of the silica gel but did not plug the pores.
Using a column of the modified gel, thyroglobulin (mol. wt. 66430 Da), ovalburnin
(44000 Da), and ribonuclease (13700 Da) were separated in order of decreasing
molecular weight.

In the case of smaller proteins, however, strong interactions between analytes and

polymer chains limit the use of polymer brushes in size-exclusion chromatographic

columns.64 Yoshikawa, et al. studied the interactions of proteins with ‘ poly(2-

15

hydroxyethyl methacrylate) (poly(HEMA), Figure 1.1.(vii)) modified silica as well as
the effects of the poly(HEMA) on size-exclusion chromatography.60 Their findings
suggest that the interactions of poly(HEMA) brushes with proteins are minimal on the
outermost surface but more prominent inside the brushes. Thus, for large proteins that
cannot penetrate the brushes, the separation is dominated by size-exclusion. However,
for smaller proteins, which can enter the brushes, both size and adsorption affect the
separation. The ability to use both size-exclusion and adsorption may allow some
separations that are not possible with size-exclusion alone.
1.2.1.-e. Polymer brush modified substrates as protein microarrays

In the last decade, microarrays of antibOdies and enzymes have emerged as
important tools for rapid, parallel analyses of a wide range of analytes.46‘65 In the
immobilization of proteins in arrays, however, unwanted non-specific adsorption often
lowers the signal-to-background ratio and also generates false-positive identifications.“
Thus an efficient array substrate should demonstrate specificity towards the desired
protein in appropriate areas and at the same time show minimal non-specific interactions.

Recent studies suggest that surfaces modified with polymer brushes are more
efficient substrates for protein microarrays than nonpolymeric self-assembled monolayers

(SAMs) containing the same functional groups.“68

Brushes are advantageous over
monolayers for several reasons. First, as shown in section l.2.1-b., brushes have a high
protein-binding capacity that can enhance the sensitivity of protein arrays. A second
asset of protein arrays formed with polymer brushes is that the three-dimensional

structure of swollen brushes should allow ready access to binding sites. In the case of

non-polymeric self-assembled monolayers, immobilized antibodies/enzymes lie flat on

16

 

the surface and are not highly accessible to the antigen/substrate molecules. For instance,
silicon wafers modified with copolymer brushes containing poly(2-methacryloyloxyethyl
phosphorylcholine) (poly(MPC)) (Figure l.1.(viii)) and poly(GMA) can immobilize 4
times more antibody Fab’ fragments than a SAM containing epoxy groups};7 Moreover,
the antibody fragments immobilized on the polymer brushes show ~6 times higher
activities than antibodies attached to epoxysilane films. This increased activity suggests
that the antibody fragments immobilized in polymer brushes are more accessible to
antigens than the antibodies immobilized on the monolayers.

Hydrophilic polymer brushes are especially attractive substrates for antibody
arrays because water-swollen films typically show low non-specific protein adsorption.
Poly(ethylene glycol) (PEG) (Figure l.1.(ix)) 6’7 brushes are especially recognized as
biocompatible materials that resist protein adsorption, but poly(AA),50 poly(HEMA)}r’O‘69
and poly(carboxybetaine methacrylate) (poly(CBMA), Figure 1.1.(x)) 8 brushes also
exhibit low non-specific interactions and can be fimctionalized to allow binding of
proteins. Zhang and coworkers modified gold films with poly(CBMA) brushes using
surface-initiated ATRP and showed that these brushes prevent the non-specific
adsorption of fibrinogen, lysozyme and human chorionic gonadotropin (hCG).8
Immobilization of anti-hCG on N-hydroxysuccinimide (NHS)-modified poly(CBMA)
brushes (Scheme 1.6.) results in specific binding of hCG while maintaining resistance to
non-specific protein binding.

Similarly, Tugulu, et al. utilized glass slides modified with poly(HEMA),
poly(oligo(ethy1ene g1ycol)methacrylate), or poly(MPC), for synthesizing nonfouling'

films for protein microarrays. The presence of the polymer brushes prevents non-specific

l7

 

protein binding and at the same time, immobilization of 06-benzylguanine onto the
brushes results in chemoselective immobilization of OC-alkylguanine-DNA-
alkyltransferase (AGT) fusion proteins (Scheme 1.7.).69

Wang et al. developed a chitosan-g-methyl-PEG-coated poly(dimethyl siloxane)
(PDMS) microchip to minimize non-specific protein adsorption.70 The methyl-PEG units
provide hydrophilic domains and minimize non-specific adsorption of biomolecules. A
fluorescence image of the polymer-coated microfluidic channels was acquired after
exposure to fluorescein isothiocyanate (FITC)-labeled bovine serum albumin (BSA) for a
period of 24 h. The image shows no detectable fluorescence, indicating effective
suppression of BSA adsorption on this microchip.70 On the other hand, a PDMS
microchip without polymer modification showed bright fluorescence after exposure to

FITC-BSA for 24 h.

 

. $H3
AU \/[’CH2_(I3}BBT

 

 

 

 

C:
d 0 (EH3 II
‘CHz—CHz—u+—CH2—CH2—c—o-
CH3
Poly(CBMA) Brush
1. NHS/EDC
2. Anti-hCG
.-j~ (le3
Au vl-CHz-(PlfiBr
,C=o

 

 

 

0 C“3 l?

\CHZ-CHz—N+_CH2‘CH2-C‘NH‘Q
CH3

1 hCG .

 

i“3
Au («kHz—char
. [0:0

0 9H3 9
\CH2_CH2—N+_CH2—CH2‘C—NH‘“
CH3

Scheme 1.6. A schematic illustration of anti—human chorionic gonadotropin (anti-

 

 

 

 

hCG) immobilization on N-hydroxysuccinimide (NHS)-modified poly(CBMA)
brushes. Immobilization of anti-hCG results in specific binding of hCG protein.
Poly(CBMA) brushes were modified with NHS in the presence of l-[3-
(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC). Redrawn from

Zhang. Z.; Chen. S. F.; Jiang. S. Y. ,Biomacromolecules 2006, 7. 3311-3315.

19

Scheme 1.7. Fabrication of protein-functionalized poly(oligo(ethylene
g1ycol)methacrylate) brushes: (3) grafting of ATRP initiator and surface-initiated
ATRP of oligo(ethylene glycol)methacrylate; (b) activation of hydroxyl groups with
p-nitrophenyl chloroformate; (c) functionalization with 06-(4-(13-amino-2,5,8,l l-
tetraoxatridecyl)oxymethylbenzyl)guanine and quenching of residual p-nitrophenyl
chloroformate groups; ((1) immobilization of O6-alkylguanine-DNA-alkyltransferase
(AGT) fusion protein on benzylguanine surfaces. Redrawn from Tugulu, 8.; Arnold,

A.; Sielaff, 1.; Johnsson, K.; Klok, H. A. Biomacromolecules 2005, 6, 1602-1607.

20

Vie/V0”
m
CuBr,CuBr2,bpy

H20, MeOH,
60°C

 

 

(a) 0

 

o
N /LI>LO\/>\O/\/NH2
1 </ k 3
n \N NH;

 

 

   
   

 

OZNOOYO 0Y0
CIA COMET

THF, Et3N, RT
('3)

 

,DMAP, DMF M

2. ”ZN/V0”, DMF, RT

 

0
, N 0\/\
Protein of </ / IN 0
interest N \NXNH "H
H 2 o

 

21

In addition to minimizing non-specific interactions, substrates for protein
microarrays should also prevent denaturation of protein molecules. Ober and coworkers
developed a lithographic method (Scheme 1.8.) to produce protein patterns with minimal
denaturation and a low level of non-specific interactions.“ A patterned PEG surface was
back-filled with ATRP initiators, and poly(AA) brushes were grown from the initiator—
containing regions. F [TC-labeled BSA was covalently immobilized on poly(AA) brushes

that were activated with NHS/l-[3-(dimethylamino)propyl]-3-ethylcarbodiimide

uv light

_/_\_

Photoreslst V _—| m {——

PEG PEG PEG
Silicon Silicon Silicon

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02 plasma to — Strip
etch PEG a F photoresist —‘l m I—'

 

 

 

 

 

 

 

 

 

 

 

a #> _
Silicon Silicon
Surface-initiated
Back-fill with polymerization of Na
ATRP initiator acrylate

 

 

 

 

 

 

Silicon Silicon

 

 

Scheme 1.8. Patterning of PEG and poly(AA) brushes on a silicon surface. Redrawn
from Dong, R.; Krishnan, S.; Baird, B. A.; Lindau, M.; Ober, C. K.

Biomacromolecules 2007, 8, 3082-3092.

22

hydrochloride (EDC). The fluorescence image of these films showed well-defined,
bright BSA patterns. In contrast, the PEG modified regions were dark due to minimal
attachment of BSA. Thus this method can yield protein microarrays with low non-
specific binding.“

Proper orientation of an enzyme or antibody is required to maintain biological
activity and should be taken into account when developing a substrate for immobilizing
biomolecules. Randomly oriented proteins frequently show decreases in activity due to
the inaccessibility of the active site. Control over protein orientation is generally
achieved by (a) changing the surface charge,72 or (b) using site-specific immobilization
through biotin-streptavidin interactions or thiol-diSulfide interchange reactions between

67'69'73’74 Iwata and coworkers utilized well-

polymer brushes and protein molecules.
defined block copolymer brushes consisting of poly(MPC) and poly(GMA) on silicon
wafers to immobilize antibody Fab’ fragments in a defined orientation.”74 The
orientation of the antibody fragments was defined by derivatizing the GMA units with
pyridyl disulfide and immobilizing the antibodies via thiol-disulfide interchange reactions
(Scheme 1.9.). Increases in the length of poly(GMA) units resulted in increased loading
of the antibody fragments due to the availability of more binding sites. The fluorescence
intensity after reaction of immobilized antibody fragments with FITC-labeled antigen
also increased with an increasing length of poly(GMA) units because of the increased
loading of the antibody fragments. Polymer brushes without pyridyl disulfide moieties
were also used to immobilize antibody Fab’ fragments via reaction with the epoxy groups

on GMA units. The fluorescence intensity arising from binding of labeled antigen to

antibodies linked to unmodified poly(GMA) blocks was 20 times lower than that

23

Scheme 1.9. Immobilization of F ab' fragments in defined orientation on poly(MPC)-
b-poly(GMA) brushes. The proper orientation was achieved by (a) introduction of
pyridyl disulfide units onto the polymer brushes followed by (b) a thiol-disulfide
interchange reaction between thiol groups in Fab' fragments and pyridyl disulfide
units in the polymer chain. The activity of the immobilized antibody fragments was
investigated by (c) reaction with F [TC-labeled mouse immuglobulin G. Redrawn

from Iwata, R.; Satoh, R.; Iwasaki, Y.; Akiyoshi, K. Colloid Surface B 2008, 62, 288-

298.

24

 

a it‘s,
giO—FCHZT 1 m 1 CHZ‘ ?-—}'Br

'0 H2
o/C\ o/
Cl)/C§O 0C?
0 H CH3
”ZC\C/o\g/0\C/C§N '+ 0,3 Poly(MPC)-b-poly(GMA)
H2 I_ H2 [Cl
0 CH3

(a) l 1. Dithiothreitol
2. 2,2-Dithiodipyridine and 2-thiopyridone

 

3 W310H3

(.0. C ‘ l iCHZ‘tI'J'i—Br

,9 m 'C\" H H2 :2 9“ N’ l
O/C\\\0 0% O/ng/C\S/ \g/fi\C/S\S \
| 5H OH H2

H c
o 2 ”3
H2C\ /O\F,/O\ /C\ I-vaH
c c N -3
(1)_H2 I HCI

Fab” fra ment
(b) H21 HMB g

F ITO-conjugated mouse lgG
(C)

I l W3 I 3

a) _ l l _ -..}

Q CHzT 1m [CHZCE ”Er H H OH
6 ¢C\ / 2 CZ |

 

2 s
N /C§O O O C\g/C\S/ \$/CH:\C/ \S
o , H
l H 01-13 OH OH 2
0
”2R /0\H/0\ /c2\ 1+, CH3
ii i C 'lcr
2 0' H2 CH3

25

obtained using antibodies attached to pyridyl-disulfide-modified brushes. Presumably,
the random orientation of antibodies attached to unmodified poly(GMA) results in a
lower activity.
1.2.1-f. Polymer brushes for enzyme immobilization

Enzymatic reactions in non-aqueous media are also important in industrial
applications,”76 but enzymes frequently do not show sufficient activity in non-aqueous
solvents. To overcome this problem enzymes have been coated with surfactants77 and
immobilized on microspheres78 and in polymer brush-modified membranes.79’8'
Enzymes immobilized in polymer brush-modified membranes show higher activity than

9- . .
7 8' This 1s because

enzymes coated with surfactants or immobilized on microspheres.
convective flow brings the substrates to the immobilized enzymes and minimizes mass-
transport limitations. Goto and coworkers immobilized Iipases via ion-exchange inside a
porous polyethylene hollow fiber (Figure 1.3.). The anion-exchange sites were created
by radiation-induced grafting of poly(GMA) brushes followed by reaction of the GMA

units with diethylamine.80

The immobilized lipase showed 23-fold higher activity than
the native lipase (suspended in substrate solution) in the esterification reaction between
lauric acid and benzyl alcohol. The grafted poly(GMA) acts as a hydrophobic surfactant
to stabilize the enzyme and enhance activity. Moreover, reuse of the immobilized lipase

three times in a batch reactor over a period of 24 h resulted in no loss in activity, whereas

there was a 75% decrease in the activity of native lipase under similar conditions.8|

26

     

Grafted polymer brush
(protection from
non-aqueous
environment)

 

Hollow fiber

   
   
      
       

Micropore Immobilized lipase

  

Figure 1.3. Schematic representation of lipase immobilized in a porous hollow-fiber
polyethylene membrane. The pores of the membrane were modified with poly(GMA)
brushes followed by reaction with diethylamine. Lipase irmnobilization occurred via
ion-exchange interactions, and the lipase-containing membranes were used to study
the esterification reaction between lauric acid and benzyl alcohol. Redrawn from
Goto, M.; Kawakita, H.; Uezu, K.; Tsuneda, S.; Saito, K.; Goto, M.; Tamada, M.;

Sugo, T. Journal of the American Oil Chemists Society 2006, 83, 209-213.

1.2.2. Polymer brush-based capture of proteins for analysis by mass spectrometry
The identification and analysis of the proteins associated with specific diseases is

a major opportunity in analytical chemistry. Identification of phosphorylated proteins, in

particular, is important because changes in phosphorylation states can cause various

82

diseases including cancer. Mass spectrometry is ofien employed for the identification

27

and analysis of phosphorylated proteins, but the low abundance of the phosphorylated
proteins and suppression of signals by non-phosphorylated proteins make such analyses
challenging. One way to overcome this problem is to develop methods for selective
capture of the proteins of interest from a pool of unwanted proteins. A recent review
describes a number of techniques for phosphopeptide enrichment such as immobilized
metal-affinity chromatography (IMAC), reversible covalent binding, and metal oxide
affinity chromatography.82

Capture of peptides directly in polymer brushes on a MALDI plate is attractive
for high-throughput analysis of moderately complex samples. Dunn and coworkers used
poly(HEMA) brushes in on-plate enrichment of phosphopeptides for analysis by matrix-
assisted laser desorption/ ionization mass spectrometry (MALDI—MS) (Figure 1.4.).83
Au-coated Si wafers were modified with poly(HEMA)-nitrilotriacetate (NTA)-Fe(III)
brushes, and small volumes (~lpL) of digest containing phosphopeptide, non-
phosphopeptides and salts were spotted on these films. After incubation, the films were
washed and dried, and matrix was added for MALDI-MS analysis. In a specific example,
mass spectra of a B-casein digest were obtained with and without enrichment on
poly(HEMA)-NTA-Fe(III) brushes. Enrichment on the brushes yielded a 30-fold
increase in the signal intensities of several phosphorylated peptides relative to the
conventional MALDI-MS analysis. Additionally, the enrichment procedure resulted in
signals from several phosphopeptides that were undetectable in conventional MALDI-
MS, as well as decreased signals for nonphosphopeptides. Finally, enrichment decreased
the detection limit to 15 fmol for the peptide with m/z 2062 in a B-casein digest. The

effectiveness of polymer brushes in the enrichment of phosphopeptides is likely due to

28

if if.

Poly(HEMA)-NTA-Fe(lll) brushes on gold plate

 

Add sample solution containing
phosphopeptides, O
non-phosphopeptides and salts A

 

‘3‘ ‘A‘ we “case
A A A §C§lemiw

A A A

lA A A A A A AJ

Add matrix

 

Rinse and dry

as +

Waste solution

 

Figure 1.4. Protocol for phosphopeptide enrichment using poly(HEMA)— NTA-

Fe(III) films on Au MALDI plates.

29

the high density of peptide-binding sites, as 60 nm-thick poly(HEMA)-NTA-Fe(III)
brushes have a binding capacity of ~0.6 pg/cm2 for phosphoangiotensin. Remarkably,
these brushes showed ~70% recovery of a synthetic monophosphopeptide and 100%
recovery of a diphosphopeptide. In contrast, MALDI plates modified with a monolayer
of NTA-Fe(III) gave a monophosphopeptide recovery of only ~9%.

1.2.3. Polymer brushes for capillary chromatography and electrochromatography

84,85

Polymer brushes may also find use in capillary chromatography and open-

tubular capillary electrochromatography (OT-CEC),86’87

where the polymers are grafted
on the inner walls of the capillary and separation is achieved without the need for packing
of the column. The advantages of using polymer brushes in capillary chromatography
and OT-CEC include optimization of separation efficiency by controlling the polymer
brush thickness as well as the ease of derivatization of the polymer chains for specific
binding. Huang and Wirth found that polyacrylamide coated capillaries enhance protein

64 Miller and coworkers used substituted

separations by decreasing surface adsorption.
poly(HEMA) brushes as stationary phases in OT-CEC and demonstrated that
derivatization of the brushes by appropriate reagents allows for separation of a wide
range of molecules.86 Specifically, derivatization of poly(HEMA) brushes with octanoyl
chloride or ethylenediamine (Scheme 1.10.) facilitated separation of a series of phenols
and anilines, which could not be separated using bare silica capillaries or underivatized
brushes. They studied separation of three amines, aniline, 4-nitroaniline and 3, 5-
dichloroaniline, using a bare silica column as well as a capillary coated with

ethylenediamine-derivatized poly(HEMA). These amines could not be resolved using

bare silica columns whereas the use of polymer-coated capillaries resulted in full

30

 

 

 

 

 

 

 

 

 

 

 

? (:1 TE
OH
1'. *— Cl/Slrj‘(CH2)1TO\n)<Br \O 0
c _ : ('3
A OH 18h _ \
Initiator-modified film
HEMA
CuCl/CuBrz, bpy
b
(Si\/((3"l2)11O/N></:I//\>r:rOH
Poly(HEMA)
CDl'
18h
..5? CH O
\S|i/( 2)11\O Br
0\ n
—— o o /
//\N
N\//
CDl-activated poly(HEMA) \g/
ethylenediamine (en)
__ 12 h
\o
—‘ H
en- poly(HEMA) ‘\/0\n/N\/\NH2
O

*CDI: 1,1’-carbonyldiimidazole
Scheme 1.10. Attachment of trichlorosilane initiator to a silica capillary surface,
ATRP of HEMA from the immobilized initiator, activation of poly(HEMA) by 1,1 '-
carbonyldiimidazole (CD1), and derivatization with ethylenediamine (en). Redrawn
from Miller, M. D.; Baker, G. L.; Bruening, M. L. Journal of Chromatography A

2004, 1044, 323-330.

resolution. Although the peaks were broad, they were fully separated. Moreover,
increasing the thickness of the coatings enhanced the resolution for several aniline pairs
presumably due to increase in effective stationary phase to mobile phase volume ratio.
1.2.4. Selective protein purification in polymer brush-modified membranes
1.2.4-a. Membrane absorbers for protein purification via affinity interactions
Purification is often the bottleneck step in producing proteins for therapeutic or
research purposes. Perhaps the most powerful method for protein purification is affinity
adsorption in which immobilized ligands interact specifically with an affinity tag that is
genetically engineered into the protein of interest. Modified surfaces selectively bind the
proteins containing the affinity tags, whereas other cellular proteins can be washed away.

88,89

The most common affinity tags are polyhistidine, streptavidin,90 glutathione-S-

1 and maltose binding protein.” The use of polyhistidine tags allows

transferrase,9
purification by IMAC where selectivity is usually based on the interaction of the
polyhistidine with a Ni2+ complex immobilized on silica beads or in a gel (Figure 1.5.).
However, drawbacks to IMAC include slow diffusion of macromolecules into porous
beads, difficulties in packing large columns, relatively high pressure drops, and long

separation times.93’94

These drawbacks become particularly important in large scale
separations.

Membrane absorbers95 have the potential to provide more rapid affinity
purification than column-based methods because convective flow, rather than diffusion,
brings proteins to binding sites in membrane pores. Convective rinsing of pores may also

89

help to remove non-specifically adsorbed proteins. Additionally, the development of

32

Q

 
 
    

  

HN/:N

:11 Eff: .21 WE:

LN\

His,3 -tagged
protein

Ni2+

O I \ -~‘ _ z ’,\‘1\ ‘0 .
N, o Aminobutyl
o NTA-Ni2+
complex
0
HN

r’]\2:\

 

Figure 1.5. His-tagged protein binding to a NTA-Ni2+ modified surface.

disposable membranes for one time use would avoid challenges with crossover between
samples.44

Unfortunately, typical protein-absorbing membranes suffer from low binding
capacities relative to porous beads. In order to increase the capacity of membrane
absorbers, the membrane pores can be modified with polymer brushes (Figure
1.6.).20‘89'96‘98 The brushes are attractive because they have multiple protein binding sites
and can be easily modified with ligands for specific biological recognition. A wide range
of polymeric and inorganic membranes such as aluminafo‘89 silica,99 PVDF,98 nylon,100
polyethersulfone,22 regenerated cellulose,'°' and polyethylene'02 have been modified with
polymer brushes to develop protein-absorbing membranes with high protein-binding
capacity as well as selectivity towards the protein of interest. The ability of polymer

brushes to enhance protein binding to membranes depends greatly on both membrane

33

geometry and the polymer brush. Sun et al. showed that modification of microporous
alumina membranes with poly(HEMA) and subsequent derivatization of the
poly(HEMA) brushes with NTA-Cu2+ gives membranes with a BSA-binding capacity as
high as 130 mg/cm3 of membrane (~95 mg of BSA per g of membrane), which is several
fold higher than the capacities of other protein absorbing membranesls'zo'99 This high
capacity stems in part from both the relatively small pore diameter (0.2 pm) in the
membrane and the thickness of the polymer brushes. Binding capacity should increase as

pore size decreases, but unfortunately so does resistance to flow.

Protein binding
polymer brush

Protein 0
-——-—>

   

/'

Porous membrane

Figure 1.6. Protein binding to a polymer brush inside a membrane pore.

34

1.2.4-b. Membrane absorbers for protein purification via ion-exchange interactions

Membrane absorbers are also being used for protein immobilization via ion-
exchange interactions. The separation is achieved by differential absorption of charged
proteins at oppositely charged membrane surfaces. For instance, Kawakita and
coworkers formed anion-exchange membranes through modification of porous glass with
poly(GMA)-diethyl amine brushes. The resulting membranes showed a BSA binding
capacity of 12 mg protein per g of membrane.99 Kumar, et al. grafted
poly(vinylbenzyltrimethylammonium chloride) (Figure 1.1.(xi)) brushes onto cotton
cellulose and achieved an equilibrium binding capacity of 40 mg BSA per g of

3 Ulbricht and coworkers formed cation-exchange microfiltration

membrane.IO
membranes by photo-grafting copolymers of acrylic acid and a cross-linking monomer,
methylene bisacrylamide, inside the pores of polypropylene membranes.3"45 The
incorporation of cross-links within the grafted polymer layers led to higher dynamic
protein-binding capacities than grafting of a linear polymer, even when the amount of
fimctional groups was same in both cases.45 Overall, the poly(acrylic acid-co-methylene
bisacrylamide)-modified membranes exhibited a lysozyme binding capacity of 60
mg/cm3, which is 30 times higher than the binding capacity of an unmodified membrane.
Husson and coworkers reported that regenerated cellulose membranes modified with poly
_ (AA) brushes via 1 h of ATRP show static lysozyme binding capacities of 99 mg/mL.‘01
This capacity is 2-3 times higher than the capacity of commercial Sartobind C

membranes. However, the Sartobind C membranes are 40 times more permeable because

of a larger pore size.

35

OH NHC H R
1. Electron beam 1. NH2C2H4R 2 4

 

 

‘W 'l'_'
o
o

 

 

     

 

 

a , ‘_I'——'|
2. GMA 2. H20 OH OH
PE fiber GMA fiber Amphoteric fiber
Inner diameter 2.0 mm R Fiber name
Outer diameter 2.0 mm COOH AC(X)*diO|
Porosity 70% PO3H2 AP(X)-di0|
Pore diameter 0.36 pm 303H AS(X)-di0|

 

 

 

Scheme 1.11. Modification of a porous high-density polyethylene hollow fiber
membrane with poly(GMA) brushes and subsequent reaction of the brushes with 3-
aminopropionic acid (AC), (2-aminoethyl)phosphonic acid (AP), or 2-aminoethane-l-
sulfonic acid (AS). Redrawn from Iwanade, A.; Umeno, D.; Saito, K.; Sugo, T.

Biotechnology Progress 2007, 23, 1425-1430.

Iwanade and coworkers modified porous hollow fiber membranes with
poly(GMA) brushes and reacted the epoxide groups of the polymer with ampholite
molecules containing amino and anionic groups such as carboxylic acids (Scheme
1.11.).'04 The resulting membranes showed multilayer protein binding for lactoferrin,
cytochrome C, and lysozyme. In a specific example, membranes containing (2-
aminoethyl)phosphonic acid as an ampholite had equilibrium adsorption capacities of 130
mg/g, 150 mg/g and 190 mg/g for lysozyme (feed concentration of 0.2 g/L), cytochrome
C (feed concentration of 0.5 g/L) and lactoferrin (feed concentration of 1.0 g/L),

respectively. Moreover the elution efficiency was >99%. However, mutual repulsion

36

between the anionic groups caused extension of the polymer brush and a 3-fold decrease
in the flux relative to unmodified membranes. In fact, decreased membrane permeability
due to extended polymer brushes is the major disadvantage of grafting charged polymer
brushes into porous membranes. To overcome this drawback, Iwanade and coworkers
suggested ionic cross-linking of the negatively charged polymer brushes with divalent
cations.105

Growth of polymer brushes in membranes can also help to decrease the dispersity
of pore sizes. A relatively narrow pore-size distribution is important for obtaining narrow
breakthrough curves. Singh and coworkers tuned the ion-exchange capacity and the pore
size of commercially available microporous PVDF membranes using controlled
polymerization of poly(2-vinylpyridine) (Figure 1.1.(xii)) inside membrane pores."8
Growth of the brushes decreased the width of the pore-size distribution and prevented
premature breakthrough of proteins in the largest pores. Controlling the polymerization
time also provided control over the average pore size of the membranes and the ion-
exchange capacity.
1.3. Outline of the dissertation

This dissertation aims at developing affinity membranes for high-capacity protein

binding as well as rapid and selective purification of His-tagged proteins. Chapter 2
shows that ATRP affords controlled grth of polymer brushes inside porous alumina
membranes without clogging the pores. Growth of poly(HEMA) brushes inside porous
alumina membranes and subsequent functionalization of poly(HEMA) with NTA-Ni2+
complexes allows rapid, highly selective purification of His-tagged proteins. Gel

electrophoresis reveals that the purity of His-tagged ubiquitin eluted from these materials

37

is >99%, even when the initial solution contains 10% bovine serum or a 20-fold excess of
BSA. Moreover, the binding capacity of the membrane is at least 5-fold greater than that
for membranes reported in the literaturezof‘l‘99 Separations can be completely performed
in 30 min or less and membranes are fully reusable.

Unfortunately, purification of His-tagged proteins using porous alumina
membranes is limited to relatively simple solutions and low flow rates because of a
limited pore size. Polymeric membranes, on the other hand, can have larger pore sizes
than porous alumina, which should allow rapid purification with more complex solutions.
In Chapter 3, I discuss the use of porous nylon membranes modified with poly(HEMA)-
NTA-Ni2+ brushes for purification of His-tagged cellular retinaldehyde binding protein
directly from a cell lysate with purities that are at least as good as those obtained with
Ni2+ columns.

Nevertheless, nylon membranes modified with poly(HEMA)-NTA-Ni2+ brushes
have a relatively low protein binding capacity (25 mg protein/cm3 of the membrane),
perhaps because the organic solvents employed in the brush synthesis and derivatization
partially damage the membrane structure. Chapter 4 of this dissertation focuses on a
completely aqueous procedure for grth of polymer brushes inside polymeric
membranes. The aqueous process avoids the use of organic solvents that may dissolve or
corrupt porous substrates. This chapter describes the use of aqueous layer-by-layer
adsorption of polyelectrolyte macroinitiators and subsequent aqueous ATRP from these
immobilized initiators for successfiil, aqueous grth of poly(HEMA) brushes on

polymeric substrates.

38

Despite the successful grth of poly(HEMA) brushes with aqueous initiation,
the creation of protein-binding brushes requires conversion of hydroxyl groups in
poly(HEMA) to carboxylic acid moieties. The formation of acid groups involves
reaction of poly(HEMA) with succinic anhydride in an organic solvent for several hours.
Chapter 5 describes surface-initiated aqueous ATRP of an acidic monomer, 2-
(methacryloyloxy)ethyl succinate (MES), that allows rapid, one-step synthesis of acidic
polymer brushes. This procedure avoids the need to react the brush with succinic
anhydride in an organic solvent. Also, poly(MES) brushes and their derivatives exhibit
high protein-binding capacities. Through FTIR and ellipsometry studies, I show that
poly(MES) brushes are capable of binding many monolayers of BSA as well as
lysozyme.

Modification of polymeric membranes with poly(MES) should allow for high-
capacity purification of His-tagged proteins directly from a cell lysate. In Chapter 6, I
describe the growth, derivatization, and characterization of poly(MES) brushes inside the
pores of nylon membranes to form protein absorbers. Analysis of the protein
breakthrough curves give a high binding capacity of 80 :l: 2 and 118 :l: 8 mg protein per
cm3 of the membrane for BSA and lysozyme respectively. Finally these membranes are
utilized to purify His-tagged proteins from cell lysate.

In the last chapter, I will present the conclusions drawn from my research and

some proposed future work.

39

1.4. References

(1)
(2)

(3)

(4)
(5)
(6)

(7)

(8)
(9)

(10)

(11);

(12)

(13)
(14)
(15)
(16)

(17)

(18)

Milner, S. T. Science 1991, 251, 905-914.

Bhat, R. R.; Tomlinson, M. R.; Wu, T.; Genzer, J. Advances in Polymer Science
2006, 198, 51-124.

Huang, W.; Kim, J.-B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002, 35,
1175-1179.

Ulbricht, M.; Yang, H. Chem. Mater. 2005, 17, 2622-2631.
Brown, H. R.; Char, K.; Delinc, V. R. Macromolecules 1990, 23, 3383-3385.

Andruzzi, L.; Senaratne, W.; Hexemer, A.; Sheets, E. D.; Ilic, B.; Kramer, E. 1.;
Baird, B.; Ober, C. K. Langmuir 2005, 21, 2495-2504.

Ma, H. W.; Hyun, J. H.; Stiller, P.; Chilkoti, A. Advanced Materials 2004, 16,
338-341.

Zhang, Z.; Chen, S. F .; Jiang, S. Y. Biomacromolecules 2006, 7, 3311-3315.

Parsonage, E.; Tirrell, M.; Watanabe, H.; Nuzzo, R. G. Macromolecules 1991, 24,
1987-1995.

Kent, M. S. Macromolecular Rapid Communications 2000, 21, 243-270.

Jordan, R.; Graf, K.; Riegler, H.; Unger, K. K. Chemical Communications 1996,
1025-1026.

Huang, W. X.; Skanth, G.; Baker, G. L.; Bruening, M. L. Langmuir 2001, 17,
1731-1736.

Prucker, 0.; Riihe, J. Macromolecules 1998, 31, 592-601.
Prucker, 0.; Riihe, J. Macromolecules 1998, 31, 602-613.
Jones, D. M.; Brown, A. A.; Huck, W. T. S. Langmuir 2002, 18, 1265-1269.

Yamamoto, S.; Ejaz, M.; Tsujii, Y.; Fukuda, T. Macromolecules 2000, 33, 5608-
5612.

Bao, Z. Y.; Bruening, M. L.; Baker, G. L. Journal of the American Chemical
Society 2006, 128, 9056-9060.

Bruening, M. L.; Dotzauer, D. M.; Jain, P.; Ouyang, L.; Baker, G. L. Langmuir
2008, 24, 7663-7673.

40

(19)

(20)

(21)

(22)

(23)

(24)

(25)
(26)
(27)
(28)
(29)

(30)

(31)
(32)

(33)
(34)
(35)
(36)

(3 7)

Huang, W. X.; Baker, G. L.; Bruening, M. L. Angewandte Chemie-International
Edition 2001, 40, 1510-1512.

Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Chemistry of Materials 2006,
18, 4033-4039.

Chen, X. Y.; Annes, S. P.; Greaves, S. J.; Watts, J. F. Langmuir 2004, 20, 587-
595.

Jain, P.; Dai, J.; Grajales, S.; Saha, S.; Baker, G. L.; Bruening, M. L. Langmuir
2007, 23, 11360-11365.

Advincula, R. C.; Brittain, W. J.; Caster, K. C.; Riihe, J. Polymer Brushes:
Synthesis, Characterization, Applications, 2004.

Edmondson, S.; Osborne, V. L.; Huck, W. T. S. Chemical Society Reviews 2004,
33, 14-22.

Minko, S. Polymer Reviews 2006, 46, 397-420.

Zhao, B.; Brittain, W. J. Macromolecules 2000, 33, 342-348.

Choi, I. S.; Langer, R. Macromolecules 2001, 34, 5361-5363.

Huang, X. Y.; Wirth, M. J. Analytical Chemistry 1997, 69, 4577-4580.

Juang, A.; Scherman, O. A.; Grubbs, R. H.; Lewis, N. S. Langmuir 2001, 17,
1321-1323.

Moon, J. H.; Swager, T. M. Macromolecules 2002, 35, 6086-6089.
Ulbricht, M.; Yang, H. Chemistry of Materials 2005, 17, 2622-2631.

Yang, H.; Ulbricht, M. Macromolecular Materials and Engineering 2008, 293,
419-427.

Dauginet-Da Pra, L.; Demoustier-Champagne, S. Polymer 2005, 46, 1583-1594.
Wang, J. S.; Matyjaszewski, K. Macromolecules 1995, 28, 7901-7910.
Matyjaszewski, K.; Xia, J. H. Chemical Reviews 2001, 101, 2921-2990.

Kato, M.; Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules 1995,
28,1721-1723.

’Mitsukami, Y.; Donovan, M. S.; Lowe, A. B.; McCormick, C. L. Macromolecules

2001, 34, 2248-2256.

41

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)
(46)
(47)

(48)

(49)

(50)

(51)

(52)

(53)

(54)

Boyer, C.; Bulmus, V.; Liu, J. Q.; Davis, T. P.; Stenzel, M. 11.; Bamer-Kowollik,
C. Journal of the American Chemical Society 2007, 129, 7145-7154.

Husemann, M.; Morrison, M.; Benoit, D.; Frommer, K. J .; Mate, C. M.; Hinsberg,
W. D.; Hedrick, J. L.; Hawker, C. J. Journal of the American Chemical Society
2000, 122, 1844-1845.

Chung, I.-D.; Britt, P.; Xie, D.; Harth, E.; Mays, J. Chemical Communications
2005, 1046-1048.

Tsarevsky, N. V.; Pintauer, T.; Matyjaszewski, K. Macromolecules 2004, 37,
9768-9778.

Kim, J.-B.; Huang, W.; Miller, M. D.; Baker, G. L.; Bruening, M. L. Journal of
Polymer Science Part A: Polymer Chemistry 2003, 41, 386—394.

Odian, G. Principles of Polymerization; Fourth ed.; Wiley-Interscience, 2004, pp
316-317.

Singh, N.; Chen, Z.; Tomer, N.; Wickramasinghe, S. R.; Soice, N.; Husson, S. M.
Journal of Membrane Science 2008, 311, 225-234.

Yusof, A. H. M.; Ulbricht, M. Journal of Membrane Science 2008, 311, 294-305.
Dai, J. H.; Baker, G. L.; Bruening, M. L. Analytical Chemistry 2006, 78, 135-140.
Sage, L. Analytical Chemistry 2004, 76, l37A-142A.

Xu, F. J .; Cai, Q. J .; Li, Y. L.; Kang, E. T.; Neoh, K. G. Biomacromolecules 2005,
6,1012-1020.

Cullen, S. F.; Liu, X.; Mandel, I. C.; Himpsel, F. J .; Gopalan, P. Langmuir 2008,
24, 913-920.

Dai, J. H.; Bao, Z. Y.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L.
Langmuir 2006, 22, 4274-4281.

Arica, M. Y.; Yilmaz, M.; Yalcin, E.; Bayramoglu, G. Journal of
Chromatography B-Analytical Technologies in the Biomedical and Life Sciences
2004, 805, 315-323.

Bayramoglu, G.; Ekici, G.; Besirli, N.; Arica, M. Y. Colloids and Surfaces a-
Physicochemical and Engineering Aspects 2007, 310, 68-77.

Su, C. K.; Chiang, B. H. Process Biochemistry 2006, 41, 257-263.

Huang, J. 8.; Han, B. Z.; Yue, W.; Yan, H. Journal of Materials Chemistry 2007,
17, 3812-3818.

A”)

1‘:

(55)

(56)

(57)

(58)

(59)

(60)

(61)

(62)
(63)

(64)

(65)

(66)

(67)

(68)

(69)

(70)

(71)

Xu, C. J.; Xu, K. M.; Gu, H. W.; Zhong, X. F.; Guo, Z. H.; Zheng, R. K.; Zhang,
X. X.; Xu, B. Journal of the American Chemical Society 2004, 126, 3392-3393.

Huang, J. S.; Li, X. T.; Zheng, Y. H.; Zhang, Y.; Zhao, R. Y.; Gao, X. C.; Yan, H.
S. Macromolecular Bioscience 2008, 8, 508-515.

Cai, C. F.; Bakowsky, U.; Rytting, E.; Schaper, A. K.; Kissel, T. European
Journal of Pharmaceutics and Biopharmaceutics 2008, 69, 31-42.

Savina, I. N.; Galaev, 1. Y.; Mattiasson, B. Journal of Molecular Recognition
2006, 19, 313-321.

Tsujii, Y.; Ohno, K.; Yamamoto, S.; Goto, A.; Fukuda, T. Advances in Polymer
Science 2006, 197, 1-45.

Yoshikawa, C.; Goto, A.; Tsujii, Y.; Ishizuka, N.; Nakanishi, K.; Fukuda, T.
Journal of Polymer Science Part a-Polymer Chemistry 2007, 45, 4795-4803.

Kanazawa, H.; Matsushima, Y.; Okano, T. T emperature—responsive
chromatography; Marcel Dekker: New York, 2001; Vol. 41 , pp 311-336.

Kikuchi, A.; Okano, T. Progress in Polymer Science 2002, 27, 1165-1193.

Kobayashi, J.; Kikuchi, A.; Sakai, K.; Okano, T. Analytical Chemistry 2003, 75,
3244-3249.

Huang, X. Y.; Doneski, L. J.; Wirth, M. J. Analytical Chemistry 1998, 70, 4023-
4029.

Phizicky, E.; Bastiaens, P. I. H.; Zhu, H.; Snyder, M.; Fields, S. Nature 2003, 422,
208-215.

Kusnezow, W.; Jacob, A.; Walijew, A.; Diehl, F.; Hoheisel, J. D. Proteomics
2003, 3, 254-264.

Iwata, R.; Satoh, R.; Iwasaki, Y.; Akiyoshi, K. Colloids and Surfaces B-
Biointerfaces 2008, 62, 288-298.

Pirri, G.; Chiari, M.; Damin, F.; Meo, A. Analytical Chemistry 2006, 78, 3118-
3124.

Tugulu, S.; Arnold, A.; Sielaff, 1.; Johnsson, K.; Klok, H. A. Biomacromolecules
2005, 6, 1602-1607.

Wang, A. J.; Xu, J. J.; Chen, H. Y. Journal of Chromatography A 2007, 1147,
120-126.

Dong, R.; Krishnan, S.; Baird, B. A.; Lindau, M.; Ober, C. K. Biomacromolecules
2007, 8, 3082-3092.

43

(72)

(73)

(74)

(75)

(76)

(77)
(78)
(79)

(80)

(81)

(82)

(33)

(84)

(35)

(86)

(87)
(88)

(89)

Chen, S. F.; Liu, L. Y.; Zhou, J .; Jiang, S. Y. Langmuir 2003, 19, 2859-2864.
Peluso, P.; Wilson, D. S.; Do, D.; Tran, H.; Venkatasubbaiah, M.; Quincy, D.;
Heidecker, B.; Poindexter, K.; Tolani, N.; Phelan, M.; Witte, K.; Jung, L. S.;
Wagner, P.; Nock, S. Analytical Biochemistry 2003, 312, 113-124.

Iwasaki, Y.; Omichi, Y.; Iwata, R. Langmuir 2008, 24.

Bastida, A.; Sabuquillo, P.; Armisen, P.; Femandez-Lafuente, R.; Huguet, J.;
Guisan, J. M. Biotechnology and Bioengineering 1998, 58, 486-493.

Saxena, R. K.; Ghosh, P. K.; Gupta, R.; Davidson, W. S.; Bradoo, S.; Gulati, R.
Current Science 1999, 77, 101-115.

Fadnavis, N. W.; Deshpande, A. Current Organic Chemistry 2002, 6, 393-410.
Schilke, K. F.; Kelly, C. Biotechnology and Bioengineering 2008, 101, 9-18.

Kobayashi, S.; Yonezu, S. J.; Kawakita, H.; Saito, K.; Sugita, K.; Tamada, M.;
Sugo, T.; Lee, W. Biotechnology Progress 2003, 19, 396-399.

Goto, M.; Kawakita, H.; Uezu, K.; Tsuneda, S.; Saito, K.; Goto, M.; Tamada, M.;
Sugo, T. Journal of the American Oil Chemists' Society 2006, 83, 209-213.

Goto, M.; Okubo, T.; Kawakita, H.; Uezu, K.; Tsuneda, S.; Saito, K.; Goto, M.;
Tamada, M.; Sugo, T. Biochemical Engineering Journal 2007, 3 7, 159-165.

Dunn, J. D.; Reid, G. E.; Bruening, M. L. Mass Spectrometry Reviews (in press).

Dunn, J. D.; Igrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L.
Analytical Chemistry 2008, 80, 5727-5735.

Schleimer, M.; Pirkle, W. H.; Schurig, V. Journal of Chromatography A 1994,
679, 23-34.

Idota, N.; Kikuchi, A.; Kobayashi, J.; Akiyama, Y.; Okano, T. Langmuir 2006,
22, 425-430.

Miller, M. D.; Baker, G. L.; Bruening, M. L. Journal of Chromatography A 2004,
1044, 323-330.

Zou, H. F .; Ye, M. L. Electrophoresis 2000, 21, 4073-4095.

Groll, J.; Haubensak, W.; Ameringer, T.; Moeller, M. Langmuir 2005, 21, 3076-
3083.

Jain, P.; Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Biomacromolecules
2007, 8, 3102-3107.

(90)

(91)

(92)

(93)

(94)

(95)

(96)

(97)

(98)

(99)

(100)

(101)

(102)

(103)

(104)

(105)

Groll, J .; Amirgoulova, E. V.; Ameringer, T.; Heyes, C. D.; Rocker, C.; Nienhaus,
G. U.; Moller, M. Journal of the American Chemical Society 2004, 126, 4234-
4239.

Draveling, C.; Ren, L.; Haney, P.; Zeisse, D.; Qoronfleh, M. W. Protein
Expression and Purification 2001, 22, 359-366.

Boi, C.; Cattoli, F.; Facchini, R.; Sorci, M.; Sarti, G. C. Journal of Membrane
Science 2006, 273, 12-19.

Zou, H. F .; Luo, Q. 2.; Zhou, D. M. Journal of Biochemical and Biophysical ;
Methods 2001, 49, 199-240.

Che, S.; Liu, Z.; Ohsuna, T.; Sakamoto, K.; Terasaki, 0.; Tatsumi, T. Nature
2004, 429, 281-284.

Brandt, S.; Goffe, R. A.; Kessler, S. B.; Oconnor, J. L.; Zale, S. E. Bio-
Technology 1988, 6, 779-782.

Kawai, T.; Saito, K.; Lee, W. Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences 2003, 790, 131-142.

Endres, H. N.; Johnson, J. A. C.; Ross, C. A.; Welp, J. K.; Etzel, M. R.
Biotechnology and Applied Biochemistry 2003, 3 7, 259-266.

Singh, N.; Husson, S. M.; Zdyrko, B.; Luzinov, 1. Journal of Membrane Science
2005, 262, 81-90.

Kawakita, H.; Masunaga, H.; Nomura, K.; Uezu, K.; Akiba, 1.; Tsuneda, S.
Journal of Porous Materials 2007, 14, 387-391.

Castilho, L. R.; Deckwer, W. D.; Anspach, F. B. Journal of Membrane Science
2000, I 72, 269-277.

Singh, N.; Wang, J .; Ulbricht, M.; Wickramasinghe, S. R.; Husson, S. M. Journal
of Membrane Science 2008, 309, 64-72.

Kubota, N.; Kounosu, M.; Saito, K.; Sugita, K.; Watanabe, K.; Sugo, T.
Biotechnology Progress 1997, 13, 89-95.

Kumar, V.; Bhardwaj, Y. K.; Jamdar, S. N.; Goel, N. K.; Sabharwal, S. Journal of
Applied Polymer Science 2006, 102, 5512-5521.

Iwanade, A.; Umeno, D.; Saito, K.; Sugo, T. Biotechnology Progress 2007, 23,
1425-1430.

Sasagawa, N.; Saito, K.; Sugita, K.; Kunori, S.; Sugo, T. Journal of
Chromatography A 1999, 848, 161-168.

45

This chapter is adapted from our published work in
Biomacromolecules (Jain, P.; Sun, L.; Dai, J.; Baker, G. L.;
Bruening, M. L., Biomacromolecules 2007, 8, 3102-3107).

Some of the research was done with Dr. Lei Sun.

46

Chapter 2
High-capacity purification of His-tagged proteins by alumina

membranes containing functionalized poly(HEMA) brushes

2.1. Introduction

The expansion of recombinant protein expression has generated a great need for
rapid and convenient methods of protein purification."2 Typical purifications involve a
series of steps, the most important of which frequently rely on affinity binding. Affinity
methods are based on specific interactions between immobilized ligands and an affinity
tag (e.g., polyhistidine,3 g1utathione-S-transferrase,4 streptavidin,5 or maltose binding
protein6) that is appended to the protein of interest. Proteins containing an affinity tag are
selectively bound to the chromatographic matrix, while other cellular proteins are washed
away.7

Affinity purification has several assests such as ligand stability, high protein
loading, mild elution conditions, simple regeneration and low cost.8 However, this
technique often presents a bottleneck in the purification process because of slow diffusion
of proteins into the pores of chromatographic gels, which leads to long separation times.
Difficulties in packing large columns and relatively high pressure drops, are also

drawbacks to column-based affinity separations?“

These limitations will be particularly
important for large-scale separations. (The use of commercially available Gravity-flow

and Spin-trap columns allows some reduction in purification times for small-scale

separations”).

47

Protein-absorbing membranesl3 can overcome diffusional limitations in protein
separations because convective flow through membrane pores provides rapid mass

transport to binding sites.”‘'8

Moreover, scale-up of membrane separations simply
involves increasing membrane area, which should avoid the challenges of packing large
columns and provide low pressure drops. Unfortunately, however, the low internal
surface area of membrane absorbers (when compared to porous beads) yields a relatively
low binding capacity. To overcome this problem, membranes are modified with polymer
brushes that have multiple protein-binding sites."”'25

We aim to develop affinity membranes for high-capacity protein binding as well
as rapid and selective purification of His-tagged proteins (Figure 2.1.). Previously, Sun
et al. demonstrated that grth of poly(HEMA) from initiators bound to a porous
alumina membrane and subsequent functionalization of the poly(HEMA) with NTA-Cu2+
complexes yields a remarkable binding capacity of more than 100 mg BSA/cm3 of

23 The high binding capacity is due to the strong interaction of NTA-Cu2+

membrane.
complex with histidine residue on proteins. However, since many proteins have exposed
histidine residues, NTA-Cu2+ complexes are non-selective and thus not suitable for
protein purification. The NTA-Ni2+ complex, on the other hand, is highly selective for

His-tagged proteins due to the relatively weak interaction of NTA-Ni2+ with histidine

81011118, thereby requiring a polyhistidine tag for efficient binding.3‘26

48

.ouoa 28582: a 022:

secs EEmEboa 85328 $7392 s.. 9 528m Bmwaufi co 86:5 .3 2:9...

 

swan uoN=m>th
9: 2 96:5
£995 vcmmfihf

oceanEoE
35.5.“.

. swan A<2w158
uoN=m>tou
+“2.52

49

This chapter shows that the use of alumina membranes modified with
poly(HEMA)-NTA-Ni2+ brushes allows rapid and highly selective purification of His-
tagged proteins. Gel electrophoresis reveals that the purity of His-tagged ubiquitin
(HisU) eluted from these membranes is greater than 99%, even when the initial solution
contains 10% bovine serum or a 20-fold excess of BSA. The binding capacity of the
membrane is as high as 120 mg HisU/cm3 of membrane, which is at least S-fold greater
than that for membranes reported in the literature.23’24’27 Moreover, separations can be
completely performed in 30 min or less and membranes are fully reusable.

2.2. Experimental section
2.2.1. Materials

AnodiscTM porous alumina membranes with 0.2 um-diameter surface pores were
obtained from Fisher Scientific. Dimethylformamide (DMF, anhydrous, 99.8%), 11-
mercaptoundecanol (97%), 2-bromoisobutyryl bromide (98%), CuCl (99.999%), CuBrz
(99%), 2,2’-bipyridyl (bpy, 99%), EDC, NHS, 4-dimethylaminopyridine (DMAP),
ethylenediamine tetraacetic acid (EDTA), imidazole (99%), TWEEN-20 surfactant, BSA,
lysozyme, ubiquitin, N-terrninal histidineb tagged ubiquitin (HisU), and myoglobin were
used as received fiom Sigma Aldrich. Fetal bovine serum was obtained from HyClone.
NiSO4-5H20 (Columbus Chemical), NaHzPO4 (CCI), NazHPO4 (Aldrich), Na, Na—
bis(carboxymethy1)-L-lysine hydrate (Fluka, aminobutyl NTA), succinic anhydride
(Matheson Coleman & Bell), and Coomassie protein assay reagent (Pierce) were also
used without purification. HEMA (Aldrich, 97%, inhibited with 300 ppm hydroquinone
monomethyl ether) was purified by passing it through a column of activated basic

alumina (Aldrich), and trichlorosilane initiator ( l 1 -(2-bromo-2-

50

methyl)propionyloxy)undecyltrichlorosilane) was synthesized according to a literature
procedure.28 Buffers were prepared using analytical grade chemicals and deionized
(Milli-Q, 18.2 M!) cm) water.
2.2.2. Polymerization of HEMA in porous alumina membranes

The procedure for polymerizing HEMA inside alumina membranes was reported
previously.23 Briefly, the alumina membrane was sandwiched inside a membrane cell
(Millipore, Swinnex 25), and a solution of the trichlorosilane initiator, (l 1-(2-bromo-2-
methyl)propionyloxy)undecyltrichlorosilane, was passed through the membrane followed
by subsequent rinsing. Polymerization of HEMA from the immobilized initiator occurred
by circulating a degassed solution containing 15 mL of purified HEMA, 15 mL methanol,
82.5 mg (0.825 mmol) of CuCl, 54 mg (0.24 mmol) of CuBrz, and 320 mg (2.04 mmol)
of bpy through the initiator-modified membrane for 1 hour. (The use of a mixed halide
system sometimes provides better control over polymerization29‘30). After
polymerization, the membrane was cleaned with flowing ethanol (20 mL), deionized
water (Milli-Q, 18.2 M!) cm, 20 mL), and acetone (20 mL).
2.2.3. Polymerization of HEMA on Au substrates

Au-coated silicon wafers were coated with a mercaptoundecanol monolayer that
was subsequently allowed to react with 2-bromoisobutyryl bromide as described
previously.”3 ' Polymerization of HEMA from these substrates occurred as described in
section 2.2.2. except that the substrate was simply immersed in a polymerization solution
that was kept in a glove bag, and water was used instead of methanol to give thicker
films.”

2.2.4. Poly(HEMA) derivatization and protein immobilization

51

The derivatization procedure was also described previously and is shown in
Scheme 2.1.23 However, in the present case the aminobutyl NTA derivatized membrane
was exposed to a 0.1 M NiSO4, rather than 0.1 M CuSO4 solution and rinsed with buffer
(20 mM phosphate buffer, pH 7.2). After loading of membranes with Ni”, a solution
containing pure protein or a mixture of proteins (in 20 mM phosphate buffer, pH 7.2) was
then pumped through the membrane using a peristaltic pump, and the permeate was
collected for analysis at specific time intervals. Subsequently, the membrane was rinsed
with 20 mL pH 7.2 washing buffer (20 mM phosphate buffer containing 0.1% Tween-20
surfactant and 0.15 M NaCl) and 20 mL phosphate buffer, and protein was then eluted
using 5-10 mL of a solution containing 20 mM Sodium phosphate, 0.5 M NaCl, and 0.5
M imidazole at pH 7.4. N12“: was later eluted using a 50 mM EDTA solution (pH 7.2),
and the poly(HEMA)-NTA film was recharged with Ni2+ prior to reuse.

To prepare derivatized films for reflectance Fourier transform infrared (FTIR)
characterization, a poly(HEMA) film on a gold substrate was treated in a similar
procedure by immersing the substrate in appropriate solutions and rinsing with solutions
(20 mL each) fi'om a pipette.

2.2.5. Film characterization methods

F TIR spectra of films on gold-coated wafers were obtained with a Nicolet Magna-
IR 560 spectrometer containing a PIKE grazing angle (80°) accessory, and film
thicknesses were measured using a rotating analyzer ellipsometer (model M-44; J.A.
Woollam) at an incident angle of 75°, assuming a film refractive index of 1.5.
Ellipsometric measurements were performed on at least three spots on a film. A

UV/ozone-cleaned gold-coated wafer was used as a background for reflectance FTIR

52

spectra. Film grth inside alumina membranes was verified using transmission FTIR
spectroscopy (Mattson Galaxy Series 3000) with an air background.
2.2.6. Determination of the amount of coordinated Ni“ in the membrane

A calibration curve was determined by measuring the absorbance of NiSO4
standard solutions in 50 mM EDTA (pH 7.2) using a Varian Spectra AA-200 atomic
absorption spectrophotometer, and a sample solution was obtained by eluting Ni2+ from a
poly(HEMA)-NTA-Ni2+-coated membrane with 5.0 mL of 50 mM EDTA (pH 7.2). The
amount of Ni2+ in the stripping solution was calculated from its absorbance using the
calibration curve.
2.2.7. Protein quantification

To determine the concentration of protein bound to poly(HEMA)brushes in a
membrane, 50 uL of eluent was added to 2.95 mL of a solution of Coomassie reagent,
and the mixture was shaken a few times and allowed to react for 5 min at room
temperature. The UV/vis absorbance spectra of these solutions were then obtained with a
Perkin-Elrner UVN is (model Lambda 40) spectrophotometer. Calibration curves for the
absorbance of BSA, HisU and myoglobin solutions at 595 nm were prepared using a
series of protein solutions (concentration range of 100 pg to 1 mg of protein per mL) that
were mixed with Coomassie reagent. All spectra were measured against a Coomassie
reagent background.

To quantify the amount of protein bound to poly(HEMA) brushes on gold-coated
Si, the method reported previously by Dai and coworkers was employed.19 Briefly, a

calibration curve was obtained by plotting the ellipsometric thickness of spin-coated

53

BSA, myoglobin, lysozyme, or ubiquitin films against the reflectance FTIR absorbance
of their amide 1 band.

2.2.8. Determination of protein purity by sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE)

The protein solutions were analyzed by SDS-PAGE with a 16% cross-linked
separating gel and a 4% cross-linked stacking gel (acrylamide). Protein bands were
visualized using a standard silver staining procedure.10
2.3. Results and discussion
2.3.1. Characterization of poly(HEMA)-derivatized membranes

To form brush-modified membranes, poly(HEMA) was first grown from ATRP
initiators that were immobilized within porous alumina via silanization.23 The brushes

were derivatized as shown in Scheme 2.1., with reactant solutions being flowed through

the membrane using a peristaltic pump.

54

Scheme 2.].

adsorption.

1% 1H3
—-—o—c-(|:10H2—g+nsr
CH3 [0:0

0

\
ofio DMAP CHZ'CHZOH
55 °C,3h

(Succinic DMF
anhydnde) h

1"“
Hz-(IZtBr
09.0 o 0

II II
sz—CHz—o—c—CHz-CHz—c-OH

 

1) NHS, EDC, 30 min

 

CHZCOO'
HZNWCW‘
(aminobutyl NTA)
v pH 10.2
o 0'
-C ,0
KN/‘C',
0H3 O‘
CH2_(.|:_E‘n'Br 0-0
Po (”3
0. 9 11
CH2"CH2“O—C"CH2—CH2‘C—NH
O
Ni2+ 2:..0-
- _ - OH
O‘C_<O\.‘N2 2
KN,” i 0H2
(EH3 ,0-
CHz-(ElgBr ..
Po
0 R 0

sz-CHz—o—c-CHz—CHz-i‘l—NH

Derivatization of poly(HEMA) with NTA-Ni” prior to protein

55

Transmission IR spectra of membranes (Figure 2.2.) after polymerization and each
derivatization step provide evidence for formation of poly(HEMA), subsequent reaction

with succinic anhydride (SA) and finally, successfiJl linking of aminobutyl NTA to these

 

l

Absorbance

1850 1750 1650 1550 1450
Wavenumbers (cm'1)

 

 

 

Figure 2.2. Transmission FTIR spectra of an alumina membrane modified with
poly(HEMA) brushes before (a) and afier the following sequential steps: (b) reaction

with succinic anhydride, (c) activation with EDC/NHS, (d) reaction with aminobutyl

NTA, and (e) exposure to 0.1 M NiSO4.

polymer brushes. These transmission IR spectra are consistent with previous spectra of
poly(HEMA) films grown on Au-coated Si.23 To prepare poly(HEMA)-NTA-Ni2+ films
in porous alumina, we first passed a solution of succinic anhydride in DMF through a
poly(HEMA)-derivatized membrane to create free -COOH groups in the film. We then
immersed the membrane in pH 9.9 buffer for 15 min and rinsed with ethanol and acetone

before taking a transmission IR spectrum. The peak at ~174O cm'l doubled in intensity

56

after reaction with succinic anhydride due to ester formation [Figure 2.2., spectrum b],
suggesting that a large fraction of the —OH groups of poly(HEMA) reacted with succinic
anhydride. The smaller peak at ~1594 cm‘I in spectrum (b) most likely results from
newly formed carboxylic acid groups that are deprotonated.

Passing a 0.1 M mixture of EDC and NHS in water (pH~4.9) through the
membrane converted —COOH groups to succinimidyl esters. Peaks due to the
succinimide ester appeared at 1817 and 1786 cm'I [Figure 2.2., spectrum c].32 The
asymmetric succinimide stretch at ~1753 cm‘l overlaps with the carbonyl band (1740
cm") of the previously formed esters, resulting in a broad peak with a height that is about
double that for the ester carbonyls present afier reaction with succinic anhydride only.

Subsequently, the EDC/NHS-activated poly(HEMA). was allowed to react with
aminobutyl-NTA. Afier this reaction, the membrane was immersed in pH 9.9 buffer for
15 min and rinsed with ethanol and acetone. This reaction resulted in a loss of the
absorbance due to the active ester, and a shitting of the peak at 1753 cm'1 back to 1740
cm'I [Figure 2.2., spectrum d]. The new absorbance at 1680 cm'1 suggests NTA
immobilization, as it likely results from a combination of absorbance due to the
carboxylate groups of NTA and amide bonds formed between succinic anhydride and
NTA. The broad peak around 1600 cm'1 could be due to carboxylate groups from either
NTA or hydrolyzed active esters.

Immersion of membranes modified with NTA-derivatized poly(HEMA) in 0.1 M
NiSO4 and subsequent rinsing with water gave immobilized NTA-Ni2+ that should
selectively bind His-tagged proteins. However, there was no dramatic change in the IR

spectrum [Figure 2.2., spectrum e] of membranes afier Ni2+ complexation, but the peak

57

at 1600 cm’1 seems to move to around 1630 cm'1 and sharpen. This spectral change could
occur because spectrum ((1) was obtained after immersion in pH 9.9 buffer, while the
membrane represented in spectrum (e) was not immersed in this buffer to avoid
precipitation of Ni(OH)2, which may occur at high pH. XPS data for a poly(HEMA)-
NTA-Ni2+ film on a gold substrate confirmed the presence of Ni2+ in these films and
showed a Ni:N ratio of 0.5 : 1, which is consistent with our previous results and
indicative of one Ni2+ per NTA moiety.'9'23 We also determined the amount of Ni2+ in
the membrane by passing 5 mL of 50 mM EDTA through a modified membrane and
subsequently analyzing the EDTA solution for Ni2+ using atomic absorption
spectroscopy. This procedure revealed that 1421:] umol Ni2+ was present in a 2 cm-
diameter, 60 um-thick modified membrane. Assuming that every NTA moiety bound a
Ni2+ ion, and that each hydroxyl group of poly(HEMA) was derivatized with one NTA
moiety, these results suggest that prior to derivatization the poly(HEMA) consisted of 12
nm-thick annuli inside the 230 nm-diameter pores in the alumina. (This calculation also
assumes a poly(HEMA) density of 1 g/cm3 and a membrane porosity of 50%).
2.3.2. HisU binding to poly(HEMA)-NTA-Ni2+ brushes on gold-coated silicon
substrates

We first examined absorption of His-tagged proteins in films formed on Au-
coated Si to allow characterization of binding using ellipsometry and reflectance FTIR
spectroscopy. These films were grown from monolayers of initiator as reported
previously.33 In initial studies, absorption of HisU occurred during a 2 h immersion of
the poly(HEMA)-NTA-Ni2+-modified gold substrate in a stirred solution containing 0.01

mg/mL HisU in pH 7.2 phosphate buffer. The film was then rinsed with 20 mL washing

58

buffer (20 mM phosphate buffer containing 0.1% Tween-20 surfactant and 0.15 M
NaCl), immersed in phosphate buffer (pH 7.2) for 15 min, and finally rinsed with ethanol.
To examine whether protein binds to poly(HEMA)-NTA-Ni2+, the reflectance FTIR
spectrum of the film prior to protein exposure was subtracted from the corresponding
spectrum measured after protein exposure and rinsing. The subtracted spectra of films
immersed in 0.01 mg/mL HisU show strong amide 1 (1680 cm'l) and amide II (1545 cm'

1) bands (Figure 2.3., spectrum a), indicative of extensive protein binding. (The small

 

0.02

G)

0

C

(U

2 a

O

(D

.D

< b

1”.

C

 

 

 

1850 1750 1650 1550 1450

Wavenumbers (cm")
Figure 2.3. Subtracted reflectance FTIR spectra of protein immobilized on
poly(HEMA)-NTA-Ni2+ brushes after exposure of the films to (a) 0.01 mg/mL HisU
or (b) 0.1 mg/mL BSA or (c) 0.1 mg/mL ubiquitin. The spectra were obtained by
subtracting the spectrum of poly(HEMA)-NTA-Ni2+ from that of poly(HEMA)-NTA-
Ni2+-protein (both films were rinsed with buffer and ethanol prior to the

measurement).

59

negative peak at 1740 cm"1 likely result from deprotonation of some —COOH groups.) In
contrast to HisU, the subtracted spectrum of a poly(HEMA)-NTA-Ni2+ coating exposed
to 0.1 mg/mL BSA for 6 h (Figure 2.3., spectrum b) shows no amide absorbances, even
though the concentration of BSA in the exposure solution was 10-fold more than the
concentration used for HisU. To further demonstrate that histidine tags are crucial for
binding to poly(HEMA)-NTA-Ni2+ brushes, we immersed a film in a 0.1 mg/mL
ubiquitin solution in phosphate buffer (pH 7.2) and stirred the solution for 6 h. No
ubiquitin binding was detected (Figure 2.3., spectrum c). Thus, these films exhibit no
detectable non-specific adsorption and should maintain the high selectivities that are
available through adsorption of His-tagged proteins to NTA-Ni2+ complexes.

Previously, Dai et al. correlated the subtracted-spectrum amide absorbance of
BSA with the amount of BSA bound to poly(HEMA)-NTA-Cu2+ films using calibration
curves of ellipsometric thickness versus amide absorbance for films of pure, spin-coated
BSA.l9 However, obtaining a calibration curve by spin-coating of HisU is not practical
due to the high cost of this protein. To overcome this challenge, we showed that
calibration curves of thickness versus amide I absorbance are essentially independent of
the protein employed to obtain the curve (Figure 2.4.). The slopes of calibration curves
for BSA, myoglobin, lysozyme, and ubiquitin absorption were all 0.0017 absorbance
units/nm protein. The uniformity of slopes is reasonable, as the extinction coefficient for
the amide absorbance should be similar for all proteins. Given the value of 0.0017
absorbance units/nm, spectrum (a) in Figure 2.3. suggests that the poly(HEMA)-NTA-

Ni2+ film absorbed 31 nm of HisU. This corresponds to binding of 3.1 pg HisU/cm2 or

60

 

Absorbance

 

1800

Absorbance

1 700

 

 

    

y =0.0017 x
R2 = 0.9923

 

 

50 100 150

Thickness (nm)

 

 

1 600

Wavenumbers (cm'1)

Absorbance

1 500 1400

 

I 0.02

/

 

//
1800 1700 1600 1500 1400 1800 1700

 

I 0.02

y =0.0017 x
R2 = 0.9986

 

 

 

0 50
Thickness (nm)

Wavenumbers (cm-1)

100

 

 

 

 

 

I 0.02

 

 

I 0.02

    

y =0.0017 x
R2 = 0.9938

 

 

O

50
Thickness (nm)

100

  

 

 

 

 

_./ J

 

 

 
 
  

 

 

   

—:“'_'/y
1800 1700 1600 1 500 1400
Wavenumbers (cm-1)
d
I 0.02
0.02

y =0.0017 x
3 R2 = 0.9976
g o 15 30 45 60 75
9 Thickness (nm)
8
.0
<1:

 

 

 

1600

1500 1400

Wavenumbers (cm")

Figure 2.4. Reflectance FTIR spectra of spin-coated (8) BSA, (b) myoglobin, (c)

lysozyme, and (d) ubiquitin films on Au. Films of different thicknesses were prepared

using aqueous solutions containing from 0.1 to 10 mg/mL protein solution. The insets

in the figures show the correlation between film ellipsometric thickness and amide l

absorbance (1680 cm").

61

about 10 monolayers of HisU (assuming a HisU density of l g/cm3 and a monolayer
thickness of 2.9 m”).

Figure 2.5. shows how the amount of HisU bound to poly(HEMA)-NTA-Ni2+
films on Au varies as a function of (a) concentration and (b) time. (Data were obtained
using a combination of reflectance F TIR spectroscopy and calibration curves of thickness
versus amide absorbance as described above.) Figure 2.5.(a) reveals that significant
HisU binding occurs at concentrations as low as 0.001 mg/mL, and that saturation of the
film is approached when HisU concentrations reach 0.04 mg/mL. The data in Figure
2.5.(b) demonstrate ~90% film saturation in 20 min when using a 0.05 mg/mL HisU
solution. The maximum binding capacity of these 51 nm-thick (before derivatization)
films was 40 nm of HisU or about 13 monolayers. Capacity increases with film thickness
(Figure 2.6.) but eventually plateaus when poly(HEMA) thickness reaches 115 nm. The
highest binding capacity obtained was 70 nm of HisU or the equivalent of about 23

monolayers.

62

 

 

 

 

 

 

A 5
N
g 4 ‘ . g 0
g 3 J O
O
3
.92 2 1
I
'O
1 P
S (a)
33 o . .
o 0.02 0.04 0.06
HisU concentration (mg/mL)
A 5
N
E
2 4 ' . . e e e
U) 0
3 3 +
3 O
.92 2‘
I
a 1‘
3 (b)
O
m 0

 

 

O 50 100 150 200 250
Incubation time (min)

Figure 2.5. Amount of HisU bound to poly(HEMA)-NTA-Ni2+ films as a firnction of
(a) concentration and (b) time. To obtain the data in (a), a 51 nm-thick (prior to
derivatization) poly(HEMA) film was used, and the binding capacity was determined
afler 2 h. The same poly(HEMA)-NTA film was used throughout the experiment,
with protein elution and reloading of Ni2+ after measurements at each concentration.
For (b), 0.05 mg/mL HisU and a single 51 nm poly(HEMA) film were used. To test
the reproducibility of the experiments, binding on a 51 nm film with a protein

concentration of 0.01 mg/mL was repeated three times and found to be 3.01d: 0.05

ug/cmz.

63

 

8
an
E; .’ to to
3:6
&
34‘ '
z:
1, to
C
821
CD
0 ,,i-

 

 

 

0 4o 90 120 160 200
Thickness (nm)

Figure 2.6. HisU binding capacity versus thickness of poly(HEMA) films that were
subsequently derivatized with NTA-Ni2+. The concentration of HisU in solution was
0.05 mg/mL. The binding on a 40 mm film was repeated using 3 different films and

different protein samples. The binding capacity obtained was 4.3:t0.1 ug/cmz.

2.3.3. HisU binding to poly(HEMA)-NTA-Ni2+ brushes in membranes

To test the HisU binding capacity of poly(HEMA)-NTA-Ni2+ brushes in
membranes, we pumped a solution of 0.3 mg/mL HisU in pH 7.2 phosphate buffer
through the membrane, collected the permeate over specific time intervals, and analyzed
these samples using a Bradford assay. Figure 2.7. shows the breakthrough curve for
HisU binding to poly(HEMA)-NTA-Ni2+ brushes along with similar curves for BSA and
myoglobin binding to poly(HEMA)—NTA-Cu2+-modified membranes. (Poly(HEMA)-
NTA-Ni2+ brushes do not bind a significant amount of BSA and myoglobin as shown
below.) Integration of the differences between the feed concentrations and the permeate
concentrations shown in the breakthrough curves gives binding capacities of 110 and 83

mg/cm3, respectively, for binding of BSA and myoglobin to poly(HEMA)-NTA-Cu2+-

64

0'4 . A Myoglobin

a HisU ‘
0 BSA ‘

   
 

0.3

 

0.2 ‘

0.1 -

Concentration (mg/mL)

 

 

0 5 10 15

Penneate Volume (mL)
Figure 2.7. Breakthrough curves for absorption of 0.3 mg/mL HisU, 0.3 mg/mL
BSA, and 0.35 mg/mL myoglobin in membranes modified with poly(HEMA)-NTA-
Ni2+ (HisU) or poly(HEMA)-NTA-Cu2+ (BSA and myoglobin). The permeate flow
rate was initially 2.4 mL/min and 0.9 mL/min at the end of the experiment. Solid lines

are added to guide the eyes.

modified membranes, and 120 mg/cm3 for HisU binding to a poly(HEMA)-NTA-Ni2+-
modified membrane. However, the breakthrough curves of HisU and myoglobin are
slightly sharper than that for BSA. (Flow rate measurements at a constant pressure
showed that flow rates are similar for all three proteins for a given permeate voltune).
The very sharp breakthrough curve for HisU may stem from its high affinity for the
NTA-Ni2+ or its low molecular mass (10.7 kDa) relative to myoglobin (17 kDa) and BSA
(67.5 kDa). A low molecular mass could allow rapid diffirsion of smaller proteins into
the polymer brushes, but future studies with several different His-tagged proteins are

needed to determine if this is the case.

65

The linear flow velocity through the membrane is on average about 21 cm/h. This
is about an order magnitude less than the flow rate through typical gel columns,”38 but
the capacity of the membranes per cm3 is also an order of magnitude greater than that of

”'38 ‘Still, linear velocities need to increase to make the membranes a more

gels.
competitive technology, and this can be accomplished by increasing pore size (at a
constant porosity, flow rate is proportional to the square of pore radius for assemblies of
parallel pores).

After measuring the breakthrough curve of HisU, the poly(HEMA)-NTA-Ni2+-
HisU membrane was washed with 20 mL washing buffer followed by 20 mL phosphate
buffer, and HisU was eluted with 5-10 mL elution buffer containing 0.5 M imidazole.
Analysis of the eluent using a Bradford assay and comparison of this analysis with the
capacity determined from the breakthrough curve showed that 99i1% of the bound HisU
was recovered.

To prove that poly(HEMA)-NTA-Ni2+-modified membranes selectively bind
HisU, we also investigated the binding of BSA and myoglobin to poly(HEMA)-NTA-
Ni2+ films in alumina. A poly(HEMA)-NTA—Ni2+-derivatized membrane was loaded
with 10 mL of 0.3 mg/mL BSA or 10 mL of 0.3 mg/mL myoglobin in buffer, rinsed with
washing buffer, and treated with eluent. BSA was not detected in the imidazole eluent
(<0.001 mg/cm3), while a small amount (0.05 mg/cm3) of myoglobin binding was
observed. However, the amount of myoglobin bound was only 3% of the amount of

HisU bound when using 0.3 mg/mL HisU. (Eleven of the 153 residues of myoglobin are

histidine, which could enhance the undesired adsorption of this protein to NTA-Ni“).

66

2.3.4. Separation of protein mixtures using poly(HEMA)-NTA-Ni2+ brushes on Au
substrates

We first tested the ability of Au-poly(HEMA)—NTA-Ni2+ films to purify His
tagged proteins due to ease in handling and characterization. In an initial experiment, a
gold substrate modified with poly(HEMA)-NTA-Ni2+ was immersed in 10 mL of
phosphate buffer containing BSA, myoglobin and HisU (0.05 mg/mL each), and the
solution was stirred for 2 h. The film was then rinsed with 20 mL washing buffer
followed by 20 mL phosphate buffer, followed by ethanol and dried under a stream of
nitrogen. Finally, we immersed the film in 1 mL elution buffer (20 mM phosphate
buffer, 0.5 M NaCl, and 0.5 M imidazole, pH 7.2) for 30 min to recover the bound
protein and then analyzed this eluent by electrophoresis. The gel electropherogram of the
eluent (Figure 2.8.(a), lane 6) shows no bands due to BSA or myoglobin and an intense
band for HisU, demonstrating that highly pure HisU can be obtained using this
procedure. (Lanes 1-5 are given to show positions of proteins)

In a second test of selectivity, another film was immersed in 10 mL of phosphate
buffer containing 1 mg/mL BSA and 0.05 mg/mL HisU for 2 h, and the solution was
stirred for 2 h. The film was rinsed, treated with elution buffer, and analyzed as
described above. Even though this sample contained a 20-fold excess of BSA, the eluent

showed only a band for HisU (Figure 2.8.(a), lane 8).

67

BSA

Myoglobin

HisU

 

12345678 12345678
(a) (b)

Figure 2.8. SDS-PAGE analysis (silver staining) of protein solutions and eluents
from (a) Au-poly(HEMA)-NTA—Ni2+ films and (b) poly(HEMA)-NTA-Ni“ modified
alumina membranes. Samples for both electropherograms are: lane 1, standard broad
range ladder; lane 2, mixture of BSA, myoglobin and HisU (0.05 mg/mL each); lane
3, BSA; lane 4, myoglobin; lanes, HisU; lane 6, eluent from a film (a) or membrane
(b) loaded with a mixture of BSA, myoglobin and HisU (0.05 mg/mL each); lane 7,
mixture of BSA and HisU; lane 8, eluent from a film (a) or membrane (b) loaded with

10 mL of a solution containing 1 mg/mL BSA and 0.05 mg/mL HisU.

2.3.5. Separation of protein mixtures using poly(HEMA)-NTA-Ni2+ brushes in
membranes

After obtaining promising results with Au-poly(HEMA)-NTA-Ni2+ films, we
utilized poly(HEMA)-NTA-Ni2+-modified alumina membranes to purify HisU. Three
sets of experiments were performed. In the first experiment, 6 mL of phosphate buffer
containing BSA, myoglobin and HisU (0.05 mg/mL of each protein) was passed through

the poly(HEMA)-NTA-Ni2I-derivatized membrane, and the membrane was then rinsed

68

with 20 mL washing buffer followed by 20 mL phosphate buffer. Finally, we used 5-10
mL elution buffer to recover the bound protein and analyzed this eluent by
electrophoresis. As with purification on a film, the gel electropherogram of the eluent
from the membrane (Figure 2.8. (b), Lane 6) shows no bands due to BSA or myoglobin
and an intense band for HisU. This demonstrates that one pass through the membrane is
sufficient to give >99% pure HisU. The purity was calculated assuming that the amount
of BSA in the 15 uL protein solution that was loaded on the gel is <20 ng.
Approximately 0.75 ug of HisU was loaded on the gel based on a Bradford assay of the
protein sample, which was possible because of the high purity of the eluent. Moreover,
the membranes can be used multiple times. Gel electrophoresis confirmed that the purity
of eluted HisU was not affected by reusing the membrane four times.

Afier this initial experiment, we washed the membrane with 10 mL of 50 mM
EDTA, rinsed it with water, recharged it with 0.1 M Ni”, and rinsed with 20 mL water
and 20 mL phosphate buffer. To further test the membranes, 10 mL of phosphate buffer
containing 1 mg/mL BSA and 0.05 mg/mL HisU was passed through the membrane, and
the membrane was rinsed, treated with elution buffer, and analyzed as described above.
Here also, the eluent showed only a band for HisU (Figure 2.8.(b), Lane 8) even in the
presence of 20-fold excess of BSA.

In a final experiment, 10% Bovine serum in 10 mL phosphate buffer containing
45 mM imidazole (pH 7.2) was spiked with 0.3 mg/mL HisU. (lmidazole was added to
help prevent adsorption of non His-tagged proteins.) This solution was passed through
the membrane, which was then rinsed, treated with elution buffer, and analyzed using the

procedure described above.

60

Figure 2.9. shows the SDS-PAGE analysis of the spiked bovine serum (lefi) and
the eluent from the membrane (right). Although the diluted serum contains about 4
mg/mL of protein and only 0.3 mg/mL of HisU, the electropherogram of the eluent shows
a strong band for HisU and only a very faint band of BSA. The HisU in this case is more
than 99% pure, assuming that the faint band of BSA represents <40 ng of BSA.
(Approximately 4.5 pg of HisU was loaded on the gel, as determined using a Bradford
assay of the eluent, and separate gels showed that 40 ng of BSA gave a readily visible
band.) Figure 2.9. clearly shows the selectivity of the membrane for HisU over other
proteins present in serum. Moreover, similar results were obtained upon reusing the

membrane.

 

1 2

Figure 2.9. SDS-PAGE analysis (silver staining) of 10% bovine serum spiked with
0.3 mg/mL HisU (lane 1) and the imidazole eluent fi'om an alumina-poly(HEMA)-

NTA-Ni2+ membrane loaded with this solution (lane 2).

70

2.4. Conclusions

This work demonstrates that grth of poly(HEMA)-NTA-Ni2+ brushes inside
porous alumina supports yields high-capacity membranes that selectively bind His-tagged
proteins. The membranes show a binding capacity of 120 mg HisU/cm3 of membrane
along with minimal non-specific adsorption. Moreover, 99% of the bound HisU could be
recovered, and membranes can be reused.

Studies of poly(HEMA)-NTA-Ni2+ brushes on flat, gold-coated silicon substrates
show that significant HisU binding occurs at concentrations as low as 0.001 mg/mL, and
saturation of the film is approached in 20 min or less at HisU concentrations of 0.05
mg/mL. Thus, even at His-tagged protein concentrations of 0.04 mg/mL, it should be
possible to make full use of binding sites in a few minutes. The time required to
approach saturation should be even less at higher concentrations and in membranes,
where convection through the membrane will help to overcome diffusion limitations.
However, comparison of breakthrough curves of BSA and myoglobin suggest that
binding times may increase for large proteins.

Gel electrophoresis results indicate that the membrane is selective towards HisU
even in the presence of a 20-fold excess of BSA or in 10% Bovine serum. The recovered
HisU in such cases is 99% pure. These membranes should also be effective for whole-
cell extracts provided the solution does not clog the membrane. Increases in membrane
pore sizes should help to avoid clogging and increase linear velocities, but large pores
may also require some optimization of brush thickness and density to achieve rapid

binding and high capacities.

71

Importantly, the time required for membrane-based purification included only 10
min for loading, 10 min for washing, and 5 min for elution. Thus purification of His-
tagged proteins can be achieved in less than 30 min, and further time reductions could
likely be achieved by using more permeable membrane supports and higher pressures.
Hence, these membranes are attractive for rapid, selective purification of His-tagged

proteins.

72

2.5. References:

(1)

(2)

(3)

(4)

(5)
(6)

(7)

(8)

(9)

(10)

(11)

(12)
(13)

(14)

(15)
(16)

(17)

Arnau, J.; Lauritzen, C.; Petersen, G. E.; Pedersen, J. Protein Expression and
Purification 2006, 48, 1-13.

Zhen, G. L.; Falconnet, D.; Kuennemann, B.; Voros, J.; Spencer, N. D.; Textor,
M.; Zurcher, S. Advanced Functional Materials 2006, 16, 243-251.

Stiborova, H.; Kostal, J.; Mulchandani, A.; Chen, W. Biotechnology and
Bioengineering 2003, 82, 605-61 1.

Draveling, C.; Ren, L.; Haney, P.; Zeisse, D.; Qoronfleh, M. W. Protein
Expression and Purification 2001, 22, 359-366.

Skerra, A.; Schmidt, T. G. M. Biomolecular Engineering 1999, 16, 79-86.

Cattoli, F.; Boi, C.; Sorci, M.; Sarti, G. C. Journal of Membrane Science 2006,
2 73, 2-11.

Suen, S. Y.; Liu, Y.-C.; Chang, C. S. Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences 2003, 797, 305-319.

szara, S.; Yavuz, H.; Denizli, A. Journal of Applied Polymer Science 2003, 89,
1567-1572.

Che, S.; Liu, Z.; Ohsuna, T.; Sakamoto, K.; Terasaki, 0.; Tatsumi, T. Nature
2004, 429, 281-284. .

Clairbois, A.—S.; Letoumeur, D.; Muller, D.; Jozefonvicz, J. Journal of
Chromatography B 1998, 706, 55-62.

Zou, H. F.; Luo, Q. Z.; Zhou, D. M. Journal of Biochemical and Biophysical
Methods 2001, 49, 199-240.

Luzinov, 1.; Minko, S.; Tsukruk, V. V. Soft Matter 2008, 4, 714-725.

Brandt, S.; Goffe, R. A.; Kessler, S. B.; O’Connor, J. L.; Zale, S. E. Bio-
Technology 1988, 6, 779-782.

Boi, C.; Cattoli, F .; Facchini, R.; Sorci, M.; Sarti, G. C. Journal of Membrane
Science 2006, 2 73, 12-19.

Boi, C.; Facchini, R.; Sorci, M.; Sarti, G. C. Desalination 2006, 199, 544-546.

Endres, H. N.; Johnson, J. A. C.; Ross, C. A.; Welp, J. K.; Etzel, M. R.
Biotechnology and Applied Biochemistry 2003, 3 7, 259-266.

Sorci, M.; Boi, C.; Facchini, R.; Sarti, G. C. Desalination 2006, 199, 550-552.

73

(18)

(19)

(20)

(21)

(22)

(23)

(24)
(25)
(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

Yang, H. W.; Viera, C.; Fischer, 1.; Etzel, M. R. Industrial & Engineering
Chemistry Research 2002, 41, 1597-1602.

Dai, J. H.; Bao, Z. Y.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L.
Langmuir 2006, 22, 4274-4281.

Hicke, H.-G.; Becker, M.; Paulke, B.-R.; Ulbricht, M. Journal of Membrane
Science 2006, 282, 413-422.

Kawai, T.; Saito, K.; Lee, W. Journal of Chromatography B-Analytical
Technologies in the Biomedical and Life Sciences 2003, 790, 131-142.

Singh, N.; Husson, S. M.; Zdyrko, B.; Luzinov, 1. Journal of Membrane Science
2005, 262, 81-90.

Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Chemistry of Materials 2006,
18, 4033-4039.

Ulbricht, M.; Yang, H. Chemistry of Materials 2005, 17, 2622-2631.
Yusof, A. H. M.; Ulbricht, M. Desalination 2006, 200, 462-463.

Johnson, R. D.; Arnold, F. H. Biotechnology and Bioengineering 1995, 48, 437-
443.

Kawakita, H.; Masunaga, H.; Nomura, K.; Uezu, K.; Akiba, 1.; Tsuneda, S.
Journal of Porous Materials 2007, 14, 387-391.

Matyjaszewski, K.; Miller, P. J.; Shukla, N.; lmmaraporn, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T. Macromolecules 1999, 32, 8716-8724.

Matyjaszewski, K.; Shipp, D. A.; Wang, J.—L.; Grimaud, T.; Patten, T. E.
Macromolecules 1998, 31, 6836-6840.

Huang, W. X.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,
35,1175-1179.

Kim, J.-B.; Huang, W. X.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,
35, 5410-5416.

Xiao, S.-J.; Brunner, S.; Wieland, M. Journal of Physical Chemistry B 2004, 108,
16508-16517.

Kim, J.-B.; Bruening, M. L.; Baker, G. L. Journal of the American Chemical
Society 2000, 122, 7616-7617.

Cook, W. J.; Suddath, F. L.; Bugg, C. B.; Goldstein, G. Journal of Molecular
Biology 1979, 130, 353-355.

74

(35)
(36)
(37)

(3 8)

Yusof, A. H. M.; Ulbricht, M. Journal of Membrane Science 2008, 311, 294-305.
Motornov, M.; Sheparovych, R.; Katz, E.; Minko, S. Acs Nano 2008, 2, 41-52.

Goto, M.; Okubo, T.; Kawakita, H.; Uezu, K.; Tsuneda, S.; Saito, K.; Goto, M.;
Tamada, M.; Sugo, T. Biochemical Engineering Journal 2007, 37, 159-165.

Bayramoglu, G.; Ekici, G.; Besirli, N.; Arica, M. Y. Colloids and Surfaces a-
Physicochemical and Engineering Aspects 2007, 310, 68-77.

75

Chapter 3
Purification of His-tagged proteins by polymer membranes

containing functionalized poly(HEMA) brushes

3.1. Introduction

Chapter 2 shows that the use of alumina membranes modified with poly(HEMA)-
NTA-Ni2+ brushes allows rapid and highly selective purification of His-tagged proteins.
Unfortunately, the largest pore diameter available in commercial alumina membranes is
0.25 um,"2 which limits the use of these membranes to relatively simple solutions and
low flow rates because complex solutions such as cell extracts often plug the relatively
small pores. Moreover porous alumina is expensive and frequently breaks both because
of its fragility and the pressure needed to achieve required flow rates. Polymeric
membranes, on the other hand, can have larger pore diameters than porous alumina (1-10
p.m),3’4 which should facilitate rapid purification with more complex solutions. Higher
permeabilities through membranes with larger pores will also allow for lower pressures
during separations and the use of thicker membranes that can bind more protein.

This chapter describes the growth of poly(HEMA) brushes inside porous polymer
(nylon and PVDF) membranes, and the use of nylon membranes modified with
poly(HEMA)-NTA-Ni2+ brushes to purify His-tagged cellular retinaldehyde binding
protein (CRALBP) from a cell extract. The purities of CRALBP eluted from these

membranes are at least as good as those obtained with commercial Ni2+ columns.

3.2. Experimental section
3.2.]. Materials

Hydroxylated (LoProdyne® LP) nylon membranes with 1.2 and 5.0 [rm-diameter
surface pores were obtained from Pall Corporation, and hydrophilic PVDF membranes
with a nominal 0.45 pm pore size were purchased from Millipore Corporation. 2-
Bromoisobutyryl bromide (98%), CuCl (99.999%), CuBrz (99%), bpy (99%), EDC,
NHS, DMAP, EDTA, imidazole (99%), TWEEN-ZO surfactant, and BSA were used as
received from Sigma Aldrich. NiSO4-5H20 (Columbus Chemical), NaH2P04 (CCI),
NazHPO4 (Aldrich), aminobutyl NTA (Fluka), succinic anhydride (Matheson Coleman &
Bell), and Coomassie protein assay reagent (Pierce) were also used without purification.
Tetrahydrofuran (THF, Jade Scientific Inc., anhydrous, 99%) was distilled and stored
over molecular sieves. HEMA (Aldrich, 97%, inhibited with 300 ppm hydroquinone
monomethyl ether) was purified by passing it through a column of activated basic
alumina (Aldrich), and trichlorosilane initiator (11-(2-bromo-2-
methyl)propionyloxy)undecyltrichlorosilane) was synthesized according to a literature
procedure.5 Buffers were prepared using analytical grade chemicals and deionized
(Milli-Q, 18.2 M!) cm) water.
3.2.2. Polymerization of HEMA in porous polymer membranes

A polymer membrane was cleaned with UV/ozone (Boekel model 135500) for 10
min, and placed inside a home-built Teflon cell. Subsequently a 1 mM solution of the
trichlorosilane initiator in 20 mL of anhydrous THF was circulated through the
membrane for 2 h at a flow rate of 3 mL/min, followed by rinsing with 20 mL of ethanol.

Polymerization of HEMA from the immobilized initiator occurred by circulating a

77

degassed solution containing 15 mL of purified HEMA, 15 mL water, 82.5 mg (0.825
mmol) of CuCl, 54 mg (0.24 mmol) of CuBrz, and 320 mg (2.04 mmol) of bpy through
the initiator-modified membrane for 1 hour inside a glove bag (flow rate of 1 mL/min).
(The use of a mixed halide system sometimes provides better control over
polymerization“). After polymerization, the membrane was cleaned with flowing
ethanol (20 mL) followed by 20 mL of deionized water (Milli-Q, 18.2 MQ cm, 20 mL).
(Acetone should not be used with nylon membrane as it partially damages the membrane
structure).

3.2.3. Poly(HEMA) derivatization and protein immobilization

Chapter 2 described the derivatization procedure, however, in the present case
succinic anhydride was dissolved in THF instead of DMF because DMF partially
damages nylon membranes. Additionally, triethylarnine was used instead of DMAP as
the latter does not dissolve in THF. Moreover, the reaction took place at room
temperature unlike with alumina membranes where heating to 55 °C was required
(Scheme 3.1).

To study BSA binding, a solution of BSA in 20 mM phosphate buffer (pH 7.2)
was pumped through the poly(HEMA)-NTA-Cu2+ modified membrane using a peristaltic
pump, and the permeate was collected for analysis at specific time intervals.
Subsequently, the membrane was rinsed with 20 mL of pH 7.2 washing buffer (20 mM
phosphate buffer containing 0.1% Tween-20 surfactant and 0.15 M NaCl) followed by 20
mL of phosphate buffer. The protein was then eluted using 5-10 mL of a solution

containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M imidazole at pH 7.4. Cu2+

78

\0 0%
0-8Ii.(CH2)fi 0 Br
I

0\
HEMA
CuCI, CuBrz, bpy
RT, 1 h
9 9H3 H.
—-—o—c-c|:1c:H2—crn8r
CH3 /C=o
Oi
CHz'CHon
O
DUO Triethyl amine
RT, 3 h
(Succinic THF
anhydnde)
i“
Hz-qt‘sr
OlCto O O

\ II II
CH2‘CH2‘0_C-CH2'CH2—CTOH

1) NHS, EDC, 30 min
CHZCOO‘
HZN 000‘

(aminobutyl NTA)
pH 10.2

r 3) Cu2+ O

 

H3 N (l)-
/
CHZ-ClaBr ll
Cr: 0
o’ O 0 o

\ H l
CHz-CHz-O-C-CHz-CHz-C-NH
Scheme 3.1. Schematic illustration of the growth and derivatization of poly(HEMA)

brushes inside polymer (nylon or PVDF) membranes.

79

was later eluted using a 50 mM EDTA solution (pH 7.2), and the poly(HEMA)-NTA film
was recharged with Cu2+ prior to reuse.
3.2.4. Characterization methods

Film grth inside polymer membranes was verified using attenuated total
reflectance (ATR) FTIR spectroscopy (Perkin Elmer Spectrum One Instrument, air
background) as well as field-emission scanning electron microscopy (FESEM, Hitachi 8-
470011 equipped with an EDAX Phoenix energy dispersive X-ray spectrometer system,
acceleration voltage of 15V).
3.2.5. Protein quantification

To determine the amount of protein eluted from poly(HEMA) brushes in a
membrane, 50 pL of eluent was added to 2.95 mL of a solution of Coomassie reagent,
and the mixture was shaken a few times and allowed to react for 5 min at room
temperature. The UV/vis absorbance spectra of these solutions were then obtained with a
Perkin-Elmer UVN is (model Lambda 40) spectrophotometer. A calibration curve for the
absorbance of BSA solutions at 595 nm was prepared using a series of protein solutions
(concentration range of 100 pg to 1 mg of protein per mL) that were mixed with
Coomassie reagent. All spectra were measured against a Coomassie reagent background.

To quantify the amount of protein bound to poly(HEMA) brushes on gold-coated
Si, the method reported by Dai and coworkers was employed.8 Briefly, a calibration
curve was obtained by plotting the ellipsometric thickness of spin-coated BSA,
myoglobin, lysozyme, or ubiquitin films against the reflectance FTIR absorbance of their

amide 1 band (for more details refer to page 60).

80

3.2.6. Determination of protein purity by sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE)

The protein solutions were analyzed by SDS-PAGE with a 16% cross-linked
separating gel and a 4% cross-linked stacking gel (acrylamide). Protein bands were
visualized using standard silver staining9 or coomassie blue staining10 procedures.

3.3. Results and discussion
3.3.1. Characterization of poly(HEMA)-derivatized membranes

To form brush-modified membranes, poly(HEMA) was grown from ATRP
initiators that were immobilized via silanization within the porous polymer membrane.”
The brushes were derivatized as shown in Scheme 3.1., with reactant solutions being
circulated through the membrane using a peristaltic pump. Figure 3.1. shows the ATR-
FTIR spectra of (a) a pristine membrane and a similar membrane after (b) polymerization
of HEMA, and subsequent reaction with (c) succinic anhydride, ((1) activation with
NHS/EDC and finally, (6) derivatization with aminobutyl NTA. A bare nylon membrane
shows amide I and amide II peaks at 1630 and 1530 cm’1 respectively (Figure 3.1.,
spectrum (a)). The small peak at 1720 cm'l might be due to carbonyl groups introduced
during hydroxylation of the membrane (however we are not sure as the procedure for
hydroxylation is not disclosed by the vendors). The growth of poly(HEMA) is evident by
the increase in the carbonyl peak intensity at 1720 cm'1 (Figure 3.1., spectrum (b)). The
difference in intensity of amide peaks before and after derivatization is due to differences
in the pressure applied to hold the membrane against the ATR crystal. We applied a

lower pressure to the polymer-modified membrane to prevent any damage, and hence the

81

amide peaks of the modified membrane are less intense than those for the bare

membrane.

 

 
     

| 0.1
e
U
C
W
e C
8
.n b
<
Poly(HEMA)
C=O
a

  

 

 

 

1900 1700 1500 1300 1100
Wavenumbers (cm‘1)

Figure 3.1. ATR-FTIR spectra of a hydroxyl functionalized nylon membrane before
(a) and after the following sequential steps: (b) formation of poly(HEMA) brushes
inside the membrane; (c) reaction with succinic anhydride; (d) activation with

EDC/NHS; (e) reaction with aminobutyl NTA.

82

The peak at ~1720 cm'l further increased in intensity afier reaction with succinic
anhydride due to ester formation (Figure 3.1., spectrum (c), the membrane swells after
reaction with succinic anhydride, thus it is difficult to quantify the increase in ester peak
intensity.) Passing a mixture of EDC and NHS in water through the membrane converted
—COOH groups to succinimidyl esters. Peaks due to the succinimide ester appeared at
1810 and 1776 cm'1 (Figure 3.1., spectrum (d)). Subsequently, the EDC/NHS-activated
poly(HEMA) was allowed to react with aminobutyl-NTA. This reaction resulted in a loss
of the absorbance due to the succinimide ester, (Figure 3.1., spectrum (e)). NTA
immobilization is difficult to characterize because of the nylon amide peaks, but there is
an absorbance due to the carboxylate groups of NTA (~1680 cm") that appears under the
amide peaks.

SEM images corroborate the growth of poly(HEMA)-NTA-Cu2+ brushes in the
polymer membranes. The image of a pristine membrane (Figure 3.2.(a)) contains open
pores, whereas modified pores (Figure 3.2(b)) appear less open, presumably because
they are covered with a polymer film. The energy dispersive X-ray (EDAX) spectrum of
a membrane cross-section shows the presence of copper ions throughout the sample (the
thickness of the nylon membrane is ~110 pm, Figure 3.2(c)), confirming that the growth

and derivatization of poly(HEMA) was not limited to the membrane surface.

83

 

 

 

(C)

Figure 3.2. SEM images of (a) a bare nylon membrane with a 1.2 pm nominal

filtration cutoff, (b) a similar membrane modified with poly(HEMA)-NTA-Cu2+ and

(c) a cross-sectional image of the membrane.

3.3.2. BSA binding to poly(HEMA)-NTA-Cu2+ brushes in membranes

To test the protein binding capacity of poly(HEMA)-NTA-Cu” brushes in
polymer membranes, we pumped a 1 mg/mL solution of BSA in pH 7.2 phosphate buffer
through the membrane, collected the permeate over specific time intervals, and analyzed

these samples using a Bradford assay. Figure 3.3. shows the breakthrough curve for

84

BSA binding to a poly(HEMA)-NTA-Cu2+-modified nylon membrane (1.2 pm).
Integration of the differences between the feed concentrations and the permeate
concentrations in the breakthrough curves gives a binding capacity of 25 mg/cm3. This
binding capacity is only 20% of that in alumina membranes modified with poly(HEMA)-
NTA-Cu2+ (see Chapter 2) and thus is difficult to determine. Similar experiments with

PVDF-poly(HEMA)-NTA-Cu2+ membranes yield a BSA binding capacity of only 15

 

 

mg/cm3.
1.0 -
0.9 -
E 0.8 -
B: 0.7 -
E 0.6 -
C
.3 0.5 4
g 0.4 4
C
8 0.3 -
8 0.2 -
0 0.1 -
0 b—l l l 7 I l l T F l

 

012345678910
Permeate Volume (mL)
Figure 3.3. Breakthrough curve for absorption of 1 mg/mL BSA in a nylon

membrane (1.2 pm pore diameter) modified with poly(HEMA)-NTA-Cu2+. The

permeate flow rate was 0.77 mL/min. The solid line is a guide to the eyes.

After measuring the breakthrough curve of BSA, the poly(HEMA)-NTA-Cu2+-
BSA membrane was washed with 20 mL washing buffer followed by 20 mL phosphate
buffer, and the bound BSA was eluted with 5-10 mL EDTA. Analysis of the eluent using
a Bradford assay and comparison of this analysis with the capacity determined from the

breakthrough curve showed that >94% of the bound BSA was recovered. (Most likely,

85

essentially all of the protein eluted from the membrane, but the uncertainty in the amount
of binding determined from the breakthrough curve only allows us to say that >94% of
the protein was eluted.)

The low BSA binding capacity in polymer membranes relative to alumina
membranes may be due to the larger pore size in the polymer membranes and the
inability of protein to penetrate thick brushes, or less polymer grafting in the polymer
substrates. To examine whether the amount of polymer grafted in the membrane affects
protein binding capacity, we increased the polymerization time. However, increasing the
polymerization time from 1 h to 2 h did not lead to an increase in BSA binding capacity.
Presuming that the longer polymerization time results in more polymer growth, this result
suggests that the interior of long brushes is not accessible for protein binding. Longer
brushes may also block some pores to decrease binding capacity. We also tried
increasing the nominal pore size of the membrane from 1.2 to 5 pm. However, the
protein binding capacities for the two types of membranes were similar. Nevertheless,
the 5 pm membranes are advantageous over 1.2 pm membranes because of an increase in
permeability. After modification with poly(HEMA)-NTA-Cu2+, flow rate at 6.9 X 104
Pascal (10 psig) was 1.5 mL/min and 3.0 mL/min for 1.2 pm and 5 pm membranes,
respectively .

3.3.3. Purification of His-Tagged CRALBP using nylon membranes modified with
poly(HEMA)-NTA-Ni“ brushes

As discussed in chapter 2, alumina membranes modified with poly(HEMA)-NTA-
Ni2+ brushes are promising for purification of His-tagged proteins. However it is difficult

to purify complex mixtures with alumina membranes due to the small pore diameter (0.25

86

pm). Nylon membranes, on the other hand, have larger pore sizes (1.2-5.0 pm) than
alumina so they can potentially purify His-tagged proteins directly from a cell extract.
We tested the performance of poly(HEMA)-NTA-Ni2+-modified nylon membranes
towards purification of His-tagged CRALBP (36 kD) that was over-expressed in E. colil2
(Dr. James Geiger and Dr. Xiaofei J ia kindly provide the cell extracts.) The cells were
lysed by sonication, centrifuged at 4 °C, and 1.25 mL of the supernatant was added to
3.75 mL of 20 mM phosphate buffer (containing 10 mM imidazole and 300 mM NaCl,
pH 7.2). This solution was pumped through the poly(HEMA)-NTA—Niz+ modified nylon
membrane in an amicon cell at 6.9 X 104 Pascal (10 psig) at room temperature. (The
flow rate was 0.12 mL of extract/min.) Subsequently, the membrane was rinsed with 20
mL pH 7.2 washing buffer I (20 mM phosphate buffer containing 0.1% Tween-20
surfactant and 0.15 M NaCl) followed by 20 mL pH 7.2, washing buffer 11 (20 mM
phosphate buffer containing 45 mM imidazole and 0.15 M NaCl), and protein was then
eluted using a solution containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M
imidazole at pH 7.2. The flow rate increased to >1 mL/min during elution at 6.9x104
Pascal.

Figure 3.4.a shows the electropherograms of the cell extract (lane 1) and the
eluate from the membrane (lane 2). The eluate contains remarkably pure protein (~99%
pure based on the faintness of all other bands). Unfortunately, we cannot establish the
protein binding capacity or elution efficiency in this case because the concentration of
CRALBP in the cell extract is unknown. We reused the membrane twice without a
change in protein purity or flow rate. Future studies should further establish how many

times these membranes can be reused without affecting protein purity.

87

.. ..4—CRALBP

 

(a) (b)

Figure 3.4. SDS-PAGE analysis (silver staining) of an extract from E. coli containing
over-expressed His-tagged CRALBP (Lanes 1 in gels (a) and (b)); and CRALBP
purified from these extracts using flow-through (gel (a), lane 2) and immersion
methods (gel (b), lane 2). In the flow-through method, the cell extract, washing and
elution solutions were flowed through the nylon-poly(HEMA)-NTA-Ni2+ membrane

whereas in the immersion method, the membrane was immersed in the solutions.

To determine whether flow through the membrane plays an important role in
obtaining high protein purity, we attempted to isolate His-tagged CRALBP by simply
immersing a membrane in the cell extract, washing, and elution solutions. Figure 3.4.b.
shows that the protein isolated by the immersion method is much less pure than the
protein purified by flowing the cell extract, washing and elution solutions through the
membrane. This result suggests that that the flowing solution removes non-specifically
adsorbed proteins from the membranes, which may be a major advantage of membranes
over columns because the nanopores in resins will not be rinsed with flow.

We also studied the effect of the imidazole concentration in the loading buffer on

the purity of CRALBP eluted from membranes. Imidazole may inhibit the unwanted

88.

binding of histidine-containing proteins to the Ni2+ complexes. In one experiment, 10
mM imidazole was added to the cell extract before loading on the membrane, and in
another experiment the concentration of imidazole was halved (i.e. 5 mM imidazole was
added). The electropherograms in Figure 3.5. show that the higher concentration of
imidazole in the cell extract results in a significantly higher CRALBP purity in the eluate.
However, too much imidazole in the binding buffer is undesirable because it will
compete with His-tagged protein for the binding sites and decrease the protein binding

capacity.

   

nt.“ <

{(339

if}.
1 2 3

Figure 3.5. SDS-PAGE analysis (silver staining) of an extract from E. coli containing
over-expressed His-tagged CRALBP (Lane 1) and CRALBP purified from cell
extracts containing 10 mM (Lane 2) and 5 mM imidazole (Lane 3). CRALBP was
purified with a poly(HEMA)-NTA-Ni2+-modified nylon membrane (1.2 pm nominal
pore size) using procedures described in the text. Solutions were flowed through the

membrane.

89

The goal of this research is to develop affinity membranes for purification of His-
tagged proteins in order to overcome the limitations in column-based protein separations.
To establish how well our membranes perform, we first compared the performance of
poly(HEMA)-NTA-Ni2+-modified nylon membranes with commercial spin-trap columns,
in the purification of His-tagged CRALBP. Spin-trap columns are small scale
purification “columns” (100 pL resin in each tube) where solutions are passed through a
small amount of resin using a centrifuge.13 They have a protein binding capacity of 750
pg per column.

In CRALBP purification with the spin-trap columns, 1.25 mL of cell-free extract
containing over-expressed His-tagged CRALBP was added to 3.75 mL of 20 mM
phosphate buffer (pH 7.2) that contained 10 mM imidazole and 300 mM NaCl. The
solution was loaded onto the spin-trap column (in 500 pL fractions, 10 times) and the
column was centrifuged each time. This was followed by rinsing with 20 mL pH 7.2
washing buffer I (20 mM phosphate buffer containing 0.1% Tween-20 surfactant and
0.15 M NaCl) and 20 mL pH 7.2 washing buffer 11 (20 mM phosphate buffer containing
45 mM imidazole and 0.15 M NaCl). The protein was then eluted using a 600 pL
solution containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M imidazole at pH
7.2. The electropherogram of the eluate (Figure 3.6., lane 3 and 4) shows a high protein

purity similar to that obtained with membrane purification.

9O

CRALBP

 

1 2 3 4

Figure 3.6. SDS-PAGE analysis (silver staining) of an extract fi‘om E. coli containing
over-expressed His-tagged CRALBP (Lane 1) and CRALBP purified from the cell
extracts using poly(HEMA}NTA-Ni2+-modified membranes (Lane 2) and spin-trap
columns (Lanes 3 and 4). 8 pL of eluate was loaded in lane 3 whereas 15 pL was

loaded onto lanes 2 and 4.

We further compared membrane-based isolation of CRALBP with purification
using a relatively large scale Ni-NTA column (Qiagen, Ni-NTA-Agarose, 6xHis-tagged
protein purification kit”). These columns have a protein binding capacity of 50 mg per
mL of resin. Gel electrophoresis of the eluate shows that the membranes were
comparable to the columns interms of purity and time of purification. However these
membranes have lower protein-binding capacities compared to the column. Thus, in
order to compete with the column-based purification, we need to increase the protein-
binding capacities of these membranes as well as decrease the time of purification, which
can be accomplished with the use of larger pore sized membranes (e.g. 5 pm). Chapter 6
demonstrates that new methods of membrane formation offer significantly higher

capacities and much shorter purification times.

91

3.4. Conclusions

This work demonstrates that poly(HEMA)-NTA-Ni2+-modified porous nylon
membranes can isolate His-tagged proteins directly from cell extracts. Gel
electrophoresis results indicate that the high purity obtained with membranes is partly due
to flow through the membrane that removes non-specifically adsorbed proteins.
lmportantly, the purity of the His-tagged protein is comparable to that obtained with
commercial spin-trap columns and large columns. However, the membranes have
potential advantage over commercial columns in terms of separation rates and higher
capacities. Hence, these modified membranes are attractive for selective purification of

His-tagged proteins.

92

3.5. References:

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)
(13)

(14)

http://www.whatman.com/PRODAnoporeInorganicMembranes.aspx (accessed
July 9, 2009).

Cai, C. F.; Bakowsky, U.; Rytting, B.; Schaper, A. K.; Kissel, T. European
Journal of Pharmaceutics and Biopharmaceutics 2008, 69, 31-42.

Kawakita, H.; Masunaga, H.; Nomura, K.; Uezu, K.; Akiba, 1.; Tsuneda, S.
Journal of Porous Materials 2007, 14, 387-391.

Coad, B. R.; Kizhakkedathu, J. N.; Haynes, C. A.; Brooks, D. E. Langmuir 2007,
23,11791-11803.

Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immarapom, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T. Macromolecules 1999, 32, 8716-8724.

Matyjaszewski, K.; Shipp, D. A.; Wang, J. L.; Grimaud, T.; Patten, T. E.
Macromolecules 1998, 31 , 6836-6840.

Huang, W. X.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,
35,1175-1179.

Dai, J. H.; Bao, Z. Y.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L.
Langmuir 2006, 22, 4274-4281.

Wray, W.; Boulikas, T.; Wray, V. P.; Hancock, R. Analytical Biochemistry 1981,
118, 197-203.

Bertolini, M. J.; Tankersley, D. L.; Schroeder, D. D. Analytical Biochemistry
1976, 71, 6-13.

Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Chemistry of Materials 2006,
18, 4033-4039.

Jia, X.; Michigan State University, 2008; Vol. Ph.D. Dissertation.

Hsu, C. Y.; Lin, H. Y.; Thomas, J. L.; Wu, B. T.; Chou, T. C. Biosensors &
Bioelectronics 2006, 22, 355-363.

Zhang, L.; Li, W.; Zhang, A. Progress in Chemistry 2006, 18, 939-949.

93

This chapter is adapted fi'om our published work in Langmuir
(Jain, P.; Dai, J .; Grajales, S.; Saha, S.; Baker, G. L.; Bruening, M.
L., Langmuir 2007, 23, 11360-11365). Some of the research was

done with Dr. J inhua Dai.

94

Chapter 4

Completely aqueous procedure for growth of polymer brushes

on polymeric substrates

4.1. Introduction

Chapter 2 demonstrated the high protein-binding capacities that can be achieved
by modifying porous alumina membranes with polymer brushes. However, as noted in
chapter 3, pore sizes in alumina membranes are limited to 0.25 pm, and this greatly limits
both membrane permeability and the viscosity of solutions that can be processed.
Development of more practical brush-modified membranes requires growth of brushes in
polymer substrates that contain micron-sized pores. However, when creating such
membranes, synthetic methods must afford both fine control over the rate of chain growth
and compatibility with the porous membrane supports.

As explained previously, the first step in the synthesis of polymer brushes by
ATRP is attachment of initiators to a surface, which is often realized by immersion of a
hydroxyl-functionalized substrate in an anhydrous organic solvent containing an initiator
precursor such as 2-bromoisobutyryl bromide or trichlorosilane."5 Unfortunately, many
polymer membrane materials such as polysulfone and polyethersulfone (PES)‘5 are
incompatible with the organic solvents used for initiator attachment. For such substrates,
organic solvents should be completely avoided in the brush synthesis to preserve the pore
structure of the membrane. Moreover robust membranes such as nylon are also affected
by the use of organic solvents. As shown in the chapter 3, nylon membranes modified

with poly(HEMA)-NTA-Ni2+ brushes have a relatively low protein binding capacity (25

95

mg protein/cm3 of the membrane), perhaps because the organic solvents employed in the
brush synthesis and derivatization partially damage the membrane structure.

This chapter presents a general, method for immobilizing ATRP initiators on
polymer films and membranes via layer-by-layer adsorption of macroinitiators from
water. Subsequent aqueous ATRP from these immobilized initiators yields polymer
brushes on polymeric substrates (Scheme 4.1.), and the entirely water-based process
avoids the use of organic solvents that may dissolve or corrupt porous, polymeric
substrates. Initiator immobilization relies on adsorption of a polycationic macroinitiator
(PMI) recently described by Armes and coworkers.7’9 To facilitate PMI adsorption, we
first prime the surface through adsorption of multilayer polyelectrolyte films comprised

of poly(styrene sulfonate) (PSS) and poly(diallyldimethylammonium chloride)

Polycationic
macroinitiator

   
    

 

Monomer
. , ;. Polymer brush
: CUCI’CUBrZ' hp? Polyelectrolyte
film

Polymer substrate

 

 

+ +1
/N\ 0 ”K
Br
PDADMAC PSS Polycationic macroinitiator

Scheme 4.1. Growth of polymer brushes by ATRP from macroinitiators adsorbed in a

membrane pore.

96

(PDADMAC). Polycationic or polyanionic initiators readily adsorb onto oppositely
charged surfaces through electrostatic interactions, as shown previously in adsorption on

0 This work builds on these

silica beads,8 silicon,9 and modified-colloidal particles.l
previous studies to demonstrate macroinitiator adsorption and subsequent polymerization
on PES films and membranes, which have a much different surface chemistry and
morphology than previous substrates employed for macroinitiator adsorption and
subsequent ATRP. Most importantly, the macroinitiator immobilization and
polymerization occur from water so damage to PES substrates caused by the use of
organic solvents is avoided. The adsorbed macroinitiators readily initiate polymerization
on PBS surfaces to allow growth of 100 nm-thick poly(HEMA) brushes in 30 min, and
the procedure can be applied to porous PES membranes.
4.2. Experimental section
4.2.1. Materials

3-mercaptopropionic acid (MPA), ll-mercapto-l-undecanol (MUD), 2-
bromoisobutyryl bromide, CuCl (99.999%), iodomethane (99%), PSS (Mw~70,000),
PDADMAC (Mw~150,000), CuBrz (99.999%), 2,2'-bipyridyl (299%, bpy), and 2,2'-
azobis(2-methylpropionitrile) (AIBN) were obtained from Sigma-Aldrich and used as
received. HEMA (97%, Aldrich) and 2-(dimethylamino)ethyl methacrylate (98%,
DMAEMA, Aldrich) were purified before use by passing the monomer through a column
of activated basic alumina (Spectrum). PES membranes with a nominal 0.45 pm cutoff

were obtained from GE Osmonics (Cat. #: SO4WP02500). PES coatings on Au-coated Si

wafers (200 nm of gold sputtered on 20 nm of Cr on Si(100)) were prepared by

97

dissolving a PES membrane (17 mg) in 10 mL CH2C12 and spin-coating the solution (1.0
mL) onto a substrate (l.1x2.4 cm) at a spin rate of 500 rpm.
4.2.2. Characterization methods

NMR Spectra were collected on a Varian Gemini-300 spectrometer. Polymer
molecular weights were determined by GPC (Gel Permeation Chromatography) at 35°C
using two PL-gel 10 mm mixed-B columns in series and an Optilab rEX differential
refiactometer (Wyatt Technology Co.) as a detector. THF was the eluting solvent at a
flow rate of 1 mL/min. All samples were filtered through a 0.2 pm Whatrnan
polytetrafluoroethylene (PTFE) syringe filter prior to GPC analysis. Films were
examined by reflectance FTIR spectroscopy and ellipsometry after each step leading to
brush growth. Reflectance FTIR spectra were acquired with a Nicolet Magna-IR 560
spectrophotometer containing a PIKE grazing angle (80°) attachment, and a UV/O3-
cleaned gold slide served as a background. Film thicknesses were determined using a
rotating analyzer ellipsometer (model M-44; J. A. Woollam) at an incident angle of 75°
and assuming a film refractive index of 1.5. The films were imaged in tapping-mode
with a Dimension 3100 scanning probe microscope equipped with a Nanoscope IIIA
control station. Film growth inside PES membranes was verified using transmission
FTIR spectroscopy (Mattson Galaxy Series 3000) with an air background as well as field-
emission scanning electron microscopy (FESEM, Hitachi S-470011, acceleration voltage
of 15 kV).
4.2.3. Determination of the internal surface area of a PES membrane

The internal surface area of these membranes was determined using N2 adsorption

measurements. N2 adsorption/desorption isotherms were obtained on a Micromeritics

98

ASAP 2010 Sorptometer using static adsorption procedures at -196 °C. 15 Pieces of PES

membranes (255 mg) were degassed for 48 hrs at 80 °C and 1023 torr prior to analysis.

The BET surface area was calculated to be 1800 cmz/membrane from the linear part of

the BET plot according to IUPAC recommendations.ll

4.2.4. Synthesis of 2-(2-bromoisobutyryloxy)ethyl acrylate
2-(2-bromoisobutyryloxy)ethyl acrylate (BIEA) was synthesized according to a

modified literature proceduren'l3

2-Hydroxyethyl acrylate (11.6 g, 0.1 mol) and
triethylamine (11.2 g, 0.11 mol), dried over 4-A molecular sieves, were dissolved in dry
dichloromethane (150 mL). The mixture was cooled in an ice bath, and a 15 mL solution
of 2-bromoisobutyryl bromide (24 g, 0.10 mol) in dry dichloromethane was added drop-
wise. After the addition was complete, the ice bath was removed, and the reaction
mixture was stirred at room temperature for 3h. The solution was filtered to remove
triethylammonium bromide and the salt was washed with dichloromethane (100 mL).
The organic solutions were combined and washed sequentially with 250 mL of 0.1 M
HCl, 250 mL saturated N3HCO3, and 250 mL water. After drying the organic phase over
anhydrous Na2804, the solvent was removed in vacuum to provide 24.2 g of BIEA in
92% yield. 1H NMR (Figure 4.1., CDCl3): 8 6.41 (dd, J = 18 Hz, J = 2.2 Hz, 1H), 6.11
(dd, J= 18 Hz, J= 18 Hz, 1H), 5.84 (dd, J= 11 Hz, J: 2.2 Hz, 1H), 4.40 (s, 4H), 1.90 (s,

6H).

99

 

 

 

 

h H>=$=O
0 d
d g .
O
0
Br d
H3C CH3
9
l h g
l lLlll I ll;-
7 6 5 4 3 2 1 0

Figure 4.1. 1H NMR (300 MHz, CDCl3) of 2-(2-bromoisobutyryloxy)ethyl acrylate

(BIEA).

4.2.5. Synthesis of poly(2-(dimethylamino)ethyl methacrylate-co-Z-(Z-
bromoisobutyryloxy)ethyl acrylate) (poly(DMAEMA-co-BIEA), the precursor of the
polycationic initiator

DMAEMA (2.97 g, 18.9 mmol), BIEA (1.45 g, 5.47 mmol), and AIBN (82 mg,
0.5 mmol) were added to 5 mL of dry THF. The mixture was degassed via three freeze-
pump-thaw cycles, and then polymerization was carried out at 60 °C with stirring for 2 h.
The highly viscous polymer mixture was diluted with 15 mL of THF, and the polymer
was precipitated into pH 11 water. After filtration, the polymer was dried under vacuum,
re-dissolved in 15 mL THF, and precipitated into hexane. Filtration and drying under
vacuum at room temperature gave 2.1 g of the copolymer. The lH NMR spectrum of the
polymer is shown in Figure 4.2, and indicates approximately 20% BIEA in the

copolymer.

100

N backbone CH, CH2
0 a \ a e
Br c,d b ,4

f' 5 * 4 3 2 1 0

 

 

 

backbone CH, CH2

 

 

 

 

Figure 4.2. 1H NMR spectra of poly(DMAEMA-co-BIEA) (top, acetone-do + trace
D20 as solvent) and its quatemized product poly(TMAEMA-co-BIEA) (bottom, D20
as solvent). The corresponding structures and proton assignments are given in the

figure. * indicates resonances from solvents.

4.2.6. Synthesis of poly(2-(trimethylammonium iodide)ethyl methacrylate-co-BIEA)
(poly(TMAEMA-co-BIEA)), the polycationic initiator
Poly (DMAEMA-co—BIEA) (1.71 g) was dissolved in 25 mL THF, and 1.0 mL of

CH31 was added to the stirred solution at room temperature. Within 2 min, the reaction

101

 

mixture became turbid with a butter-like color. After 1 h of stirring, the solution was
added dr0pwise to vigorously stirred hexane to precipitate the polymer as a fine powder.
Washing with hexane and drying under vacuum at room temperature for 12 h provided
2.52 g of polymer. The 1H NMR spectrum of the polymer is shown in Figure 4.2.
4.2.7. Polymer brush synthesis

Poly(HEMA) and poly(DMAEMA) were grown from poly(TMAEMA-co-BIEA)-
modified Au and PES surfaces. In the case of Au surfaces, an Au—coated silicon wafer
was UV/ozone cleaned and immersed overnight in a 5 mM MPA solution in ethanol to
form a MPA self-assembled monolayer. This substrate was then alternatively immersed
in aqueous solutions of 0.02 M PDADMAC (containing 0.5 M NaCl) and 0.02 M PSS
(containing 0.5 M NaCl) for 5 min, with a 1-min water rinse between each immersion.
(Polymer concentrations are given with respect to the repeating unit, and the pH of these
solutions was ~7.0). After deposition of a (PDADMAC/PSS)2 film, the substrate was
immersed in a solution of poly(TMAEMA-co-BIEA) (2.0 mg/mL in water) for 10 min,
rinsed with water, and dried under a stream of N2. PES substrates (both membranes and
spin-coated films) were modified similarly but with slightly thicker polyelectrolyte films,
(PSS/PDADMAC)4PSS. (Deposition on polymers begins with PSS instead of
PDADMAC because PSS adsorption on PES should be more favorable due to
hydrophobic interactions.) For poly(TMAEMA-co-BlEA)/PSS multilayer films on gold,
we used the same deposition solutions and times as above.

Brush growth from adsorbed initiators followed procedures reported previously.3
Briefly, 15 mL of monomer (HEMA or DMAEMA) was mixed with 15 mL of water in a

Schlenk flask, and this solution was degassed via three freeze-pump-thaw cycles.

102

u...‘ “~14

 

Catalyst and ligand were added, and the solution was subjected to one additional freeze—
pump-thaw cycle. The molar ratio of the reagents was monomerzCuClzCuBr2szy =
50:1:0.3:2.6. In a glovebag, initiator-modified substrates were immersed in the degassed
solution, and polymerization was allowed to proceed at room temperature for the desired
time. Polymer-coated substrates were then removed from the glovebag and rinsed
sequentially with ethanol and water.

4.3. Results and discussion

4.3.1. Synthesis of polycationic macroinitiator

The macroinitiator, poly(TMAEMA-co-BIEA), is closely related to that reported
by Armes et al.8‘[7 The primary difference is the method used to incorporate the ATRP
initiator into the macroinitiator. Armes used a post-polymerization strategy, acylating a
HEMA-DMAEMA copolymer, while we copolymerized DMAEMA with 2-(2-
bromoisobutyryloxy)ethyl acrylate (BIEA), a monomer capable of initiating ATRP.
Quaternization of the resulting copolymer is the final step in both syntheses.

The copolymer intermediate, poly(DMAEMA-co-BIEA), is insoluble in water but
easily dissolves in acetone and THF. The mole fraction of BIEA in the copolymer
determined from the 1H NMR integration ratios (Figure 4.2., top) was ~20%, in
reasonable agreement with the initial ratio of monomers in the polymerization solution
(DMAEMA/BIEA = 3.46/1). The broad resonances in the 1H NMR spectra imply a high
molecular weight for the copolymer, but GPC data acquired in THF and calibrated with
polystyrene standards reveal a bimodal molecular weight distribution with molecular
weights <3000 g/mol (Figure 4.3.). Since GPC separates molecules on the basis of their

size in solution, the data may indicate that poly(DMAEMA-co-BIEA) has a smaller

103

 

 

 

 

 

A

9 e

9 e
a: 9 e
g N3 ’-
8 0 o
in f e
g e
.9
o f
.3
3 M

16 17 . 18 19

Elution time (min)
Figure 4.3. GPC trace of poly(TMAEMA-co-BIEA). The sample was run at 35 0C
using two PL gel 10 mm mixed-B columns in series, and THF as the eluting solvent at
a flow rate of 1 mL/min. The samples were filtered through a 0.2 pm Whatrnan PTF E
syringe filter 'prior to GPC analysis. The arrow indicates the elution time for the
lowest polystyrene standard (2727 g/mol) used for calibrating the columns. The

elution time for pure solvent is ~20 minutes.

hydrodynamic radius than polystyrene of comparable molecular weight. Thus, using
polystyrene standards may underestimate the true molecular weight of poly(DMAEMA-
co-BIEA). In addition, the AIBN-initiated free radical copolymerization of DMAEMA
with BIEA may provide branched rather than linear polymers since the bromine atom in
BIEA is activated (a to a carbonyl) and can act as an internal chain transfer agent.

Bromine atom abstraction by initiator or growing polymer chains is expected to cause

104

1a

branching from the BIEA segments resulting in a more compact copolymer. However,
the extent of chain transfer is difficult to characterize by 1H and '3 C NMR. When
growing polymer chains abstract bromine from BIEA segments, the polymer chains are
terminated in a-bromocarbonyls, which will have reactivities comparable to that of
BIEA. Solutions of poly(DMAEMA-co-BIEA) have limited stability and eventually
become insoluble, precluding extensive characterization of the copolymer; we suspect
cross-linking via intra and intermolecular quatemization of DMAEMA with BIEA.
Poly(TMAEMA-co-BIEA), the quatemized polymer, is stable and water soluble, but
insoluble in common organic solvents. The NMR spectrum of poly(TMAEMA-co-
BIEA) shows downfield shifting of the methyl groups attached to nitrogen (a' in Figure
4.2.), indicating successful quatemization. The NMR spectrum of this polymer is
essentially the same as that reported by Armes and coworkers for a similar macroinitiator
prepared by acylation of the hydroxyl groups after polymerization. To verify that the a-
bromo ester survived the quatemization step, we precipitated an aqueous solution of the
polymer in acetone and analyzed the dried polymer using X-ray photoelectron
spectroscopy (Figure 4.4.). The atom ratio of Br to N was 0.28, which is reasonably
consistent with 20% BIEA in poly(TMAEMA-co-BIEA). Thus, the initiating moiety is
at least temporarily stable to the quatemization process and to exposure to water.14
However, over time it appears that the Br is slowly abstracted. In the NMR spectrum of a
10-month old sample of poly(TMAEMA-co-BIEA), the peak due to the methyl groups
adjacent to the Br, e in Figure 4.2., split into two peaks, suggesting that about half of the

Br had been extracted (Figure 4.5.). Still, as shown below, lO-month old macroinitiator

was capable of effectively initiating polymerization.

105

 

Counts/s (x10‘)
N on A on c» ‘1 co co

A

 

1M
_ Atomic Concentration Table

I MNN1
l3p1
N KLL
OMNN
|3d5
C15

013

 

 

 

 

 

Br3p3 - Br3p1 - 8er

C13 N15 015 Br3d 13d5
[0.314] [0.499] [0.733] [1.149] [6.302]
78.22 12.82 I 15.90 0.78 2.28

EMS
- Br3s

l

 

 

O

1000 900 800 700 600 500 400 300 200

Binding Energy

Figure 4.4. XPS elemental analysis of quatemized poly(DMAEMA-co-BIEA).

106

 

‘

 

 

 

 

 

 

 

 

X
0 0
2‘” °'
d 0 b' ._
N*~ l'
/
O? a'\a'
, Br
6 e
e
9109'
ll ’"
I!
5 4 3 2 1 0
99m

Figure 4.5. 1H NMR spectrum of a 10-month old sample of poly(TMAEMA-co-
BIEA) (D20 as solvent), showing evolution of the u—methyl resonances (e', e") . The
corresponding structures and proton assignments are given in the figure. * indicates

resonances from solvents.

4.3.2. Polymer brush growth from surfaces modified with poly(TMAEMA-co-
BIEA)

The first step in growing polymer brushes from poly(TMAEMA-co-BIEA) is
adsorption of the macroinitiator on a surface. To examine the adsorption properties of
this polymer, we fabricated multilayer initiator-containing films by alternating deposition
of poly(TMAEMA-co-BIEA) and PSS on a MPA-modified Au surface. Reflectance
FTIR spectra of these films (Figure 4.6.) show a linear increase in the absorbance at

1730 cm’l (attributed to the ester carbonyl groups of poly(TMAEMA-co-BIEA)) as a

107

fmetion of the number of deposited bilayers, confirming controlled macroinitiator
adsorption. Trends in absorbances due primarily to CH3 and CH2 bands (around 1490
cm") in poly(TMAEMA-co-BIEA) as well as several PSS vibrations (1200-1225, 1040,
and 1010 cm") are also consistent with regular film growth. Plots of ellipsometric
thickness versus the number of deposited bilayers showed a grth rate of 4.3

nm/bilayer.

 

 

 

[0.005

d)

0

C

1‘!

9 J

o

m

5 .n=4 1"
.n=3 l
41:2 ' '
~n=1 /\
'n=0

 

1900 1700 1500 1300 1100 900
Wavenumbers (cm-1)

Figure 4.6. Reflectance FTIR spectra of poly(TMAEMA-co-BIEA)/

[PSS/poly(TMAEMA -co-BIEA)]n films deposited on Au-MPA substrates (n=0-4).

We investigated the ability of absorbed poly(TMAEMA-co-BIEA) multilayer
films to initiate polymerization by growing poly(HEMA) brushes from such films via
aqueous ATRP for 2 h. The appearance of broad hydroxyl peaks (3500-3300 cm") and
large ester carbonyl peaks at 1730 cm'1 in the reflectance FTIR spectra of these films
(Figure 4.7.) confirmed the grth of poly(HEMA) from the macroinitiator-modified

surfaces.

108

 

i.1"..".

 

0.1

 

 

 

0

O n:4

g N

n

3 ":3 M

m N

g n=2 A , ll
n=0

— A *!

 

 

 

 

3800 3300 2800 2300 1800 1300 800
Wavenumbers (cm")
Figure 4.7. Reflectance FTIR spectra poly(HEMA) films grown from
poly(TMAEMA-co-BIEA)/ [PSS/poly(TMAEMA-co-BIEA)]n films deposited on Au-

MPA substrates (n = 0-4).

Moreover, both the absorbances due to poly(HEMA) and the ellipsometric
thickness of the poly(HEMA) films (Figure 4.8.) increased monotonically with the
number of PSS-macroinitiator bilayers. The enhanced polymer grth with more
PSS/poly(TMAEMA -co-BIEA) bilayers suggests that initiators throughout the film can
initiate brush growth. We should note that these brushes were polymerized from
multilayer films that were prepared using macroinitiator that was about a year old, and
about half of the Br groups may be removed from the initiator under such conditions
(Figure 4.5.). Still, films as thick as 160 nm can be obtained in just 2 h when
polymerizing from poly(TMAEMA-co-BIEA)/[PSS/poly(TMAEMA -co-BIEA)I4 films.
(Unless specified, all other experiments reported here were performed within a few

months of initiator synthesis.)

109

 

 

200

..x
0'1
0

Brush thickness (nm)
0|

0

0

 

 

0
0 1 2 3 4 5

Number of PSS-Initiator bilayers
Figure 4.8. Poly(HEMA) brush ellipsometric thickness versus number of bilayers for
growth from poly(TMAEMA-co-BlEA)/[PSS/poly(TMAEMA -co-BIEA)]n films
deposited on Au—MPA substrates (n=0-4). The thickness of the

initiator/polyelectrolyte layer was subtracted from the total thickness of the film.

We examined the kinetics of polymerization using a single macroinitiator layer
adsorbed on a (PDADMAC/PSS)2 film (Figure 4.9.). Brush thickness increased with
polymerization time to give films as thick as 200 nm in just 3 hours, but decreasing
grth rates over time suggest some termination. As expected, control experiments with
(PDADMAC/PSS)2 films devoid of initiator layers showed no polymer growth,
confirming that the polyelectrolyte layers are incapable of initiating polymerization in the

absence of the macroinitiator.

110

-__. .. .ant mixuawfilfl

 

 

250
E 200- e
I ' °
8 .
.2
F. 1004 0 °
13 e

e

a 501 e

 

 

 

O 50 100 1 50 200
Polymerization time (min)

Figure 4.9. Poly(HEMA) brush thickness versus polymerization time for grth from
Au-MPA-[PDADMAC/PSS]2-poly(TMAEMA-co-BIEA) films (circles) and from Au-
MUD-BribBr films (diamonds). The thickness of the initiator layer was subtracted

from the total thickness of the film.

To compare the initiation performance of the adsorbed poly(TMAEMA-co-BIEA)
to that of typical initiator monolayers, we also synthesized poly(HEMA) brushes from Au
surfaces modified by a self-assembled monolayer of mercaptoundecanol that was
subsequently reacted with 2-bromoisobutyryl bromide (BribBr). The initiating motifs of
the BribBr-modified monolayer and the adsorbed poly(TMAEMA-co-BIEA) are the
same (2-bromoisobutyryl ester). The poly(HEMA) growth rate from adsorbed
poly(TMAEMA-co-BIEA) is comparable to or greater than that of the initiator
monolayer, as shown in Figure 4.9. This might seem surprising considering the fact that
only 20% of the repeat units in poly(TMAEMA-co-BIEA) contain the initiating motif.
However, poly(TMAEMA-co-BIEA) likely adsorbed in a coiled form, which allows for

deposition of a thick (~45 A) layer with a reasonably high areal density of initiation

lll

 

sites.'5‘16 Given the thickness of the poly(TMAEMA-co-BIEA) layer and its
composition, the total amount of transferable Br in the adsorbed macroinitiator is
essentially the same as what we would expect for a fillly derivatized self-assembled
monolayer.3 Moreover, most of the initiation sites in the monolayer film will not be
active due to steric constraints.l7 Such constraints should be less demanding in the more
three-dimensional poly(TMAEMA-co-BIEA) layer.

Film surface morphologies were examined using Atomic Force Microscopy
(AF M) before and after polymerization. Figure 4.10.a. and 4.10.b. clearly show changes
in surface roughness after polymerization. The root mean square roughness increased
from 3.6 nm for a Au-MPA-(PDADMAC/PSS)2-poly(TMAEMA-co-BIEA) film to 14.7
nm after poly(HEMA) grth for 2h. This topology may be indicative of patchy
initiation because of the use of aged initiator in this case. However, the peak to trough
distance of 51 nm is still considerably less than the film thickness of about 110 nm.
Moreover, roughness may be a strong function of solvent treatment.
4.3.3. Polymer brush growth on PES

We selected PES as a representative material to demonstrate the use of the
macroinitiator for synthesizing brushes on polymer substrates that can be formed into
membranes. PES is widely used in membrane manufacturing because of its resistance to
high temperature, acid, and base, but it has a very weak tolerance to organic solvents.'8
For convenience of characterization, we first dissolved a PES membrane in
dichloromethane and spin-coated this solution on Au slides to obtain a PES film (~40
nm). A PSS(PDADMAC/PSS)4 polyelectrolyte film was then deposited on the PES,

followed by adsorption of poly(TMAEMA-co-BIEA).

112

 

 

pm

   

0.6

 
   

0.4
0.2

x 0.200 pm/div
2 40.000 nmldiv

Figure 4.10. AFM Images of (a): a Au-MPA-(PDADMAC/PSS)2-poly(TMAEMA-
co-BIEA) film; (b): (a) + a poly(HEMA) brush. Note that these films were prepared

using a l-year old macroinitiator.

113

 

Reflectance F TIR spectroscopy (Figure 4.11.) confirmed each step in the initiator
attachment and grth of polymer from these substrates. The peaks at 1040 and 1010
cm'1 (Figure 4.ll.b., under the thick arrow) stem from the adsorbed PSS, and indicate
adsorption of PSS(PDADMAC/PSS)4, while the ester carbonyl absorbance at 1730 cm"1
(Figure 4.ll.c., under the arrow) provides evidence for deposition of the

poly(TMAEMA-co-BIEA). A 2 h polymerization of HEMA from these surfaces yielded

 

a ~ 250 nm increase in the film’s ellipsometric thickness, while the corresponding

polymerization of 2-(dimethylamino)ethyl methacrylate resulted in grth of a 40 nm-

 

thick brush. Large increases in carbonyl absorbances (Figure 4.1 1.d. and Figure 4.] Le.)

'°" will.
M e M

7:“ ': mm

3400 2900 2400 1900 1400 900

Wavenumbers (cm‘)

 

 

 

 

 

Absorbance

O

 

 

 

 

 

Figure 4.11. Reflectance FTIR of (a): spin-coated PES on Au; (b): (a) +
PSS(PDADMAC/PSS)4; (c): (b) + poly(TMAEMA-co-BIEA); (d): (c) +

poly(DMAEMA) brush; and (e): (c) + poly(HEMA) brush.

114

also confirmed the grth of the brushes.

Lastly, we synthesized polymer brushes directly on porous PES membranes. In
one case, the PES membrane was coated with a PSS(PDADMAC/PSS)4/poly(TMAEMA-
co—BIEA) film before poly(HEMA) brush growth, and the polymerization was performed
for l h. Membranes were exposed to deposition and polymerization reagents by simple
immersion in the solution. In a second case, to insure modification throughout the
interior of the membranes, we used only a single PSS/poly(TMAEMA-co-BIEA) bilayer
because PSS(PDADMAC/PSS)4 bilayers blocked the pores at some point in the
membrane as judged by our inability to flow water through these membranes using a
peristaltic pump. The time of polymerization employed to grow polymer inside the pores
was also reduced to 15 minutes to avoid blocking, and all solutions were flowed through
the membrane during both initiator deposition and polymerization. Poly(HEMA) grth
in both cases is evident from a new peak at 1724 cm'1 in transmission FTIR spectra
(Figure 4.12.b. and 4.12.c., under the arrow).

We quantified the amount of poly(HEMA) brushes in PES membranes both
gravimetrically and by calibrating the transmission IR absorbance of poly(HEMA)-
modified membranes using spectra of pure poly(HEMA) in KBr pellets (Figure 4.13.).
Results from these two methods match reasonably well. For example, the sample shown
in Figure 4.12.b. has a poly(HEMA) ester carbonyl absorbance of 0.48, which
corresponds to 1.7 mg poly(HEMA) per membrane according to an infrared calibration
curve, while the direct mass measurement gave 2.3 mg poly(HEMA) in the membrane.
The slightly larger value obtained gravimetrically could be attributed to incomplete

removal of water and monomer/catalyst residues after polymerization. Assuming a

115

 

 

 

Absorbance
o-
‘—

 

 

 

2000 1 500 1 000 500

Wavenumbers (cm")

Figure 4.12. Transmission FTIR spectra of (a) a bare PES membrane, (b) 3 PES
membrane modified with a PSS(PDADMAC/PSS)4/poly(TMAEMA-co-BIEA) film
and 1 h polymerization of poly(HEMA) (growth on the membrane surface), and (c) a
PBS membrane modified with a PSS/poly(TMAEMA-co-BIEA) bilayer and 15 min

polymerization of poly(HEMA) (grth in the membrane pores).

porosity of about 80%, 1.7 mg of poly(HEMA) in a PBS membrane is equivalent to about
5% of the initial open membrane volume. However, in this case much of the
polymerization likely occurred on the membrane surface. Similarly for the sample shown
in Figure 4.12.c, the poly(HEMA) ester carbonyl absorbance of 0.41 corresponds to 1.45
mg poly(HEMA) per membrane according to an infrared calibration curve, while the
direct mass measurement gave 1.5 mg poly(HEMA) in the membrane. The unmodified
membrane had an exposed diameter of 20 mm, a thickness of ~110 pm, and a N2

adsorption-based internal surface area of 1800 cm2. Assuming this pore surface area to

116

 

be flat, 1.5 mg of poly(HEMA) would correspond to brushes with a thickness of ~8.3 nm,
which is reasonable given the short polymerization time (15 min for polymerization

inside the pores).

 

  

y = 0.2823 x
R2 = 0.9936

C=O absorbance
at 1730 cm 1
O
CO

 

 

 

0 1 2 3 4 5 6 7
Amount of poly(HEMA) in KBr
pellet (mg/cm?)
Figure 4.13. Calibration curve of the absorbance at 1730 cm'1 (attributed to the ester
carbonyl of poly(HEMA) versus the areal concentration of poly(HEMA) in KBr
pellets. Poly(HEMA) KBr pellets were prepared by thoroughly mixing different

amounts of poly(HEMA) and KBr (the total amount of the mixture was 100 mg, and

the face of the pellet had an area of 1.54 cmz).

To investigate the morphology of poly(HEMA) grown within the pores of PES
membranes, we dissolved away the PES framework using dichloromethane and examined
the remaining material using SEM. Poly(HEMA) is partially cross-linked due to
transesterification and thus insoluble in dichloromethane, andthe FTIR spectrum of the
material remaining after an overnight immersion in CH2C12 showed predominantly

poly(HEMA). The SEM image of the dissolved PES-poly(HEMA) membrane shows

117

 

what appears to be to be a replica of the large pores of a bare PES membrane (Figure
4.14.). The small pores are not present in the dissolved sample, however, which could be

due to bridging of pores by poly(HEMA) or collapse of the structure during dissolution

and swelling of poly(HEMA).

   

Figure 4.14. SEM image of (a) a bare PES membrane and (b) 3 PES membrane

dissolved in dichloromethane after derivatization with poly(HEMA).

4.4 Conclusions
4 To conclude, layer-by-layer deposition of polyelectrolytes including at least one
layer of a macroinitiator allows growth of polymer brushes from polymer supports in an
entirely aqueous procedure, circumventing utilization of organic solvents that may

dissolve or corrupt substrates.

118

 

4.5. References

(1)

(2)

(3)

(4)

(5)
(6)

(7)
(8)

(9)

(10)
(11)

(12)

(13)
(14)

(15)

(16)
(17)

Tsujii, Y.; Ohno, K.; Yamamoto, S.; Goto, A.; Fukuda, T. Advances in Polymer
Science 2006, 197, 1-45.

Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immarapom, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.;
Pakula, T. Macromolecules 1999, 32, 8716-8724.

Huang, W.; Kim, J.-B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002, 35, F:
1175-1179. 1

Edmondson, S.; Osborne, V. L.; Huck, W. T. S. Chemical Society Reviews 2004, i
33, 14-22. .1

 

Matyjaszewski, K.; Xia, J. Chemical Reviews 2001, 101, 2921-2990.

Baker, R. W. Membrane Technology and Applications; Second ed.; John Wiley &
Sons Ltd., 2004.

Chen, X.; Armes, S. P. Advanced Materials 2003, 15, 1558-1562.

Chen, X. Y.; Armes, S. P.; Greaves, S. J.; Watts, J. F. Langmuir 2004, 20, 587-
595.

Edmondson, S.; Vo, C.-D.; Armes, S. P.; Unali, G.-F. Macromolecules 2007, 40,
5271-5278.

Fulghum, T. M.; Patton, D. L.; Advincula, R. C. Langmuir 2006, 22, 8397-8402.

Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.;
Rouquerol, J .; Siemieniewska, T. Pure and Applied Chemistry 1985, 57, 603-619.

Matyjaszewski, K.; Gaynor, S. G.; Kulfan, A.; Podwika, M. Macromolecules
1997, 30, 5192-5194.

Zhai, G.; Kang, E. T.; Neoh, K. G. Macromolecules 2004, 3 7, 7240-7249.

Boyer, C.; Bulmus, V.; Liu, J. 0.; Davis, T. P.; Stenzel, M. H.; Barner-Kowollik,
C. Journal of the American Chemical Society 2007, 129, 7145-7154.

Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromolecular Rapid
Communications 2000, 21 , 319-348.

Dubas, S. T.; Schlenoff, J. B. Macromolecules 1999, 32, 8153-8160.

Bao, Z.; Bruening, M. L.; Baker, G. L. Macromolecules 2006, 39, 5251-5258.

119

(18)

Nicolas, J.; Mantovani, G.; Haddleton, D. M. Macromolecular Rapid
Communications 2007, 28, 1083-111 1.

120

 

 

This chapter is adapted from ' our published work in
Macromolecules (Jain, P.; Dai, J.; Baker, G. L.; Bruening, M. L.,

Macromolecules 2008, 41 , 8413-8417).

121

Chapter 5
Rapid synthesis of functional polymer brushes by surface-
initiated atom transfer radical polymerization of an acidic

monomer

5.1. Introduction

The previous chapter describes aqueous methods for growth of poly(HEMA)
brushes inside polymer membranes. However, our procedure for creation of protein-
binding membranes requires conversion of hydroxyl groups in poly(HEMA) to
carboxylic acid moieties. As explained in chapters 2 and 3, the formation of acid groups
involves reaction of poly(HEMA) with succinic anhydride in an organic solvent for
several hours (Scheme 5.1.a), which partially damages the membrane structure and leads
to low protein binding capacities. This chapter describes surface-initiated aqueous ATRP
of an acidic monomer, 2-(methacryloyloxy)ethyl succinate (MES), that allows direct
synthesis of poly(carboxylic acid) brushes, hence avoiding the need to react the brush
with succinic anhydride in an organic solvent. Interestingly, this method yields the same
polymer formula as obtained after reaction of poly(HEMA) with succinic anhydride

(Scheme 5.1.b).

122

 

 

 

0 CH3

 

ll .
3+CHzio—c-c—Br
11 CH,

Initiator-modified substrate

H26 ,CH3

Sc
' HEMA

O¢C\ _ ( )

o—Cl-12——CH2 OH

(a)

H20, CuCl, CuBrz, bpy

 

V

(3 CH3 (EH3
S+CH2)o—c-¢—l—CH2—c—+—Br
11 an n

6
¢
0 \O-CHz-CH2-OH

O 0 0
U ,DMAP,DMF,3h

(Succinic anhydride)

[,9 CH3 (EH3
S+CH2)o-c-¢—[—CH2-c—[—-Br
11 CH3 '1

c
/
o/\

0

II II
O-CH2—CH2-O—C-CH2-CH2-C-OH <-

(b)

 

HZCQHYCHC,
/C O O
0/ \o—CHz—CHZ-o—b-CHZ-CHz—b-0H
(MES)
CuCI, CuBerigand, DMF
HZO/NaOH

O

 

Scheme 5.1. Polymerization of (a) HEMA or (b) MES from initiator immobilized on

a gold substrate. Derivatization of the poly(HEMA) with succinic anhydride gives

brushes with the same composition as poly(MES).

There are a number of other methods for synthesizing poly(carboxylic acid)

brushes because they have potential applications in cell adhesion and growth,"2

immobilization of biomacromolecules,3'5 synthesis of shell-crosslinked micelles?‘7 and

entrapment of nanoparticles and catalysts.8 Synthesis of polyacid brushes frequently

includes surface-initiated anionic polymerizations"l3 or ATRP'4"8 of an ester-containing

monomer and subsequent hydrolysis. For instance, Boyes et al. synthesized polyacid

123

 

brushes through grth and hydrolysis of poly(tert-butyl acrylate) and
poly(methacrylate) films.l6 Unfortunately, these strategies require multiple steps
including harsh acidic conditions for deprotection. Techniques such as

19,20 and

photopolymerization are capable of directly forming acrylic acid (AA)
methacrylic acid (MAA) films,21 however the use of irradiation or high temperatures for
initiation might be incompatible with opaque or temperature-sensitive substrates.
Additionally, conventional free radical techniques lead to extensive polymerization in
solution and require exhaustive rinsing to remove physisorbed polymers. The use of
controlled polymerization methods can potentially overcome these limitations.”23
Among the strategies for controlled polymerization, ATRP is attractive due to its mild
reaction conditions (room temperature in many cases), tolerance to impurities, and use of

readily available catalysts, initiators and monomers.”26

Perhaps most importantly,
surface-initiated ATRP results in minimal solution polymerization and allows control
over polymer thickness by variation of polymerization time. Thus, ATRP is attractive for
creating films in complex geometries such as the pores of membranes and for tailoring
film architectures to maximize protein adsorption.

Here, we describe the rapid synthesis of carboxylic acid-containing polymer
brushes using surface-initiated ATRP of an acidic monomer, MES. The combination of
this monomer and highly active ATRP catalysts, e.g., Cu(I) complexed with
1,1,4,7,]0,10-hexamethyltriethylenetetramine (HMTETA), allows rapid formation of

polyacid brushes from a surface. ATRP from initiators immobilized on Au-coated Si

wafers yields films with an ellipsometric thickness of 120 nm in less than 15 min. To the

124

 

1""

best of our knowledge, this is the first example of direct ATRP of protonated acidic
monomers that is capable of yielding such thick films.

Several papers reported growth of polymer brushes via surface-initiated ATRP of
acid monomers in aqueous solution, but most polymerization rates and/or thicknesses

were low.”3 '

Using ATRP from immobilized initiators, Sankhe et al. polymerized
itaconic acid and MAA from Au, but film grth ceased after the formation of 10 nm-
thick coatings.32 The addition of salt to polymerization solutions allowed more
controlled growth of poly(itaconic acid) and poly(MAA) films, but the growth rates were
only 0.2 nm/h and l nm/h, respectively.33 Other groups used a similar strategy to grow
~40 nm-thick poly(AA) brushes.3 1‘34 Tugulu and coworkers reported that direct, aqueous
polymerization of sodium methacrylate from initiators on Si can yield 300 nm thick
poly(MAA) in 3 h.35 Using the same conditions, we tried growing poly(MAA) brushes
from Au surfaces that were modified by immersion in a solution containing a disulfide
initiator, (BrC(CH3)2COO(CH2)nS)2. These polymerizations yielded inhomogeneous
films that were visibly rough and not smooth enough for ellipsometric characterization.
Similar results occurred on Si modified with initiators. In contrast, the polymerization of
MES in its protonated form provides visibly homogeneous films, and ellipsometry of
these films yields data that are characteristic of smooth coatings.

As mentioned in prior chapters, polyacid films are attractive as protein
binders3’20’36'38 for applications such as protein purification by affinity adsorption,36'394'
immunoassays,42 enzymatic reactions,3 and analyses with antibody arrays."’3‘46 In these

applications, the formation of thick polymer brushes is vital for achieving high binding

capacities. This chapter also shows that a 55 nm-thick poly(MES) film absorbs the

125

 

 

equivalent of ~70 monolayers (14.4i0.3 pg/cmz) of lysozyme, and derivatization of
poly(MES) with metal-ion complexes allows binding of large amounts of protein via
metal-affinity interactions. The binding capacities are similar to those of poly(AA)
brushes prepared by polymerization of tert-butyl acrylate and subsequent hydrolysis, but
the poly(MES) synthesis is a one-step, aqueous process.
5.2. Experimental section
5.2.1. Materials

ll-mercaptoundecanol (97%), DMF (anhydrous, 99.8%), 2-bromoisobutyryl
bromide (98%), CuBr (99.999%), CuBr2 (99%), bpy (99%), 1,1,4,7,10,10-
hexamethyltriethylenetetramine (HMTETA), 1 ,4,8,1 l-tetraaza- 1 ,4,8,1 1-

tetramethylcyclotetradecane (Me4Cyclam), 4,4'-dinonyl-2,2'-bipyridy1 (anbpy), EDC,

NHS, 4-(bromomethyl)benzoic acid (97%), column packing for removing hydroquinone

and monomethyl ether hydroquinone (MEHQ, cat. no. 311332), TWEEN-20 surfactant,
BSA and lysozyme were used as received fi'om Sigma-Aldrich. CuSO4-5H2O (CCI),
NaOH (Spectrum), NaH2PO4 (CCI), Na2HPO4 (Aldrich) and aminobutyl NTA (F luka)
were also used without purification. HEMA (Aldrich, 97%, inhibited with 300 ppm
hydroquinone monomethyl ether) was purified by passing it through a column of
activated basic alumina (Aldrich). In most cases, MES (Aldrich, inhibited with 750 ppm
MEHQ), MAA (Aldrich, 99%, inhibited with 250 ppm MEHQ), AA (Aldrich, 99%,
inhibited with 200 ppm MEHQ) were used as received, but where noted, the inhibitor
was removed by passing the monomer through a column of inhibitor removal packing.
The disulfide initiator, (BrC(CH3)2COO(CH2)11S)2,47 and tris[2-

(dimethylamino)ethyl]amine (MC6TREN)48 were synthesized as described previously.

126

 

II. ..
1.!

Buffers were prepared using analytical grade chemicals and deionized water (Milli-Q,
18.2 M!) cm).
5.2.2. Polymerization of MES on Au substrates

Au—coated wafers (200 nm of sputtered Au on 20 nm of sputtered Cr on Si
wafers) were cleaned with UV-ozone for 15 min, immersed in a 1 mM ethanolic solution
of the disulfide initiator, (BrC(CH3)2COO(CH2)nS)2, for 24 h, rinsed sequentially with
ethanol and water, and dried under a stream of N2. These initiator-modified substrates
were then transferred to a N2-filled glove bag where polymerization was carried out at
ambient temperature.

To prepare most polymerization solutions, 10 mL of a mixture of neat monomer
and 1 M aqueous NaOH (1:1, v/v) was first degassed with three freeze-pump-thaw
cycles. A 1 mL solution of DMF containing CuBr, CuBr2 and ligands was similarly
degassed, and in a N2-filled glove bag, this solution of catalyst was mixed with the
monomer/NaOH solution. Finally, the initiator-coated substrate was immersed in the
polymerization solution, and after the desired time, the coated wafers were removed from
the glove bag, immediately sonicated in DMF for 10 min, rinsed sequentially with
ethanol and water, and dried under a stream of N2. Final concentrations of the different
catalysts in the various polymerization solutions were: CuBr (2 mM), CuBr2 (1 mM),
HMTETA (6 mM); CuBr (2 mM), CuBr2 (1 mM), Me4Cyclam (2 mM), anbpy (1 mM);
and CuBr (2 mM), CuBr2 (1 mM), Me6TREN (6 mM).”’49 In a few cases, we halved the
concentrations of HMTETA and Me6TREN ligands and achieved similar film
thicknesses."”’SO For the system containing bpy, the concentrations were CuBr (0.55

mM), CuBr2 (0.16 mM) and bpy (1.56 mM). The use of a bpy system containing 6 mM

127

...r ...:r- 1- 11:3
z.”

 

"1......

bpy, 2 mM CuBr, and 1 mM CuBr2 resulted in very thin (< 25 nm) films after 2 h of
polymerization. The pH of the polymerization solutions was ~5.0.

When examining polymerization rates as a function of the amount of NaOH
added to the polymerization solution, the monomer and catalyst concentrations were kept
constant. For these studies, neat monomer, aqueous solutions of 1 M and 5 M NaOH,
deionized water, and catalyst solution containing CuBr, CuBr2, and HMTETA in DMF
were degassed separately with three freeze-pump-thaw cycles and transferred to a N2-
filled glove bag. Solutions were prepared by mixing 5 mL of MES with varying amounts
of NaOH solution and water in order to achieve the desired molar ratios of NaOH to
MES. The total volume in each case was 10 mL. One milliliter of degassed catalyst
solution was then mixed with the polymerization solution, and the initiator-coated films
were immersed in this solution for 2 h. The polymer-coated wafers were removed from
the glove bag, immediately sonicated in DMF for 10 min, rinsed sequentially with
ethanol and water, and dried under a stream of N2.

5.2.3. Derivatization of carboxylic acid groups and protein immobilization

The carboxylic acid groups of poly(MES) were activated by immersing the films
in a solution containing NHS (0.1 M) and EDC (0.1 M) in water for 30 min, and the films
were rinsed sequentially with water and ethanol (20 mL each) and dried with N2. The
NHS-modified films were immersed for 1 h in an aqueous solution of aminobutyl NTA
(0.1 M, pH 10.2), rinsed with 20 mL of water and dried with N2. Finally, the NTA-Cu2+
complex was formed by immersing the coated wafer in 50 mM CuSO4 for 2 h, rinsing the
substrate sequentially with water and ethanol (20 mL each) and drying with N2. To

immobilize BSA, the poly(MES)-NTA-Cu2+ films were immersed in a solution of 1.0

128

 

‘1... l

,

.v‘ ' ~

'I :53"-
1

mg/mL BSA in phosphate buffer (20 mM, pH 7.2) for 18 h. The films were then rinsed
with 20 mL washing buffer (phosphate buffer containing 150 mM NaCl and 0.1%
Tween-20, pH 7.2) followed by 20 mL of phosphate buffer and 20 mL ethanol. Films
were dried under a stream of N2.

For lysozyme-binding studies, poly(MES) wafers were immersed in a 1 mg/mL
solution of lysozyme in phosphate buffer for 3 h. The films were then rinsed with 20 mL
of washing buffer followed by 20 mL of phosphate buffer and 20 mL of ethanol. Films
were dried under a stream of N2. Before characterizing the poly(MES)-NTA-Cu2+ and
poly(MES)-NTA-Cu2+-protein films by reflectance FTIR spectroscopy and ellipsometry,
the films were immersed in phosphate buffer for 15 min followed by rinsing with 20 mL
ethanol and drying under a stream of N2
5.2.4. Quantification of protein binding

To quantify the amount of protein bound to poly(MES) brushes on Au—coated Si
wafers, the method mentioned in Chapter 2 was employed. Briefly, a calibration curve
was obtained by plotting the ellipsometric thickness of spin-coated BSA or lysozyme
films against the reflectance FTIR absorbance of their amide I band (for more details
refer to page 60). The amide absorbance of lysozyme or BSA adsorbed to poly(MES) or
poly(MES)-NTA-Cu2+ films was then compared to the calibration curve to obtain the
thickness added due to protein adsorption. These results were confirmed by ellipsometric
studies.

5.2.5. Kinetics of solution polymerization of MES, MAA and HEMA
Solution polymerizations of MES and MAA were performed using 4-

(bromomethyl)benzoic acid as the initiator.30 For these studies, neat monomer (with or

129

F3

"fill... -l -5-

Al‘—

 

without inhibitor), 1 M aqueous NaOH, and catalyst solution in DMF were degassed
separately using three freeze-pump-thaw cycles and transferred to a N2-filled glove bag.
Initiator (20 mM) was dissolved in 1 mL of 1 M NaOH. In another vial, neat monomer
(5 mL) was mixed with 4 mL of 1 M NaOH, and 1 mL of degassed catalyst solution
containing CuBr (2 mM), CuBr2 (1 mM), and HMTETA (6 mM) in DMF was added.
The initiator and monomer solutions were mixed together, and a 0.5-mL aliquot of this
mixture was transferred to an NMR tube containing 50 pL of D20. (Before adding the
polymerization solution, the NMR tube was purged with N2 for 30 min and then kept in a
N2-filled glove bag for at least 3 h to remove oxygen). The time difference between the
start of polymerization and the start of NMR data acquisition was 8 min.
5.2.6. Instrumentation

1H NMR Spectra were collected on a Varian Inova-300 spectrometer, and
reflectance FTIR spectra of films on Au-coated wafers were obtained with a Nicolet
Magna-IR 560 spectrophotometer containing a PIKE grazing angle (80°) accessory. A
UV/ozone-cleaned Au—coated wafer was used to obtain a background spectrum. Film
thicknesses were determined using a rotating analyzer ellipsometer (model M-44; J .A.
Woollam) at an incident angle of 75°, assuming a film refractive index of 1.5.
Ellipsometric measurements were performed on at least three spots on a film.
5.3. Results and Discussion
5.3.1. Kinetics of surface-initiated MES polymerization

Scheme 5.2. outlines the synthesis of poly(MES) brushes on Au-coated Si wafers

and the derivatization of these brushes with metal-ion complexes that bind proteins.

130

 

8 9H3
StCHfirO-C-C—Br
11

 

 

CH3
c B /c e / (1H3
U r U r2 :
ligand/DMF H20 <9
020
H O/N OH O: 9 l?
2 RT8 CH2-CH2-O-C-CH2-CH2-C-OH
v 2-(Methacryloyloxy)ethyl succinate (MES)
H3
CH2-eta
C:
’ ° 1? it
\CH2-CH2-O-C-CH2~CH2-C-OH
1) 0H
0 N 0
U ,EDC, H20, 30 min, RT
1’2)aminobutyl NTA, H20, pH 10.2, 1 11, RT
,0'
9H3 Nb \0.
CH2-(Erna 90-
/ 2O
0. i? S? 0
CHz‘CHz’O‘C-CHz'CHz—C’NH
3) Cu”, 2h, RT 0
4) BSA, 18 h, RT ‘EL ,rNH
Otc_,0 (1)/N /
J j,:<:.ju~N \ ®
9H3 N ,6 kNH
CH2-CifiBr \\
cs 0
o’ O 9 0

‘CHZCHz-o-c-CHz-CHz-b—NH

Scheme 5.2. Synthesis of protein-binding poly(MES) brushes on Au surfaces.

Growth and modification of poly(MES) brushes on Au- coated Si wafers were
characterized with reflectance FTIR spectroscopy using a clean wafer as background

(Figure 5.1.). The peak at ~1740 cm'1 in spectrum (a) is assigned to the ester carbonyl

131

 

 

.,. ._ _
.
kc- ...-

" h—& I‘-

 

 

 

Absorbance

m (is

 

 

 

I T I T

1 900 1700 1 500 1300

Wavenumbers (cm-1)

Figure 5.1. Reflectance F TIR spectra of a poly(MES) brush on a Au-coated Si wafer
(a) before and after the following sequential steps: (b) activation with EDC/NHS; (c)

reaction with aminobutyl NTA.

groups of poly (MES). Exposing the film to a 0.1 M mixture of EDC and NHS in water
converted —COOH groups to succinimidyl esters. Formation of the succinimide ester
resulted in new peaks at 1817 and 1786 cm'1 (Figure 5.1., spectrum b), and an increase
in the absorbance around 1753 cm". Subsequently, the EDC/NHS-activated poly(MES)
was allowed to react with aminobutyl-NTA. This reaction resulted in a loss of the
absorbances due to the active ester, and a shifting of the peak at 1753 cm'1 back to 1740
cm‘1 (Figure 5.1., spectrum c). The new absorbance at 1680 cm“1 suggests NTA
immobilization, as it likely results from a combination of absorbance due to the
carboxylate groups of NTA and amide bonds formed between poly(MES) and NTA. The
broad peak around 1600 cm'1 could be due to carboxylate groups from either NTA or
hydrolyzed active esters. Alter immersion of NTA-derivatized poly(MES) films in 0.1 M

CuSO4, there was no dramatic change in the IR spectrum.

132

. = _ v-
u‘ u‘
w.

 

 

 

We examined the kinetics of room-temperature, surface-initiated MES
polymerization using several catalyst systems (Figure 5.2.). Usually, ATRP maintains a
low concentration of active radicals to provide control over molecular mass and
polydispersities, and the rate of ATRP is low. However, the use of HMTETA or
Me4Cyclam/anbpy as ligands for the Cu catalyst systems yields unusually rapid film

growth and high film thicknesses. The decline in film growth rate with time for these

 

 

 

200
T
T
3160 ;t I
A ...:‘h
311201,! 1‘ 4
'3 A
a 'P
3 80 z“ I.
a ‘A l
0
E 40E 1
l- T
lei-.I I I
0": .

 

0 20 40 6'0 80 100120140
Polymerization time (min)
Figure 5.2. Evolution of ellipsometric thickness with time for surface-initiated
polymerization of MES using HMTETA (diamonds); Me4Cyclam)/anbpy
(triangles); Me6TREN (circles) and bpy (squares) catalyst systems. The room-
temperature, aqueous polymerizations occurred on Au-coated Si, and each point
represents a different film. The diamonds and circles show the average of 3

independent runs, and the error bars correspond to the standard deviation.

133

 

 

systems suggests a relatively high radical concentration that leads to rapid polymerization
as well as some termination. Compared to HMTETA and Me4Cyclam/anbpy catalyst
systems, polymerizations using Me6TREN as the Cu ligand were more controlled, as
evidenced by a nearly linear increase in thickness with time for the first hour of
polymerization. Even with the MebTREN catalyst system, however, the brush thickness
was >120 nm after 90 min of polymerization. In contrast, polymerizations using bpy as
the catalyst ligand yielded <25 nm-thick poly(MES) films. As proposed by
Matyjaszewski,“ multidentate ligands like Me4Cyclam, Me6TREN and HMTETA may
complex the cupric species more efficiently than bpy, shifting the equilibrium towards
the Cu(II) species and providing a higher radical concentration and faster polymerization
than bpy-Cu catalysis. Previous studies demonstrated that Me4Cyclam, Me6TREN and
HMTETA Cu complexes are highly active catalysts for solution ATRP at ambient
temperatrrre,52’53 however MeéTREN provides better control over polymerizations due to
a higher deactivation rate.5 ' ’53’54
5.3.2. Comparison of ATRP of MES, MAA and AA

To determine whether the high rates of film growth are partly due to the relatively
high reactivity of MES, we grew films of poly(AA) and poly(MAA) using the
CuBr/CuBr2/HMTETA catalyst system. After 2 h of polymerization, poly(AA) and
poly(MAA) brush thicknesses were <15 nm and 60 nm, respectively, while the
poly(MES) thickness was >170 nm. Notably the rates of polymerization of MES and
HEMA are similar (Figure 5.3.), suggesting that the methacrylates have an inherently

faster polymerization rate than methacrylic and acrylic acid. Unfortunately, methyl

134

 

 

 

methacrylate and ethyl methacrylate are not soluble in water, so we could not investigate

the polymerization of these monomers under similar conditions.

 

 

 

 

 

200
s160" ..
__. <>
“- A
3 ’ o A
£120« ‘23:;
° 00
3 on
0 801A
2A
.5 A. 000 ’6‘ k.)
u to ..
E 40* e
l- (90
‘0.
Ole)... - -

 

 

0 20 40 60 80 100120140
Polymerization time (min)

Figure 5.3. Evolution of ellipsometric film thickness with time for surface-initiated
polymerization of (a) MES inhibited with 750 ppm of MEHQ (filled diamonds) (b)
MES without inhibitor (hollow diamonds) (c) HEMA without inhibitor (hollow
triangles) (d) MAA inhibited with 250 ppm of MEHQ (filled circles) and (e) MAA
without inhibitor (hollow circles). Polymerizations were performed at room
temperature on initiator-modified Au substrates in aqueous solutions using
CuBr/CuBr2/I-IMTETA as a catalyst. Each point represents a different film.
Polymerization of MES or MAA occurred in 10 mL of a mixture of neat monomer and
l M aqueous NaOH (1:1, v/v) whereas for HEMA, polymerization took place in 1:1,

v/v monomer and deionized water.

135

We should note that during most polymerizations of MES and MAA, the inhibitor
(MEHQ) was still present. As shown in Figure 5.3., the initial rate of MAA
polymerization increased in the absence of inhibitor, but film growth stopped after 30
min so the overall MAA film thickness (<60 nm) was the same as that obtained in the
presence of inhibitor. MES, on the other hand, gave films that were ~15% thicker when
polymerized in the presence of inhibitor. Lower initial radical concentrations in the
presence of inhibitor could lead to less termination and thicker films with MES. In any
case, in the presence or absence of inhibitor, poly(MES) films are much thicker than
poly(MAA) films.

To test whether the observed kinetic behavior during film growth is specific to
surface-initiated polymerization, we examined solution-phase ATRP of MES and MAA
under similar conditions. Figure 5.4. shows that both in the presence and absence of
inhibitor, MES polymerizes much more rapidly than MAA. Polymerization of inhibitor-
free MAA resulted in only 4.2% conversion to poly(MAA) (Figure 5.4., hollow circles),
and no detectable MAA polymerization occurred in the presence of inhibitor (Figure
5.4., filled circles). In contrast, polymerization of inhibitor-free MES reached >99.5%
conversion in about 7 h. Thus, the faster grth of poly(MES) than poly(MAA) from
surfaces is likely a direct result of the different reactivities of the two monomers.

Interestingly, unlike polymerization from a surface, the solution phase
polymerization of MES is faster without inhibitor. This difference between solution and
surface polymerization of MES presumably occurs because there are many more

initiators in solution than on the substrate, so termination by radical recombination is less

136

 

 

 

important in solution. Additionally, radicals are likely in closer proximity on the surface
than in solution.26
Figure 5.5. and Figure 5.6. show the NMR spectra of the polymerization

solutions for MES and MAA polymerization respectively.

 

100 W

00
o
000
°o
is

O)
O
o
000
°<>
O
O
'5

 

Percent monomer conversion

 

 

3“! "“.‘.

0 100 200 300 400 500

Polymerization time (min)

Figure 5.4. Percent monomer conversion as a function of time for solution
polymerization of MES without inhibitor (hollow diamonds), MES inhibited with 750
ppm of MEHQ (filled diamonds), MAA without inhibitor (hollow circles), and MAA
inhibited with 250 ppm of MEHQ (filled circles). Polymerizations were studied using

NMR at room temperature.

137

 

~ I At time = 8 min
a‘ H ‘00 HOD

CH2 e c,d

 

 

 

 

 

H. ,0“: ll At time = 8 11 a' a

ll ‘1'
/ It \

O (I) /' \ l \\
[CH2 . / \V/ \d
\O HOD b/

0‘0, 6.0 5.9 5.8 5.7 5.6 5.5 5.4 5.3 5.2

 

 

 

 

 

 

Figure 5.5. lH NMR spectra of an MES polymerization solution 8 min (1, top) and 8
h (11, bottom) after the addition of initiator. The polymerization solution contained
neat MES (without inhibitor), 1 M NaOH in H20, and CuBr, CuBr2, and HMTETA in
DMF. The sodium salt of u-bromo-p-toluic acid was used as the initiator. 0.5 mL of
polymerization solution and 50 pL D20 were added to the NMR tube. The inset
contains an expanded view of the region of the spectrum containing peaks due to the

alkene protons.

138

 

 

    

 

 

 

 

 

 

 

 

 

[I] At time = 8 min (I) HOD
a‘ a
a H‘ [CH3 1’ HOD ii A
p=c\ ,
a‘ H 70:0 5.7 5.5 5.3 5.1 4.9 4.7 4.5
HO
a‘ a l b
.L_
8 0 7 0 6 0 5.0 4 0 3 0 2 0 1 0
HOD
[II] At time = 8 h
a‘ a b
. 1 ill ll
' I l I I I l I
8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0
PM“

Figure 5.6. 1H NMR spectra of an MAA polymerization solution 8 min (1, top),
and 8 h (11, bottom) after the addition of initiator. The polymerization solution
contained neat MAA (without inhibitor), 1 M NaOH in H20, and CuBr, CuBr2,
and HMTETA in DMF. The sodium salt of a-bromo-p-toluic acid was used as the
initiator. 0.5 mL of polymerization solution and 50 pL D20 were added to the
NMR tube. The inset contains an expanded view of the region of the spectrum

containing peaks due to the alkene protons.

139

 

5.3.3. Effect of NaOH on MES polymerization

We also examined film grth as a function of the amount of NaOH added to the
polymerization solution. Figure 5.7. shows that film thicknesses are highest when most
of the MES in solution is in the protonated form (little NaOH is added), but even with a

1:1 ratio of NaOH to MES, films with thicknesses >100 nm can form. Thus, both 1'"

 

300
250 «P

200 . o

 

1.‘ -

150 1

100 .

501

Thickness of poly(MES) (nm)
0

 

 

 

o . . . - .
0 0.2 0.4 0.6 0.8 1.0 1.2

[NaOH] I [MES]
Figure 5.7. Thickness of poly(MES) brushes as a function of the molar ratio of NaOH
to MES added to the polymerization solution. The brushes were grown using a 2 h,

room temperature polymerization with a CuBr/CuBr2/HMTETA catalyst system.

protonated and deprotonated MES can polymerize, but the protonated form polymerizes
faster. This is in agreement with some previous reports that showed that the rate of
polymerization of acidic monomers decreases with increasing pH.55'56 The decrease in
rate at high pH presumably occurs because the electrostatic repulsion between growing

chains and monomers decreases the rate at which monomers can reach confined radicals.

l. 40

At NaOH to MES ratios <0.08, the polymerization solution is turbid, and when no
NaOH is added, the solution consists of two phases, the denser of which contains MES
and most of the catalyst, as discerned by its blue color. Even two-phase polymerizations,
when no NaOH is added to the reaction mixture, give film thicknesses similar to those
obtained when adding small amounts of NaOH to the solution. (The substrate sits in the
MES-rich phase during polymerization.) However, some water in the MES phase is
necessary for rapid brush growth,57 as polymerization from a solution containing only
monomer and catalyst gives film thicknesses of 25 nm after 2 h of polymerization.
5.3.4. Protein binding to poly(MES) brushes and their derivatives

To demonstrate the utility of poly(MES). brushes, we examined their ability to
bind proteins. Initially, we immersed poly(MES) films on Au-coated Si in 1 mg/mL
solutions of lysozyme in 20 mM phosphate buffer (pH 7.2) for 3 h. After removal of the
film from solution and rinsing, reflectance F TIR spectroscopy allowed determination of
the amount of bound lysozyme using a procedure we developed previously4 (for more
details refer to page no. 60). Briefly, we prepared a linear calibration curve of
ellipsometric thickness versus reflectance FTIR amide absorbance (1670 cm") for spin-
coated lysozyme films on Au—coated Si wafers. Using the calibration curve and the FTIR
spectra that reveal immobilized lysozyme (e.g., Figure 5.8.c), we calculated protein
binding capacity with the assumption that the lysozyme film has a density of 1 g/cm3.
Remarkably, a 55 nm poly(MES) film binds 14.4i0.3 pg lysozyme/cmz, which is
equivalent to ~70 monolayers of lysozyme in the brushes (assuming a monolayer
thickness of 2 nm).58 Ellipsometric measurements also show that the film thickness

increases from 55 nm to 205 nm after lysozyme adsorption. This binding capacity is

141

 

higher than the 38 monolayers of lysozyme reported to bind to sulfonated poly(glycidyl
methacrylate) coatings,36 and is comparable to the ~80 lysozyme monolayers found to

bind to poly(AA) brushes prepared by hydrolysis of poly(tert-butyl acrylate).4

 

 

 

 

0.1
8 c
C
W
'E
O
m
.D
< b
M
1 900 1 700 1 500 1 300

Wavenumbers (cm-1)

Figure 5.8. Reflectance FTIR spectra of a (a) poly(MES) film (b) poly (MES)-
lysozyme film and (c) immobilized lysozyme, which was obtained by subtracting (a)
from (b). The 55 nm thick poly(MES) film was immersed in 1.0 mg/mL lysozyme

solution for 3 h and then rinsed with buffers and ethanol.

142

 

 

\E.
' I’.‘ — .-_

 

 

Building on these promising results, we investigated the adsorption of BSA in
poly(MES) brushes modified with NTA-Cu2+ complexes (Scheme 5.2.). 2 This
adsorption occurs through a metal-affinity interaction between BSA and the Cu(II)
complex. We immersed poly(MES)-NTA-Cu2+ films overnight in a solution containing
1 mg/mL BSA and then thoroughly rinsed these films with buffers and solvent. As
seen in Figure 5.9., 55 nm and 85 nm poly(MES) films derivatized with NTA-Cu2+ had
BSA binding capacities of 6.8 and 7.2 pg/cmz, respectively, which is equivalent to 17
and 18 monolayers of BSA in poly(MES) brushes (assuming a monolayer thickness of
4 nm).59 These high binding capacities and the fact that BSA binding initially increases
with brush thickness suggest that binding occurs both at the film-solution interface and
inside the brushes. However steric hindrance to binding may result in the plateau in

adsorption capacity at thicknesses >60 nm (Figure 5.9.). The poly(MES) binding

 

Bound BSA (pg/cmz)
A
0

 

 

 

0 20 4O 60 80 100
Poly(MES) thickness (nm)
Figure 5.9. BSA binding capacity as a function of poly(MES) film thickness. The

poly(MES) films were derivatized with NTA-Cu2+ complexes, and the amount of

bound BSA was determined using reflectance FTIR spectroscopy.

143

 

capacity is similar to that reported previously for poly(AA), but the poly(MES) can be
synthesized in a rapid, one-step procedure.4
5.4. Conclusions

Surface-initiated aqueous ATRP enables rapid grth of poly(MES) brushes
under gentle conditions that should allow formation of films on a wide range of
substrates. Both on a surface and in solution, polymerization of MES occurs much faster
than polymerization of MAA, presumably because methacrylates are more reactive than
MAA or AA. HEMA, another water-soluble methacrylate, shows brush growth rates
similar to those of MES. Moreover, poly(MES) brushes and their derivatives are capable

of binding many monolayers of BSA as well as lysozyme.

144

 

 

5.5. References:

(1)
(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)
(12)

(13)

(14)

(15)

(16)

Singh, N.; Cui, X.; Boland, T.; Husson, S. M. Biomaterials 2007, 28, 763-771.

Ghosh, P.; Amirpour, M. L.; Lackowski, W. M.; Pishko, M. V.; Crooks, R. M.
Angewandte Chemie International Edition 1999, 38, 1592-1595.

Cullen, S. P.; Liu, X.; Mandel, I. C.; Himpsel, F. J.; Gopalan, P. Langmuir 2008,
24, 913-920.

Dai, J.; Bao, Z.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L. Langmuir
2006, 22, 4274-4281.

Tsarevsky, N. V.; Braunecker, W. A.; Vacca, A.; Gans, P.; Matyjaszewski, K.
Macromolecular Symposium 2007, 248, 60-70.

Bfitiin, V.; Lowe, A. B.; Billingham, N. C.; Armes, S. P. Journal of the American
Chemical Society 1999, 121, 4288-4289.

Zhang, 0.; Remsen, E. E.; Wooley, K. L. Journal of the American Chemical
Society 2000, 122, 3642-3651.

Xu, J .; Bhattacharyya, D. Industrial & Engineering Chemistry Research 2007, 46,
2348-2359.

Balastre, M.; Li, F.; Schorr, P.; Yang, J. C.; Mays, J. W.; Tirrell, M. V.
Macromolecules 2002, 35, 9480-9486.

Mykytiuk, J.; Armes, S. P.; Billingham, N. C. Polymer Bulletin 1992, 29, 139-
145.

Zhang, H.; Riihe, J. Macromolecular Rapid Communications 2003, 24, 576-5 79.

Haddleton, D. M.; Crossman, M. C.; Dana, B. H.; Duncalf, D. J.; Heming, A. M.;
Kukulj, D.; Shooter, A. J. Macromolecules 1999, 32, 2110-2119.

Rannard, S. P.; Billingham, N. C.; Armes, S. P.; Mykytiuk, J. European Polymer
Journal 1993, 29, 407-414.

Bao, Z.; Bruening, M. L.; Baker, G. L. Journal of the American Chemical Society
2006, 128, 9056-9060.

Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immarapom, B.; Gelman, A.;
Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffrnann, H.;
Pakula, T. Macromolecules 1999, 32, 8716-8724.

Boyes, S. G.; Akgun, B.; Brittain, W. J.; Foster, M. D. Macromolecules 2003, 36,
9539-9548.

145

 

(17)

(18)
(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)
(31)

(32)

(33)

Treat, N. D.; Ayres, N.; Boyes, S. G.; Brittain, W. J. Macromolecules 2006, 39,
26-29.

Biesalski, M.; Riihe, J. Macromolecules 2003, 36, 1222-1227.

Hu, S.; Ren, X.; Bachman, M.; Sims, C. E.; Li, G. P.; Allbritton, N. L. Analytical
Chemistry 2004, 76, 1865-1870.

Ulbricht, M.; Yang, H. Chemistry ofMaterials 2005, 17, 2622-2631.

Rahane, S. B.; Floyd, J. A.; Metters, A. T.; Kilbey, S. M., 11 Advanced Functional
Materials 2008, 18, 1232-1240.

Matyjaszewski, K. Overview: fundamentals of controlled living radical
polymerization, ACS Symposium Series 1998, 685, 2-30.

Matyjaszewski, K. Comparison and classification of controlled/living radical
polymerizations, ACS Symposium Series 2000, 768, 2-26.

Chung, I.-D.; Britt, P.; Xie, D.; Harth, B.; Mays, J. Chemical Communications
2005, 1046-1048.

Tsarevsky, N. V.; Pintauer, T.; Matyjaszewski, K. Macromolecules 2004, 3 7,
9768-9778.

Kim, J.-B.; Huang, W.; Miller, M. D.; Baker, G. L.; Bruening, M. L. Journal of
Polymer Science Part A: Polymer Chemistry 2003, 41 , 386-394.

Keoshkerian, B.; Georges, M. K.; Boils-Boissier, D. Macromolecules 1995, 28,
6381-6382.

Vo, C.-D.; Iddon, P. D.; Armes, S. P. Polymer 2007, 48, 1193-1202.

Ashford, E. J.; Naldi, V.; O'Dell, R.; Billingham, N. C.; Armes, S. P. Chemical
Communications 1999, 1285- l 286.

Wang, X. S.; Jackson, R. A.; Armes, S. P. Macromolecules 2000, 33, 255-257.

Singh, N.; Wang, J .; Ulbricht, M.; Wickramasinghe, S. R.; Husson, S. M. Journal
of Membrane Science 2008, 309, 64-72.

Sankhe, A. Y.; Husson, S. M.; Kilbey, S. M., II Macromolecules 2006, 39, 1376-
1383.

Sankhe, A. Y.; Husson, S. M.; Kilbey, S. M., 11 Journal of Polymer Science Part
A: Polymer Chemistry 2007, 45, 566-575.

146

 

 

 

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)
(42)

(43)

(44)

(45)

(46)

(47)

(48)
(49)

(50)

(51)

Dong, R.; Krishnan, S.; Baird, B. A.; Lindau, M.; Ober, C. K. Biomacromolecules
2007, 8, 3082-3092.

Tugulu, S.; Barbey, R.; Harms, M.; Fricke, M.; Volkmer, D.; Rossi, A.; Klok, H.-
A. Macromolecules 2007, 40, 168-177.

Kawai, T.; Saito, K.; Lee, W. Journal of Chromatography 8: Analytical
Technologies in the Biomedical and Life Sciences 2003, 790, 131-142.

Singh, N.; Husson, S. M.; Zdyrko, B.; Luzinov, 1. Journal of Membrane Science
2005, 262, 81-90.

Kurosawa, S.; Aizawa, H.; Talib, Z. A.; Atthoff, B.; Hilborn, J. Biosensors and
Bioelectronics 2004, 20, 1165-1176.

Jain, P.; Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Biomacromolecules
2007, 8, 3102-3107.

Roper, D. K.; Lightfoot, E. N. Journal of Chromatography A 1995, 702, 3-26.
Lee, W. C.; Lee, K. H. Analytical Biochemistry 2004, 324, 1-10.

Arenkov, P.; Kukhtin, A.; Gemmell, A.; Voloshchuk, S.; Chupeeva, V.;
Mirzabekov, A. Analytical Biochemistry 2000, 278, 123-131.

MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760-1763.

Phizicky, E.; Bastiaens, P. I. H.; Zhu, H.; Snyder, M.; Fields, S. Nature 2003, 422,
208-215.

Zhu, H.; Bilgin, M.; Bangham, R.; Hall, D.; Casamayor, A.; Bertone, P.; Lan, N.;
Jansen, R.; Bidlingrnaier, S.; Houfek, T.; Mitchell, T.; Miller, P.; Dean, R. A.;
Gerstein, M.; Snyder, M. Science 2001, 293, 2101-2105.

Dai, J. H.; Baker, G. L.; Bruening, M. L. Analytical Chemistry 2006, 78, 135-140.

Shah, R. R.; Merreceyes, D.; Husemann, M.; Recs, 1.; Abbott, N. L.; Hawker, C.
J .; Hedrick, J. L. Macromolecules 2000, 33, 597-605.

Ciampoli.M; Nardi, N. Inorganic Chemistry 1966, 5, 1150.

Huang, W. X.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,
35, 1175-1179.

Zheng, Y.; Michigan State University, 2007; Vol. Ph.D. Dissertation.

Matyjaszewski, K. Current Organic Chemistry 2002, 6, 67-82.

147

 

 

(52)

(53)

(54)

(55)
(56)

(57)

(58)

(59)

Pintauer, T.; Matyjaszewski, K. Coordination Chemistry Reviews 2005, 249,
1155-1184.

Tsarevsky, N. V.; Matyjaszewski, K. Journal of Polymer Science Part A: Polymer
Chemistry 2006, 44, 5098-5112.

Xia, J. H.; Gaynor, S. G.; Matyjaszewski, K. Macromolecules 1998, 31, 5958-
5959.

Niwa, M.; Date, M.; Higashi, N. Macromolecules 1996, 29, 3681-3685.
Matyjaszewski, K.; Xia, J. H. Chemical Reviews 2001, 101, 2921-2990.

Wang, X.-S.; Lascelles, S. F.; Jackson, R. A.; Armes, S. P. Chemical
Communications 1999, 1 817-1 818.

Tsuneda, S.; Shinano, H.; Saito, K.; Furusaki, S.; Sugo, T. Biotechnology
Progress 1994, 10, 76-81.

Carter, D. C.; He, X. M.; Munson, S. H.; Twigg, P. D.; Gemert, K. M.; Broom,
M. B.; Miller, T. Y. Science 1989, 244, 1195-1198.

148

’W. .1
.a .

...—...n. _ 3.113

ll-y

 

J
'U.

Chapter 6
His-tagged protein purification with high capacity affinity

membranes containing functionalized poly(MES) brushes

6.]. Introduction

Chapter 3 demonstrates the use of polymer brush-modified nylon membranes for
purification of His-tagged proteins from cell extracts. Unfortunately, nylon membranes
modified with poly(HEMA)-NTA-Ni2+ brushes have a relatively low protein binding

capacity (25 mg protein/cm3

of membrane), perhaps because the organic solvents
employed in the brush synthesis and derivatization partially damage the membrane
structure. To avoid the use of organic solvents, we created a completely aqueous
procedure for growth of polymer brushes inside polymeric membranes (chapter 4).
However, despite the successful polymerization of HEMA with aqueous initiation from
surfaces, the creation of protein-binding brushes requires conversion of hydroxyl groups
in poly(HEMA) to carboxylic acid moieties, and this involves reaction of poly(HEMA)
with succinic anhydride in an organic solvent that may damage the membrane. To
overcome this problem, we developed the surface-initiated aqueous ATRP of an acidic
monomer, MES, as a rapid, one-step route to polyacid brushes (see chapter 5 for details).
This procedure avoids the need to react the brush with succinic anhydride in an organic
solvent. Also, poly(MES) brushes and their derivatives exhibit high protein-binding
capacities. Thus, modification of polymeric membranes with poly(MES) should allow

for high-capacity purification of His-tagged proteins directly from a cell extract. This

chapter describes the growth, derivatization, and characterization of poly(MES) brushes

149

 

inside porous nylon membranes, and the use of these modified membranes as protein
absorbers. Protein breakthrough curves show remarkable HisU, BSA and lysozyme
binding capacities of 85, 80 a; 2 and 118 :t 8 mg protein per cm3 of membrane,
respectively. (The lysozyme binds to poly(MES) via ion-exchange interactions, whereas
BSA binds to poly(MES)-NTA-Cu2+ via affinity interactions). Most importantly, the
poly(MES)-NTA-Ni2+-modified membranes allow isolation of His-tagged CRALBP
directly fiom a cell extract.
6.2. Experimental section
6.2.1. Materials

Hydroxylated (LoProdyne® LP) nylon membranes with 1.2 and 5.0 pm-diameter
surface pores were obtained from Pall Corporation; nylon microfiltration membranes
(non-hydroxylated, 1.2 pm) were received from GE Water & Process Technologies; and
regenerated cellulose membranes (RC 60, 1.0 pm) were purchased from Whatrnan. All
membranes were cut into 25 mm diameter discs prior to modification or use. 2-
Bromoisobutyryl bromide (98%), CuBr (99.999%), CuBr2 (99%), MES (Aldrich,
inhibited with 750 ppm MEHQ), EDC, NHS, EDTA, imidazole (99%), TWEEN-20
surfactant, lysozyme, and BSA were used as received from Sigma Aldrich. MeoTREN
(ATRP Solutions), CuSO4-5H2O (CCI), NiSO4-5H20 (Columbus Chemical), NaH2PO4
(CCI), Na2HPO4 (Aldrich), aminobutyl NTA (F luka), NaOH (Spectrum), and Coomassie
protein assay reagent (Pierce) were also used without purification. Tetrahydrofuran
(THF, Jade Scientific Inc., anhydrous, 99%) was distilled and stored over molecular
sieves. Trichlorosilane initiator (l l -(2-bromo-2-

methyl)propionyloxy)undecyltn'chlorosilane) was synthesized according to a literature

150

 

procedure.l Buffers were prepared using analytical grade chemicals and deionized
(Milli-Q, 18.2 MQ cm) water.
6.2.2. Initiator attachment in porous polymer membranes
6.2.2.-a. Initiator attachment in hydroxylated (LoProdyne® LP) nylon membranes
A LoProdyne® LP nylon membrane was cleaned with UV/ozone (Boekel model
135500) for 10 min, and placed inside a home-built Teflon cell. Initiator attachment
occurred by circulating 1 mM trichlorosilane initiator in 20 mL of anhydrous THF
through the membrane for 2 h at a flow rate of 3 mL/min, followed by subsequent rinsing
with 20 mL of ethanol at the same flow rate.
6.2.2.-b. Initiator attachment in non-hydroxylated nylon membranes
Prior to modification with initiator, the surface amide groups in the non-
hydroxylated nylon membranes (GE Water & Process Technologies) were activated
according to a procedure described by Xu and coworkers.2 Briefly, 30 membranes were
immersed for 12 h in a 60 °C solution containing 50 mL formaldehyde and 1 mL of 85%
(w/v) phosphoric acid. This resulted in conversion of the nylon membrane surface to N-
methylol polyamide (or nylon-OH) (Scheme 6.1.(a)). After the reaction, the activated
membranes were washed with copious amounts of water and dried overnight under
vacuum. Initiator attachment in these membranes occurred as described above by
circulating 1 mM trichlorosilane initiator in 20 mL of anhydrous THF through the

membrane for 2 h and rinsing with 20 mL of ethanol (Scheme 6.1.(b)).

151

 

 

....
(a) l HCHO

:N—CH2—OH

 

Cl
l O
(b) ,Si‘(CH2)17 Br
Cl \
Cl 0

Trichlorosilane initiator

 

\

 
 

' N-CH —o 9 ll (EH3
2 ‘si—(—CH2—)o—c—c—8r
‘ ' O\ 11 CH3
MES, NaOH
(c) CuBr, CuBr2, ligand
RT, 1 h
i-N-CH —o\c1) (ll 9H3 (1H3
2 \ .
. s.—(—CH2—)o——c—cl:——[—CH2—c|:—]—;—Br
. o\ 11 CH3 6
/
0/ \O—CH2—CH2—O—ICl-CH2—CH2—-|Cl-OH

Scheme 6.1. Schematic illustration of (9) activation of the amide groups in a nylon
membrane to introduce hydroxyl functionalities (b) trichlorosilane initiator
immobilization on the hydroxylated membrane and (c) polymerization of MES from

initiator-modified membrane.

152

6.2.3. Polymerization of MES in porous nylon membranes

To prepare polymerization solutions, 10 mL of a mixture of neat MES monomer
and l M aqueous NaOH (1:1, v/v) was first degassed with three freeze-pump-thaw
cycles. A 1 mL solution of DMF containing CuBr (2mM), CuBr2 (1 mM), and
Me6TREN (6 mM)3'5 was similarly degassed, and in a N2-filled glove bag, this solution
of catalyst was mixed with the monomer/NaOH solution. Polymerization of MES
occurred in the glove bag by circulating this solution through the initiator-modified
membrane at a flow rate of 0.9 mL/min (Scheme 6.1.(c)). Unless mentioned otherwise,
the polymerization time was I h. After polymerization, the membrane was cleaned by
passing ethanol (20 mL) followed by 20 mL of deionized water (Milli-Q, 18.2 MQ cm,
20 mL) through the membrane at a flow rate of 1-2 mL/min. (Acetone should not be
used with nylon membrane as it partially damages the membrane structure).
6.2.4. Poly(MES) derivatization and protein binding

Chapter 5 describes the derivatization procedure, which is shown in Scheme 5.2.
Briefly, the carboxylic acid groups of poly(MES) were activated by circulating an
aqueous solution containing NHS (0.1 M) and EDC (0.1 M) through a poly(MES)-
modified nylon membrane for 1 h. This was followed by rinsing sequentially with 20 mL
of deionized water and 20‘ mL of ethanol through the membrane. An aqueous solution of
aminobutyl NTA (0.1 M, pH 10.2) was then flowed through the NHS-modified
membranes for l h, and the membrane was subsequently rinsed with 20 mL of water.
Finally, the NTA-Cu2+ (or Ni2+) complex was formed by circulating aqueous 0.1 M
CuSO4 (or NiSO4) through the membrane for 2 h followed by rinsing with water followed

by ethanol (20 mL each). The membrane was dried with N2 prior to protein binding.

153

 

To study lysozyme binding, a solution of lysozyme (1 mg/mL) in 20 mM
phosphate buffer (pH 7.2) was pumped through the poly(MES)-modified membrane
using a peristaltic pump (flow rate ~ 1 mL/min), and the permeate was collected for
analysis at specific time intervals. Subsequently, the membrane was rinsed with 20 mL
of pH 7.2 washing buffer 1 (20 mM phosphate buffer containing 0.1% Tween-20
surfactant and 0.15 M NaCl) followed by 20 mL of phosphate buffer. The protein was
then eluted using 5-10 mL of 20 mM phosphate buffer (pH 7.2) containing 1 M
potassium thiocyanate.

For BSA binding, a membrane modified with poly(MES)-NTA-Cu2+ was used. A
solution of 1 mg/mL BSA in 20 mM phosphate buffer (pH 7.2) was pumped through the
poly(MES)-NTA-Cu2+-modified membrane using a peristaltic pump at various flow rates
(section 6.3.2.-c.), and the permeate was collected for analysis at specific time intervals.
Subsequently, the membrane was rinsed with 20 mL of pH 7.2 washing buffer I followed
by 20 mL of phosphate buffer. The protein was then eluted using 5-10 mL of a solution
containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M imidazole at pH 7.2. Cu2+
was later eluted using a 50 mM EDTA solution (pH 7.2), and the poly(MES)-NTA film
was recharged with Cu2+ prior to reuse.

Prior to purification of His-tagged CRALBP (36 kD) that was over-expressed in
E. coli, the cells were lysed with sonication and centrifuged at 4 °C (Dr. James Geiger
kindly provided the cell extracts.) Supernatant (1.25 mL) was added to 3.75 mL of 20
mM, pH 7.2 phosphate buffer that contained 10 mM imidazole and 300 mM NaCl. This
solution was pumped through the poly(MES)-NTA-Ni2+ modified nylon membrane in an

amicon 8010 cell at a pressure less than 6.9x103 Pascal (l psig) at room temperature. The

154

.1 "n - ...

 

 

flow rate was 1.2 mL of extract/min. Subsequently, the membrane was rinsed with 20
mL of washing buffer I followed by 20 mL of washing buffer 11 (pH 7.2, 20 mM
phosphate buffer containing 45 mM imidazole and 0.15 M NaCl), and protein was eluted
using a pH 7.2 solution containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M
imidazole at pH 7.2.

In CRALBP purification with the spin-trap column, 1.25 mL of cell-free extract
containing over-expressed His-tagged CRALBP was added to 3.75 mL of 20 mM
phosphate buffer (pH 7.2) that contained 10 mM imidazole and 300 mM NaCl. The
solution was loaded onto the spin-trap column (in 500 pL fractions, 10 times) and the
column was centrifuged. This was followed by rinsing with 20 mL pH 7.2 washing
buffer I and with 20 mL pH 7.2 washing buffer 11. The protein was then eluted using a
600 pL solution containing 20 mM sodium phosphate, 0.5 M NaCl, and 0.5 M imidazole
at pH 7.2.

6.2.5. Characterization of brush growth and derivatization

Film growth on polymer membranes was verified using attenuated total
reflectance (ATR) FTIR spectroscopy (Perkin Elmer Spectrum One Instrument, air
background) as well as field-emission scanning electron microscopy (FESEM, Hitachi S-
470011 equipped with an EDAX Phoenix energy dispersive X-ray spectrometer system,

acceleration voltage of 15V).

155

' r.‘-"?‘v1 _':_‘fl._1

 

6.2.6. Protein quantification

To determine the amount of protein eluted from the modified membrane, 50 pL of
permeate was added to 2.95 mL of a solution of Coomassie reagent, and the mixture was
shaken a few times and allowed to react for 5 min at room temperature. The UV/vis
absorbance spectra of these solutions were then obtained with a Perkin-Elmer UV/V is
(model Lambda 40) spectrophotometer. A calibration curve for the absorbance of
lysozyme, BSA and HisU at 595 nm was prepared using a series of protein solutions
(concentration range of 100 pg to 1 mg of protein per mL) that were mixed with
Coomassie reagent in a 50 pL to 2.95 mL ratio. All spectra were measured against a
Coomassie reagent background.
6.2.7. Determination of protein purity by SDS-PAGE

The protein solutions were analyzed by SDS-PAGE with a 16% cross-linked
separating gel and a 4% cross-linked stacking gel (acrylamide). Protein bands were
visualized using standard silver staining" or coomassie blue staining7 procedures.
6.3. Results and discussion
6.3.1. Characterization of poly(MES)-derivatized membranes

To form brush-modified membranes, poly(MES) was grown from ATRP initiators
that were immobilized via silanization within the porous polymer membrane (Scheme
6.1.b.). The brushes were derivatized as shown in chapter 5, Scheme 5.2., with reactant
solutions being circulated through the membrane using a peristaltic pump. Figure 6.1.
shows the ATR-FTIR spectra of a nylon membrane (a) before and after (b)
polymerization of MES, (c) subsequent reaction with NHS/EDC, and (d) derivatization

with aminobutyl NTA. The IR-spectrum of the bare nylon membrane contains dominant

156

 

I‘M-"II

 

amide I and amide II peaks at 1630 and 1533 cm“, respectively (Figure 6.1., spectrum
(a)). The grth of poly(MES) is evident by the appearance of a small carbonyl peak at
1723 cm“1 (Figure 6.1., spectrum (b)). Passing a mixture of EDC and NHS in water
through the membrane converted -COOH groups to succinimidyl esters. Peaks due to
the succinimide ester appeared at 1810 and 1779 cm'1 (Figure 6.1., spectrum (e)).
Subsequently, the EDC/NHS-activated poly(MES) was allowed to react with aminobutyl-
NTA. This reaction resulted in a loss of the absorbance due to the succinimide ester,

(Figure 6.1., spectrum (d)).

 

 

0'05 amide I and II

I l

succinimide

Absorbance

 

 

 

 

 

1900 1800 1700 1600 1500 1400 1300
Wavenumbers (cm")

Figure 6.1. ATR-FTIR spectra of a hydroxyl functionalized nylon membrane before
(a) and after the following sequential steps: (b) formation of poly(MES) brushes
inside the membrane; (c) activation with EDC/NHS; ((1) reaction with aminobutyl

NTA.

157

Using a constant pressure of 6.9 x 104 Pascal (10 psig), we also monitored the
changes in pure water flux through the membrane before and after each derivatization
step. A bare nylon membrane with a 1.2 pm nominal filtration cut off shows a pure water
flux of 71 i 7 mL/cm2 min, but after modification with poly(MES), the flux drops to 14 i
3 mL/cm2 min presumably due to the resistance to the flow of water by highly swollen
poly(MES) brushes. Derivatization of poly(MES) with NHS gives an increase in water
flux to 56 :1: 2 mL/cm2 min, which is consistent with a decreased brush swelling after
formation of the succinimide ester. The hydrophilicity of the brushes increases after
immobilization of aminobutyl NTA, and the water flux decreases to 0.38 :t 0.22 mL/cm2
min. In addition to being hydrophilic, the swollen poly(MES)-NTA brushes have a
higher molecular mass than poly(MES), which may also lead to decline in permeability.
Finally after immobilization of Cu2+ or Ni“, the flux increases to 8.6 :t 3 mL/cm2 min
indicating that the metal ion immobilization decreases the swelling of polymer brushes.
Nevertheless, the flux through poly(MES)-NTA-Cu2+ or poly(MES)-NTA-Ni2+
membrane is only 12% of that through a bare nylon membrane, suggesting that the
modified polymer swells in water and occupies a significant fraction of the membrane
volume. If flow in these spongy membranes could be decreased by the Hagen-Poiseuille
law where the flow rate at a constant pressure is proportional to the pore radius to the
fourth power, the 88% drop in flux relative to a bare membrane would correspond to a

40% drop in pore radius.

158

 

:3

SEM images also corroborate the growth of poly(MES)-NTA-Cu2+ brushes in the
polymer membranes. The image of a pristine membrane (Figure 6.2.(a)) contains many
open pores, whereas modified pores (Figure 6.2.(b)) appear much less open, presumably

because they are covered with a polymer film.

 

Figure 6.2. SEM images of (a) a bare nylon membrane with a 1.2 pm nominal

filtration cutoff and (b) a similar membrane modified with a poly(MES)-NTA-Cu2+

film.

6.3.2. Protein binding to polymer brush-modified membranes
6.3.2.-a. Protein binding capacity in nylon membranes with two different pore sizes
We first studied lysozyme binding to poly(MES) brushes in nylon membranes
(LoProdyne® LP) with nominal pore sizes of 1.2 pm and 5.0 pm. In this case, binding to
the poly(MES) brushes occurs via ion-exchange interactions. Solutions containing 1
mg/mL of lysozyme in pH 7.2 phosphate buffer were pumped through the membrane,
and permeate aliquots were analyzed using a Bradford assay. Figure 6.3. shows the
breakthrough curves for lysozyme binding to poly(MES)-modified nylon membranes

with different pore sizes. The lysozyme binding capacities, as determined by integration

159

 

1.0 e 00.0050'0

O
0.9 0
O O
a; 0.8 .
B 0.7 . .
g 0.6 .
5 0.5
a: e
g 0.4
5 0.3 ~¢ 0
8
8 0.2
0.1 '
O ”O.

0 2 4 6 810121416
Permeate Volume (mL)

Figure 6.3. Breakthrough curves for absorption of 1 mg/mL lysozyme in poly(MES)-
modified nylon membranes with 1.2 pm (circles) and 5.0 pm (diamonds) nominal
filtration cutoffs. The permeate flow rates through the 1.2 pm and 5.0 pm membranes

were 1.0 mL/min and 1.1 mL/min respectively.

of the differences between the feed concentrations and the permeate concentrations in the
breakthrough curves, are 110 mg/cm3 and 45 mg/cm3 for nylon membranes with 1.2 pm
and 5 pm nominal cutoffs, respectively.

After obtaining the breakthrough curves for lysozyme, the poly(MES)-modified
membranes were washed with 20 mL 1 buffer followed by 20 mL phosphate buffer, and
the protein was eluted in 5-10 mL of 20 mM phosphate buffer (pH 7.2) containing 1 M
potassium thiocyanate. Analysis of the eluent (using a Bradford assay) showed a
lysozyme binding capacity of 118 i: 8 mg/cm3 and 51 d: 5 mg/cm3 for 1.2 pm and 5 pm

membranes respectively. These values agree well with capacities determined from

160

 

 

 

breakthrough curves and suggest that essentially all of the lysozyme was eluted from the
membrane.

To examine BSA binding, we derivatized the 1.2 pm and 5.0 pm nylon-
poly(MES) membranes with NTA-Cu2+ complexes as described in section 6.2.4. and
pumped a 1 mg/mL solution of BSA in pH 7 .2 phosphate buffer through the membranes.
The BSA presumably binds to the brushes via interaction between histidine residues in
the protein and the Cu2+ complexes in the polymer brush. Analyses of permeate aliquots

gave the breakthrough curves in Figure 6.4., and these curves imply BSA binding

1.0 e
0.9 " .
0.8 ,

0.7 0 °

0.6 .

0.5 ‘

0.4

0.3 .

0.2

0.1 -

Oe
012345678910

Permeate Volume (mL)

Concentration (mg/mL)

Figure 6.4. Breakthrough curve for absorption of BSA in nominal 1.2 pm (circles)
and 5.0 pm (diamonds) nylon membranes modified with poly(MES)-NTA-Cu2+

brushes. The flow rates of the 1 mg/mL BSA solution through the 1.2 pm and 5.0 pm

membranes were 1.1 mL/min and 0.98 mL/min, respectively.

161

a ' .'L.'.".\‘IG'.
.* ""11

 

13"

 

capacities of 65 mg/cm3 and 20 mg/cm3 for 1.2 pm and 5 pm nominal pore-sized nylon
membranes, respectively. After obtaining the breakthrough curves, the poly(MES)-NTA-
Cu2+-BSA membranes were washed with 20 mL washing buffer I followed by 20 mL
phosphate buffer, and the bound BSA was eluted with 5-10 mL of EDTA. Analyses of
the eluents using a Bradford assay showed BSA binding capacities of 80 :1: 2 mg/cm3 and
24 i 4 mg/cm3 for 1.2 pm and 5 pm membranes respectively. The 20% higher capacities
obtained with eluent analysis rather than breakthrough curves likely reflect the higher
accuracy inherent in analyzing single eluent solutions rather than multiple solutions for
breakthrough curves. The breakthrough curve analysis also involves of the difference
between two similar concentrations as the. membrane becomes saturated, which is
inherently imprecise.

Both the lysozyme and BSA binding capacities indicate that 5 pm membranes
have lower protein binding capacities than 1.2 pm membranes. This is not surprising
given the more open pore structure of a 5 pm membrane relative to a 1.2 pm membrane
(Figure 6.5.). The 5 pm membranes likely have a lower amount of polymer brush per
membrane volume and, hence, a lower binding capacity per membrane volume. Smaller
volume fi‘actions of brushes in the larger membrane pores are also consistent with the
decreases in pure water flux alter grth of the poly(MES). At constant pressure
(6.9x103 Pa), water flux through a 5.0 pm membrane modified with poly(MES) is ~35%
of the flux through a pristine membrane (flux decreased from 99 :1: 7 mL/cm2 min to 35 i
0.6 mL/cm2 min after modification with poly(MES)), whereas as discussed in section
6.3.1., water flux through a 1.2 pm membranes modified with poly(MES) is only 20% of

that through an unmodified membrane.

1. 62

 

 

   

Figure 6.5. SEM images of pristine nylon membranes with [(a), (c)] 1.2 pm nominal

filtration cutofl‘s, and [(b), (d)] 5.0 pm nominal filtration cutoffs.

163

6.3.2.-b. Protein binding capacity as a function of polymerization time

Using both breakthrough curves and eluent analysis, we also examined how the
lysozyme binding capacity varied with the time employed in MES polymerization in 1.2
pm nylon membranes. Figure 6.6. shows that the lysozyme binding capacity increases
with increasing MES polymerization times up to 1 h of polymerization and then decrease
with longer reaction times. The longer polymerization times should result in thicker
polymer brushes (see Figure 5.2. in chapter 5), so the initial increase in protein binding
with polymerization time most likely results fi‘om more protein binding sites in thicker

brushes.8 However increasing the polymerization time beyond 1 h leads to decreased

 

140
g: 120- i
8
g 100- 1
° 1
516‘
:5 E 80-

0
es, [
-° E 60-
Ev
E 40- i i
,, 5
>4
-' 20-

0 I I r I I I

 

 

 

0 20 40 6O 80 100 120 140
Polymerization time (min)
Figure 6.6. Lysozyme binding capacity as a function of MES polymerization time.
Each point represents a different membrane and shows an average of three

independent runs. The error bars correspond to the standard deviation.

164

 

protein binding, which may occur because the polymer brush is so thick and crowded that
the interior of the brush is no longer accessible for protein binding.8 Longer brushes may
also block some pores to decrease binding capacity. Nevertheless, a l h polymerization
yields membranes with a lysozyme binding capacity of 118 :t 8 mg/cm3. This capacity is
2 to 4-fold higher than the binding capacities of commercial ion-exchange membranes?“
6.3.2.-c. Protein binding capacity as a function of flow rate

One of the potential advantages of membrane absorbers over column-based
separations with nanoporous resins is that convection through membrane pores should
minimize diffusion limitations and lead to rapid binding and short purification times.
However, if slow diffusion into polymer brushes or slow binding kinetics limit the rate of
binding, this advantage may be negated. Slow diffusion into brushes is likely to be more
problematic for larger proteins, so we examined protein binding as a function of flow rate
for BSA absorption in a membrane containing poly(MES)-NTA-Cu2+ brushes. Figure
6.7. shows the breakthrough curves for this system at three different flow rates (0.3
mL/min, 0.8 mL/min, 1.1 mL/min). Within the limits of experimental error, the protein
binding capacity is independent of flow rate over this range (capacity is ~ 60 mg/cm3).
These results are in agreement with the findings of Knudsen and coworkers, who reported
that the dynamic capacity of cation-exchange membranes for antibody purification
remained constant even for a 50-fold increase in flow rate.12 Moreover, increasing the
flow rate does not lead to earlier protein breakthrough, although admittedly breakthrough
is relatively rapid at all three flow rates. Thus, poly(MES)-NTA-Cu2+ modified

membranes can be used at a flow rate of 1 mL/min without compromising the protein

binding capacity.

165

 

1.0 - ‘ K Flow rate = 0.3 mUmin
4-Flow rate = 0.8 mL/min
K Flow rate = 1.1 mUmin

   

Concentration (mg/mL)
O
OI

012345678910

Permeate Volume (mL)
Figure 6.7. Breakthrough curves for absorption of 1 mg/mL BSA in poly(MES)-
NTA-CuZI-modifed nylon membranes (1.2 pm nominal filtration cutoff) at flow rates
of 0.3 mL/min (triangles), 0.8 mL/min (squares), and 1.1 mL/min (circles). Curves are

added to guide the eye.

6.3.2.-d. Protein binding in membranes with different compositions

To test whether the high protein binding capacity is specific to LoProdyne® LP
nylon membranes, we also studied other nylon and regenerated cellulose membranes
modified with polymer brushes. Nylon membranes from GE Water & Process
Technologies (nominal pore sizes of 1.2 and 5.0 pm) were hydroxylated as described in
section 6.2.2.-b prior to initiator attachment via the resulting hydroxyl groups and
subsequent polymerization of MES. Poly(MES)-modified GE membranes showed
lysozyme binding capacities of 122 mg/cm3 and 50 mg/cm3 for 1.2 pm and 5 pm

membranes respectively. After modification with NTA-Cu”, the BSA binding capacities

166

 

 

were 82 mg/cm3 and 23 mg/cm3 for 1.2 pm and 5 pm membranes, respectively. Thus,
the protein binding capacities with these membranes are comparable to those obtained
with LoProdyne® LP nylon membranes (see section 6.3.2.-3.).

In the case of regenerated cellulose, the membranes were initially modified either

by reaction with the trichlorosilane initiator (as described above) or using a method‘

described by Singh and coworkers'3 in which 2-bromoisobutyryl bromide reacts with
hydroxyl groups. With either method, the membranes showed excessive swelling and
shrinking that resulted in cracking. Thus we were unable to examine protein binding to
these materials.

6.3.3. HisU binding to poly(MES)-NTA-Ni“ brushes in membranes

”'15 and one

Polyhistidine is the most common affinity tag for protein purification,
of the most important goals of this work is the development of polymeric membranes
with a high capacity for binding of His-tagged proteins. HisU served as the model
protein for determining the binding capacities for His-tagged proteins because it is readily
available in high purity. Figure 6.8. shows the breakthrough curve for HisU absorption in
a nylon (1.2 pm nominal pore size) membrane modified with poly(MES)-NTA-Ni”.
Integration of the differences between the feed concentration and the permeate
concentrations in the breakthrough curve gives a binding capacities of 60 mg/cm3. After
measuring the breakthrough curve of HisU, the poly(MES)-NTA-Ni2+-HisU membrane
was washed with 20 mL washing buffer I followed by 20 mL phosphate buffer, and HisU
was eluted with 5-10 mL elution buffer containing 0.5 M imidazole. Analysis of the

eluent using a Bradford assay showed a binding capacity of 85 mg of HisU per cm3

membrane. The uncertainty in the amount of binding determined from the breakthrough

167

. -J5?‘5:._I- —

 

0.3 . g
0.25

0.2 e
0.15

0.1

Concentration (mg/mL)

0.05
e

0
0 .-. 0. . . .
0 1 2 3 4 5 6 7 8 9 10

Permeate Volume (mL)
Figure 6.8. HisU breakthrough curve during passage of a 0.3 mg/mL HisU solution
through a poly(MES)-NTA-Ni2+-modifed nylon membranes (Loprodyne 1.2 pm
nominal filtration cutoff). The permeate flow rate through the membrane was 0.34

mL/min.

curve likely explains that the binding capacity is 25% higher when determined with
eluent rather than breakthrough analysis. Nonetheless, the binding capacity of the
membrane is as high as 85 mg HisU/cm3 of membrane, which is at least 6-fold greater

than that for affinity membranes reported in the literature.'6"9 The capacity is also higher

than that of commercial IMAC resins (maximum reported capacity of 50 mg/mL resin),20 '

but we need to test the binding capacity for high molecular weight His-tagged proteins as

well.
6.3.4. Purification of His-Tagged CRALBP from cell extracts
The above results show excellent binding of His-tagged proteins by membranes

modified with poly(MES)-NTA-Ni”, but they do not demonstrate whether the

168

 

 

membranes can isolate His-tagged protein from cell extracts. As discussed in chapter 3,
nylon membranes modified with poly(HEMA)-NTA-Ni2+ brushes are promising for
purification of His-tagged proteins directly from a cell extract, but in that case binding
capacities were low, presumably because organic solvents damaged the membrane. With
improved aqueous syntheses, we examined the performance of nylon membranes
modified with poly(MES)-NTA-Ni2+ in the purification of His-tagged CRALBP (36 kD)
that was over-expressed in E. coli. Figure 6.9. shows the gel electropherograms of the
cell extract (gels (a) and (h), lane 1) and the eluent from the membrane (gel (a), lane 2).
The eluent contains remarkably pure protein. Unfortunately, we cannot establish the

protein binding capacity or elution efficiency in this case because the concentration of

 

(a) (b)
Figure 6.9. SDS-PAGE analysis (coomassie blue staining) of an extract from E. coli
containing over-expressed His-tagged CRALBP (gels (a) and (h), lane 1) and
CRALBP purified from the cell extracts using a poly(MES)-NTA-Ni2+-modified

membrane (gel (a), lane 2) and a spin-trap column (gel (h), lane 3).

169

 

 

CRALBP in the cell extract is unknown.

The goal of this research is to develop affinity membranes for purification of His-
tagged proteins in order to overcome limitations in column-based protein separations. To
compare the performance of poly(MES)-NTA-Ni2+-modified nylon membrane with
commercial spin-trap columns (for details on spin-trap columns refer to chapter 3), His-
tagged CRALBP was purified using both systems. The electropherogram of the eluent
from a spin-trap column (Figure 6.9. gel (h), lane 2) shows high protein purity similar to
that obtained with membrane purification. However, membrane-based purification
included only 6 min for loading, 5 min for washing, and 2-3 min for elution. Thus
purification of His-tagged proteins can be achieved in less than 15 min. On the other
hand, spin-trap systems are small scale “columns” with a maximum loading of 600 pL.
Thus purification of 5-10 mL of cell extract requires several loading cycles, which
increases the time and labor required for the separation. Another advantage of the
membrane-based separation is a high capacity (2.9 mg protein/membrane) compared to a
spin-trap column (750 pg/column). Hence, these membranes are attractive for rapid,
selective purification of His-tagged proteins.

Future studies will focus on reusability of the membranes and comparison of
membrane-based purification with relatively large scale Ni-NTA column in terms of
purity, protein binding capacity and time of purification.

6.4. Conclusions

This work demonstrates that growth of poly(MES)-NTA-Ni2+ brushes inside

porous nylon supports yields high-capacity membranes that can selectively purify His-

tagged proteins directly from cell extracts. Brush-modified membranes show a binding

170

 

 

capacity of 118 i 8 mg lysozyme/cm3 of membrane and 85 mg HisU/cm3 of membrane
along with minimal non-specific adsorption. Lysozyme binding capacity increases with
polymerization times up to 1 h and then decreases at longer polymerization times,
suggesting that if brushes are too long, binding sites become inaccessible.

Gel electrophoresis results indicate that membranes modified with poly(MES)-
NTA-Ni2+ can effectively isolate His-tagged CRALBP from a cell-lysate. Importantly
these membranes have a pure water flux of 8.6 at 3 mL/cm2 min at 6.9 x 104 Pascal (10
psig) and support a flow rate of 1.2 mL diluted cell extract/min at a pressure less than 6.9
X 103 Pascal (l psig). Moreover, the purity of the His-tagged protein is comparable to
that obtained with commercial spin-trap columns. The membranes have potential
advantages over spin-trap and other columns in terms of time and ease of purification as
well as binding capacity. Hence, these modified membranes are attractive for high-

capacity, selective purification of His-tagged proteins.

171

 

6.5. References:

(1)
(2)

(3)

(4)

(5)

(6)

(7)

(8)
(9)
(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

Ciampoli.M; Nardi, N. Inorganic Chemistry 1966, 5, 1150.

Xu, F. J.; Zhao, J. P.; Kang, E. T.; Neoh, K. G.; Li, J. Langmuir 2007, 23, 8585-
8592.

Jain, P.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Macromolecules 2008, 41,
8413-8417.

Bao, Z.; Bruening, M. L.; Baker, G. L. Journal of the American Chemical Society
2006, 128, 9056-9060.

Huang, W. X.; Kim, J. B.; Bruening, M. L.; Baker, G. L. Macromolecules 2002,
35,1175-1179.

Wray, W.; Boulikas, T.; Wray, V. P.; Hancock, R. Analytical Biochemistry 1981,
118, 197-203.

Bertolini, M. J.; Tankersley, D. L.; Schroeder, D. D. Analytical Biochemistry
1976, 71, 6-13.

Bhut, B. V.; Husson, S. M. Journal of Membrane Science 2009, 33 7, 215-223.
Zhao, B.; Brittain, W. J. Macromolecules 2000, 33, 342-348.

Juang, A.; Scherman, O. A.; Grubbs, R. H.; Lewis, N. S. Langmuir 2001, 17,
1321-1323.

Wang, J.; Dismer, F .; Hubbuch, J.; Ulbricht, M. Journal of Membrane Science
2008, 320, 456-467.

Knudsen, H. L.; Fahrner, R. L.; Xu, Y.; Norling, L. A.; Blank, G. S. Journal of
ChromatographyA 2001, 907, 145-154.

Singh, N.; Wang, J .; Ulbricht, M.; Wickramasinghe, S. R.; Husson, S. M. Journal
of Membrane Science 2008, 3 09, 64-72.

Groll, J .; Haubensak, W.; Ameringer, T.; Moeller, M. Langmuir 2005, 21, 3076-
3083.

Gaberc-Porekar, V.; Menart, V. Chemical Engineering & Technology 2005, 28,
1306-1314.

Kawakita, H.; Masunaga, H.; Nomura, K.; Uezu, K.; Akiba, 1.; Tsuneda, S.
Journal of Porous Materials 2007, 14, 387-391.

Nova, C. J. M.; Paolucci-Jeanjean, D.; Belleville, M. P.; Barboiu, M.; Rivallin,
M.; Rios, G. Journal of Membrane Science 2008, 321, 81-89.

172

 

"E- v
' ""‘f- I
my.

 

 

(18)

(19)

(20)

Ma, Z. W.; Masaya, K.; Ramakrishna, S. Journal of Membrane Science 2006,
282, 23 7-244.

Bayramoglu, G.; Celik, G.; Arica, M. Y. Colloids and Surfaces a-
Physicochemical and Engineering Aspects 2006, 28 7, 75-85.

Ionov, L.; Stamm, M.; Diez, S. Nano Letters 2005, 5, 1910-1914.

173

u film

 

In _

 

 

Chapter 7

Summary and future work

7.1. Research summary

The research described in this dissertation aims at developing polymer brush-
.modified affinity membranes for high-capacity protein binding as well as rapid and
selective purification of His-tagged proteins. In Chapter 2, I discussed growth of
poly(HEMA) brushes inside porous alumina membranes and subsequent
functionalization of the poly(HEMA) with NTA-Ni” complexes for rapid, highly
selective purification of His-tagged proteins. Gel electrophoresis revealed that the purity
of HisU eluted from these materials is >99%, even when the initial solution contains 10%
bovine serum or a 20-fold excess of BSA. Moreover, the binding capacity of the
membrane is at least S-fold greater than that for membranes reported in the literature."3
Separations can be completely performed in 30 min or less and membranes are fully
reusable. Unfortunately, purification of His-tagged proteins using porous alumina
membranes is limited to relatively simple solutions and low flow rates because of a
limited pore size. Polymeric membranes, on the other hand, can have larger pore sizes
than porous alumina, which should allow rapid purification with more complex solutions.
In Chapter 3, we demonstrated the use of porous nylon membranes modified with
poly(HEMA)-NTA-Ni2+ brushes for selective purification of His-tagged CRALBP
directly from a cell lysate. The resulting CRALBP has a purity that is at least as good as
that obtained with Ni"2+ columns. Unfortunately, nylon membranes modified with
poly(HEMA)-NTA-Ni2+ brushes have a relatively low protein binding capacity (25 mg

protein/cm3 of the membrane), perhaps because the organic solvents employed in the

174

 

 

 

 

brush synthesis and derivatization partially damage the membrane structure. To avoid
the use of organic solvents that may dissolve or corrupt porous substrates, we developed
a completely aqueous procedure for growth of polymer brushes inside polymeric
membranes. In Chapter 4, we discussed the use of aqueous layer-by-layer adsorption of
polyelectrolyte macroinitiators and subsequent aqueous ATRP from these immobilized
initiators for successful grth of poly(HEMA) brushes on polymeric substrates.

Despite the growth of poly(HEMA) brushes with aqueous initiation, the creation
of protein-binding brushes requires conversion of hydroxyl groups in poly(HEMA) to
carboxylic acid moieties, which involves reaction of poly(HEMA) with succinic
anhydride in an organic solvent for several hours. In Chapter 5, we demonstrated that
surface-initiated aqueous ATRP of an acidic monomer, MES, allows a rapid, one-step
synthesis of polyacid brushes that avoids the need to react the brush with succinic
anhydride in an organic solvent. Also, poly(MES) brushes and their derivatives exhibit
high protein-binding capacities. F TIR spectroscopy and ellipsometry studies showed that
poly(MES) brushes are can bind the equivalent of many monolayers of BSA as well as
lysozyme.

In Chapter 6, we described the formation of protein absorbers through the grth
and derivatization of poly(MES) brushes inside porous nylon membranes. These
membranes exhibit protein binding capacities of 80 :L- 2 and 118 i 8 mg protein per cm3
of the membrane for BSA and lysozyme, respectively. Finally, we showed that the

poly(MES)-NTA-Ni2+-modified membranes can selectively and rapidly purify His-

tagged CRALBP directly from a cell extract.

175

7.2. Future work

A limitation to our current method for creating protein-absorbing nylon
membranes is that the initiator attachment requires THF, which makes it difficult to
extend the procedure to other polymer membranes. Many polymer membrane materials
such as polysulfone and polyethersulfone (PES) are incompatible with the organic
solvents used for initiator attachment. Thus, for such substrates organic solvents should
be completely avoided in the brush synthesis to preserve the pore structure of the
membrane. To overcome this problem, we propose the use of aqueous layer-by-layer

adsorption of polyelectrolyte macroinitiators and subsequent aqueous ATRP from these

Polycationic

ma r ' "
c omltlator MES ‘

CuBr, CuBr2,ligand

 

Polymer
substrate

 

 

Poly(MES) brush

8 it .5

:/'\T

SO
3 j’Br

PDADMAC PSS

 

V

Polyelectrolytes Polycationic macroinitiator

Scheme 7.]. Schematic illustration of growth of poly(MES) brushes by ATRP from

macroinitiators adsorbed in a membrane pore.

176

 

immobilized initiators for growth of poly(MES) brushes (Scheme 7.1). We have
successfiilly used this method for aqueous growth of poly(HEMA) brushes in PES
membranes (described in Chapter 4), however, we still need to study polymerization of
MES using the macroinitiator.

Chapter 6 shows that poly(MES)-NTA-Ni2I-modified nylon membrane are
promising for rapid, high-capacity purification of His-tagged CRALBP directly fi'om a
cell extract. However to generalize this work, it is important to test the performance of
these membranes for purification of a variety of overexpressed His-tagged proteins.
Future studies should also determine the reusability and elution efficiency of these
membranes. To examine the elution efficiency, we propose spiking a cell extract (devoid
of over-expressed His-tagged protein) with a known amount of HisU. Purification of this
solution using our modified membranes should provide information about the HisU
binding capacity as well as elution efficiency.

Chapters 3 and 6 showed that the membrane-based purification is comparable to
isolation with conventional IMAC resins in terms of purity. To establish the utility of the
membranes, we need to further compare membrane-based purification with IMAC resins
in terms of time and ease of purification. Moreover, a challenge with traditional IMAC
chromatography is that certain recombinant proteins are contaminated, even after
purification. For example, it is difficult to purify proteins like small nuclear RNA
activating protein complex (SNAPc),4‘6 mannose 6-phosphate glycoprotein,7 and GroEL-
GroES chaperonin complex8 using a single-step chromatographic separation. In such
cases, highly abundant or “sticky” proteins are present along with the recombinant

protein.8 The binding of sticky proteins to IMAC columns occurs due to affinity for the

177

resin material or the presence of surface clusters of histidine residues that bind to metal
complexes. For example, metal-binding lipocalin,9 glucosamine-6-phosphate synthase,‘0
and peptidoylproline cis-trans isomerase” have high affinity for metal-binding sites and
are often co-purified with His-tagged proteins during IMAC. On the other hand, proteins
like Hsp60 are more likely to bind to the sepharose of typical resins through hydrophobic

interactions.8

In fact in certain cases, the level of Hsp60 is higher than that of
recombinant protein.8 It is important to compare the performance of membranes with the
IMAC columns for purification of these proteins. We expect that proteins that bind to the
resin will be less abundant in the membrane-purified solutions whereas the proteins
binding to Ni2+ should be contaminants with both membranes and columns.

In chapter 6, we showed that a poly(MES)-NTA-Ni2+ membrane can bind 85 mg
of HisU per cm3 of the membrane, which is equivalent to ~29 mg of protein per
membrane (assuming a membrane thickness of 110 pm (Figure 3.2 (b)) and a membrane
diameter of 2 cm). Also, the flow rate for cell extract containing over-expressed
CRALBP was 1.2 mL/min at a pressure less than 6.9 x 103 Pascal (l psig). Based on
these results, we expect that purification of up to 2.9 mg protein can be achieved rapidly
using a ‘membrane-based syringe filter’ (Figure 7.1.(a)). A poly(MES)-NTA-Ni2+-
modified membrane can be placed in the membrane holder attached to a syringe (Figure
7.1.(b)), and the cell extract, washing, and elution solutions can be sequentially passed
through the membrane. The advantage of using this system over conventional small-
scale columns would be time and ease of purification. Moreover a membrane-based

syringe filter would have a 4-fold higher protein binding capacity than a commercial spin

trap column (2.9 mg protein/membrane vs. 750 pg protein/column).

178

Membrane

   

(a) (b)
Figure 7.1. (a) Syringe filter with a disposable syringe and (b) a membrane in the

holder.

To purify more than 2.9 mg of protein, we can stack several poly(MES)-NTA-
Ni2+ membranes in a holder and pass the cell extract through the stack of membranes
(Figure 7.2.). This should increase the amount of protein binding, but it might increase
the time of purification or the applied pressure. In this case, we will need to optimize the

binding capacity and time of purification as a function of applied pressure.

179

 
  
  
 
  

Cell extract containing overexpressed
His-tagged protein

Membrane stack

Membrane holder

 

 

 

 

Peristaltic pump Collection vial

Figure 7.2. Purification of His-tagged protein fi'om a cell extract using a stack of

poly(MES)-NTA-Ni2+ membranes.

7.3 Potential Impact

Protein purification is vital in biomedical research and the development and
manufacture of therapeutic peptides and proteins. Typical purifications involve a series
of steps, the most important of which fi'equently relies on affinity binding. Unfortunately,
affinity methods often present a bottleneck in the purification process because of slow
diffusion of proteins into the pores of chromatographic gels. Protein-absorbing
membranes can overcome this challenge because convective flow through membrane
pores provides rapid mass transport to binding sites. The research described in this
dissertation shows that porous polymer membranes modified with poly(MES)-NTA-Ni2+
brushes allows rapid and highly selective purification of His-tagged proteins with

separations achieved in 15 min or less. Moreover the membranes coated with polymer

180

brushes have binding capacity as high as 85 mg HisU/cm3 of membrane, which is several
folds greater than that for membranes reported in the literature. Potentially, the
‘membrane-based syringe filters’ are promising for rapid, small scale purification of His-
tagged proteins and should be advantageous over conventional small scale purification
columns in terms of binding capacity, time and ease of purification. Moreover, scale-up
of membrane separations through stacking of membranes, should avoid the challenges of
packing large columns, and the pressure drops in membranes can be very low compared

to a column-based separation.

18]

 

7.4. References:

(1) Sun, L.; Dai, J. H.; Baker, G. L.; Bruening, M. L. Chemistry of Materials 2006,
18, 4033-4039.

(2) Ulbricht, M.; Yang, H. Chemistry of Materials 2005, 17, 2622-2631.

(3) Kawakita, H.; Masunaga, H.; Nomura, K.; Uezu, K.; Akiba, 1.; Tsuneda, S.
Journal of Porous Materials 2007, 14, 387-391.

(4) Lai, H. T.; Kang, Y. S.; Stumph, W. E. FEBS Letters 2008, 582, 3734-3738.

(5) Emran, F.; Florens, L.; Ma, B. C.; Swanson, S. K.; Washbum, M. P.; Hernandez,
N. Gene 2006, 377, 96-108.

 

(6) Hanzlowsky, A.; Jelencic, B.; Jawdekar, G.; Hinkley, C. S.; Geiger, J. H.; Henry,
R. W. Protein Expression and Purification 2006, 48, 215-223.

(7) Sleat, D. E.; Della Valle, M. C.; Zheng, H.; Moore, D. F.; Lobel, P. Jaurnal of
Proteome Research 2008, 7, 3010-3021.

(8) Bolanos-Garcia, V. M.; Davies, 0. R. Biochimica Et Biophysica Acta-General
Subjects 2006, I 760, 1304-1313.

(9) David, G.; Blondeau, K.; Schiltz, M.; Penel, S.; Lewit-Bentley, A. Journal of
Biological Chemistry 2003, 278, 43728-43 735.

(10) Obmolova, G.; Badet-Denisot, M. A.; Badet, B.; Teplyakov, A. Journal of
Molecular Biology 1994, 242, 703-705.

(11) Hottenrott, S.; Schmnann, T.; Pluckthun, A.; Fischer, G.; Rahfeld, J. U. Journal of
Biological Chemistry 1997, 272, 15697-15701.

182