PROPERTIES AND APPLICATIONS OF SELF-ASSEMBLED BIOMOLECULES IN
NANOSTRUCTURED BIOMIMETIC INTERFACES
By
ANGELINES A. CASTRO FORERO

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Chemical Engineering
2011

ABSTRACT
PROPERTIES AND APPLICATIONS OF SELF-ASSEMBLED BIOMOLECULES IN
NANOSTRUCTURED BIOMIMETIC INTERFACES
By
ANGELINES A. CASTRO FORERO
Living cells are able to synthesize a wide range of biomolecules whose precisely
controlled size, shape, charge, linkage, molecular weight, functional groups, and binding affinity
enable self-assembly. These molecular building blocks, including amino acids, carbohydrates,
lipids, and nucleic acids, form macromolecular structures whose functional properties often
cannot be matched by man-made materials. To harness the advantages of biomolecules in the
design of new materials, controlled methods of fabrication are needed. This dissertation focuses
on advancing the understanding of the fabrication process of two biomaterials – phospholipids
and self-organizing peptides.
Phospholipids, the most abundant components of cell membranes, spontaneously form
lipid bilayers in aqueous solutions. Synthetic lipid bilayers in the form of small unilamellar
vesicles (SUV) and supported bilayer lipid membranes (sBLM) are widely used to study
membrane-mediated processes and to mimic natural membranes in surface-based devices. The
two preferred methods to fabricate SUVs are extrusion and sonication. The goal of this study was
to evaluate whether the method of SUV fabrication method has any influence on selected key
bilayer properties. The results confirmed that the morphology, average size and distribution of
SUVs varied with the fabrication method. However, the molecular-scale behavior of the bilayer,

as studied by fluorescence lifetime, anisotropy, and translational diffusion measurements, was
independent of the fabrication method.
Some peptides self assemble into biologically important macromolecular structures, and
thus represent excellent candidates for bottom-up fabrication of nanostructured biomimetic
materials. The objective of this study was to determine whether synthetic analogues of the
conductive pili produced by the metal-reducing bacterium Geobacter sulfurreducens could be
fabricated in-vitro. Type IV pili are homopolymers of a pilin subunit (PilA) or pilin that
polymerize via hydrophobic interactions to form pili. The resulting synthetic protein nanowires
could be potentially used to develop novel nanotechnologies, including nanoelectronic devices.
Using genetic engineering and protein expression tools, a method to mass-produce pilA peptides
was developed. In-vitro assembly of recombinant pilin subunits resulted in the formation of
filaments with properties similar to native G. sulfurreducens pili, including a similar diameter,
tendency to aggregate into tangled bundles, and electrical conductivity. This achievement offers
a potential cost-effective method to mass-produce protein nanowires for commercial
applications.
Collectively, this research provided important insights into the use of self-assembling
biomolecules for bottom-up fabrication of functional and nanostructured biomimetic interfaces.
It demonstrated that the liposome preparation method only affects physical properties of the
vesicles (average size and distribution) but not the dynamics of rotational and translational
diffusion of the bilayers. It also enabled production of recombinant pilin subunits into synthetic
nanowires having properties similar to the native pili of G. sulfurreducens.

Copyright by
ANGELINES A. CASTRO FORERO
2011

To God, who never forsakes me,
and
To my parents

v

ACKNOWLEDGEMENTS
Finishing this dissertation would have not been possible without the help and support of
many people. To them I would like to convey my gratitude. I would like to thank my advisor
Robert Y. Ofoli and co-advisor R. Mark Worden for the opportunity to join their labs and for
their help in becoming an independent researcher. I wish to thank Ilsoon Lee for his time and
suggestions while serving as a member of my graduate committee.
This dissertation is the result of several collaborative efforts. Therefore, I want to thank
the former and current graduate students with whom I worked closely: Aaron Greiner, Monique
Lapinski, and Rebecca Steidl. I also want to thank undergraduate students Teresa DeLuca, Ben
Kremkow, Rafael Goseer, and Logan Matthews for their help and for sharing the long-standing
tradition of ice-cream Friday.
I would like to thank those who helped me with different aspects of my research, Stuart
Tessmer and Shannon Demlow for their collaboration with the scanning tunneling microscopy
measurements; Alicia Pastor and Carol Flegler for their invaluable help with the electron
microscopy techniques; Daniel Jones for training me on the use of mass spectrometry techniques
and experimental suggestions; Lee Alexander and William Kittleman for their help with protein
expression and purification; Joe Leykam and Sanela Lampa-Pastirk for the research discussions
and insights; and Hazel-Ann Hosein for her training and help with the atomic force microscopy.
Special thanks to Gary J. Blanchard for his guidance and support through out my studies.
The joy and enthusiasm he has for research was contagious and his upbeat personality made it
very enjoyable working with him. I am deeply grateful to Gemma Reguera for her enthusiasm,
encouragement, selfless support, and useful advice during my research. Thank you for trusting

vi

me with such an important project, and for the excellent example of a successful woman in
science.
Finally, my sincere gratitude goes to my parents, family, and friends for their love,
patience, words of encouragement, and prayers during the last couple of years.

vii

TABLE OF CONTENTS
LIST OF TABLES...................................................................................................................xi
LIST OF FIGURES................................................................................................................xii
NOMENCLATURE............................................................................................................... xvi
LIST OF ACRONYMS ........................................................................................................xvii
1. INTRODUCTION ............................................................................................................... 1
1.1. Motivation................................................................................................................... 1
1.2. Problem Description.................................................................................................... 2
1.3. Background................................................................................................................. 6
1.3.1.
Model Lipid Membranes: Liposomes and Supported Bilayer Lipid Membranes.. 6
1.3.2.
Extracellular Electron Transfer in Geobacter Bacteria ........................................ 8
1.3.3.
Geobacter Pili: Structure and Function ............................................................... 9
1.4. Organization of this Dissertation ............................................................................... 13
1.4.1.
Characterization of Liposomes Prepared by Different Techniques: Measurement
of Rotational and Translational Diffusion of an Embedded Chromophore........................ 13
1.4.2.
Mass-Production and In Vitro Assembly of Protein Nanowires Using
Recombinant Pilins From Geobacter Sulfurreducens....................................................... 14
1.4.3.
Study of Pilin Modifications to Enable Interfacing with Electrodes................... 14
2. CHARACTERIZATION OF LIPOSOMES PREPARED BY DIFFERENT
TECHNIQUES: MEASUREMENT OF ROTATIONAL AND TRANSLATIONAL
DIFFUSION OF AN EMBEDDED CHROMOPHORE ....................................................... 17
2.1. Introduction............................................................................................................... 17
2.2. Materials and Methods .............................................................................................. 19
2.2.1.
Preparation of Liposomes by Extrusion............................................................. 19
2.2.2.
Preparation of Liposomes by Sonication ........................................................... 20
2.2.3.
Dynamic Light Scattering ................................................................................. 20
2.2.4.
Transmission Electron Microscopy................................................................... 20
2.2.5.
Steady-State Spectroscopy................................................................................ 21
2.2.6.
Time Correlated Single Photon Counting.......................................................... 21
2.2.7.
Formation of Supported Bilayer Lipid Membranes ........................................... 22
2.2.8.
Fluorescence Recovery After Pattern Photobleaching ....................................... 22
2.3. Results and Discussion .............................................................................................. 24
2.3.1.
Liposome Morphology, Mean Diameter, and Size Distribution......................... 24
2.3.2.
Fluorescence Lifetimes and Rotational Diffusion in Liposomes........................ 24
2.3.3.
Translational Diffusion in Supported Bilayer Lipid Membranes........................ 28
2.4. Conclusions............................................................................................................... 30

viii

3. MASS-PRODUCTION AND IN-VITRO ASSEMBLY OF PROTEIN NANOWIRES
USING RECOMBINANT PILINS FROM GEOBACTER SULFURREDUCENS............... 39
3.1. Introduction............................................................................................................... 39
3.2. Materials and Methods .............................................................................................. 42
3.2.1.
Recombinant DNA techniques.......................................................................... 42
3.2.2.
Expression of Recombinant Pilins..................................................................... 43
3.2.3.
Purification of Recombinant Pilins ................................................................... 45
3.2.4.
Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) .... 46
3.2.5.
Western Blot Analysis ...................................................................................... 47
3.2.6.
Matrix Assisted Laser Desorption-Ionization-Time of Flight Mass Spectrometry
(MALDI-TOF MS).......................................................................................................... 47
3.2.7.
Circular Dichroism (CD) .................................................................................. 48
3.2.8.
In-vitro Assembly of Recombinant Pilins into Pilus Fibers ............................... 49
3.2.9.
Transmission Electron Microscopy (TEM) ....................................................... 50
3.2.10. Atomic Force Microscopy (AFM)..................................................................... 50
3.2.11. Scanning Tunneling Microscopy (STM) ........................................................... 50
3.3. Results and Discussion .............................................................................................. 51
3.3.1.
Production of Recombinant Pilins..................................................................... 51
3.3.2.
Characterization of Secondary Structure of Pilin Subunits ................................ 58
3.3.3.
In-vitro Assembly of Recombinant Pilins into Pili Fibers.................................. 63
3.3.4.
Conductivity Assessment of Recombinant PilA19 Fibers via STM.................... 64
3.4. Conclusions............................................................................................................... 65
4. STUDY OF PILIN MODIFICATIONS TO ENABLE INTERFACING WITH
ELECTRODES....................................................................................................................... 92
4.1. Introduction............................................................................................................... 92
4.2. Materials and Methods .............................................................................................. 94
4.2.1.
Interfacing of Recombinant PilA19 Monomer with Gold Electrodes ................. 94
4.2.2.

Expression and Purification of Recombinant PilA19-A20C Subunits ................ 95

4.2.3.

Interfacing Recombinant PilA19-A20C Pilins with Gold Electrodes ................. 96

4.2.4.
Interfacing Recombinant PilA19-A20C Fibers with Gold Electrodes................. 97
4.2.5.
Ellipsometry ..................................................................................................... 98
4.2.6.
Scanning Electron Microscopy (SEM).............................................................. 98
4.2.7.
Cyclic Voltammetry (CV)................................................................................. 99
4.2.8.
Quartz Crystal Microbalance (QCM) .............................................................. 100
4.3. Results and Discussion ............................................................................................ 101
4.3.1.
Interfacing Recombinant PilA19 Monomers with Gold Electrodes .................. 101
4.3.2.

Production of Cysteine-Modified Recombinant Pilin: PilA19-A20C ............... 103

4.3.3.

Interfacing Recombinant PilA19-A20C Pilins with Gold Electrodes ............... 104

4.3.4.
Interfacing Recombinant PilA19-A20C Fibers with Gold Electrodes............... 106
4.4. Conclusions............................................................................................................. 107

ix

5. CONCLUSIONS .............................................................................................................. 116
REFERENCES ..................................................................................................................... 120

x

LIST OF TABLES
Table 2.1 DLS results for extruded and sonicated liposomes for two systems: (1) 98 mol%
DOPC and 2 mol% NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBDa
PC. Data are means ± SD of three samples. ............................................................................. 31
Table 2.2 Ratio of pre-exponential factors (Ai) and fluorescence lifetimes (τi) of NBD-PC
fluorophores in sonicated and extruded liposomes comprising (1) 98 mol% DOPC and 2 mol%
a
NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC. Data represent
means ± SD of three liposome solutions.................................................................................... 31
Table 2.3 Reorientation times (τOR), cone angles (θ0), and wobbling diffusion coefficients (DW)
of NBD-PC fluorophores in sonicated and extruded liposomes comprising (1) 98 mol% DOPC
and 2 mol% NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC.
Data represent means ± SD of three liposome solutions. ........................................................... 32
Table 2.4 Translational diffusion coefficients (DL) and mobile fractions for NBD-PC in sBLMs
prepared with sonicated and extruded liposomes comprising (1) 98 mol% DOPC and 2 mol%
NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC. Data represent
a
means ± SD of three liposome solutions. ................................................................................. 32
Table 3.1 Restriction enzymes and primers used for cloning of pilAn gene into expression
vectors. ..................................................................................................................................... 67
Table 3.2 Bacterial strains used to express recombinant pilin subunits...................................... 68
Table 3.3 Affinity resin and composition of buffers and solutions used during affinity
chromatography purification of recombinant pilin subunits....................................................... 69
Table 3.4 Analysis of CD data using the CONTINLL algorithm. The various secondary
structures determined are α-helix (regular αR and distorted αD), β-strands (regular βR and
distorted βD), turns and unordered. The NRMSD parameter indicates the goodness of fit. Values
below 0.1 are considered a good fit........................................................................................... 70

xi

LIST OF FIGURES
Figure 1.1 Structural comparison of a) Pseudomonas aeruginosa PAK pilin and b) Geobacter
pilin. c) Structural alignment of PAK and Geobacter pilins. The αβ-loop is colored magenta, the
D-region is colored green, and the disulphides are colored cyan. The crystal structure of PAK
pilin was resolved by Craig et al.[49] and is available at the Protein Data Bank. The structure of
Geobacter pilin presented here was calculated by a computational model (Gemma Reguera,
personal communication, October 7, 2011). For interpretation of the references to color in this
and all other figures, the reader is referred to the electronic version of this dissertation............. 16
Figure 2.1 Steady-state solution phase absorption and emission spectra of vesicles containing
2% NBD-PC and 98% DOPC. Both spectra have been normalized for clarity of presentation. .. 33
Figure 2.2 Schematics of the experimental setup for time correlated single photon counting
(TCSPC)................................................................................................................................... 34
Figure 2.3 Schematics of the experimental set-up for fluorescence recovery after pattern
photobleaching (FRAPP). ......................................................................................................... 35
Figure 2.4 Transmission electron microscopy images of liposomes formed by extrusion and
sonication: a) TEM image of extruded vesicles without cholesterol; b) TEM image of sonicated
vesicles without cholesterol. ..................................................................................................... 36
Figure 2.5 Size distribution obtained by dynamic light scattering showing diameters of
liposomes as a function of method of preparation and lipid composition: a) extruded DOPC
liposomes, b) extruded liposomes containing DOPC and cholesterol, c) sonicated DOPC
liposomes, and d) sonicated liposomes containing DOPC and cholesterol. ................................ 37
Figure 2.6 Representative FRAPP recovery curve and residual plot (below) for DOPC/NBD-PC
supported bilayer lipid membranes reconstituted from extruded liposomes. Similar recovery
curves were obtained for bilayer lipid membranes reconstituted from sonicated liposomes.
Photobleaching time was 300 ms. ............................................................................................. 38
Figure 3.1 Purification steps used for each expression system. ................................................. 71
Figure 3.2 Analysis of hydrophobicity of pilin subunits and effect of amino acid truncation. Kyte
and Doolittle hydrophaticity [87] plot using a 19-residue segment, amino acid sequence and
GRAVY index of PilA and PilAn truncations. .......................................................................... 72
Figure 3.3 Expression and purification of 6×His-PilA20. a) 16% tricine SDS-PAGE and b)
western blot analysis of cell pellet (A), soluble fraction of cell lysate (B), and insoluble fraction
of cell lysate (C); before (-) and after (+) induction. .................................................................. 73
Figure 3.4 Periplasmic expression and purification of MBP-PilA20. a) 12% glycine SDS-PAGE
and b) western blot analysis of crude extract (A) before (-) and after (+) induction; soluble
xii

fraction of cell lysate (B), and elution fraction from amylose column (C); c) MALDI-TOF MS
analysis of elution fraction from amylose column. Theoretical molecular weight of the
periplasmic MBP-PilA20 fusion is 46,902.6 Da. ....................................................................... 74
Figure 3.5 Cytoplasmic expression and purification of MBP-PilA20. a) 12% glycine SDS-PAGE
and b) western blot analysis of crude extract (A) before (-) and after (+) induction; soluble
fraction of cell lysate (B), and elution fraction from amylose column (C); c) MALDI-TOF MS
analysis of elution fraction from amylose column. Theoretical molecular weight of the
periplasmic MBP-PilA20 fusion is 47,033.9 Da. ....................................................................... 75
Figure 3.6 Cleavage efficiency of PilA20 subunit from MBP tag mediated by Factor Xa
o

protease. Elution fractions from amylose column were incubated with Factor Xa protease at 4 C
o

and 23 C. Samples were analyzed by 7.5% glycine SDS-PAGE after 24 h and 48 h of reaction.
Gels show elution sample untreated with protease as control (A), elution treated with 0.5 µg/mL
(B), and elution treated with 1.0 µg/mL (C). ............................................................................. 76
Figure 3.7 Production of PilA20 subunits after treatment with Factor Xa protease. MALDI-TOF
MS analysis of a) sample untreated with protease used as a control, and b) sample treated with
protease. MS results indicate two peptides are produced during cleavage. The peptide with MW
of 4,527.4 Da corresponds to the PilA20 subunit. c) N-terminal sequencing of the peptide with
MW of 3,563.3 Da. Sequencing indicated that non-specific cleavage of the pilin subunit
occurred.................................................................................................................................... 77
Figure 3.8 Expression of various truncations of pilin subunits using a CBD fusion tag (CBD+
PilAn). Expression was determined by analyzing cell pellet samples before (A ) and after (A )
induction using a 12% glycine SDS-PAGE. .............................................................................. 78
Figure 3.9 Cleavage efficiency of CBD tag from PilAn subunits after incubation with 50 mM
o

o

DTT at 4 C and 23 C. During cleavage, pilin monomers are released but CBD remains bound to
the chitin beads. Chitin beads were boiled with 1% SDS to release bound protein, which was
analyzed by 7.5% glycine SDS-PAGE. Control (no DTT in incubation buffer) (A), cleavage
efficiency after 24 h (B), 48 h (C), and 72 h (C). The appearance of a lower molecular weight
band indicates the cleavage....................................................................................................... 79
Figure 3.10 Presence of pilin monomers in elution from chitin column after cleavage from CBD
tag determined by 16-20% Tricine SDS-PAGE......................................................................... 80
Figure 3.11 MALDI-TOF MS analysis of elution from chitin column after cleavage. MS
analysis confirmed the molecular weight of the four truncated pilins. Theoretical molecular
weights of the truncated pilins are PilA10: 5,431 Da, PilA19: 4,595 Da, PilA20: 4,524 Da, and
PilA22: 4,314 Da....................................................................................................................... 81

xiii

Figure 3.12 Surfactant-dependent folding of recombinant PilA19 subunits at pH 3.8. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 3.8, 50 mM Na2SO4 with and without sodium
dodecyl sulfate (SDS). .............................................................................................................. 82
Figure 3.13 Surfactant-dependent folding of recombinant PilA19 subunits at pH 7.0. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 7.0, 50 mM Na2SO4 with and without sodium
dodecyl sulfate (SDS). .............................................................................................................. 83
Figure 3.14 Surfactant-dependent folding of recombinant PilA19 subunits at pH 3.8. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 3.8, 50 mM Na2SO4 with and without octyl
β-D-glucopyranoside (OG). ...................................................................................................... 84
Figure 3.15 Surfactant-dependent folding of recombinant PilA19 subunits at pH 7.0. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 7.0, 50 mM Na2SO4 with and without octyl
β-D-glucopyranoside (OG). ...................................................................................................... 85
Figure 3.16 Effect of sodium dodecyl sulfate (SDS) on the conformation of PilA19 subunits.
Molar ellipticities at pH 3.8 and 7.0 are plotted for three wavelengths, 192, 208, and 222 nm, as a
function of SDS concentration. ................................................................................................. 86
Figure 3.17 Effect of octyl β-D-glucopyranoside (OG) on the conformation of PilA19 subunits.
Molar ellipticities at pH 3.8 and 7.0 are plotted for three wavelengths, 192, 208, and 222 nm, as a
function of SDS concentration. ................................................................................................. 87
Figure 3.18 Effect of SDS on the percentage of regular α-helix of PilA19 subunits at pH 3.8 and
7.0. Percentage of regular α-helix was calculated using the CONTINLL algorithm. Lines
connecting data points are for visual aid only............................................................................ 88
Figure 3.19 Transmission electron microscopy images of recombinant pili fibers. Samples were
stained with uranyl acetate. TEM JEOL 2200FS (JEOL Inc., Boston, MA) operated at an
accelerating voltage of 100 kV.................................................................................................. 89
Figure 3.20 Atomic force microscopy images of recombinant pili deposited on mica. Tapping
mode in air................................................................................................................................ 90
Figure 3.21 Scanning tunneling microscopy image of a recombinant PilA19 fiber at an applied
voltage of 0.05 V. ..................................................................................................................... 91
Figure 4.1 Structural comparisons of wild type and recombinant pilins of Geobacter
sulfurreducens. Structure of a) wild type PilA protein; b) recombinant 19 amino acid truncation
PilA19 protein; c) recombinant cysteine-modified 19 amino acid truncation PilA19-A20C protein,
with the added cysteine colored turquoise. In each structure, the tyrosines are colored light blue.
d) Sequence alignment of wild type and recombinant pilins of Geobacter sulfurreducens. ...... 108
xiv

Figure 4.2 Formation of an undecanethiol self-assembled monolayer on a gold electrode. Cyclic
voltammograms of bare gold (black line) and undecanethiol SAM (red dashed line). Data
recorded at room temperature in 100 mM sodium phosphate buffer at pH 7.0 containing100 mM
-1
NaCl, and 5 mM K3[Fe(CN)6] at a potential scan rate of 100 mV s . The insert shows the
thickness of the undecanethiol monolayer measured by ellipsometry. ..................................... 109
Figure 4.3 Scanning electron microscopy images of gold substrates after deposition of an
undecanethiol SAM followed by incubation with PilA19 monomers in 20 mM sodium phosphate
pH 7.0, 100 mM NaCl, 10 mM SDS. Accelerating voltage 5 kV............................................. 110
Figure 4.4 Energy Dispersive X-ray (EDS) microanalysis of fibers present on gold substrates
modified by deposition of an undecanethiol SAM, followed by incubation with PilA19
monomers in 20 mM sodium phosphate pH 7.0, 100 mM NaCl, 10 mM SDS. Accelerating
voltage 1 kV. .......................................................................................................................... 111
Figure 4.5 Expression and purification of recombinant PilA19-A20C subunits: a) 12% SDS-

+

PAGE and western blot analysis of cells before (A ) and after (A ) induction. Anti-chitin serum
was used for western blot. b) MALDI-TOF mass spectrometry and c) 10-20% tricine gel of
elution fraction after purification............................................................................................. 112
Figure 4.6 Deposition of PilA19-A20C monomers on a gold electrode. Quartz crystal
microbalance (QCM) was used to monitor changes in frequency and dissipated energy after the
addition of pilin monomers. .................................................................................................... 113
Figure 4.7 Deposition of PilA19-A20C subunits on a gold electrode. Cyclic voltammograms of
bare gold (black line) and undecanethiol SAM (red dashed line). Data recorded at room
temperature in 100 mM sodium phosphate buffer at pH 7.0 containing 100 mM NaCl, and 5 mM
-1
K3[Fe(CN)6] at a potential scan rate of 100 mV s . Insert shows the thickness of PilA19-A20C
monolayer measured by ellipsometry...................................................................................... 114
Figure 4.8 In-vitro assembly of PilA19-A20C fibers. The transmission electron microscopy
images were collected at an accelerating voltage of 100 kV. ................................................... 115

xv

NOMENCLATURE
a

Ronchi ruling stripe periodicity

A

absorbance

Ai

ith pre-exponential factor

Dw

wobbling diffusion coefficient

D

translational diffusion coefficient

f(t)

time-dependent post-bleach fluorescence

III(t)

time-dependent light emission intensity polarized parallel to incident beam

I⊥(t)

time-dependent light emission intensity polarized perpendicular to incident beam

l

path length of cuvette

m

mobile fraction

r(0)

zero time anisotropy

r(t)

induced orientational anisotropy

r(∞)

infinite time anisotropy

t

time

τ

fluorescence lifetime

τHR

reorientation time

θ0

The semi-angle of the confining cone volume of the chromophore

[θ]

mean residue molar ellipticity

xvi

LIST OF ACRONYMS
ATP

adenosine triphosphate

BLM

bilayer lipid membrane

CBD

chitin binding domain

CD

circular dichroism

CFD

constant fraction discriminator

CMC

critical micelle concentration

CP-AFM

conductive probe atomic force microscopy

CV

cyclic voltammetry

DI

deionized

DLS

dynamic light scattering

DOPC

1-palmitoyl-2-hydroxy-sn-glycero-3-phosphocholine

DO

dissolved oxygen

DTT

dithiothreitol

EDS

energy dispersive X-ray spectroscopy

EDTA

ethylenediaminetetraacetic acid

FRAPP

fluorescence recovery after pattern photobleaching

GRAVY

grand average of hydropathicity index

HEPES

4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid

HM

Hagem medium

HOPG

highly oriented pyrolytic graphite

HPLC

high performance liquid chromatography

IMPACT

intein mediated purification with an affinity chitin-binding tag

xvii

IPTG

isopropyl-1-thio-D-galactopyranoside

KCl

potassium chloride

LB

Luria Bertani

MALDI

matrix assisted laser desorption ionization

MBP

maltose binding protein

MCP-PMT

microchannel plate photomultiplier tube

MgSO4

magnesium sulfate

MS

mass spectrometry

NaCl

sodium chloride

NBD-PC

1-oleoyl-2-[12-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-snglycero-3-phosphocholine nitro-benzo derivative

NTA

nitrilotriacetic acid

OD

optical density

OG

octyl β-D-glucopyranoside

PCR

polymerase chain reaction

PID

proportional–integral–derivative

PMSF

phenylmethylsulfonyl fluoride

PVDF

polyvinylidene fluoride

QCM

quartz crystal microbalance

SAM

self-assembled monolayer

sBLM

supported bilayer lipid membrane

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

SEM

scanning electron microscopy

STM

scanning tunneling microscopy
xviii

SUV

small unilamellar vesicles

TAC

time-to-amplitude converter

TBST

Tris-buffered saline solution and Tween 20

TCSPC

time-correlated single photon counting

TEM

transmission electron microscopy

TFA

trifluoroacetic acid

TOF

time of flight

xix

Chapter 1.
INTRODUCTION

1.1.

Motivation
Nanotechnology refers to the creation, manipulation and utilization of materials in the

submicron range [1, 2]. Even though nanoscale objects have found applications in electronics,
materials science and medicine, questions regarding their potential adverse health and
environmental effects remain unanswered [2]. An emerging alternative is to look at how nature
tackles highly complex problems and apply nature-inspired or biomimetic solutions to
nanotechnology applications. The intersection between nanotechnology and biology is
sometimes referred to as Bionanotechnology.
The customizable nature of many biological molecules makes them attractive for the
manufacturing of nanostructured materials whose properties can be tailored for the desired
purpose [3]. Nature uses a relatively small set of molecules to assemble impressive arrays of
hierarchical structures with highly specific functions [3, 4]. These nature-derived building blocks
include amino acids, carbohydrates, lipids, and nucleic acids [3]. Biological building blocks are
synthesized with precise control of sequence, linkage, molecular weight, functionality, size and

1

shape [5]. In addition, some biomolecules can also self-assemble with nanoscale precision. For
example, phospholipids form bilayers with a thickness of 4 to 5 nm and deoxynucleotides
polymerize via phosphodiester bonds into DNA strands, with complementary strands using
hydrogen bonding to form a double helix with a diameter of 2 nm [5]. These properties cannot
easily be duplicated in synthetic materials and still have the added value of enabling sustainable
production, biodegradability and low toxicity [6].
On the other hand, nanostructured materials and interfaces can be used as biomimetic
model systems to study cellular processes in a controlled and simplified environment void of
biological noise. Moreover, such biomimetic interfaces also find applications in biosensor
design, drug delivery, and biocatalysis [1, 7]. The key to these applications is the synthesis of the
biomaterials, which requires the use of non-disruptive methods that preserve the original
biological function, and sufficient understanding of the biomaterial properties so they can be
manipulated and further diversified or optimized [3, 5].
1.2.

Problem Description
The projects presented in this thesis are focused on two biomaterials: phospholipids and

amino acids. Phospholipids have been widely investigated because they readily self-assemble
into structures (vesicles, micelles, and bilayers) [5]. Vesicles and bilayers mimic the cell
membrane environment and serve as model membranes in the study of cellular processes [8, 9].
In addition, model membranes act as functional elements in a variety of applications including
biosensors, drug delivery and screening, and production of biocatalytic interfaces [8, 10]. Amino
acids are the building blocks for peptides and proteins. The same set of amino acids is used to
build a diverse array of structures including membrane proteins, enzymes and fibrous proteins
[4]. Self-organizing peptides and proteins have received attention in recent years because of their
2

potential for bottom-up fabrication of nanostructured interfaces. In addition, using genetic
engineering tools and functionalization, proteins can be tailored to serve specific applications
[11].
Inasmuch as the method used to synthesize biomaterials can affect their properties, the
first research objective was to evaluate the effect of the method of preparation of small
unilamellar vesicles (SUV) on selected bilayer properties. Phospholipids are a class of lipids and
the main component of cell membranes [12]. They have a hydrophilic head and two hydrophobic
tails. In aqueous solution, most diacyl phospholipids spontaneously form bilayer structures that
close themselves into vesicles [12]. The method of lipid dispersion will determine whether the
vesicles are unilamellar or multilamellar. When dry lipid is reconstituted in an aqueous solution,
multilamellar vesicles (MLV) are usually formed. MLV are onion-like structures of concentric
bilayers separated by narrow aqueous channels [13, 14]. Sonication and extrusion are the most
common methods to produce small unilamellar vesicles from MLV [13, 15]. Small unilamellar
vesicles, also called liposomes, can be used to form supported bilayer lipid membranes (sBLM).
The transition from multilamellar to unilamellar vesicles can be achieved by either
sonication or extrusion. Sonication uses acoustic energy, whereas extrusion uses pressure-driven
flow through a polycarbonate membrane [13, 16]. It is conceivable that the physical and
molecular scale properties of liposomes prepared by the two techniques would be different,
making it difficult to directly compare model BLMs reconstituted from sonication and extrusion.
In this study, we evaluated how the method of liposome preparation affected the morphology,
average size, and size distribution of liposomes. In addition, we studied the rotational and
translational diffusion of a tethered chromophore in model membranes reconstituted from
vesicles prepared by sonication and extrusion.

3

The second objective of this research was to exploit the self-organizing properties of
some peptides to construct conductive protein nanotubes. Specifically, we focused our
investigations on the conductive hair-like protein appendages or pili produced by the metalreducing bacterium Geobacter sulfurreducens [17]. Pili are 5 to 10 nm in diameter, and are
proteinaceous appendages anchored in the outer membrane of a wide range of Gram-negative
bacteria [18]. Pili are homopolymers of the pilin peptide or PilA, which assemble via
hydrophobic interactions to make the pilus filament [19, 20]. The pilus of G. sulfurreducens is of
special interest for nanotechnology applications because, in addition to its self-assembly
capabilities, it is also conductive [17, 21]. The pili of G. sulfurreducens function as electrical
conduits to transfer metabolically-generated electrons from inside the cell to extracellular
electron acceptors such as Fe(III) oxides [17] or other cells in electroactive biofilms [22]. These
attractive features, self-assembly and conductivity, make the pili of G. sulfurreducens ideal
candidates for applications in the design and fabrication of biobased electronic devices.
Key to the application of conductive pili in bioelectronics is a better understanding of
how the pilin peptides self-assemble and their mechanism of electron transfer. To accomplish
this, we first focused on producing the pilin monomer, or PilA peptide, in sufficient quantities to
enable investigation of its assembly. Current methods of pilin production are based on shearing
pili filaments from the outer surface of living cells, removal of non-pili outer surface proteins,
and depolymerization of the pili into pilins [23]. Production of pilins directly from Geobacter
cells is a challenging process because it is time consuming, the efficiency is low, and
reproducibility of the protocols is difficult. The pilin subunits are also highly hydrophobic, which
greatly complicates their depolymerization from the pili [23]. However, the proteinaceous nature
of the PilA peptides makes them potentially amenable to genetic engineering. Furthermore,

4

recombinant protein expression techniques could be used to mass-produce the PilA peptide in
heterologous hosts. We hypothesized that a combination of genetic engineering and recombinant
expression approaches would enable the production of large quantities of PilA monomers in a
controlled environment, thus eliminating the interference of other proteins from Geobacter cell
envelope.
We investigated several strategies to produce G. sulfurreducens pilins as soluble proteins.
The native PilA peptide is very hydrophobic, which makes their expression in soluble form
challenging. Thus, step-wise reductions in the hydrophobic amino acids at the N-terminus were
carried out to prepare truncated versions of the peptide that retained the redox-active amino acids
(at the C-terminus). The solubility of the recombinant peptides was increased by engineering
Escherichia coli strains that express the truncated peptides fused to soluble proteins. These
expression systems were also selected to enable the purification and cleavage of the peptide from
its carrier protein. Several expression systems, purification, and cleavage conditions were
investigated and optimized. The most efficient system that enabled the mass production of pilin
peptides carrying the shortest truncation was selected. Prior to developing various assembly
strategies, the secondary structure of the recombinant pilin was characterized to confirm that
their native structure had been retained, since this is required for self-assembly, Specifically, we
studied the in-vitro assembly of recombinant pilins in solution and interfaced them with solid
substrates.
The results presented here are stepping stones towards using lipid bilayers and selfassembled peptides and proteins as materials for nanotechnology applications. Important
questions regarding the methods used to fabricate these biomaterials have been addressed. We
have established that the two common methods used to prepare liposomes, sonication and

5

extrusion, only affect the physical properties of the bilayers (average size and size distribution).
The dynamics of rotational and translational diffusion of bilayers are not influenced by the
manner in which they were prepared. Consequently, the extensive body of work reported in the
literature regarding bilayer diffusion can be directly compared regardless which of the two
methods of liposome preparation was used.
We have also developed a method for the mass production of pilin monomers from G.
sulfurreducens. Using genetic engineering and recombinant protein expression, a soluble
truncated version of the pilin peptide PilA was produced. The recombinant pilins were assembled
into pili-like filaments that retained their conductive properties. Various strategies for interfacing
recombinant pilins and pili with gold substrates were also investigated. Findings from these
studies provide novel insights into their biological assembly and their mechanism of
conductivity. Furthermore, the possibility of mass-producing and self-assembling recombinant
pilins into conductive protein nanotubes is critical for the development of nanostructured
interfaces using a bottom-up fabrication approach. In the long term, in vitro-assembled pili can
be integrated with other redox-active components of G. sulfurreducens such as cytochromes to
develop biomimetic interfaces. Such interfaces can facilitate the study of the bacterium’s
complex electron transfer machinery for applications in bioenergy and bioremediation.
1.3.

Background

1.3.1. Model Lipid Membranes: Liposomes and Supported Bilayer Lipid Membranes
Cell membranes, which consist of a lipid bilayer containing proteins and other biological
species, serve as a boundary between cells and intracellular organelles [24, 25]. Lipids present in
biological membranes are usually amphipathic. The hydrophobic moieties form the membrane

6

core, and the polar head groups interact with the surrounding water. Proteins can be associated
with the membrane through electrostatic interactions, or embedded to various degrees in the
hydrophobic core of the membrane [26]. Cell membranes provide the surface or interface where
many biological reactions and processes take place. They participate in (1) exclusion of toxins
and certain molecules from the cell, (2) accumulation of nutrients, (3) energy transduction, (4)
cell locomotion, (5) reproduction, (6) signal transduction, and (7) interactions with other cells
[24, 26].
Synthetic organic membranes can be used as models to mimic biological membranes in a
controlled environment [25, 27]. They have been used to study membrane-mediated processes
and to investigate properties and functions of membrane proteins [10]. They are very promising
for applications such as diagnostic devices, biosensors, and drug delivery [28]. Vesicles, also
known as liposomes, and supported bilayer lipid membranes (sBLMs) are two model systems of
biological membranes [8, 29].
Phospholipids are the most abundant lipids in cell membranes. Consequently, they are
widely used to prepare model bilayers [25]. When phospholipids are added to aqueous media,
they spontaneously form bilayers because their pair of fatty acid chains does not pack well in a
micellar structure. Hydration of a phospholipid dry film usually results in formation of
multilamellar lipid vesicles (MLVs) [13, 16]. These onion-like structures have concentric
bilayers with low volumes of liquid trapped between them [13]. Small unilamellar vesicles
(SUVs) can be prepared directly from MLVs by various methods, mainly sonication and
extrusion [13, 30]. In sonication, acoustic energy is used to break large multilamellar vesicles
into smaller unilamellar ones [16]. The duration of sonication determines the size of the resulting
liposome [31]. In extrusion, phospholipids are forced through a polycarbonate membrane.

7

Extrusion produces liposome populations with lower polydispersities than sonication because the
liposome size is determined by the pore size of the membrane used [32].
Supported bilayer lipid membranes (sBMLs) are planar bilayers adsorbed on a suitable
solid support [24]. Lipids arranged in a sBLM retain the lateral fluidity observed in living
membrane systems. The water layer separating the bilayer from the support provides the
lubrication necessary to preserve the lateral fluidity of both lipid leaflets [26]. Supported bilayer
lipid membranes are formed essentially by the Langmuir-Blodgett protocols or vesicle fusion.
Langmuir-Blodgett techniques deposit each monolayer separately from the liquid-air interface of
a trough onto a hydrophilic substrate [24, 33]. Vesicle fusion, first reported by McConnell and
coworkers [27, 34], is a simpler and more reproducible one-step way to transform unilamellar
vesicles into sBLMs. With this method, vesicles from solution are adsorbed onto a hydrophilic
substrate, and then undergo rupture and fusion to form a planar bilayer [27, 34].
1.3.2. Extracellular Electron Transfer in Geobacter Bacteria
Members of the Geobacteraceae family are anaerobic bacteria that are ubiquitous in
subsurface environments [35]. Geobacter species gain energy in the form of adenosine
triphosphate (ATP) by transferring electrons to extracellular metal while oxidizing organic
compounds to CO2 [36, 37]. Electron acceptors used by these bacteria include soluble metals
(e.g. Fe(III) citrate and U(VI)) and insoluble metals (e.g. Fe(III) Mn(IV) oxides) [38]. Geobacter
sulfurreducens serves as a model for the Geobacteraceae because it is available in pure culture,
has a sequenced genome and is amenable to genetic manipulation [22, 39].
Geobacter bacteria have attracted interest for their application in bioenergy and
bioremediation. Their ability to oxidize aromatic hydrocarbons into environmentally innocuous

8

CO2 shows promise for the bioremediation of petroleum contaminants [40, 41]. Geobacter
bacteria also reductively precipitate metals and radionuclides such as vanadium and uranium, a
process that could be harnessed for the environmental restoration of contaminated groundwater
and sediments [42-44]. Finally, Geobacter can be used to harvest energy as electricity in devices
known as microbial fuel cells, which fully oxidize electron donors using only an electrode as
electron acceptor and without the addition of electron mediators [41, 45].
1.3.3. Geobacter Pili: Structure and Function
G. sulfurreducens produces hair-like proteinaceous filaments called pili (singular: pilus),
which emanate from one side of the cell. The pili of G. sulfurredunces are classified as type IV,
which reflects the conservation of the N-terminal amino acid sequence of the pilin monomer
involved in secretion and assembly [17, 46]. Type IV pili are 5 to 10 nm in diameter and can
reach several microns in length. They are helical polymers composed mainly of a single peptide
subunit known as the pilin (PilA) [17, 20]. Type IV pili are classified together on the basis of
common features shared by the pilins subunits. Type IV pilins are usually ∼ 150 amino acidslong 15-20 kDa proteins [47, 48]. They comprise a conserved hydrophobic N-terminal region
that adopts an α-helical structure and a variable C-terminus. The first 25 residues of the Nterminal region carry a conserved N-terminal segment with an N-methylated phenyalanine in the
first position, whereas the variable C-terminus contains a conserved disulphide bond [19, 47].
The G. sulfurreducens pilin contains the N-terminus segment and methylated phenalanine
characteristic of type IV pilins but is shorter (61 amino acids-long) and lacks the variable Cterminus and disulphide bond [17].
Understanding the mechanism for type IV pilins assembly has proved to be extremely
challenging because of the lack of structural data and the large number of proteins involved in
9

pilus assembly [47]. Crystallization studies have revealed the structures of only two full-length
type IV pilin: Neisseria gonorrhoeae GC pilin and Pseudomonas aeroginosa PAK pilin; and
three truncated pilins lacking 28 residues from the hydrophobic N-terminus: PAK pilin, P.
aeruginosa strain K122-4 pilin and Vibrio cholerae TcpA pilin [47]. Structural information of
the type IV pilus at atomic resolution is limited due to the thin diameters and smooth surfaces of
the filament [47]. To date, structural data of G. sulfurreducens pilus and pilin protein are not
available. Difficulties in de-polymerizing the filament and the insolubility of the pilin subunits
make crystallization and nuclear magnetic resonance (NMR) experiments extremely difficult.
The structural data available have shown that type IV pilins share a similar architecture:
the N-terminal end ∼ 53 residues which form an extended α-helix, α1; the globular head domain;
the αβ-loop located on one side of the globular domain; the anti-parallel β-sheet; and the Dregion, which is delineated by conserved cysteines (Figure 1.1) [19, 47]. The assembly of the
type IV pilin peptide into a pilus is mediated by interactions between the conserved ∼ 53 residues
of the N-terminal α-helices of the pilin subunits, which form a hydrophobic helical bundle in the
filament core [47, 49].
By contrast, the Geobacteraceae pilins are shorter (∼ 60 amino acids, ∼ 6.5 kDa),
narrower in diameter (3-5 nm), and form a distinct phylogenetic line of descent within the
bacterial type IV pilins [17]. Homology models of the pilin of G. sulfurreducens also show a
unique structure among type IV pilins. As shown in Figure 1.1, a structural model of the pilin of
G. sulfurreducens in reference to the structure of the model pilin from P. aeruginosa strain K (or
PAK) confirms the presence of the conserved ∼ 53 residues N-terminal α-helix and part of the
αβ-loop. This conserved hydrophobic helix at the N-terminus controls inner membrane insertion,
signal processing and pilin polymerization, and forms the central helical core of the pilus

10

filament [17, 39]. Therefore, the assembly of Geobacter pilins is predicted to be similar to that of
other type IV pilins.
The assembly of pilins involves a large number of proteins. At least 10 proteins have
been shown to be involved in the assembly of the TcpA pilus of Vibrio cholerae and 18 in P.
aeruginosa. The role of these proteins remains poorly understood [19, 50]. The pilus is
polymerized (protrusion) and depolymerized (retraction) using powerful molecular motors that
hydrolyze ATP to gain energy for pilin polymerization and depolymerization [19, 51]. Two
clusters in the genome of G. sulfurreducens are homologous to those encoding components of
the pilus apparatus in other Gram-negative bacteria [17]. The first cluster encodes the pilin
subunit (PilA), and proteins homologous to PilB (pilin polymerization), PilT (pilus retraction),
PilC (cytoplasmic membrane assembly protein), and regulatory proteins modulating pili
expression (the PilS sensor and the PilR response regulator) [17]. The second cluster encodes
PilQ (the outer membrane pilus secretin), PilP (the PilQ-stabilizing lipoprotein), PilO (pilin
glycosylation), and the two assembly proteins PilN and PilM. The gene encoding PilD (prepilin
peptidase) is found in an unlinked position in the genome of G. sulfurreducens [52].
Regulation of pilin expression and assembly also shares conserved mechanisms with
other type IV pili. Pilin expression is under control of the two-component signal transduction
system integrated by the PilR (response regulator) and PilS (sensor) proteins [37] whereas pili
assembly is growth-regulated [17, 43]. Pili are assembled during growth on insoluble Fe(III)
oxides but not on soluble Fe(III) [17]. They are also formed when cells are grown with fumarate
o

as an electron acceptor at suboptimal temperatures (25 C), suggesting that suboptimal growth,
rather than the presence of Fe(III) oxides, controls the assembly of the pili [17, 45].

11

The dynamic protrusion and retraction of type IV pili also enables pili-producing cells to
translocate on surfaces, a motile behavior termed twitching motility. However, this type of
surface motility has never been demonstrated in G. sulfurreducens [17]. Furthermore, a mutant in
the retraction pilus motor, PilT, had no defect when grown on Fe(III) oxides [17]. Several
observations pointed to a possible direct role of Geobacter pili in electron transfer to
extracellular electron acceptors. First, the pili were required for growth with insoluble Fe(III) and
Mn(IV) oxides as sole electron acceptors but not with soluble Fe(III). These results indicated that
pili might serve as a conduit between the cell and Fe(III) oxides that are heterogeneously
dispersed in the environment [17]. Second, the pili are also required to achieve cell stacking on
the anode electrode of microbial fuel cells and maximum power generation [22]. These results
indicated that the pili mediate long range electron transfer to extracellular electron acceptors [17]
and across anode biofilms [22].
Reguera et al. first elucidated the role of Geobacter pili in direct electron transfer through
conductive measurements [17]. Conductive-probe atomic force microscopy (CP-AFM) was used
to probe the conductivity of the pili of G. sulfurreducens. Pili and other proteins were
mechanically sheared off the outer membrane of G. sulfurreducens and deposited on highly
ordered pyrolytic graphite (HOPG) [17]. A voltage was applied to the AFM tip to measure the
distal conductivity of the proteins in the sample. The pilus filament of G. sulfurreducens showed
a strong current response to the voltage, whereas non-pilin proteins had no detectable current.
The results showed that the pili of G. sulfurreducens are highly conductive [17]. This is
consistent with the pili being the electrical connection between the cell and the surface of the
Fe(III) or Mn(IV) oxides [17].

12

Recent studies have also provided evidence in support of conductivity along the length of
the pili G. sulfurreducens [21]. Pili of G. sulfurreducens strain KN400 was sheared from the
outer surface of living cells and placed between a non-conductive 50 mm gap of two gold
electrodes and dried in a desiccator. A control sample was prepared using a KN400 strain where
the gene coding the PilA protein was deleted. Networks of wild type pili had a conductivity 10100 times higher than the control of PilA-deficient mutant or buffer alone. Conductivity was
attributed to the pili only because denaturation of cytochromes present in the sample after heat
treatment did not affect the results [21]. However, the effect of heat treatment in pili stability and
conductivity and the potential doping of the sample with metals released from the denatured
cytochromes were never assessed.
1.4.

Organization of this Dissertation

1.4.1. Characterization of Liposomes Prepared by Different Techniques: Measurement of
Rotational and Translational Diffusion of an Embedded Chromophore
Chapter 2 describes the effect the method of preparation has on selected properties of
bilayers. This project was carried out in collaboration with Dr. Gary J. Blanchard, Dr. Monique
M. Lapinski (a Ph.D. student in the Department of Chemistry at the time) and Dr. Aaron J.
Greiner (a Ph.D. student at the time in the Department of Chemical Engineering and Materials
Science at the time). Each of the collaborators made equal contributions to the work, which dealt
with the effect of sonication and extrusion on selected properties of reconstituted liposomes and
supporter bilayer lipid membranes. Liposome morphology, average size and size distribution
were determined by transmission electron microscopy (TEM) and dynamic light scattering
(DLS). Fluorescense lifetimes and rotational diffusion coefficients of fluorophores embedded in

13

liposomes were measured by time-correlated single photon counting (TCSPC). Translational
diffusion of sBLMs reconstituted from liposomes was evaluated by fluorescence recovery after
pattern photobleaching (FRAPP).
1.4.2. Mass-Production and In Vitro Assembly of Protein Nanowires Using Recombinant
Pilins From Geobacter Sulfurreducens
Chapter 3 describes the heterologous expression of truncated pilins and their assembly
into synthetic pili that retained the conductive properties of native pili from G. sulfurreducens.
This project was done in collaboration with Dr. Gemma Reguera from the Department of
Microbiology and Molecular Genetics and her Ph.D. student Rebecca Steidl. Rebecca and I
collaborated on the engineering of truncated versions of the pilin peptide. Expression of the
recombinant pilin peptide was investigated using the QIAexpressionist system, pMAL system,
and IMPACT system. Expression and purification conditions were analyzed using SDS-PAGE,
western blot, and MALDI-TOF mass spectrometry. The secondary structure of the recombinant
pilin was studied using circular dichroism. The in-vitro assembly of truncated pilins was
investigated and the resulting fibers imaged by TEM and atomic force microscopy (AFM). Distal
conductivity of in-vitro pili was assessed using scanning tunneling microscopy (STM). The STM
measurements were done in collaboration with Dr. Stuart Tessmer from the Department of
Physics and Ph.D student Shannon Demlow from the Department of Electrical and Computer
Engineering.
1.4.3. Study of Pilin Modifications to Enable Interfacing with Electrodes
Chapter 4 describes several strategies investigated to interface recombinant pilins with
gold electrodes. In one strategy, the effect of interaction of recombinant pilins with a
14

hydrophobic alkanethiol monolayer was investigated. In another strategy, a cysteine-modified
pilin peptide was engineered to facilitate interfacing with gold electrodes. Expression,
purification, and assembly of the cysteine-modified monomers were studied. Film thicknesses
were measured by ellipsometry. The insulating behavior of monolayers was evaluated using
cyclic voltammetry (CV). Real time deposition of cysteine-modified pilins was studied using
quartz crystal microbalance (QCM). In-vitro assembly of recombinant pilins was determined by
TEM and scanning electron microscopy (SEM).

15

Figure 1.1 Structural comparison of a) Pseudomonas aeruginosa PAK pilin and b) Geobacter
pilin. c) Structural alignment of PAK and Geobacter pilins. The αβ-loop is colored magenta, the
D-region is colored green, and the disulphides are colored cyan. The crystal structure of PAK
pilin was resolved by Craig et al.[49] and is available at the Protein Data Bank. The structure of
Geobacter pilin presented here was calculated by a computational model (Gemma Reguera,
personal communication, October 7, 2011). For interpretation of the references to color in this
and all other figures, the reader is referred to the electronic version of this dissertation.

16

Chapter 2.
CHARACTERIZATION OF LIPOSOMES PREPARED BY DIFFERENT
TECHNIQUES:

MEASUREMENT

OF

ROTATIONAL

AND

TRANSLATIONAL DIFFUSION OF AN EMBEDDED CHROMOPHORE

2.1.

Introduction
Cell membranes comprise a variety of components such as lipids, proteins and

carbohydrates embedded into a lipid bilayer. Membranes mediate or participate in various lifesustaining cellular processes such as synthesis of lipids and proteins, energy transference,
molecular transport, and enzymatic activity and cell-cell interactions [24, 25]. The use of model
systems to emulate the interactions occurring in natural environments is an active area of
research. Model bilayers have been used to simplify the study of cell membrane-related functions
[10, 53]. Moreover, model bilayers have found applications in electrochemical or optical
biosensors, drug screening and delivery, and production of catalytic interfaces [10, 53].
Small unilamellar vesicles (SUV), also known as liposomes, and supported bilayer lipid
membranes (sBLM) are two of the most widely used model systems [8, 24]. The two preferred
methods for liposome preparation are extrusion and sonication [15]. In extrusion, a lipid
suspension is forced through a polycarbonate membrane with a well-defined pore size to produce
17

vesicles with a characteristic diameter near the pore size of the membrane used in preparing
them. The main advantage of extrusion is that it produces liposomes with a reproducible mean
vesicle size and narrow size distribution [32]. In sonication, lipid suspensions are mixed using
acoustic energy from either a bath or a probe tip sonicator. The induced pressure breaks up the
larger multilamellar vesicles in the sample to form smaller unilamellar vesicles. The time over
which lipid solutions are sonicated determines the size of the vesicles, with the smallest radius
for phospholipid vesicles being in the range of 10.25 ± 0.55 nm, independent of the phospholipid
hydrocarbon chain length [54]. Sonication is less time consuming than extrusion, but the mean
diameter and size distribution are controlled by the duration of the process, which makes vesicle
sizes less reproducible from batch to batch [31, 55].
The objective of this work was to compare small unilamellar vesicles (SUV) prepared by
bath sonication and extrusion, to determine if the method of preparation has any effect on
selected bilayer properties. This is an important issue because it bears on whether experimental
data acquired for bilayer lipid membranes formed by the two methods can be compared directly.
We were interested in characterizing the morphology of liposomes, as well as the molecular
environment and mesoscopic properties of the lipid bilayers reconstituted from them.
Transmission electron microscopy (TEM) was used to study liposome morphology, and to
evaluate whether liposomes were unilamellar or multilamellar. The average size and size
distribution were determined by dynamic light scattering (DLS). The molecular scale and
mesoscopic properties were evaluated by measuring the rotational motion of a chromophore
incorporated into the liposomes using time correlated single photon counting (TCSPC). The
translational diffusion in a supported planar bilayer formed by liposome fusion was determined
by fluorescence recovery after pattern photobleaching (FRAPP). TCSPC experiments were done

18

in collaboration with Dr. Gary J. Blanchard of the Department of Chemistry.
2.2.

Materials and Methods

2.2.1. Preparation of Liposomes by Extrusion
Phospholipids 1,2-dioleoyl-sn-glycero-3-phosphocoline (DOPC), 1-oleoyl-2-[12-[(7nitro-2-1,3-benzoxadiazol-4-yl)amino]dodecanoyl]-sn-glycero-3-phosphocoline

(18:1-12:0

NBD-PC), and cholesterol were purchased from Avanti Polar Lipids Inc. (Alabaster, AL) and
used as received. The composition of the liposomes studied was 1 mM lipid solution containing
DOPC and NBD-PC at a ratio 98:2 mol/mol or DOPC/cholesterol/NBD-PC at a ratio 68:30:2
mol/mol/mol. The chloroform solvent was evaporated from the lipid mixture, and the dry film
was hydrated using HEPES buffer (Sigma-Aldrich, St. Louis, MO) containing NaCl (Fisher
Scientific, Fair Lawn, NJ). The buffer (10 mM HEPES, 150 mM NaCl, pH 7.4) was prepared
using purified water from a Milli-Q Plus water purification system (Millipore, Bedford, MA).
The mixtures were processed five times through a freeze-thaw-vortex cycle to ensure complete
mixing of the sample constituents. Each cycle consisted of immersion in liquid nitrogen for 5
o

min, followed by immersion in a 60 C water bath for 5 min, and then vortexing the thawed
sample for 2 min. After the freeze-thaw-vortex cycles were completed, the solutions were
extruded through two 400 nm pore diameter polycarbonate membrane filters using a mini
extruder (Avanti Polar Lipids Inc., Alabaster, AL). The pre-extruded suspension was then
extruded 11 times through two polycarbonate membranes (Avanti Polar Lipids Inc., Alabaster,
AL) with a nominal pore diameter of 100 nm. All extrusions were performed at room
temperature.

19

2.2.2. Preparation of Liposomes by Sonication
Lipid solutions with the same compositions as those used to prepare extruded liposomes
were used to form liposomes by sonication. The chloroform-solvated lipid mixtures were dried
o

under nitrogen and placed under vacuum at − 45 C for 3 h. The dried lipid was hydrated in
buffer (10 mM HEPES, 150 mM NaCl, pH 7.4) and sonicated for 20 minutes using a bath
ultrasonic cleaner (Branson 1510, Branson Ultrasonic Corporation, Danbury, CT). Fresh
liposome solutions were prepared prior to every experiment.
2.2.3. Dynamic Light Scattering
The mean diameter and size distributions of liposome mixtures of DOPC/NBD-PC with
and without cholesterol were determined by dynamic light scattering (DLS) using a Protein
Solutions DynaPro-MS/X system (Wyatt Technology Corporation, Santa Barbara, CA).
Liposome solutions were diluted by a factor of 10 with buffer and placed in a polyethylene
cuvette for measurement.
2.2.4. Transmission Electron Microscopy
A 10 µL aliquot of each sample was deposited on a Formvar nickel-coated grid and
stained with a solution containing 2% uranyl acetate in water. Images were acquired using a
transmission electron microscope (TEM, JEOL 100CX, Peabody, MA) operated at an
accelerating voltage of 100 kV.

20

2.2.5. Steady-State Spectroscopy
Steady-state excitation and emission spectra (Figure 2.1) were acquired on vesicle
samples for the purpose of characterizing the NBD fluorophore band positions. We used an
emission spectrometer (Spex Fluorolog 3, Edinson, NJ) for all measurements, set to a spectral
bandpass of 3 nm for both excitation and emission monochromators.
2.2.6. Time Correlated Single Photon Counting
Fluorescence lifetime and rotational diffusion of a tethered NBD chromophore embedded
in the liposomes were measured with time correlated single photon counting (TCSPC). The
experimental setup for TCSPC has been described in detail elsewhere [56, 57]. A schematic of
the TCSPC setup is shown in Figure 2.2. Briefly, it consists of a light source, CW mode-locked
Nd:YAG laser (Coherent Antares 76-s, Santa Clara, CA) that produces 1064 nm pulses for 100
ps at a 76 MHz repetition rate. The third harmonic of this light is used to excite a cavity-dumped
dye laser (Coherent 702-2, Santa Clara, CA) operated with Stilbene 420 dye (Exciton, Dayton,
OH) at 460 nm. The average output of the dye laser is approximately 25 mW, with 5 ps pulses
and a 4 MHz repetition rate. The beam is then split into a reference beam and sample beam,
which can be frequency-doubled based on the excitation wavelength of the chromophore. The
sample beam travels through a set of neutral density filters into a vertical polarizer before
o

reaching the sample. Emitted light is collected at 90 from the incident excitation beam. A Glano

Taylor prism is used to polarize the emitted light 54.7 with respect to the vertical excitation
o

polarization for fluorescence lifetime measurements. Polarizations parallel (0 ) and
o

perpendicular (90 ) to the vertically polarized incident light are collected for calculating

21

anisotropy. Sample emission is detected with a microchannel plate photomultiplier tube (MCPPMT, Hamamatsu R3809U-51, Bridgewater, NJ). The electronics utilized to temporarily resolve
the fluorescence transients include a constant fraction discriminator (CFD, Tennelec 454, Oak
Ridge, TN) and a time-to-amplitude converter (TAC, Tennelec 864, Oak Ridge, TN). The
collection wavelength (545 nm, 10 nm detection bandwith) and polarization were computer
controlled using LabVIEW 7.1 code.
2.2.7. Formation of Supported Bilayer Lipid Membranes
The reconstitution of supported bilayer lipid membranes was done by vesicle spreading
on fused silica slides. Fused silica microscope slides (75 mm × 25 mm × 1 mm) were purchased
from Technical Glass Products, Inc. (Painesville, OH). The substrates were cleaned by bath
sonication (Branson 1510, Branson Ultrasonic Corporation, Danbury, CT) in detergent solution
o

for 20 minutes, rinsed with deionized (DI) water, baked at 160 C for 4 h, and plasma treated
(Harrick Plasma, Ithaca, NY) with oxygen under vacuum (200 mTorr) for 10 min immediately
before bilayer deposition. Supported bilayers were deposited in a custom-made flow cell
described elsewhere [58]. The flow cell was initially washed with buffer, followed by 1 h
incubation with liposome solution and a final buffer wash to remove any unadsorbed liposomes.
All experiments were carried out at room temperature.
2.2.8. Fluorescence Recovery After Pattern Photobleaching
This technique, created by Smith and McConell [59], measures translational mobility of
fluorescently labeled lipids or proteins in natural or model supported lipid membranes [60, 61].
An image of alternating bright and dark stripes is projected onto the sample by passing an

22

expanded laser beam through a Ronchi ruling. A periodic variation of fluorescence intensity as a
function of space is created after permanent photobleaching of the fluorescent molecules [59].
Monitoring the subsequent fluorescence recovery provides a measure of lateral transport
(diffusion) of unbleached fluorophores into the bleached area [60, 62].
The experimental setup used to perform FRAPP has been described elsewhere and is
illustrated in Figure 2.3 [58]. Briefly, it consists of a custom made 1 mL flow cell attached to a
double syringe pump (Harvard Apparatus 551382, Holliston, MA), an inverted microscope
(Zeiss Axiovert 135M, Carl Zeiss, Thornwood, NY) with a 32X objective lens (Carl Zeiss,
Thornwood, NY), an argon ion laser (95, Lexel Laser, Fremont, CA), a side on photomultiplier
tube (Hamamatsu R4632, Bridgewater, NJ), a photon counter (SR400, Stanford Research
Systems, Sunnyvale, CA), and a fast preamplifier (SR445, Stanford Research Systems,
Sunnyvale, CA). The 488 nm laser line was directed through a 5X beam expander (Edmund
Optics, Inc. Barrington, NJ). A system of optical flats was used to switch between a low intensity
(20 µW) monitoring beam for continuous fluorescence detection and a high intensity (500 mW)
beam for photobleaching of fluorophores. A Ronchi ruling (50 lines per inch, Edmund Optics,
Inc., Barrington, NJ) placed in the back image plane of the microscope creates a fringe pattern on
the sample plane with a periodicity of 25 µm. An aperture placed in the image plane in front of
the PMT was used to restrict the observation area and prevent unbleached fluorophores from
outside the bleached patterned area from diffusing into the observation region during
fluorescence recovery measurements. The illuminated area was 200 µm in diameter, while the
observed area was 75 µm.

23

2.3.

Results and Discussion

2.3.1. Liposome Morphology, Mean Diameter, and Size Distribution
Transmission electron microscopy (TEM) images showed that liposomes prepared by
bath sonication and extrusion have a unilamellar structure (Figure 2.4). Dynamic light scattering
allowed us to determine the average size and size distribution of the liposomes. The mean
diameter of liposomes formed by extrusion with or without cholesterol was 112 ± 6 nm (Table
2.1 and Figure 2.5). Thus, cholesterol does not appear to affect the size of liposomes prepared by
extrusion. This is not surprising because the size of extruded vesicles should be fairly close to the
pore size of the polycarbonate membrane used to prepare them (100 nm). Liposomes prepared by
bath sonication had a mean diameter of 28 ± 2 nm without cholesterol and 64 ± 3 nm with
cholesterol (Table 2.1), indicating a significant influence of cholesterol on vesicle size. We also
observed that liposomes formed by sonication exhibited a bimodal size distribution, with or
without cholesterol. Previous reports have shown that liposomes formed by sonication generally
have a bimodal distribution [55]. Liposomes formed by extrusion also had a bimodal size
distribution that was more pronounced in cholesterol-containing liposomes (Figure 2.5). We
were not able to determine if a compositional difference exits between the two populations of
cholesterol-containing liposomes. The physical size and distribution information is not
necessarily correlated with the molecular scale properties of the bilayer structures.
2.3.2. Fluorescence Lifetimes and Rotational Diffusion in Liposomes
TCSPC permits measurement of time constants from the exponential decay of fluorescent
light and the time-resolved orientational anisotropy [63]. Fluorescence lifetime data are related to
the local environment experienced by the chromophore [64]. Anisotropy measurements estimate
24

the confined volume in which a chromophore rotates, as well as the rate of rotation within a
volume using a hindered rotor model.
Fluorescence anisotropy decay calculations relate the degree of polarization to the
rotational movement of the chromophore. It monitors the time dependence of the transition of a
system from a biased to a random orientation [65]. Induced orientational anisotropy r(t) is
defined as:

I ( t ) " I# ( t )
r( t ) = II
III ( t ) + 2I# ( t )

(2.1)

where III(t) and I⊥(t) are the emission intensities polarized parallel and perpendicular to the

!
vertically polarized incident excitation pulse, respectively [66, 67].
Kinosita et al. and Lipari and Szabo [63, 68] expanded the basic theory of anisotropy
decay developed by Wahl to the hindered rotation of fluorophores embedded in membranes [65].
This analysis is known as the hindered rotor model. A fluorophore in solution can assume all
possible orientations with equal probability; therefore the infinite-time anisotropy will decay to
zero. Since the NBD chromophore was tethered to a lipid in a bilayer lipid membrane, it does not
have the structural freedom to fully re-randomize its orientation within its environment following
excitation, which gives rise to an infinite time anisotropy [69, 70]. For these systems, the
function r(t) is given by
% t (
r( t ) = r(") + ( r(0) # r(")) exp'#
*
& $ HR )

(2.2)

where r(∞) is the infinite time anisotropy and r(0) is the zero time anisotropy. The value of r(0)
!
is determined by the angle between the excited and emitted transition dipole moments of the

chromophore, and r(∞) serves as a gauge for the orientational confinement that the chromophore

25

experiences. The reorientation time (τHR) is related to the orientational relaxation of the
chromophore within its confined volume, and is approximated by [67]

" HR =

2
7# 0
24Dw

(2.3)

where Dw is the wobbling or rotational diffusion coefficient. The semi-angle that defines the
!
confined cone volume of the chromophore (θ0) is related to the zero and infinite time

anisotropies by [67]
r(")
= 0.5cos # 0 (1+ cos # 0 )
r(0)

(2.4)

The liposome systems studied in the present work showed a fluorescent population decay
!
that can be fitted with a two-component exponential mode. The results indicate that NBD

experienced at least two environments that do not exchange populations on the time scale of the
excited-state relaxation. This behavior has been reported for a chromophore like NBD tethered to
phospholipids in a variety of environments. The fluorescence lifetime of NBD is sensitive to the
polarity of its immediate environment resulting in a shorter lifetime for NBD chromophores
residing in a more polar environment, closer to the polar phospholipid headgroup [71-73]. It has
been reported for NBD chromophores attached to an acyl chain that the chromophore can “loop
back” to position itself closer to the polar head group of the phospholipids [72].
The observation of two fluorescence lifetimes suggests that the fluorophore existed in
two discrete populations during the timescale of the measurement. The proposed environments
for NBD are only separated by a few angstroms. However our results showed a difference of
more than 4.5 ns between the two fluorescence lifetimes measured, which is consistent with the
fluorophore existing in a dielectric gradient that changes substantially over the length scale of the

26

fluorophore itself. It has been reported that the fluorescence lifetimes of identical molecules can
be affected by variations in the local dielectric environment around the molecule [64]. For that
reason, we interpreted our results to mean that the fluorophore effectively existed in only one
population and experienced one fluorescence lifetime. Since two fluorescence lifetimes were
measured experimentally, this is an indication that the fluorophore experienced a dielectric
gradient.
The two fluorescent lifetimes were essentially the same for all vesicle systems examined,
with τ1 ≈ 2 ns and τ2 ≈ 6 ns (Table 2.2). We note that the relative contributions of the two
lifetime components depend to some degree on vesicle composition for the sonicated vesicles
(0.05 in the absence of cholesterol versus 0.3 with cholesterol, as opposed to 0.3 for both
extruded vesicles; Table 2.2). This likely led to the large differences in the ratio of preexponential factors [A1/(A1 + A2) and A2/(A1 + A2)] obtained for sonicated vesicles,
particularly in the absence of cholesterol. However, owing to the structural complexity of these
vesicles and the comparatively limited information contained in the lifetime data, it would be
speculative to suggest that either the structure of the vesicle or the positional distribution of the
NBD chromophores accounts for the observed differences in pre-exponential factors.
Using the hindered rotor model, we determined the reorientation time (τHR), the cone
angle (θ0), and the wobbling diffusion coefficient (Dw). The cone angles were the same within
experimental uncertainty, regardless of the method of liposome preparation. We noticed that
liposomes containing cholesterol, either extruded or sonicated, have a cone angle smaller than
liposomes without (∼ 40

o

o

versus ∼ 60 ). These results are in agreement with previous

observations that addition of cholesterol decreases bilayer fluidity [74]. Our data point to a

27

slightly longer reorientation time (τHR) and a more restricted volume (θ0) for the cholesterolcontaining system. From these data, we found Dw to be the same for extruded and sonicated lipid
bilayers: Dw ≈ 100 MHz for liposomes without cholesterol and Dw ≈ 45 MHz with cholesterol. It
is interesting to note that there is an apparent difference in τHR for the cholesterol-containing
liposomes. This difference in time constants, while suggestive of a different environment formed
by the two methods, cannot be taken simply at face value. The observed time constant is related
to both the confining cone angle and the “wobbling” diffusion coefficient, Dw. The cone angles
recovered for these two liposomes are the same within experimental uncertainty, and when the
fundamental quantity Dw is extracted from the θ0 and τHR data, we find that there is no
statistically discernible difference for liposomes formed by extrusion versus those by sonication.
It is also important to emphasize that, in all of the anisotropy data, we do not find any
measurable differences between the molecular environments of the liposomes formed by
extrusion and sonication. While there is a slight difference in the fast component of the NBD
fluorescence lifetime for sonicated versus extruded liposomes containing cholesterol, this finding
is not reflected in the anisotropy decay dynamics, which are arguably a more sensitive probe of
the local environment of the chromophore.
2.3.3. Translational Diffusion in Supported Bilayer Lipid Membranes
Supported bilayer lipid membranes (sBLM) were reconstituted through vesicle
adsorption and fusion onto fused silica substrates. A representative fluorescence recovery plot of
a DOPC supported bilayer containing NBD-PC as a probe molecule is shown in Figure 2.6. The

28

translational diffusion coefficients and mobile fractions of fluorophores in sBLMs can be
extracted from the recovery data by using the model developed by Starr and Thompson [75]
* $ ' $
$
2 '
2 '.
,
m
8
& " 4 # Dt ) + 1 exp&" 36# Dt ),
f (t ) = f (0) + 1" f (0) +1" & ) exp
/
& )
&
2 ) 9
2 ),
2
, %# 2 ( &
a
a
%
(
%
(0
-

[

]

(2.5)

where f(t) is the time-dependent post-bleach fluorescence (t > 0) normalized with respect to the
! pre-bleach fluorescence (t < 0), with t = 0 being the time of the bleach pulse, D the translational

diffusion coefficient of the chromophore within the bleached area, m the mobile fraction, and a
the stripe periodicity of the Ronchi ruling. This equation assumes one mobile species
(fluorophores diffusing in the fluid sBLM) and one immobile species (fluorophores confined in
intact, unruptured vesicles adsorbed to the surface). The data were fit with a LevenbergMarquardt algorithm for non-linear least squares using OriginPro 7.5 software (OriginLab
Corporation, Northampton MA).
The data were fit to equation 2.5, and the values for the diffusion coefficient and mobile
fraction were determined (Table 2.4). Results indicate that D and m are the same for a given
supported bilayer composition, and independent of the method of liposome preparation.
Therefore, the differences in the size of liposomes formed by sonication and extrusion do not
affect the fluidity of the resulting sBLM. These results are in agreement with a previous quartz
crystal microbalance (QCM) study showing that liposomes ranging in size from 25 to 200 nm
can form viable bilayer lipid membranes [12]. As expected, the presence of cholesterol in the
bilayer served to make the chromophore less mobile (i.e., the BLM is less fluid). This result is
fully consistent with the anisotropy decay data presented earlier, as well as previous findings that
addition of cholesterol to a phospholipid bilayer decreases the translational diffusion of the
phospholipids [76].

29

2.4.

Conclusions
This project focused on comparing selected properties of liposomes prepared by bath

sonication and extrusion. The results demonstrated that, while the mean diameter and size
distributions of the liposomes depended on the manner by which they were prepared, their
molecular and nanoscale organization, studied through fluorescence lifetime, anisotropy decay,
and translational diffusion measurements, are independent of the method of preparation.
Liposomes formed by the two schemes were characterized in solution and in supported bilayer
lipid membranes formed by liposome fusion on a fused silica substrate. The effect of cholesterol
on the properties of the liposomes and sBLM was also evaluated. The findings indicate that
whether liposomes are formed by sonication or extrusion does not affect the molecular scale
organization of these systems. The molecular scale organization of the bilayers is determined by
interactions between the constituent species, which themselves do not depend on the manner in
which the bilayer is formed. Consequently, all other factors being equal, data on bilayer systems
formed by sonication can in fact be compared directly to those on systems formed by extrusion.

30

Table 2.1 DLS results for extruded and sonicated liposomes for two systems: (1) 98 mol%
DOPC and 2 mol% NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBDa
PC. Data are means ± SD of three samples.

Composition

Mean diameter sonicated
vesicles (nm)

(1) DOPC/NBD-PC

112 ± 6

28 ± 2 and 196 ± 10

(2) DOPC/Cholesterol/NBD-PC
a

Mean diameter extruded
vesicles (nm)

112 ± 6 and 600 ± 30

64 ± 3 and 259 ± 13

Uncertainties in the data are estimated to be ± 5% of the average values.

Table 2.2 Ratio of pre-exponential factors (Ai) and fluorescence lifetimes (τi) of NBD-PC
fluorophores in sonicated and extruded liposomes comprising (1) 98 mol% DOPC and 2 mol%
a
NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC. Data represent
means ± SD of three liposome solutions.

Composition

τ1 (ns)

A2/(A1 + A2)

τ2 (ns)

(1) Extruded
DOPC/NBD-PC

0.44 ± 0.02

1.85 ± 0.10

0.56 ± 0.02

6.01 ± 0.06

1) Sonicated
DOPC/NBD-PC

0.06 ± 0.01

1.93 ± 0.26

0.94 ± 0.01

6.10 ± 0.07

(2) Extruded
DOPC/Chol/NBD-PC

0.27 ± 0.01

1.67 ± 0.09

0.73 ± 0.01

6.04 ± 0.08

(2) Sonicated
DOPC/Chol/NBD-PC
a

A1/(A1 + A2)

0.31 ± 0.03

2.30 ± 0.18

0.69 ± 0.03

5.91 ± 0.10

Lifetime data are best fit results for the function f ( t ) = A1 exp("t #1 ) + A2 exp("t # 2 ) .

!

31

Table 2.3 Reorientation times (τOR), cone angles (θ0), and wobbling diffusion coefficients (DW)
of NBD-PC fluorophores in sonicated and extruded liposomes comprising (1) 98 mol% DOPC
and 2 mol% NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC.
Data represent means ± SD of three liposome solutions.

θ0 (

o

)

Composition

τOR (ns)

(1) Extruded
DOPC/NBD-PC

3.03 ± 0.22

62 ± 5

113 ± 27

1) Sonicated
DOPC/NBD-PC

2.94 ± 0.09

57 ± 5

98 ± 21

(2) Extruded
DOPC/Chol/NBD-PC

3.59 ± 0.28

42 ± 3

44 ± 10

(2) Sonicated
DOPC/Chol/NBD-PC

3.00 ± 0.22

37 ± 2

41 ± 8

DW (MHz)

Table 2.4 Translational diffusion coefficients (DL) and mobile fractions for NBD-PC in sBLMs
prepared with sonicated and extruded liposomes comprising (1) 98 mol% DOPC and 2 mol%
NBD-PC and (2) 68 mol % DOPC, 30 mol% cholesterol, and 2 mol% NBD-PC. Data represent
a
means ± SD of three liposome solutions.
2

Composition

DL (µm /s)

Mobile fraction

(1) Extruded
DOPC/NBD-PC

2.24 ± 0.25

0.78 ± 0.18

1) Sonicated
DOPC/NBD-PC

2.02 ± 0.15

0.81 ± 0.11

(2) Extruded
DOPC/Chol/NBD-PC

1.65 ± 0.22

0.73 ± 0.10

(2) Sonicated
DOPC/Chol/NBD-PC

1.72 ± 0.04

0.69 ± 0.06

a

All values are averages over at least three different samples, and uncertainties are standard
deviations of the mean.

32

Figure 2.1 Steady-state solution phase absorption and emission spectra of vesicles containing
2% NBD-PC and 98% DOPC. Both spectra have been normalized for clarity of presentation.

33

Figure 2.2 Schematics of the experimental setup for time correlated single photon counting (TCSPC).

34

Figure 2.3 Schematics of the experimental set-up for fluorescence recovery after pattern photobleaching (FRAPP).

35

Figure 2.4 Transmission electron microscopy images of liposomes formed by extrusion and
sonication: a) TEM image of extruded vesicles without cholesterol; b) TEM image of sonicated
vesicles without cholesterol.

36

Figure 2.5 Size distribution obtained by dynamic light scattering showing diameters of
liposomes as a function of method of preparation and lipid composition: a) extruded DOPC
liposomes, b) extruded liposomes containing DOPC and cholesterol, c) sonicated DOPC
liposomes, and d) sonicated liposomes containing DOPC and cholesterol.

37

Figure 2.6 Representative FRAPP recovery curve and residual plot (below) for DOPC/NBD-PC
supported bilayer lipid membranes reconstituted from extruded liposomes. Similar recovery
curves were obtained for bilayer lipid membranes reconstituted from sonicated liposomes.
Photobleaching time was 300 ms.

38

Chapter 3.
MASS-PRODUCTION

AND

IN-VITRO

ASSEMBLY

OF

PROTEIN

NANOWIRES USING RECOMBINANT PILINS FROM GEOBACTER
SULFURREDUCENS

3.1.

Introduction
Bionanotechnology is an emerging area of research that applies biological principles or

integrates biological components to advance nanotechnology and molecular biology. Biological
molecules possess intrinsic characteristics that make them attractive for the fabrication of high
performance nanodevices [1, 77]. In particular, proteins, nucleic acids, and phospholipids have
precisely-controlled structures, specific functionalities, and capabilities for self-assembly [5], and
they are all being actively investigated to design new interfaces for biosensors, drug delivery, and
catalytic applications [77, 78].
The protein nanotubes or pili produced by the bacterium Geobacter sulfurreducens show
promise for the development of nanodevices. Not only do they have self-assembly capabilities but
they are also electrically conductive [17]. Although these protein nanowires can be purified from
the native organism, yields are low because only a few nanowires are produced per cell and
because, even under optimized culturing conditions, these organisms cannot be grown to the high
39

cell densities often required in industrial settings [23]. Furthermore, culturing the native organism
under pili-inducing conditions requires strict anaerobiosis, controlled temperature and specialized
culturing vessels, which increases the cost associated with their natural production. These
considerations greatly limit practical applications and commercialization of the protein nanowires.
We are interested in studying possible applications of protein nanowires in the
development of novel nanostructured devices. This requires the development of cost-effective
methods for pili production at high yields. Pili are homopolymers of a pilin subunit (PilA) that
self-assemble and disassemble using molecular motors that hydrolyze adenosine triphosphate
(ATP) [19, 51]. The first step to integrating pili proteins into nano-devices is to produce pilin
monomers in an efficient, reproducible, and controlled manner.
Audette et al. successfully expressed recombinant pilin subunits from Pseudomonas
aeruginosa strain K122-4 (termed ∆K122-4) by deleting 28 residues from the peptide’s highly
hydrophobic N-terminus to enhance solubility [79]. The recombinant pilins were expressed in the
periplasmic space of Escherichia coli as part of a maltose-binding (MBP)-pilA(Δ1-28) fusion
protein [79]. These studies demonstrated that recombinant protein expression can be used to
mass-produce pilins in a controlled environment. However, a similar approach in Geobacter is
challenging because the nanowire subunits are much more hydrophobic and contain amino acids
1

implicated in electron transfer in the most N-t region (e.g., a tyrosine in position 27) . Such long
truncations are also challenging because they can affect the natural self-assembly process and
generate pilus fibers with structural and functional differences compared to the native ones.
The truncated K122-4 pilin subunit aggregated into soluble high molecular weight
structures in the presence of alkyl thiol molecules, and chromatography resins such as Superdex.
1

Gemma Reguera, personal communication, October 7, 2011.
40

Transmission electron microscopy (TEM) showed that the aggregates formed hair-like structures
similar to pili. Modeling of wild type IV pili and recombinant ∆K122-4 pili predicted the outer
diameter (∼ 60 Å) to be similar. However, the recombinant ∆K122-4 pili had a wider inner
channel (∼ 20 Å), which can affect the structural integrity of the protein nanotubes. The ∆K122-4
pili were also longer than the conventionally purified wild-type P. aeroginosa pili, probably
because of the lack of pilus assembly machinery in vitro versus in vivo [48]. Further studies
demonstrated that ∆K122-4 protein nanotubes retain certain functionalities of the natural pilus
fiber such as the ability to bind to single and double-stranded DNA [80].
The ability to mass-produce biologically active recombinant pili from G. sulfurreducens
is critical towards advancing their application in the development of biosensors, biocatalysts, and
nanoelectronic devices. Based on the previous studies with pilins from P. aeruginosa, we
hypothesized that recombinant pili from G. sulfurreducens could be generated without affecting
their conductivity. However, it is challenging to truncate the pilin’s N-terminus to the degree
required to increase its solubility and recombinant expression while maintaining the amino acids
critical for self-assembly and conductivity. Furthermore, recombinant expression prevents posttranslational modification of the pilin, which could affect the self-assembly and the conductive
properties of the recombinant pili.
In the present study, the expression of recombinant pilin subunits of G. sulfurreducens
and their in vitro assembly were investigated. Because of the low solubility of the full-length
pilin, several truncated versions of PilA (termed PilAn) were engineered to enhance solubility.
Several systems were tested to optimize expression and purification of pilin monomers. Using this
approach, a truncated pilin was mass-produced as a soluble peptide. The secondary structure of
the recombinant pilin was investigated using circular dichroism (CD). We also investigated the

41

effect of addition of surfactants on pilin conformation. Finally, recombinant subunits were
assembled into fibers and imaged using transmission electron microscopy (TEM) and atomic
force microscopy (AFM.) to assess structural features, and by scanning tunneling microscopy
(STM) to investigate their conductive properties.
3.2.

Materials and Methods

3.2.1. Recombinant DNA techniques
Production of pilin subunits from G. sulfurreducens was investigated using three different
expression systems. All of the systems produced a fusion protein by attaching an affinity tag to
the pilA sequence expressed. The use of these tags simplifies purification and, in some cases,
TM

improves solubility. The systems considered were: a) QIAexpressionist

TM

hexa polyhistidine-tag (6×His-Tag) (Qiagen, Valencia, CA); b) pMAL

, which attaches a
protein fusion and

purification system, which attaches a maltose binding protein (MBP) tag (New England Biolabs
Inc., Ipswich, MA); and c) Intein-mediated purification system, which attaches a chitin binding
domain (IMPACT

TM

, New England Biolabs Inc., Ipswich, MA).

Due to the low aqueous solubility of the pilin monomer (PilA), truncated versions of PilA
(termed PilAn, were n is the number of amino acids removed from the hydrophobic N-terminus)
were engineered to enhance solubility. The truncated pilin gene was extracted from the genomic
o

o

DNA of G. sulfurreducens. PCR amplification (30 cycles, 30 s at 95 C, 20 s at 50.7 C, and 1
o

min at 72 C) was performed using 100 ng genomic DNA and Platinum® PCR SuperMix High
Fidelity (Invitrogen, Carlsbad, CA). DNA was isolated from a 2% agarose gel using

42

Zymoclean® Gel DNA Recovery Kit (Zymo Research, Irvine, CA) and ligated in-frame into the
expression vector. Table 3.1 shows the primers used to amplify the DNA and the restriction sites
used to clone the pilAn gene into the expression vectors. Plasmids were then transformed into
DH5α E. coli cells for propagation. The recombinant plasmids were confirmed by digestion and
sequencing.
3.2.2. Expression of Recombinant Pilins
Expression plasmids were transformed into competent Escherichia coli host cells (see
Table 3.2) for expression. To test expression and optimize growth conditions of recombinant
proteins, small-scale shake-flask experiments (∼ 50 mL cultures) were carried out using LuriaBertani (LB) media. Glucose at a concentration of 2 g/L was added to LB media when using
vectors from the p-MAL expression system to repress the maltose genes on the chromosome of
the E. coli cells. Antibiotics were added where appropriate to the following final concentrations:
ampicillin (Amp), 100 µg/mL; chloramphenicol (Cam), 34 µg/mL; kanamycin (Kan), 50 µg/mL.
o

In general, cells were grown at 37 C and 200 rpm and induced with isopropyl-1-thio-Dgalactopyranoside (IPTG) when OD600 was between 0.4 and 0.6. After induction, the
o

temperature was lowered to 30 C. Cells were harvested 6 h later by centrifugation at 5000 x g
o

o

for 20 minutes at 4 C, and the cell pellet was stored at -70 C.
After confirming expression and optimizing conditions, large-scale expression was
carried out in a batch reactor. In these experiments, an overnight culture of transformed E. coli
cells was used to inoculate a fed-batch bioreactor (Bioflo 3000, New Brunswick Scientific,
Edison, NJ) containing 5 L yeast-enriched Hagem medium (HM) media plus ampicillin and

43

chloramphenicol. The HM medium contained 9 g/L potassium phosphate monobasic KH2PO4, 6
g/L potassium phosphate dibasic K2HPO4, 6 g/L sodium phosphate dibasic Na2HPO4, 3 g/L
ammonium phosphate dibasic (NH4)2HPO4 , 5 g/L yeast, 1 g/L magnesium sulfate MgSO4, 10
g/L glucose, and 5 mL/L of trace elements [81]. The solution of trace elements contained 10 g/L
iron (II) sulfate FeSO4, 2.5 g/L zinc sulfate heptahydrate ZnSO4.7H2O, 1 g/L manganese sulfate
pentahydrate MnSO4.5H2O, 1 g/L cobalt (II) chloride hexahydrate CoCl2.6H2O, 1 g/L sodium
molybdate dihydrate Na2MoO4.2H2O, 0.2 g/L sodium borate decahydrate Na2B4O7.10H2O, and
5 g/L calcium chloride dihydrate CaCl2.2H2O dissolved in 5 M hydrochloric acid HCl.
Temperature, pH, glucose feeding, and dissolved oxygen were controlled with a PID
(proportional–integral–derivative) controller [82]. Cells were grown overnight (∼16 to 20 h) at
o

37 C. Glucose depletion was indicated by a spike in dissolved oxygen (DO) concentration. A
solution containing 50% (w/v) glucose and 20 g/L MgSO4 was fed on demand to replenish
depleted glucose. The pH was controlled at 7.0 via automatic addition of concentrated
ammonium hydroxide NH4OH. Dissolved oxygen (DO) was kept at 10% during the
fermentation process by increasing agitation speed from 50 to 1000 rpm, then increasing the
airflow rate, and finally by controlling the glucose feeding rate. After overnight growth,
expression of the recombinant protein was induced with 0.5 mM IPTG, and the temperature was
o

lowered to 30 C. Six hours after induction, cells were harvested by centrifugation at 5000 x g for
o

o

20 minutes at 4 C. The cell pellet was stored at -70 C. Expression of the fusion protein was

44

tested by SDS-PAGE and western blot.
3.2.3. Purification of Recombinant Pilins
Recombinant pilins were purified by affinity chromatography. The type of resin used
varied with the expression system. The QIAexpressionist
acid (Ni-NTA) resin; the pMAL

TM

TM

system used nickel-nitrilotriacetic

system used amylose resin; and the IMPACT

TM

system used

chitin beads. Figure 3.1 shows the purification steps for each system.
In all cases of cytoplasmic expression, harvested cells were resuspended in lysis buffer
(10 mL buffer per gram of cells) and lysed using a tip sonicator (Sonifier 250, Branson,
Danbury, CT). The composition of all buffers used during purification can be found in Table 3.3.
The crude extract was sonicated on ice 7 times for 10 s. Cell extract was clarified by
o

centrifugation at 20,000 x g for 30 min at 4 C. The supernatant was loaded into the affinity
column to allow binding of the target protein. Vector p4x from pMAL

TM

system produced

periplasmic expression of the MPB-PilAn fusion protein. In this case, the harvested cell pellet
was subjected to osmotic shock to release periplasmic proteins. The osmotic shock fluid was
loaded into the affinity column. After binding, a wash buffer was pumped through the column to
remove unwanted proteins that were bound non-specifically to the column. The subsequent steps
to recover the target protein varied depending on the purification system used.
TM

For the QIAexpressionist

system, His-tagged pilin was eluted from the column with a
TM

buffer containing 300 mM imidazole. For the pMAL

system, the MBP-PilAn fusion was

eluted with a buffer containing 10 mM maltose. After elution, the MBP tag was cleaved from the

45

pilin using Factor Xa protease. For the IMPACT

TM

system, the PilAn subunit was cleaved from

the chitin-bound intein tag (CBD) by incubating the column with a buffer containing 50 mM
o

o

dithiothreitol (DTT). Cleavage efficiency was tested at 4 C or 23 C for 24 to 72 h. After
cleavage, the pilin subunits were eluted from the column using a storage buffer. Production of
pilin subunit was confirmed by SDS-PAGE, and MALDI-TOF mass spectrometry.
3.2.4. Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE)
Protein expression and cleavage efficiency were evaluated using Tris-glycine (Biorad,
Hercules, CA). Cell pellets from 100 to 1000 µL of cell culture were dissolved in a loading
buffer (100 mM Tris pH 6.8, 20% glycerol, 8 M urea, 2% (w/v) sodium dodecyl (SDS), 0.02%
bromophenol blue). Liquid samples (e.g. cell lysate, osmotic shock fluid, elution fractions) were
dissolved in a 1:1 ratio using a loading buffer. Production of pilin was confirmed as bands
having the right molecular weight as the pilin peptide appeared on Tris-tricine gels (Biorad,
Hercules, CA). Samples for Tris/tricine gels were prepared by dilution with the same volume of
Tricine sample buffer ((Biorad, Hercules, CA). Gels were run in a Mini Trans-Blot cell system
(Biorad, Hercules, CA). Tris/glycine gels were run using Tris/glycine/SDS buffer (25 mM Tris,
192 mM glycine, 0.1% w/v SDS). Tris/tricine gels were run using Tris/tricine/SDS buffer (100
mM Tris, 100 mM Tricine, 0.1% w/v SDS). The gel running conditions were as follows: 30 min
at 200 V for Tris-glycine gels, and 120 min at 100 V for Tris-tricine gels. Gels were stained with
Bio-safe Coomassie (Biorad, Hercules, CA) for one hour and de-stained in water until bands
were clearly visible. Tricine gels required an additional 30 minutes fixing step with a solution of
50% methanol, 10% acetic acid in water prior to being stained.

46

3.2.5. Western Blot Analysis
Proteins were transferred from acrylamide gel to a polyvinylidene fluoride (PVDF)
membrane (Millipore, Billerica, MA) using a Trans-Blot SD semi-dry transfer cell (Biorad,
Hercules, CA). After transfer, the PVDF membrane was blocked with 5% (w/v) non-fat dry milk
in Tris buffered saline and tween 20 buffer (TBST, 20 mM Tris pH 7.6, 137 mM NaCl, 0.1%
o

o

Tween 20) for at least 1 hour at 4 C. The membrane was incubated for 1 hour at 23 C with the
primary antibody. After three washes of 5 min each with TBST buffer, the membrane was
o

incubated with the secondary antibody for 1 h at 23 C, and then the TBST washing step was
repeated. Proteins were visualized after staining with 3,3’,5,5’-tetramethylbenzidine (TMB)
(Thermo Scientific Pierce, Rockford, IL). Antibodies were dissolved in blocking solution.
Primary antibodies, anti-His and anti-MBP, were used at 1:5000 dilutions (New England Biolabs
Inc., Ipswich, MA). Secondary antibody, goat anti-mouse IgG-HRP linked, was used at 1:5000
dilutions (Sigma-Aldrich, St. Louis, MO).
3.2.6. Matrix Assisted Laser Desorption-Ionization-Time of Flight Mass Spectrometry
(MALDI-TOF MS)
A 1 µL of sample was mixed with 1 µL of 50 mM 3,5-dimethoxy-4-hydroxycinnamic
acid in 50% acetonitrile CH3CN/0.5% trifluroacetic acid (TFA) and dried on a sample plate.
Mass spectra were collected on a time-of-flight (TOF) Voyager-DE Pro-MALDI-TOF (Applied
Biosystems, Framingham, MA).

47

3.2.7. Circular Dichroism (CD)

Circular dichroism (CD) spectra were acquired on a Chirascan

TM

spectrometer (Applied

Photophysics Ltd., Leatherhead, United Kingdom). Spectra were recorded from 190 to 260 nm at
0.5 nm increments and a 5 second integration time using a quartz cuvette with a 0.1 cm path
length (Starna Cells Inc., Atascadero, CA), and peptide concentrations ranging from 65 µg/mL to
41 µg/mL. Measurements were performed in 10 mM potassium acetate buffer at pH 3.8 with 50
mM Na2SO4 and 10 mM potassium phosphate buffer at pH 7.0 with 50 mM Na2SO4. The effect
of the detergents sodium dodecyl sulfate (SDS) and octyl β-D-glucopyranoside (OG) on the
secondary structure of the pilin monomers was investigated. The spectra were baseline-corrected
and smoothed using a third order Savitsky-Golay filter. Spectral data are presented in units of
molar ellipticity per residue, [θ]. To convert CD instrument units (θ, millidegrees) into mean
residue molar ellipticity [θ] units, the following equation was used [83]:

$ " # 0.1# MRW '
)
c#l
(

[" ] = &
%

(3.1)

where c is the peptide concentration in mg/mL, l is the path length of the cuvette in cm, and
!
MRW is the mean residue weight of the sample, calculated by:

MRW =

MW
n "1

(3.2)

where MW is the peptide molecular mass in Daltons, and n is the number of amino acid residues

!
(42 for the PilA19 monomer).
The peptide concentration was determined from the difference spectrum of the protein
dissolved in 6 M guanidine hydrochloride at pH 12.5 versus pH 7.1 [84]. The baseline was

48

scanned from 320 to 270 nm with 1 mL of the pH 7.1 guanidine solution in the reference
compartment and 1 mL of the pH 12.5 guanidine solution in the sample compartment. The same
volume of peptide sample was added to each cuvette, and the difference spectrum was obtained
and baseline-corrected. The concentration was determined from absorbance at 294 nm using the
known amino acid composition of the PilA19 subunit and the reported values of the molar
extinction coefficients for tyrosine and tryptophan residues [85], using the equation below:
c=

A293
2,357Y + 830W

(3.3)

where Y is the number of tyrosines and W is the number of tryotophans, which are three and
!
zero, respectively, for the PilA19 subunit.
3.2.8. In-vitro Assembly of Recombinant Pilins into Pilus Fibers
After purification, recombinant pilin subunits were exchanged from buffer solution to a
mixture of 80% acetonitrile and 20% methanol. The exchange of the pilins from buffer to
organic solvent was done using the Oasis MAX extraction cartridges, 60 mg sorbent (Waters
Corporation, Milford, MA). The column was wetted with 1 mL of 100% acetonitrile, followed
by 1 mL wash with DI water. Before loading, the pH of the protein solution was adjusted above
10 with 1 mM NaOH. Approximately 2 mL of buffer containing pilin was passed through the
column. The column was washed with 1 mL solution of 5% NH4OH. The pilin was eluted with 1
mL solution of 20% methanol and 80% acetonitrile. Samples were dried with a SpeedVac
concentrator. The dry sample was re-suspended in filtered deionized (DI) water or 10 mM CHES
buffer at pH 10.0. The re-suspended sample was imaged by TEM and AFM to confirm the
assembly of PilA19 subunits into pili.

49

3.2.9. Transmission Electron Microscopy (TEM)
A 10 µL aliquot of sample was deposited on a Formvar nickel-coated grid, and stained
with a solution containing 2% uranyl acetate in water. Images were acquired using a JEOL 1000
CX transmission electron microscope (JEOL, Peabody, MA) operated at an accelerated voltage
of 100 kV.
3.2.10. Atomic Force Microscopy (AFM)
Pili fibers were adsorbed on freshly cleaved mica and imaged with a Nanoscope IV
Multimode scanning probe microscope (Veeco, Santa Barbara, CA) equipped with a J scanner.
Images were collected in tapping mode using commercially available silicon or silicon nitrite
cantilevers.
3.2.11. Scanning Tunneling Microscopy (STM)
We assessed the conductive properties of the recombinant pili using a table-top Nanosurf
EasyScan STM. Members of the research group of our collaborator, Dr. Stuart Tessmer, carried
out the STM measurements. This system is described in detail elsewhere [86]. Briefly, the
sample was firmly attached to a conductive cylinder that rested on a rail system. The tip was held
by a conductive clamp that also served as a tunneling current lead. The sample holder was
moved towards the fixed tip via a linear piezo motor. The sample was grounded and the voltage
was applied just to the tip. Vibration isolation was implemented by a rubber damping stack that
suppresses high frequencies and a single spring suspension stage that suppresses frequencies
lower than 50 Hz. The STM was placed inside a plastic glove bag through which a constant flow
of dry nitrogen is supplied. Even though this microscope is easy to use and can achieve atomic

50

level resolution, it has some disadvantages. Mainly, the scan range (500 nm × 500 nm) is small
and the high propensity for thermal drift makes the image not reproducible from one scan to the
next.
After corroboration of recombinant pili assembly via TEM, a 10 µL drop of the sample
was deposited on a freshly cleaved layer of highly oriented pyrolytic graphite (HOPG). The
sample was left on the HOPG for at least 30 minutes to allow pili diffusion to and adsorption
onto the surface. The HOPG substrate was then rinsed with DI water, and excess moisture was
removed using filter paper. STM topography measurements of the recombinant pili were done at
o

room temperature (∼ 23 C).

3.3.

Results and Discussion

3.3.1. Production of Recombinant Pilins
The high hydrophophicity of the nanowire pilin greatly limits its mass-production in a
heterologous host. The pilin’s hydrophobicity is illustrated using a Kyte and Doolittle [87]
hydropathicity plot (Figure 3.2). This method calculates the average hydropathicity of a moving
segment as it advances through the sequence from the amino to the carboxy terminus. It uses a
hydropathicity index scale that assigns hydrophobicity values to each amino acid in the protein
segment under consideration. Hydrophilic amino acids give negative hydropathicity values, and
hydrophobic amino acids give positive values. The plot presented was evaluated using a 19residue segment algorithm. The Kyte and Doolittle plot confirms the trans-membrane nature of
the pili. This method establishes that when the hydropathicity of a given 19-residue segment
averages a value greater than 1.6, there is a high probability that this sequence is part of a
membrane-bound protein spanning the membrane. This criterion is met up to position 18 of PilA

51

(red dashed line in Figure 3.2). Figure 3.2 also shows hydropathicity values for the truncations
investigated in this project. The GRAVY index is the average hydropathicity score for all the
amino acids in a protein, where a positive GRAVY index indicates a protein is hydrophobic,
while a negative GRAVY index indicates a protein is hydrophilic [87]. A PilA truncation of 10
amino acids yields a protein having a GRAVY index that is lower than the full length PilA, but is
still positive indicating that some hydrophobicity is still present. Other truncations, PilA19,
PilA20, and PilA22, have negative indexes indicating a change from a hydrophobic to a
hydrophilic nature.
Based on this analysis of pilin hydrophobicity, truncated versions of the full length PilA
were genetically engineered to overcome pilin hydrophobicity and improve solubility. Initially a
20-amino-acid truncation of the PilA subunit (named PilA20) was selected, and the expression
and purification conditions were investigated using the QIAexpressionist (Qiagen, Valencia, CA)
expression system. A hexa polyhistidine-tag was attached to the N-terminal PilA20 subunit. A
6×His-Tag facilitates purification via (Ni-NTA) metal chromatography but generally does not
interfere with the protein’s folding and function [88]. The theoretical molecular weight of the
6×His-PilA20 is 6,590 Da. Figure 3.3 shows SDS-PAGE and western blot analysis carried out
after expression and purification of the 6×His-PilA20. Appearance of a band with the right
molecular weight as the 6×His-PilA20 pilin confirmed the expression of the truncated pilin.
However this band appeared in induced as well as non-induced samples, which indicated leaky
expression of the recombinant peptide. In addition, several bands with molecular weight larger

52

than the 6×His-PilA20 pilin were observed in the western blot, which suggested that the pilins
might be aggregating or self-assembling.
To improve the solubility of the PilA20 subunits being expressed, a maltose binding
protein (MBP) tag was used. The MBP tag is a large protein secreted by E. coli cells that can be
expressed at high levels and helps keep proteins attached to either the N- or C- terminus soluble.
The MBP also binds strongly to amylase resin, enabling the fusion protein to be purified via
affinity chromatography [88]. Recombinant pilin from P. aeruginosa [79] was successfully
expressed as a MBP-PilA fusion protein. The pMAL

TM

protein fusion and purification system

(New England Biolabs Inc., Ipswich, MA) provide vectors where the cloned gene is inserted
down-stream from the malE gene, which encodes the MBP resulting in the expression of an
MBP fusion protein. This system offers the flexibility to use either cytoplasmic or periplasmic
expression depending on the vector used. The maltose binding protein (MBP) was attached to the
hydrophobic N-terminal end of the PilA20 subunit. Expression of the MBP-PilA20 fusion was
attempted both cytoplasmically and periplasmically. Both constructs were transformed into K12
TB1 E. coli cells.
Periplasmic expression of a MBP fusion after induction, osmotic shock, and recovery of
periplasmic proteins was confirmed by SDS-PAGE and western blot analysis (Figure 3.4). The
molecular weight of the fusion protein eluted from the amylase column was determined by
MALDI-TOF (see Figure 3.4) to be 43,031.9 Da. This value is smaller than the theoretical value
for the periplasmic MBP-PilA20 fusion (46,902.6 Da). Certain proteins have been reported to be
incompatible with trans-location to the periplasmic space, as evidenced by their degradation
during expression. Another possibility is that the intact fusion could have been degraded by

53

periplasmic proteases [89]. Protease inhibitors and protease deficient E. coli strains were used in
subsequent experiments without any changes in the outcome.
SDS-PAGE and western blot analysis after cytoplasmic expression confirmed the
production of a MBP fusion protein (Figure 3.5). After purification of the fusion, two bands that
tested positive against anti-MBP antibody were observed in the elution from the amylose column
(Figure 3.5). MALDI-TOF mass spectrum of the elution fraction showed two main peaks. The
peak at 47,042.6 Da has a molecular weight consistent with the theoretical molecular weight of
the MBP-PilA20 fusion produced in the cytoplasm (47,033.9 Da). The peak at 43,671.1 Da is
likely the result of proteolytic attack that might have taken place after the cells were lysed to
obtain the crude extract for purification.
Even though proteolysis was suspected, some of the MBP-PilA20 fusions produced via
cytoplasmic expression remained intact. Cleavage of the PilA20 subunit from the MBP tag was
o

tested using Factor Xa protease. Cleavage conditions tested included different temperatures (4 C
o

and 23 C), incubation times (24 and 48 h), and concentration of Factor Xa protease (0.5 and 1
µg/mL). SDS-PAGE characterization after cleavage showed that the reaction efficiency depends
on temperature, time, and concentration of Factor Xa (Figure 3.6). The gradual disappearance of
the upper band and thickening of the lower band indicates the pilin is being cleaved from the
o

MBP tag. Cleavage was more efficient at higher temperature (23 C), higher concentration of
Factor Xa, and longer incubation times. Almost complete cleavage was observed after 24 h when
o

done at 23 C using 1 µg/mL of Factor Xa (final concentration). Complete cleavage was not
o

observed at 4 C even after 48 h of incubation.

54

After cleavage, samples were analyzed by MALDI-TOF MS to confirm presence of the
pilin subunit. MS indicated that two peptides were being produced during cleavage (Figure 3.7).
The peak at 4,527.4 Da is consistent with the production of the PilA20 subunit (theoretical
molecular weight of PilA20 is 4,523.9 Da). The second peak has a molecular weight of 3,563.3
Da. To check whether this peptide was the result of non-specific cleavage by Factor Xa protease,
an HPLC-purified sample was sent for N-terminal sequencing of the first five amino acid
residues (MSU Research Technology Support Facility). The sequence obtained was V-AYN
(definite identification of second amino acid residue was not possible). Figure 3.7 shows where
this sequence matches the PilA20, and that a fragment starting at this point would have a
molecular weight consistent with the result obtained by MS. Modifications in reaction conditions
did not eliminate or reduce the non-specific cleavage.
Even though the pMAL expression system seemed very promising because of its
previously reported success on the expression of recombinant pilin from P. aeruginosa [79], it
did not give satisfactory results in the production of pilin from G. sulfurreducens. Degradation of
the MBP-PilA20 fusion protein produced either periplasmically or cytoplasmically was observed.
Additionally, treatment with Factor Xa protease aimed at removing the MBP tag from the PilA20
gave the undesired outcome of cleavage within the PilA20 subunit.
The third expression system considered was the Intein-mediated purification with an
affinity chitin binding tag system (IMPACT

TM

, New England Biolabs Inc., Ipswich, MA). An

advantage of this system is the use of a thiol-induced intein, which is a self-splicing protein that
excise itself from the target protein in the presence of reducing agents, eliminating the need for

55

proteases for cleavage [90]. This system was used to express various truncations (pilAn) of the
pilA gene. The objective was to identify a relatively small number of amino acids to truncate (n)
that would simultaneously give good expression of a soluble fusion in E. coli, as well as facile
purification of pilins and in-vitro self-assembly of the purified PilAn peptide to yield synthetic
pili. A chitin-binding domain (CBD) was attached to the N-terminus of the pilAn sequence.
Clones were prepared for the full length pilin (PilA) and truncations PilA10, PilA19, PilA20, and
PilA22. Constructs were transformed into Rosetta 2(DE3)pLysS E. coli cells, a strain that
enhances expression of foreign proteins by addition of tRNAs for codons rarely used in E. coli
(AGA, AGG, AUA, CUA, GGA, CCC, and CGG).
As evidenced by SDS-PAGE data, expression of the fusion protein was achieved for all
truncated constructs (CBD-PilA10, CBD-PilA19, CBD-PilA20, CBD-PilA22) but not for the full
length (CBD-PilA) (Figure 3.8). Cells expressing the CBD-PilA fusion only grew to a cell
density

½

that of the other constructs. This suggests that CBD-PilA might be toxic to E. coli

cells, probably because of the high hydrophobicity of the full length pilin.
Purification of truncated pilins was done by affinity chromatography with a column of
chitin beads. On-column cleavage of the pilin from CBD was done by incubating with 50 mM
DTT. Pilin peptides are released following cleavage, but the CBD tag remains bound to the
chitin beads. After the pilin peptides were eluted from the affinity column, the cleavage
efficiency was determined as a function of incubation time (24, 48, and 72 h) and temperature
o

o

(4 C and 23 C). In SDS-PAGE analysis, the appearance of a second band lower in molecular
weight than the CBD-PilAn fusion indicated cleavage of the PilAn monomer from the CBD tag
56

o

(Figure 3.9). Higher temperature (23 C) produced more complete cleavage for all constructs.
Cleavage of PilA19 from CBD was faster when compared to PilA10, PilA20, and PilA22, with
o

nearly complete cleavage of PilA19 observed at 23 C after 24 h of incubation. PilA10 showed
o

approximately 50% cleavage after 24 h at 23 C, and it did not seem to improve significantly
o

over time. PilA22 showed approximately 80% cleavage at 23 C after 48 h of reaction. Cleavage
o

was considerably less efficient for PilA20 as it required at least 48 h of incubation at 23 C to
show any noticeable effect, probably because the amino acid closest to the cleavage site
(isoleucine) is sub-optimal for the cleavage reaction with this expression vector. PilA10 and
PilA19 have an alanine at the cleavage site, and PilA22 has a glutamine; both of these amino
acids are known to achieve higher cleavage efficiency when using this expression vector [91].
The elution fractions were analyzed by SDS-PAGE to confirm pilin production (Figure
3.10). Prominent bands were observed for PilA19 and PilA22 monomers when cleavage was
o

o

done at 23 C and 4 C. Although some PilA20 was cleaved from the fusion protein (Figure 3.9),
soluble PilA20 was not visible in the SDS-PAGE gels, indicating low cleavage efficiency of this
construct. However, PilA20 in the elution fraction was detected by MALDI-TOF mass
spectrometry, a more sensitive technique (Figure 3.11). While significant cleavage of PilA10 was
observed (Figure 3.9), the subunits were only slightly visible in gels when cleavage was done at
o

o

4 C (Figure 3.10), and no bands could be seen when cleavage was done at 23 C. In addition, the
matrix clogged when cleavage was performed on-column, suggesting that the PilA10 subunits
57

aggregated after cleavage. The four truncated pilins have a molecular weight consistent with
their theoretical molecular weight as confirmed by MALDI-TOF MS (Figure 3.11).
The IMPACT

TM

expression system overcame the shortcomings of the other two

systems. The fusion protein exhibited high solubility and no protease was required during
cleavage. Four truncated constructs were expressed using this system. Even though the shortest
truncation (longest pilin peptide) was expressed as a soluble fusion protein (CBD-PilA10), the
PilA10 subunit showed signs of aggregation after cleavage that made it challenging to purify in
larger quantities. On the other hand, PilA19, the second shortest truncation, was expressed well
and could be recovered and purified readily, so it was selected for subsequent studies of pilin
self-assembly.
3.3.2. Characterization of Secondary Structure of Pilin Subunits
One of the architectural features shared by type IV pilin monomers is the hydrophobic Nterminal α-helical spine known as α1 [17]. This helical structure plays a significant role in pilus
assembly by forming a hydrophobic helical bundle in the filament core [47]. After production of
PilA19 subunits, an important goal was to determine whether recombinant pilin retained this
structural feature.
Two powerful techniques are generally used to determine the secondary and tertiary
structure of proteins, x-ray crystallography and nuclear magnetic resonance (NMR). These
methods provide structural information with atomic resolution [92, 93]. Although powerful,
these techniques have experimental limitations. X-ray crystallography can be applied to large
and small proteins but it requires the protein to be in crystalline form [93]. Crystallization

58

requires large amounts of protein to find optimum crystallization conditions and not all proteins
are suitable for crystallization. NMR does not need protein crystallization, but it requires large
quantities of soluble protein, extensive isotopic labeling, and considerable time. It is also limited
to proteins with a molecular weight lower than 30 kDa [94].
The experimental restrictions inherent to x-ray crystallography and NMR analysis have
increased the use of circular dichroism (CD) [93]. Circular dichroism allows the estimation of
structural features of a protein. CD does not provide atomic resolution; it does not assign specific
amino acids to a specific secondary structure [93]. However, CD can be applied to all proteins at
low concentrations (< 20 µg) under various solvent conditions [84]. CD is very useful for rapid
determination of conformational changes in protein structure resulting from temperature, pH,
heat, mutations, detergents or denaturants, or binding interactions [84, 92]. CD is based on the
difference in absorbance of right- and left-handed circularly polarized light as it passes through a
chiral molecule [83]. For proteins, different secondary conformations have characteristic CD
spectra. Disordered proteins have very low ellipticity above 210 nm and negative peaks near 195
nm. In contrast, α-helical peptides have two negative peaks at 222 nm and 208 nm, and a
positive peak at 193 nm [83, 84].
Circular dichroism (CD) experiments were carried out to evaluate the secondary structure
of PilA19 monomers. To determine whether the net charge of the PilA19 subunit would have an
effect on the peptide secondary structure, CD experiments were performed at pH values below
and above the isoelectric point (pI) (pH 3.8 and 7.0). The theoretical isoelectric point of the
PilA19 subunit is 4.86. The effect of detergents, sodium dodecyl sulfate (SDS) and octyl β-Dglucopyranoside (OG), on the conformation of the pilin monomer was investigated. SDS has
been shown to induce and stabilize secondary structure, especially α-helices [95]. SDS is
59

commonly used to mimic membranes and other hydrophobic environments [95]. OG has been
used to dissociate pili into pilins. The use of OG to depolymerize PAK pili demonstrated that OG
produced pilin monomers without disrupting the peptide secondary structure [96].
Depolymerization of PAK pilin using SDS resulted in a 6% increase in α-helix [96]. The critical
micelle concentrations (CMC) of these surfactants are ∼ 8 mM for SDS and ∼ 25 mM for OG
[97, 98].
The effect of SDS on the CD spectra of PilA19 subunits at pH 3.8 and 7.0 is shown in
Figures 3.12 and 3.13. Without detergent, the spectra suggest a disordered or unfolded peptide,
with very low ellipticity above 210 nm. The presence of SDS resulted in the characteristic
structural signature of a α-helix for all SDS concentrations and pH values tested. The effect of
OG on the CD spectra of PilA19 subunits at pH 3.8 and 7.0 is shown in Figures 3.14 and 3.15.
When OG was present at a concentration close to the CMC (25 mM), no significant change in
the shape of the spectra was observed at both pH values, indicating the protein was still in a
disordered conformation. However, an OG concentration approximately five times higher than
the CMC yielded an α-helical signature.
Since α-helical proteins exhibit a large positive peak at 193 nm, and two small negative
peaks at 208 nm and 222 nm [83], the ellipticity values at these wavelengths were plotted
(Figures 3.16 and 3.17). The changes observed after the addition of either SDS or OG followed
similar patterns at both pH values investigated. The ellipticity at 193 nm increased after the
addition of both surfactants. However, the change was more significant for SDS than for OG.
The mean residue ellipticity values at 208 nm and 222 nm became more negative when
detergents were added, with the effect larger for SDS than for OG. At pH 7.0, the values became
more negative as the concentration of SDS increased, suggesting an increase in α-helical content

60

with SDS concentration. The behavior at pH 3.8 was different, with the ellipticity values at 208
nm and 222 nm becoming more negative in the following fashion: 0 SDS > 1 mM SDS > 40 mM
SDS > 8 mM SDS.
To determine the percentage of the peptide having an α-helix conformation, the CD data
were analyzed using the program CONTINLL. This program, which is available through the
DICROWEB server [99, 100], is a modification of the CONTIN method developed by
Provencher and Glöckner [101, 102]. It uses a ridge regression algorithm to estimate the CD
spectrum of unknown proteins by comparison to a linear combination of CD spectra of N
reference proteins with known conformations [84, 103]. Relatively good predictions of α-helix
and β-sheet content can be obtained with this method [84]. The secondary structure components
determined with CONTINLL are α-helix (regular αR and distorted αD), β-strands (regular βR
and distorted βD), turns and unordered [103].
CONTINLL has been shown to give more accurate results than other CD-analysis
algorithms in the case of polypeptides or short protein fragments [84]. However, the main
limitation of the CONTINLL program is that the reference data sets currently available are
compiled from globular proteins, and do not necessarily represent the correct conformation of
peptides, fibrous, and membrane proteins [83, 84, 99].
The program evaluates the goodness of fit parameter normalized mean residue standard
deviation (NMRSD), which is defined as:

1/ 2
%
2(
'$ "exp # "cal *
NRMSD = '
*
' $ "exp 2 *
&
)

(

)

(

!
61

)

(3.4)

where θexp and θcal are the experimental and calculated ellipticity values at a specific
wavelength. A NRMSD value of less than 0.1 is generally considered a good fit [83]. In the
analysis presented here, data with NRMSD values above 0.12 were rejected. Samples with 125
mM OG at pH 3.8 and pH 7.0 have NRMSD values above 0.12. Therefore, these data were not
considered during the analysis.
Table 3.3 shows the calculated percentage of the various secondary structure components
for the conditions tested. The analysis is focused on the percentage of regular α-helix (αR). The
secondary structure calculations showed an increasing percentage of α-helix with increasing
concentrations of SDS both at pH 3.8 and 7.0. The trend was also apparent when the data were
plotted as percentage of regular α-helix versus the concentration of SDS at both pH values
(Figure 3.18). The percentage of α-helix was higher at pH 3.8 than at 7.0. These calculations are
consistent with qualitative observations based on changes in spectral shape and ellipticity values
at 193 nm, 208 nm, and 222 nm after the addition of SDS.
Collectively, the CD results suggest that SDS triggers a conformational change of the
pilin monomers that gives a more α-helical structure. This finding is consistent with previous
reports that SDS stabilizes and induces secondary structures, especially α-helices [95]. SDS
provides an environment that mimics the hydrophobic core and polar head of a membrane [95,
98]. Pili are transmembrane proteins that protrude from the outer membrane of Geobacter [17].
During pilus assembly, the hydrophobic part of pilin monomers is buried inside a helical coiledcoil structure in the filament core [17, 19]. We speculate that pilin monomers need that
membrane-like environment that SDS provides to attain a more native α-helical structure. The

62

conserved ∼ 53 residue helix at the N-terminus of type IV pilins controls inner membrane
insertion, and pilin polymerization [17, 39]. The specific value of α-helical content for the wild
type pilin (PilA) of Geobacter has not been determined yet for comparison with the experimental
data for the recombinant PilA19 subunit.
3.3.3. In-vitro Assembly of Recombinant Pilins into Pili Fibers
In previous work with P. aeruginosa, self-assembly of in-vitro expressed pilin into piluslike structures was triggered when an aqueous solution of the pilins was contacted with
hydrophobic undecanethiol or Superdex resin. This suggests that the nature of the hydrophobe
may not be important and that the hydrophobe may only serve as a nucleation point for selfassembly [48]. We have used a different strategy to assemble G. sulfurreducens pilin subunits
from solution. During pilus assembly, pilin monomers are held together by hydrophobic
interactions [47]. We hypothesized that a hydrophobic environment could induce proper
conformation and trigger assembly. Therefore, recombinant pilins were exchanged from buffer
to an organic solvent. The organic solvent mixture was removed using a roto-evaporator, and the
dry pellet was resuspended in DI water. TEM and AFM images of the resuspended pellet
(Figures 3.19 and 3.20) showed micron size clusters of hair-like filaments that were less than 10
nm in diameter and several microns in length. Several factors may have contributed to pilin selfassembly as the solvent was removed, including a lack of buffer, salt, or detergent to stabilize the
peptide solution, the peptide solubility limit being exceeded, and the presence of an air-water
phase boundary.

63

3.3.4. Conductivity Assessment of Recombinant PilA19 Fibers via STM
Scanning tunneling microscopy (STM) provides three-dimensional real space images and
also allows spatial variations in the density of states within the sample to be probed with atomic
resolution [104]. In STM, a bias voltage is applied between a sharp metal tip and a conductive
sample as the two are brought close together. When the tip and the sample are separated by a few
angstroms, the electron wave functions of the tip and the sample overlap, and a tunneling current
is generated whose magnitude depends exponentially on the sample-tip distance. A piezoelectric
control system with very high vertical resolution adjusts the position of the tip in three
dimensions (x, y, and z) in response to the current [104, 105]. When the tip scans the sample in
the x-y plane, changes in tunneling current are converted into an image and used to analyze the
sample’s topography and density of states [104, 105].
Wild type pili of G. sulfurreducens have been shown to be highly conductive [17]. STM
experiments are being performed to assess whether the recombinant pili fibers are also
conductive. Figure 3.21 shows an STM image collected at 0.05 V of a filament with a similar
size and structure to that of the native pilus nanowire of G. sulfurreducens. Previous studies of
wild type pili from G. sulfurreducens showed that the finite tip radius led to geometric
convolution of the tip and the pilus in the resulting images [86]. This resulted in vertical
flattening, decreasing the apparent height, and lateral broadening, enlarging the diameter.
The STM images obtained indicated that the recombinant pilus is electrically conductive,
with the low voltage required to obtain the measured tunneling current consistent with
conductivity. Complementary methods, such as scanning tunneling spectroscopy (STS) and CPAFM would provide a more complete characterization of the electrical properties of the
recombinant pilin fibers.
64

3.4.

Conclusions

The IMPACT

TM

expression system, which produced a CBD-PilAn fusion, was selected

for pilin subunit expression for several reasons: (1) no aggregation of the fusion protein was
observed, (2) the fusion peptide was not susceptible to degradation, and (3) no proteases were
required to cleave the fusion tag. To avoid problems associated with overexpressing hydrophobic
recombinant peptides, a variety of N-terminal end truncations (PilA10, PilA19, PilA20, and
PilA22) were evaluated for success in expression, purification and self-assembly. The PilA19
subunit was the shortest truncation that expressed well as a soluble fusion protein.
The CD spectrum for the purified PilA19 peptide in pure buffer was characteristic of a
disordered secondary structure. However, after SDS addition, the CD spectrum exhibited the
characteristic signature of an α-helix. This result is consistent with the hypothesis that the
recombinant pilin monomers require a hydrophobic, membrane-like local environment to attain
the α-helical structure characteristic of Type IV pilins.
Dissolution of purified PilA19 pilins in a hydrophobic solvent, followed by evaporation of
the solvent, triggered pilin self-assembly into filaments. As indicated by both TEM and AFM
analyses, the filaments had properties similar to native G. sulfurreducens Type IV pili, including
a similar diameter and tendency to aggregate in tangled bundles. STM data indicated that the
synthetic pili were conductive, but further studies are needed to more fully characterize the
electrical properties.
The production and purification of truncated recombinant pilin monomers and subsequent
assembly into synthetic pili demonstrated, for the first time, the feasibility of producing
65

nanostructured filaments based on conductive Geobacter pili. Conductive, synthetic pili represent
a new class of protein-based nanowires that may be suitable for a wide range of applications.

66

Table 3.1 Restriction enzymes and primers used for cloning of pilAn gene into expression vectors.
Expression system
cloning vectors

Restriction
enzymes
Rev

QIAexpressionist
R
pQE-30 UA (Amp )
pMAL

R

c4x (Amp )

Fwd
Rev

Hind III

R
p4x (Amp ) Fwd III
Hind Xmn I

IMPACT
R
PTYB11 (Amp )

Primers

Xmn I
Rev Pst I
Fwd Sap I

Rev
Fwd

5’-TGGATAGGCGGGCTTTCAAT-3’
pilA20 5’-TTAACTTTCGGGCGGATAGGT-3’
5’-CCCAAGCTTTTAACTTTCGGGCGGATAGGT-3’
pilA20 5’-ATTCCGCAGTTCTCGGCGTA-3’

5’-CCCAAGCTTTTAACTTTCGGGCGGATAGGT-3’
Rev
5’-GGTGGTCTGCAGTCATTAACTTTCGGGCGGATAGGT-3’
Fwd pilA 5’-GGTGGTTGCTCTTCCAACTTCACCCTTATCGAGCTGCT-3’
P
ilA
pilA pilA10 5’-GGTGGTTGCTCTTCCAACGCGATCATCGGTATTCTCGC-3’
5’ TTCpilA19CTT ATC GAG CTG CT 3’
ACC 5’-GGTGGTTGCTCTTCCAACGCGATTCCGCAGTTCTCGGC-3’
pilA20 5’-GGTGGTTGCTCTTCCAACATTCCGCAGTTCTCGGCGTA-3’
pilA22 5’-GGTGGTTGCTCTTCCAACCAGTTCTCGGCGTATCGTGT-3’

67

Table 3.2 Bacterial strains used to express recombinant pilin subunits.

Expression system
QIAexpressionist
p-MAL

TM

IMPACT
a

TM

TM

E. coli strain

Characteristics

M15[pREP4]

Kan

R

-

F ara Δ(lac-proAB) [Φ80dlac Δ(lacZ)M15]

K12 TB1

Rosetta 2 (DE3)pLysS

R

rpsL(Str ) thi hsdR
a

Novagen, EMD Chemicals, Gibbstown, NJ

68

-

-

-

F ompT hsdSB(rB mB ) gal dcm (DE3)
R
pLysSRARE2 (Cam )

Table 3.3 Affinity resin and composition of buffers and solutions used during affinity
chromatography purification of recombinant pilin subunits.
TM

TM

QIAexpressionist

p-MAL

TM

IMPACT

Resin

Ni-NTA

Amylose

Chitin

Lysis
Buffer

50 mM Na-phosphate pH
7.8, 0.3 mM NaCl, 10
mM βME, 2% CHAPS,
20 mM imidazole,
1mg/mL lyzozyme

20 mM Tris-HCl pH 7.4, 200
mM NaCl, 1 mM EDTA, 1
mM PMSF

20 mM Tris-HCl
pH 7.4, 100 mM
NaCl, 1 mM
EDTA, 1 mM
PMSF, 2% CHAPS

30 mM Tris-HCl pH 8.0, 20%
sucrose, 1 mM EDTA
(incubate 5-10 min, centrifuge
at 8000 × g for 20 minutes at
o

4 C) Resuspend pellet in 5

Osmotic
Shock

mM MgSO4 ice cold (stir for
10 min, centrifuge at 8000 × g
o

for 20 minutes at 4 C).
Supernatant is osmotic shock
fluid.
Wash
Buffer

50 mM Na-phosphate pH
7.8, 0.3 mM NaCl, 10
mM βME, 0.2% CHAPS,
20 mM imidazole

20 mM Tris-HCl pH 7.4, 200
mM NaCl, 1 mM EDTA

20 mM Tris-HCl
pH 9.0, 100 mM
NaCl, 50 mM DTT

Cleavage
Buffer

Elution
Buffer

20 mM Tris-HCl
pH 7.4, 600 mM or
1 M NaCl, 1 mM
EDTA

50 mM Na-phosphate pH
7.8, 0.3 mM NaCl, 10
mM βME, 0.2% CHAPS,
300 mM imidazole

20 mM Tris-HCl pH 7.4, 200
mM NaCl, 1 mM EDTA, 10
mM maltose

69

20 mM Tris-HCl
pH 9.0, 100 mM
NaCl, 1% OG

Table 3.4 Analysis of CD data using the CONTINLL algorithm. The various secondary structures determined are α-helix (regular αR
and distorted αD), β-strands (regular βR and distorted βD), turns and unordered. The NRMSD parameter indicates the goodness of fit.
Values below 0.1 are considered a good fit.
% secondary structure
Type of
secondary
structure

No detergent

1 mM SDS

8 mM SDS

40 mM SDS

25 mM OG

125 mM OG

pH 3.8

pH 7.0

pH 3.8

pH 7.0

pH 3.8

pH 7.0

pH 3.8

pH 7.0

pH 3.8

pH 7.0

pH 3.8

pH 7.0

αR

1.2

1.7

33.7

19.2

37.2

28.6

38

31.3

5.4

3.5

15.6

26.9

αD

3.0

4.9

14.9

7.6

18.4

14.5

18.3

17.0

5.7

5.2

6

14.9

βR

13.9

9.1

18.5

13.6

1.6

5.7

5.8

0.9

11.9

7.4

19.7

21.8

βD

6.6

5.1

10.2

4.1

2.7

2.9

4.1

2.6

6.2

4.1

6.8

12

Turns

12.0

11.9

22.7

9.6

13.6

12.7

14.8

14.3

12.7

10.4

13.5

24.5

Un-ordered

63.2

67.2

0

45.9

26.6

35.6

18.9

33.8

58.2

69.3

38.4

0

Total

99.9

99.9

100

100

100.1

100

99.9

99.9

100.1

99.9

100

100.1

NRMSD

0.045

0.068

0.036

0.113

0.058

0.119

0.028

0.117

0.073

0.073

0.177

0.206

70

Figure 3.1 Purification steps used for each expression system.

71

Figure 3.2 Analysis of hydrophobicity of pilin subunits and effect of amino acid truncation. Kyte and Doolittle hydrophaticity [87]
plot using a 19-residue segment, amino acid sequence and GRAVY index of PilA and PilAn truncations.

72

Figure 3.3 Expression and purification of 6×His-PilA20. a) 16% tricine SDS-PAGE and b)
western blot analysis of cell pellet (A), soluble fraction of cell lysate (B), and insoluble fraction
of cell lysate (C); before (-) and after (+) induction.

73

Figure 3.4 Periplasmic expression and purification of MBP-PilA20. a) 12% glycine SDS-PAGE
and b) western blot analysis of crude extract (A) before (-) and after (+) induction; soluble
fraction of cell lysate (B), and elution fraction from amylose column (C); c) MALDI-TOF MS
analysis of elution fraction from amylose column. Theoretical molecular weight of the
periplasmic MBP-PilA20 fusion is 46,902.6 Da.

74

Figure 3.5 Cytoplasmic expression and purification of MBP-PilA20. a) 12% glycine SDS-PAGE
and b) western blot analysis of crude extract (A) before (-) and after (+) induction; soluble
fraction of cell lysate (B), and elution fraction from amylose column (C); c) MALDI-TOF MS
analysis of elution fraction from amylose column. Theoretical molecular weight of the
periplasmic MBP-PilA20 fusion is 47,033.9 Da.

75

Figure 3.6 Cleavage efficiency of PilA20 subunit from MBP tag mediated by Factor Xa
o

protease. Elution fractions from amylose column were incubated with Factor Xa protease at 4 C
o

and 23 C. Samples were analyzed by 7.5% glycine SDS-PAGE after 24 h and 48 h of reaction.
Gels show elution sample untreated with protease as control (A), elution treated with 0.5 µg/mL
(B), and elution treated with 1.0 µg/mL (C).

76

Figure 3.7 Production of PilA20 subunits after treatment with Factor Xa protease. MALDI-TOF MS analysis of a) sample untreated
with protease used as a control, and b) sample treated with protease. MS results indicate two peptides are produced during cleavage.
The peptide with MW of 4,527.4 Da corresponds to the PilA20 subunit. c) N-terminal sequencing of the peptide with MW of 3,563.3
Da. Sequencing indicated that non-specific cleavage of the pilin subunit occurred.

77

Figure 3.8 Expression of various truncations of pilin subunits using a CBD fusion tag (CBD-PilAn). Expression was determined by
-

+

analyzing cell pellet samples before (A ) and after (A ) induction using a 12% glycine SDS-PAGE.

78

Figure 3.9 Cleavage efficiency of CBD tag from PilAn subunits after incubation with 50 mM
o

o

DTT at 4 C and 23 C. During cleavage, pilin monomers are released but CBD remains bound to
the chitin beads. Chitin beads were boiled with 1% SDS to release bound protein, which was
analyzed by 7.5% glycine SDS-PAGE. Control (no DTT in incubation buffer) (A), cleavage
efficiency after 24 h (B), 48 h (C), and 72 h (C). The appearance of a lower molecular weight
band indicates the cleavage.

79

Figure 3.10 Presence of pilin monomers in elution from chitin column after cleavage from CBD
tag determined by 16-20% Tricine SDS-PAGE.

80

Figure 3.11 MALDI-TOF MS analysis of elution from chitin column after cleavage. MS
analysis confirmed the molecular weight of the four truncated pilins. Theoretical molecular
weights of the truncated pilins are PilA10: 5,431 Da, PilA19: 4,595 Da, PilA20: 4,524 Da, and
PilA22: 4,314 Da.

81

Figure 3.12 Surfactant-dependent folding of recombinant PilA19 subunits at pH 3.8. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 3.8, 50 mM Na2SO4 with and without sodium
dodecyl sulfate (SDS).

82

Figure 3.13 Surfactant-dependent folding of recombinant PilA19 subunits at pH 7.0. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 7.0, 50 mM Na2SO4 with and without sodium
dodecyl sulfate (SDS).

83

Figure 3.14 Surfactant-dependent folding of recombinant PilA19 subunits at pH 3.8. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 3.8, 50 mM Na2SO4 with and without octyl
β-D-glucopyranoside (OG).

84

Figure 3.15 Surfactant-dependent folding of recombinant PilA19 subunits at pH 7.0. CD Spectra
of PilA19 subunits in 10 mM potassium acetate pH 7.0, 50 mM Na2SO4 with and without octyl
β-D-glucopyranoside (OG).

85

Figure 3.16 Effect of sodium dodecyl sulfate (SDS) on the conformation of PilA19 subunits.
Molar ellipticities at pH 3.8 and 7.0 are plotted for three wavelengths, 192, 208, and 222 nm, as a
function of SDS concentration.

86

Figure 3.17 Effect of octyl β-D-glucopyranoside (OG) on the conformation of PilA19 subunits.
Molar ellipticities at pH 3.8 and 7.0 are plotted for three wavelengths, 192, 208, and 222 nm, as a
function of SDS concentration.

87

Figure 3.18 Effect of SDS on the percentage of regular α-helix of PilA19 subunits at pH 3.8 and
7.0. Percentage of regular α-helix was calculated using the CONTINLL algorithm. Lines
connecting data points are for visual aid only.

88

Figure 3.19 Transmission electron microscopy images of recombinant pili fibers. Samples were stained with uranyl acetate. TEM
JEOL 2200FS (JEOL Inc., Boston, MA) operated at an accelerating voltage of 100 kV.

89

Figure 3.20 Atomic force microscopy images of recombinant pili deposited on mica. Tapping mode in air.

90

Figure 3.21 Scanning tunneling microscopy image of a recombinant PilA19 fiber at an applied
voltage of 0.05 V.

91

Chapter 4.
STUDY OF PILIN MODIFICATIONS TO ENABLE INTERFACING WITH
ELECTRODES

4.1.

Introduction
Protein assemblies, which are essential components in viruses, bacteria, eukaryotic cells,

and organisms, participate in transport of intracellular and intercellular cargo and act as
cytoskeletal scaffolds, storage containers, and molecular motors [4, 106]. These multi-subunit
complexes self-assemble into specific architectures with controlled shapes and physicochemical
properties, making them ideal for bottom-up fabrication in bionanotechnology applications [4].
Protein assemblies such as beta-amyloid fibrous proteins, viral proteins, bacterial pili and
bacterial flagella are being investigated as alternative components of nanostructured devices [77,
78].
The self-assembling capability of bacterial pili might be a useful feature in bottom-up
fabrication of interfaces [48, 80]. An added advantage of using the pili of Geobacter
sulfurreducens is their ability to transfer electrons [17], which may enable applications in
biobased electronics. The first step towards using the pili of G. sulfurreducens as a building

92

material is the mass production of pilin monomers. The successful mass-production of
recombinant PilA19 monomers and in-vitro assembly of conductive pili was discussed in Chapter
3 of this dissertation. The next step was to determine whether recombinant pilin monomers and
pili could be interfaced with gold electrodes. This is an important step towards using pilins and
pili from G. sulfurreducens for the construction of bioelectronic devices.
Several strategies to interface pilin monomers and pili with gold electrodes were
investigated. The first strategy used a hydrophobic alkanethiol self-assembled monolayer (SAM)
deposited on gold to interact with recombinant PilA19 monomers. This strategy was based on the
results reported by Audette et al. showing that recombinant pilins from Pseudomonas aeruginosa
strain K122-4 (ΔK122-4) interacted with surface-constrained hydrophobes [48]. In these
experiments, P. aeruginosa pilins were placed in contact with undecanethiol covalently attached
to maleimide-activated microtiter plates. The interaction was monitored immunologically using
polyclonal sera generated against the P. aeruginosa strain K pili. A concentration-dependent
binding of the pilin to the hydrophobe was observed [48]. However, it was unclear whether the
pilins were assembled into pili as a result of the interaction with undecanethiol or only physically
adsorbed onto the monolayer. The proposed approach is aimed at determining whether the
interaction of PilA19 subunits with an undecanethiol SAM could trigger self-organization into
pili. Cyclic voltammetry and ellipsometry were used to confirm the formation of the alkanethiol
monolayer. Scanning electron microscopy was used to determine the assembly of pilin subunits
into pili.
To investigate direct interaction between pilin subunits and gold substrates, a cysteinemodified PilA19 subunit was engineered. Several studies have shown that cysteine-

93

functionalized peptides and proteins bind to gold surfaces by interaction with the sulfur atoms.
Cysteine forms a strong bond through the interaction of SH groups and Au atoms [107]. The last
amino acid at the N-terminus of the PilA19 monomer, alanine, was replaced by cysteine (Figure
4.1). The truncation is referred to here as PilA19-A20C. PilA19-A20C pilins were produced with
the same protocol used to express and purify PilA19 pilins.
After production of PilA19-A20C monomers, interactions between pilin monomers and invitro assembled pili with gold electrodes were examined using cyclic voltammetry, ellipsometry,
quartz crystal microbalance, and scanning electron microscopy. PilA19-A20C monomers were
assembled into pili using the same protocol successfully demonstrated with PilA19 pilins.
4.2.

Materials and Methods

4.2.1. Interfacing of Recombinant PilA19 Monomer with Gold Electrodes
Gold substrates (LGA Thin Films, Santa Clara, CA) were washed with ethanol, dried
under nitrogen, and cleaned by a 30 s immersion in piranha solution (7 parts by volume
2

concentrated sulfuric acid to 3 parts by volume 30% aqueous hydrogen peroxide) . After piranha
cleaning, the substrates were washed with copious amounts of DI water and dried under nitrogen.
Substrates were incubated for 48 h at room temperature with a 1 mM undecanethiol (SigmaAldrich St. Louis, MO) solution in ethanol. Following monolayer formation, the gold substrates

2

Caution: piranha solution is an extremely strong oxidant and potentially explosive; it must be
handled with extreme care.

94

were washed with ethanol and dried under nitrogen. Formation of undecanethiol monolayer was
verified using ellipsometry and cyclic voltammetry.
Monolayer formation was followed by overnight incubation at room temperature with
PilA19 subunits dissolved in 20 mM sodium phosphate buffer pH 7.0, 100 mM NaCl, with or
without 10 mM SDS. To assess the assembly of recombinant pilins into pili, the samples were
analyzed using scanning electron microscopy (SEM).

4.2.2. Expression and Purification of Recombinant PilA19-A20C Subunits
Since the IMPACT expression system (New England Biolabs Inc., Ipswich, MA) gave
the best results for production of PilA19 subunits, this system was used to produce the cysteinemodified pilin. The pilin was expressed as a fusion of a chitin binding domain tag (CBD-PilA19A20C). Plasmid was transformed into DH5α E. coli cells for propagation. Sequence of the
expression plasmid was confirmed before transforming into competent E. coli strain Rosetta2
DE3 p(LysS). The expression was carried out in a 10 L batch reactor (Bioflo 3000, New
Brunswick Scientific, Edison, NJ). A detailed protocol for using this batch system is given in
section 3.2.2 of this dissertation. Briefly, 5 L of yeast-enriched HM media containing trace
elements, 100 µg/mL ampicilin, and 34 µg/mL chloramphenicol, was inoculated. Cells were
o

grown overnight at 37 C with periodic additions of 50% (w/v) glucose and 20 g/L MgSO4. The
pH was set to 7.0 and the dissolved oxygen (DO) was kept at 10%. Expression of the
o

recombinant protein was induced with 0.5 mM IPTG and the temperature was lowered to 30 C.
Six hours after induction, the cells were harvested by centrifugation at 5000 x g for 20 min at

95

o

o

4 C. The cell pellet was stored at -70 C. Expression of the fusion protein was tested by SDSPAGE and western blot.
The PilA19-A20C monomers were purified by affinity chromatography using chitin
beads. All the buffers used during purification are shown in Table 3.3. Harvested cells were resuspended in a lysis buffer (10 mL of buffer per gram of cells) and lysed using a tip sonicator
(Sonifier 250, Branson, Danbury, CT). The crude extract was sonicated on ice 7 times for 10 s,
o

and the cell extract was clarified by centrifugation at 20,000 x g for 30 minutes at 4 C. The
supernatant was loaded into the chitin affinity column to allow binding of the target protein.
Non-specifically bound proteins were removed from the column by increasing the concentration
of NaCl in the washing buffer. The CBD was cleaved from the PilA19-A20C pilin by rinsing the
column with 3 column volumes of incubation buffer containing 50 mM DTT. The column was
then incubated for 40 hours at room temperature. The PilA19-A20C monomers were eluted from
the column with a storage buffer. The production of pilin subunits was confirmed by SDSPAGE, and MALDI-TOF mass spectrometry. For experimental details regarding SDS-PAGE,
western blot, and MALDI-TOF analysis refer to sections 3.2.4, 3.2.5 and 3.2.6 of this
dissertation.
4.2.3. Interfacing Recombinant PilA19-A20C Pilins with Gold Electrodes
Clean gold substrates (see section 4.2.1) were incubated overnight with a solution
containing PilA19-A20C subunits. After incubation, samples were analyzed using ellipsometry,
cyclic voltammetry (CV), and scanning electron microscopy (SEM). The deposition of PilA19-

96

A20C monomers on the gold electrode was monitored in real time using quartz crystal
microbalance (QCM).

4.2.4. Interfacing Recombinant PilA19-A20C Fibers with Gold Electrodes

In-vitro assembly of PilA19-A20C subunits in solution was carried out with the same
protocol used with the PilA19 monomers. Cysteine-modified subunits were exchanged from
buffer to an organic solvent, followed by complete evaporation of the solvent, and re-suspension
of the dry pellet in DI water. Oasis MAX extraction cartridges, 60 mg sorbent (Waters
Corporation, Milford, MA) were used to exchange the pilins from buffer to organic solvent. The
column was wetted with 1 mL of 100% acetonitrile, followed by 1 mL wash with DI water.
Before loading, the pH of the protein solution was adjusted above 10 with 1 mM NaOH.
Approximately 2 mL of buffer containing pilin was passed through the column. The column was
washed with 1 mL solution of 5% NH4OH. The pilin was eluted with 1 mL solution of 20%
methanol and 80% acetonitrile. The organic solvent mixture was removed using a rotoevaporator. After pellet re-suspension in DI water, TEM was used to confirm the formation of
cysteine-modified pili.
Following assembly, the cysteine-modified pili were brought into contact with a gold
substrate. To shorten the micron long fibers, pili-containing samples were sonicated over three
different periods for comparison (5 min, 30 min, and 1 hour). After overnight incubation with the
pili samples, the gold substrates were analyzed using SEM.

97

4.2.5. Ellipsometry
Film thicknesses were determined using an M-44 rotating analyzer ellipsometer (J.A.
Woollam Co., Inc., Lincoln, NE) controlled by WVASE32 software. The incident angle was set
at 75o using 44 wavelengths of light between 414.0 and 736.1 nm. The refractive index (n) and
extinction coefficient (k) were assumed to be n = 1.5 and k = 0, respectively, for determination
of all thicknesses. Thickness measurements were performed on at least 3 spots on each substrate
and then averaged.
4.2.6. Scanning Electron Microscopy (SEM)
To prepare samples for SEM, the gold substrates were gently washed and dried by either
critical point drying or freeze-drying. In critical point drying, the water in the sample is first
replaced by a dehydration fluid. The samples were dehydrated by incubation in a graded ethanol
series (25%, 50%, 75%, 95%) for 20 min at each concentration, followed by three 10 min
changes in 100% ethanol. Samples were dried in a Balzers Model 010 critical point dryer
(Balzers Union Ltd., Balzers, Liechtenstein) using liquid carbon dioxide as transitional fluid.
Samples for freeze-drying were first frozen in liquid nitrogen, then water was sublimated during
an 18 h cycle in an EMS 750 freeze dryer (Electron Microscopy Sciences, Hatfield, PA). After
drying, samples were mounted on SEM stubs and coated with a thin layer (∼ 2 nm) of osmium
using a NEOC-AT pure osmium coater (Meiwa Shoji CO. Ltd., Osaka, Japan).
Scanning electron microscopy was performed using the secondary electron imaging (SEI)
detector of a JSM 7500F with cold field emission electron emitter SEM (JEOL Ltd., Tokyo,
Japan). An accelerating voltage of 5 kV or less was used for imaging. Energy dispersive X-ray
(EDS) microanalysis coupled to the SEM was used for elemental analysis. EDS was done using

98

an Oxford Instruments INCA system (Oxford Instruments, High Wycomb, Bucks, England),
2

software version 14.3, using a 30 mm detector crystal and an ultrathin window. An accelerating
voltage of 1 kV was used to avoid excitation of gold characteristic X-ray lines during EDS
characterization.
4.2.7. Cyclic Voltammetry (CV)
In cyclic voltammetry, the potential of the working electrode is cycled between two
values at a specific scan rate, and the current produced by the electron transfer process is
measured [108]. The cyclic voltammogram of a clean electrode in the presence of a redox couple
will display a characteristic duck shape with well-defined reduction and oxidation peaks. The
formation of a SAM on the electrode can be studied using CV. A defect free monolayer acts as
an insulating barrier that hinders electron transfer at the electrode surface, so no reduction or
oxidation peaks would be observed. However, pinholes or defects in the monolayer would allow
access of redox species to the electrode surface.
A conventional three-electrode cell consisting of the gold-modified working electrode, a
platinum auxiliary electrode, and a silver/silver chloride (Ag/AgCl) reference electrode was used
for electrochemical measurements. Cyclic voltammetry (CV) was performed using a CHI660
potentiostat (CH Instruments, Austin, TX). Measurements were carried out in 100 mM
phosphate buffer at pH 7.0, containing 100 mM NaCl and 5 mM potassium ferricyanide
(K3[Fe(CN)6]).

99

4.2.8. Quartz Crystal Microbalance (QCM)
Quartz crystal microbalance (QCM) is a very sensitive technique used to study formation
of films on surfaces in real time [109, 110]. It takes advantage of the piezoelectric properties of
quartz crystals and the relationship between the frequency of oscillation and the mass of
deposited material [110]. A research quartz crystal microbalance (RQCM, Inficon, East
Syracuse, NY) controlled by a RQCM logging data system was used to monitor crystal
frequency (f) changes due to the deposition of PilA19-A20C monomers on the gold substrate.
Changes in viscoelastic properties of the adsorbed mass are reflected in changes of the dissipated
energy (D), which was calculated from the measured resistance (R), using the following equation
[109, 110]:

D=

2
2
32 f f Zq d q RA

(4.1)

"
6

where ff is the fundamental frequency of the quartz crystal (5 × 10 Hz), Zq is the acoustic
!
6
2 -1
impedance for AT-cut quartz (8.8 × 10 kg m s ), dq is the piezoelectric strain coefficient for
-12

AT-cut quartz (3.1 × 10

-1

2

m V ), and A is the electrode area (34.19 mm ). The RQCM was

equipped with a 0.1 mL flow cell and an AT-cut polished sensing crystal, 5 MHz Ti/Au (Inficon,
East Syracuse, NY). A PHD standard infusion syringe pump (Harvard Apparatus, Holliston,
MA) was used to fill the flow cell with buffer (10 mM potassium phosphate buffer pH 7.0 with
50 mM Na2SO4). After equilibration of the crystal, the PilA19-A20C monomers were introduced
into the flow cell at a flow rate of 0.2 mL/min and changes in frequency and resistance were
monitored.

100

4.3.

Results and Discussion

4.3.1. Interfacing Recombinant PilA19 Monomers with Gold Electrodes
Interaction of recombinant pilins from P. aeruginosa in solution with undecanethiol has
been reported to trigger assembly of the subunits into pili [48]. The goal in this work was to
determine whether a hydrophobic SAM deposited on a gold electrode could trigger assembly of
recombinant PilA19 subunits into pili. The formation of the undecanethiol monolayer on gold
was confirmed by cyclic voltammetry and ellipsometry. Figure 4.2 shows the CV curves for bare
gold and the undecanethiol SAM. The bare gold voltammogram shows the characteristic duck
shape with well-defined oxidation and reduction peaks. The CV curve for the SAM shows a
significant decrease in current with no distinctive peaks, indicating that the monolayer acts as an
insulating barrier that blocks electron transfer between the electroactive species in solution and
the electrode. Ellipsometry results showed a monolayer thickness of 11.90 ± 0.720 Å (see insert
of Figure 4.2), consistent with values reported in literature [111].
Formation of the undecanethiol SAM on the gold substrate was followed by incubation
with PilA19 monomers in buffer solution with and without SDS. After incubation with the pilin
subunits, the substrates were gently washed and the remaining solution was removed by freezedrying. Samples were analyzed by SEM to determine whether the interaction with the
hydrophobic monolayer would trigger assembly of pilin subunits into pili.
SEM analysis of samples incubated with PilA19 monomers in buffer without SDS did not
show formation of any defined structures. Samples incubated with PilA19 subunits in buffer
containing 10 mM SDS showed areas where horizontal fiber-like structures were observed on the

101

surface (Figure 4.3). The filaments were not evenly distributed on the gold surface. Energy
dispersive X-ray spectroscopy (EDS) was used to determine the elements present in these fibers.
An accelerating voltage of 1 kV was used to avoid excitation of gold X-ray characteristic lines,
which would make it extremely difficult to identify other elements present in the sample. The
EDS data shown in Figure 4.4 indicated that the carbon content was higher on the fiber than on
the gold substrate.
Based on the available results, it was not possible to conclude whether the fibers imaged
using SEM correspond to pilin subunits assembled into pili. It has been reported that SDS can
form various structures when adsorbed onto a hydrophobic surface or monolayer [112, 113]. A
relation between the concentration of SDS and the resulting structures has been observed [113].
AFM images have shown that SDS adsorbed onto a hydrophobic monolayer form poorly-defined
aggregates at concentrations close to the CMC (8.1 mM), whereas long parallel aggregates
believed to be hemi-micellar structures are formed at concentrations well above the CMC (100
mM) [113]. The SDS concentration used in our experiments was 10 mM, which is above the
CMC. The increase in carbon content could be attributed to either a proteinaceous structure (pili)
or an SDS structure. For example, the structures observed in SEM images could be the result of
SDS forming aggregates upon interaction with the undecanethiol monolayer.
Control experiments could be used in the future to determine whether the observed fibers
represent assembled pili or SDS structures. However, even if the fibers are the result of pilin
monomers assembled into pili, the protocol used here did not lead to a controlled assembly
process. SEM images showed that the surface coverage was not uniform, and fibers of various
lengths were formed.

102

4.3.2. Production of Cysteine-Modified Recombinant Pilin: PilA19-A20C
To facilitate the interfacing of pilin monomers with gold substrates, a cysteine-modified
subunit was engineered. The PilA19-A20C monomer is a 19 amino acid truncation of the full
length pilin of Geobacter sulfurreducens, where the last amino acid at the N-terminal end,
alanine, was substituted with a cysteine (Figure 4.1). The PilA19-A20C subunit was expressed as
a fusion protein with a chitin-binding domain (CBD-PilA19-A20C). The purification of the
recombinant pilin was performed by affinity chromatography using chitin beads.
Figure 4.5 shows SDS-PAGE and MALDI-TOF results that demonstrate the successful
production of recombinant cysteine-modified subunit (PilA19-A20C). SDS-PAGE results
confirmed the expression of the CBD-PilA19-A20C fusion, based on the appearance of a strong
band at ~65 kDa after induction. Western blot analysis was positive when using an anti-chitin
antibody. Insert c in the figure corresponds to a tricine gel showing the two bands in the elution
fraction after purification. The elution fraction was analyzed by MALDI-TOF mass spectrometry
(Figure 4.5b), with the peak at 4,632.26 Da confirming the production of PilA19-A20C subunits.
The theoretical molecular weight of the PilA19-A20C monomer is 4,627.08 Da. The peak at
6,218.5 Da is likely a non-specifically bound protein that was not removed from the column
during the washing step. The peak at 1,606.94 Da does not show on the tricine gel because of the
low molecular weight and corresponds to an additional peptide that is cleaved from the fusion.

103

4.3.3. Interfacing Recombinant PilA19-A20C Pilins with Gold Electrodes
Several characterization techniques were used to investigate the interaction between the
cysteine-modified subunits and the gold substrates. The results suggested the formation of a
monolayer of PilA19-A20C subunits on the gold surface. Figure 4.6 shows QCM results obtained
during the deposition of cysteine-modified pilins on a gold electrode. The arrow pointing down
indicates the time when the protein solution was introduced into the flow cell. The arrow
pointing up indicates the time when the syringe pump was stopped. There is a perturbation
associated with the stoppage of the flow, which is probably thermally related, as it is present in
both the frequency and dissipation curves. The frequency shift after the introduction of the pilin
subunits is associated with a change in mass indicating the adsorption of pilin monomers on the
gold electrode. The dissipation change is related to the presence of protein on the QCM crystal
that is coupling more strongly to the fluid. A bare QCM crystal will not couple strongly to the
fluid surrounding it, but the protein will. As more protein is deposited on the crystal, the sum of
the protein-fluid interactions increases, and the dissipative contribution to the QCM signal
increases as a result.
Ellipsometry measurements indicated the formation of a film with a thickness of 35.582 ±
9.110 Å (insert of Figure 4.7). The dimensions of a full length PilA (61 amino acid residues)
from Geobacter sulfurreducens are approximately 80 Å in length and 1 Å in diameter. For the
cysteine-modified truncation (42 amino acid residues), a length of roughly 55 Å would be
expected. The thickness obtained by ellipsometry is about 65% of the expected length, which
would suggest that the subunits arranged themselves at a particular tilt angle with respect to the
surface normal. As reference, the tilt angle of an alkanethiol SAM is typically around 30 degrees

104

[111]. The film thickness and the lack of any structures under SEM examination suggest that the
assembly of pilin monomers into pili did not occur.
Figure 4.7 shows CV curves of a bare gold surface, and a gold substrate incubated
overnight with pilin subunits. Both curves display the characteristic duck shape with distinctive
oxidation and reduction peaks. There is a decrease in current density after incubation with pilin
subunits which, when compared to the bare gold substrate, would indicate a more insulating
surface. Based on the CV, QCM and ellipsometry results, this change in current flow can be
directly attributed to some coverage of pilin subunits on the gold substrate. However, the CV
data cannot be used to unambiguously address the question of whether the pilins are conductive
or not, primarly because there is no information on the uniformity of surface coverage. Thus,
there are four possible interpretations of the data. If the surface coverage is uniform, then the
pilins may be one of two things:
a) conductive species whose structural motif hinders ion transport to some extent; or
b) insulators that are thinner or less effective than, say, alkanethiol monolayers.
If the surface coverage is not uniform, then two other interpretations are possible. The pilins may
be:
c) insulators that formed an imperfect monolayer that does not completely block
electron transport; or
d) conductive but not to the same extent as the underlying gold substrate.
The pilus of G. sulfurreducens is highly conductive. Also, it was shown in Chapter 3 of
this dissertation that recombinant PilA19 fibers are conductive. While it has not been established
that the pilins of G. sulfurreducens are conductive, electrical studies have shown that peptides
with α-helical structures are efficient mediators of electron transfer [114, 115]. Considering that

105

pilins of G. sulfurreducens are α-helical peptides that assemble to form a conductive pilus, it is
conceivable that a monolayer of pilin subunits would also be conductive. Our group is planning
future studies to elucidate this.

4.3.4. Interfacing Recombinant PilA19-A20C Fibers with Gold Electrodes

The interaction of PilA19-A20C fibers with gold electrodes was also investigated. To
assemble the PilA19-A20C subunits into pili, we used the same strategy successfully applied to
the PilA19 subunits. The procedure has three main steps: exchange of the purified pilin
monomers from aqueous to organic solvent, roto-evaporation of the organic solvent, and resuspension of the dry pellet in water. Figure 4.8 shows TEM images of the microns long hair-like
fibers that were formed after following this procedure. We hypothesized that several factors may
have triggered pilin assembly including 1) the lack of buffer, salt, or detergent to stabilize the
peptide solution, 2) the peptide solubility being exceeded, and 3) the presence of an air-water
interface. The formation of fibers after the addition of a cysteine at the pilin N-terminal end is an
indication that the assembly of pilin monomers into pili may tolerate certain modifications in
sequence.
Each fiber formed should have a cysteine at the N-terminus that could interact with the
gold surface. The cysteine-modified pili fibers were subjected to sonication prior to contact with
the gold substrate to shorten the filaments. Three different sonication times were investigated: 1
min (tip sonicator), 30 min (bath sonicator), and 1 h (bath sonicator). After sonication, the gold
substrates were interacted with the previously sonicated pili containing solutions. Under SEM
analysis, none of the samples showed any evidence of attachment of the cysteine-modified pili to

106

the gold substrate. We speculated that the cysteine might be buried in the core of the helical fiber
and not exposed to the gold substrate.
4.4.

Conclusions
Several strategies were investigated in an attempt to interface pilin subunits with gold

substrates. The first strategy was to use a hydrophobic monolayer to trigger assembly of the
recombinant PilA19 monomers. Some fiber formation was observed when the pilin monomers
were dissolved in buffer containing SDS. However, this assembly scheme lacked control as the
surface coverage was not uniform and fibers of various lengths were observed. In addition, it
cannot be ruled out that the fibers observed were formed by SDS interacting with the
hydrophobic monolayer.
A cysteine-modified pilin monomer (PilA19-A20C) was expressed as a fusion using a
chitin binding domain and purified using affinity chromatography. The PilA19-A20C subunits
formed a monolayer when they interacted with gold substrates. PilA19-A20C subunits were
assembled in-vitro, indicating that pilin assembly tolerates certain changes in amino acid
sequence. The lack of attachment of cysteine-modified pili to the gold substrates suggests that
the cysteine at the N-terminal end of the fiber is buried inside the helical core and not available
for interaction with the gold substrate.

107

Figure 4.1 Structural comparisons of wild type and recombinant pilins of Geobacter sulfurreducens. Structure of a) wild type PilA
protein; b) recombinant 19 amino acid truncation PilA19 protein; c) recombinant cysteine-modified 19 amino acid truncation PilA19A20C protein, with the added cysteine colored turquoise. In each structure, the tyrosines are colored light blue. d) Sequence alignment
of wild type and recombinant pilins of Geobacter sulfurreducens.

108

Figure 4.2 Formation of an undecanethiol self-assembled monolayer on a gold electrode. Cyclic
voltammograms of bare gold (black line) and undecanethiol SAM (red dashed line). Data
recorded at room temperature in 100 mM sodium phosphate buffer at pH 7.0 containing100 mM
-1
NaCl, and 5 mM K3[Fe(CN)6] at a potential scan rate of 100 mV s . The insert shows the
thickness of the undecanethiol monolayer measured by ellipsometry.

109

Figure 4.3 Scanning electron microscopy images of gold substrates after deposition of an undecanethiol SAM followed by incubation
with PilA19 monomers in 20 mM sodium phosphate pH 7.0, 100 mM NaCl, 10 mM SDS. Accelerating voltage 5 kV.

110

Figure 4.4 Energy Dispersive X-ray (EDS) microanalysis of fibers present on gold substrates
modified by deposition of an undecanethiol SAM, followed by incubation with PilA19
monomers in 20 mM sodium phosphate pH 7.0, 100 mM NaCl, 10 mM SDS. Accelerating
voltage 1 kV.

111

Figure 4.5 Expression and purification of recombinant PilA19-A20C subunits: a) 12% SDS-

+

PAGE and western blot analysis of cells before (A ) and after (A ) induction. Anti-chitin serum
was used for western blot. b) MALDI-TOF mass spectrometry and c) 10-20% tricine gel of
elution fraction after purification.

112

Figure 4.6 Deposition of PilA19-A20C monomers on a gold electrode. Quartz crystal
microbalance (QCM) was used to monitor changes in frequency and dissipated energy after the
addition of pilin monomers.

113

Figure 4.7 Deposition of PilA19-A20C subunits on a gold electrode. Cyclic voltammograms of
bare gold (black line) and undecanethiol SAM (red dashed line). Data recorded at room
temperature in 100 mM sodium phosphate buffer at pH 7.0 containing 100 mM NaCl, and 5 mM
-1
K3[Fe(CN)6] at a potential scan rate of 100 mV s . Insert shows the thickness of PilA19-A20C
monolayer measured by ellipsometry.

114

Figure 4.8 In-vitro assembly of PilA19-A20C fibers. The transmission electron microscopy images were collected at an accelerating
voltage of 100 kV.

115

Chapter 5.
CONCLUSIONS

The properties of self-assembling biological molecules can be exploited for bottom-up
fabrication of well-defined nanostructured materials. This dissertation focused on two
biomolecules: phospholipids and self-organizing peptides. Phospholipids assemble into an array
of structures (vesicles, bilayers and micelles) that serve as models for cell membranes. Peptides
are versatile building blocks with extensive chemical, conformational and functional diversity.
The objectives of this study were to 1) assess the influence of the method of liposome
preparation on selected properties of lipid bilayers, and 2) harness the self-organizing properties
of a special class of peptides to produce synthetic protein nanowires.
To fulfill the first objective, the effect of the method of liposome preparation (sonication
or extrusion) on bilayer properties was investigated. The results indicated that the morphology,
average size, and size distribution of liposomes depends on the method of preparation. However,
the molecular-scale behavior of the liposomes, as studied by fluorescence lifetime and anisotropy
decay measurements, was independent of the method of preparation. In addition, translational
diffusion of sBLMs reconstituted from liposomes was also independent of the method of

116

liposome preparation. The presence of cholesterol in the bilayers decreased membrane fluidity in
both systems. The results indicate that the molecular scale organization of lipid bilayers is
determined by interactions between the constituent species, which does not depend on the
manner in which the bilayer is formed. Consequently, studies on lipid bilayers produced from
liposomes prepared either by sonication or extrusion can be directly compared.
To fulfill the second objective, the investigation focused on the conductive, hair-like
protein appendages or pili produced by the metal-reducing bacterium Geobacter sulfurreducens.
A combination of genetic engineering and heterologous protein expression was used to massproduce the recombinant pilin subunits. The highly hydrophobic nature of the pilin subunits
made their expression as soluble peptides very challenging. Therefore, step-wise deletions of
amino acids from the hydrophobic N-terminus were carried out to produce truncated pilins. The
PilA19 subunit was the shortest truncation (longest pilin) expressed as a soluble peptide. The
secondary structure of the PilA19 monomer was characterized using circular dichroism. The
results showed that the addition of surfactants such as SDS and octyl β-D-glucopyranoside
increased the helical content of the subunits. The assembly of type IV pilins is mediated by
interactions between the N-terminal α-helices of the pilin subunits. In-vitro assembly of the
recombinant pilins resulted in the formation of filaments with properties similar to those of
native G. sulfurreducens pili, including a similar diameter, tendency to aggregate in tangled
bundles, and electrical conductivity. Therefore, deletion of 19 amino acids from the N-terminal
end did not significantly affect the self-assembly process or the conductivity of the synthetic pili.
In an attempt to facilitate interfacing of recombinant subunits with gold electrodes, a
cysteine-modified pilin was engineered (PilA19-A20C), expressed and purified. The results

117

indicated that, while the PilA19-A20 monomers form a monolayer, they do not assemble into pili
on the gold substrate. However, the PilA19-A20C subunits self-assembled into pilus-like
filaments in solution, indicating that pilin assembly can tolerate some changes in amino acid
sequence.
The net recombinant expression, purification, and self-assembly of soluble pilin subunits
from Geobacter sulfurreducens could facilitate structural studies of PilA peptides and enable
application of these versatile structures in nanoelectronics. Conductive, protein nanowires
represent a new platform for design of novel biomimetic interfaces suitable for scientific
investigations of electron transfer mechanisms in Geobacter species, as well as development of
new technologies.

118

REFERENCES

119

REFERENCES
1.

Choi, J.W., et al., Nanotechnology in biodevices. J Microbiol Biotechnol, 2007. 17(1): p.
5-14.

2.

Villaverde, A., Nanotechnology, bionanotechnology and microbial cell factories.
Microbial Cell Factories, 2010. 9: p. 4.

3.

Dirks, A.J., et al., From (bio)molecules to biohybrid materials with the click chemistry
approach. Qsar & Combinatorial Science, 2007. 26(11-12): p. 1200-1210.

4.

de la Rica, R. and H. Matsui, Applications of peptide and protein-based materials in
bionanotechnology. Chemical Society Reviews, 2010. 39(9): p. 3499-3509.

5.

Wu, L.-Q. and G.F. Payne, Biofabrication: using biological materials and biocatalysts to
construct nanostructured assemblies. Trends in Biotechnology, 2004. 22(11): p. 593-599.

6.

Whitesides, G.M., The 'right' size in nanobiotechnology. Nat Biotech, 2003. 21(10): p.
1161-1165.

7.

Ball, P., Synthetic biology for nanotechnology. Nanotechnology, 2005. 16(1): p. R1.

8.

Reimhult, E., F. Höök, and B. Kasemo, Temperature dependence of formation of a
supported phospholipid bilayer from vesicles on SiO2. Phys Rev E Stat Nonlin Soft
Matter Phys, 2002. 66(5 Pt 1): p. 051905.

9.

Keller, C.A., et al., Formation of supported membranes from vesicles. Phys Rev Lett,
2000. 84(23): p. 5443-6.

10.

Keller, C.A. and B. Kasemo, Surface specific kinetics of lipid vesicle adsorption
measured with a quartz crystal microbalance. Biophys J, 1998. 75(3): p. 1397-1402.

11.

Cui, H., M.J. Webber, and S.I. Stupp, Self-assembly of peptide amphiphiles: From
molecules to nanostructures to biomaterials. Peptide Science, 2010. 94(1): p. 1-18.

12.

Reimhult, E., F. Hook, and B. Kasemo, Vesicle adsorption on SiO2 and TiO2:
Dependence on vesicle size. Journal of Chemical Physics, 2002. 117(16): p. 7401-7404.
120

13.

Hope, M.J., et al., Generation of multilamellar and unilamellar phospholipid vesicles.
Chemistry and Physics of Lipids, 1986. 40(2-4): p. 89-107.

14.

Weiner, N., Phospholipid Liposomes: Properties and Potential Use in Flavor
Encapsulation, in Flavor Technology. 1997, American Chemical Society. p. 210-218.

15.

Oliger, P., et al., Preparation and Properties of Vesicles Formed from Phospholipid
Analogues of N-(Phosphonoacetyl)-L-aspartate (PALA) by Sonication or Extrusion:
Transition Temperature, Particle Size, Glucose Entrapment, and 31P NMR. Langmuir,
2001. 17(21): p. 6426-6432.

16.

Lasic, D.D., The mechanism of vesicle formation. Biochem. J., 1988. 256(1): p. 1-11.

17.

Reguera, G., et al., Extracellular electron transfer via microbial nanowires. Nature,
2005. 435(7045): p. 1098-101.

18.

Collinson, S.K., et al., Purification and characterization of thin, aggregative fimbriae
from Salmonella enteritidis. J. Bacteriol., 1991. 173(15): p. 4773-4781.

19.

Craig, L. and J. Li, Type IV pili: paradoxes in form and function. Curr Opin Struct Biol,
2008. 18(2): p. 267-77.

20.

Craig, L., et al., Type IV Pilus Structure by Cryo-Electron Microscopy and
Crystallography: Implications for Pilus Assembly and Functions. Molecular Cell, 2006.
23(5): p. 651-662.

21.

Malvankar, N.S., et al., Tunable metallic-like conductivity in microbial nanowire
networks. Nat Nano, 2011. 6(9): p. 573-579.

22.

Reguera, G., et al., Biofilm and nanowire production leads to increased current in
Geobacter sulfurreducens fuel cells. Appl Environ Microbiol, 2006. 72(11): p. 7345-8.

23.

Dieckelmann, M., L.F. Roddam, and M.P. Jennings, Purification of post-translationally
modified proteins from bacteria: homologous expression and purification of histidinetagged pilin from Neisseria meningitidis. Protein Expr Purif, 2003. 30(1): p. 69-77.

121

24.

Reimhult, E., F. Höök, and B. Kasemo, Intact Vesicle Adsorption and Supported
Biomembrane Formation from Vesicles in Solution: Influence of Surface Chemistry,
Vesicle Size, Temperature, and Osmotic Pressure. Langmuir, 2002. 19(5): p. 1681-1691.

25.

Seantier, B., et al., Dissipation-enhanced quartz crystal microbalance studies on the
experimental parameters controlling the formation of supported lipid bilayers. J Phys
Chem B, 2005. 109(46): p. 21755-65.

26.

Boxer, S.G., Molecular transport and organization in supported lipid membranes.
Current Opinion in Chemical Biology, 2000. 4(6): p. 704-709.

27.

Schonherr, H., et al., Vesicle adsorption and lipid bilayer formation on glass studied by
atomic force microscopy. Langmuir, 2004. 20(26): p. 11600-6.

28.

Richter, R.P. and A.R. Brisson, Following the formation of supported lipid bilayers on
mica: a study combining AFM, QCM-D, and ellipsometry. Biophys J, 2005. 88(5): p.
3422-33.

29.

Reimhult, E., et al., A multitechnique study of liposome adsorption on Au and lipid
bilayer formation on SiO2. Langmuir, 2006. 22(7): p. 3313-3319.

30.

Liposomes methodos and protocols. 1st edition ed. Methods in molecular biology, ed.
S.C. Basu and M. Basu. Vol. 199. 2002, Totawa, NJ: Humana Press Inc.

31.

Moon, M.H. and J.C. Giddings, Size distribution of liposomes by flow field-flow
fractionation. Journal of Pharmaceutical and Biomedical Analysis, 1993. 11(10): p. 911920.

32.

Carrozzino, J.M., et al., Characterization of small unilamellar vesicles using
solvatochromic π* indicators and particle sizing. Journal of Biochemical and Biophysical
Methods, 2004. 60(2): p. 97-115.

33.

Ma, Z., et al., Improved method of preparation of supported planar lipid bilayers as
artificial membranes for antigen presentation. Microscopy Research and Technique,
2011. 74(12): p. 1174-1185.

34.

Richter, R., A. Mukhopadhyay, and A. Brisson, Pathways of lipid vesicle deposition on
solid surfaces: A combined QCM-D and AFM study. Biophysical Journal, 2003. 85(5): p.
3035-3047.

122

35.

Lovley, D.R., Cleaning up with genomics: applying molecular biology to bioremediation.
Nat Rev Micro, 2003. 1(1): p. 35-44.

36.

Methe, B.A., et al., Genome of Geobacter sulfurreducens: Metal Reduction in Subsurface
Environments. Science, 2003. 302(5652): p. 1967-1969.

37.

Juarez, K., et al., PilR, a Transcriptional Regulator for Pilin and Other Genes Required
for Fe(III) Reduction in Geobacter sulfurreducens. Journal of Molecular Microbiology
and Biotechnology, 2009. 16(3-4): p. 146-158.

38.

Reguera, G., et al., Possible nonconductive role of Geobacter sulfurreducens pilus
nanowires in biofilm formation. J Bacteriol, 2007. 189(5): p. 2125-7.

39.

Juarez, K., et al., PilR, a transcriptional regulator for pilin and other genes required for
Fe(III) reduction in Geobacter sulfurreducens. J Mol Microbiol Biotechnol, 2009. 16(34): p. 146-58.

40.

Sharma, V. and P.P. Kundu, Biocatalysts in microbial fuel cells. Enzyme and Microbial
Technology, 2011. 47(5): p. 179-188.

41.

Srikanth, S., et al., Electrochemical characterization of Geobacter sulfurreducens cells
immobilized on graphite paper electrodes. Biotechnology and Bioengineering, 2008.
99(5): p. 1065-1073.

42.

Law, N., et al., Formation of Nanoscale Elemental Silver Particles via Enzymatic
Reduction by Geobacter sulfurreducens. Appl. Environ. Microbiol., 2008. 74(22): p.
7090-7093.

43.

Cologgi, D.L., et al., Extracellular reduction of uranium via Geobacter conductive pili as
a protective cellular mechanism. Proceedings of the National Academy of Sciences of the
United States of America, 2011. 108(37): p. 15248-15252.

44.

Gregory, K.B. and D.R. Lovley, Remediation and Recovery of Uranium from
Contaminated Subsurface Environments with Electrodes. Environmental Science &
Technology, 2005. 39(22): p. 8943-8947.

45.

Debabov, V., Electricity from microorganisms. Microbiology, 2008. 77(2): p. 123-131.

123

46.

Klimes, A., et al., Production of pilus-like filaments in Geobacter sulfurreducens in the
absence of the type IV pilin protein PilA. Fems Microbiology Letters, 2010. 310(1): p.
62-68.

47.

Craig, L., M.E. Pique, and J.A. Tainer, Type IV pilus structure and bacterial
pathogenicity. Nat Rev Micro, 2004. 2(5): p. 363-378.

48.

Audette, G.F., et al., DNA-Binding Protein Nanotubes: Learning from Nature's Nanotech
Examples. Nano Letters, 2004. 4(10): p. 1897-1902.

49.

Craig, L., et al., Type IV pilin structure and assembly: X-ray and EM analyses of Vibrio
cholerae toxin-coregulated pilus and Pseudomonas aeruginosa PAK pilin. Molecular
Cell, 2003. 11(5): p. 1139-1150.

50.

Szeto, T.H., A. Dessen, and V. Pelicic, Structure/Function Analysis of Neisseria
meningitidis PilW, a Conserved Protein That Plays Multiple Roles in Type IV Pilus
Biology. Infection and Immunity. 79(8): p. 3028-3035.

51.

Burrows, L.L., Weapons of mass retraction. Mol Microbiol, 2005. 57(4): p. 878-88.

52.

Reguera, G., R.M. Worden, and S.D. Kelly, Elucidate mechanism of electron transfer of
uranium mediated by Geobacter nanowires. 2006, NIEHS: Michigan State University.

53.

Boudard, S., et al., Controlling the pathway of formation of supported lipid bilayers of
DMPC by varying the sodium chloride concentration. Thin Solid Films, 2006. 495(1-2):
p. 246-251.

54.

Cornell, B.A., et al., The lower limit to the size of small sonicated phospholipid vesicles.
Biochimica et Biophysica Acta (BBA) - Biomembranes, 1982. 690(1): p. 15-19.

55.

Woodbury, D.J., et al., Reducing liposome size with ultrasound: Bimodal size
distributions. Journal of Liposome Research, 2006. 16(1): p. 57-80.

56.

DeWitt, L., et al., Determination of ground- and excited-state isomerization barriers for
the oligothiophene 3',4'-dibutyl-2,2':5',2''-terthiophene. Journal of the American
Chemical Society, 1993. 115(25): p. 12158-12164.

124

57.

DelaCruz, J.L. and G.J. Blanchard, Understanding the Balance between Ionic and
Dispersion Interactions in Aqueous Micellar Media. The Journal of Physical Chemistry
B, 2003. 107(29): p. 7102-7108.

58.

Gajraj, A. and R.Y. Ofoli, Quantitative Technique for Investigating Macromolecular
Adsorption and Interactions at the Liquid-Liquid Interface. Langmuir, 2000. 16(9): p.
4279-4285.

59.

Smith, B.A. and H.M. McConnell, Determination of molecular motion in membranes
using periodic pattern photobleaching. Proc Natl Acad Sci U S A, 1978. 75(6): p. 275963.

60.

Chen, Y., et al., Methods to measure the lateral diffusion of membrane lipids and
proteins. Methods, 2006. 39(2): p. 147-53.

61.

Lanni, F. and B.R. Ware, Modulation detection of fluorescence photobleaching recovery.
Review of Scientific Instruments, 1982. 53(6): p. 905-908.

62.

Sprague, B.L. and J.G. McNally, FRAP analysis of binding: proper and fitting. Trends
Cell Biol, 2005. 15(2): p. 84-91.

63.

Kinosita, K., Jr., S. Kawato, and A. Ikegami, A theory of fluorescence polarization decay
in membranes. Biophys J, 1977. 20(3): p. 289-305.

64.

Kreiter, M., et al., Orientation dependence of fluorescence lifetimes near an interface.
The Journal of Chemical Physics, 2002. 117(20): p. 9430-9433.

65.

Steiner, R.F., Fluorescence Anisotropy: Theory and Applications, in Topics in
fluorescence spectroscopu, J.R. Lakowicz, Editor. 1991, Plenum Press: New York, NY.
p. 1-52.

66.

Blanchard, G.J., Picosecond spectroscopic measurement of a solvent dependent change
of rotational diffusion rotor shape. The Journal of Chemical Physics, 1987. 87(12): p.
6802-6808.

67.

Lipari, G. and A. Szabo, Effect of vibrational motion on fluorescence depolarization and
nuclear magnetic resonance relaxation in macromolecules and membranes. Biophysical
Journal, 1980. 30(3): p. 489-506.

125

68.

Lipari, G. and A. Szabo, Effect of librational motion on fluorescence depolarization and
nuclear magnetic resonance relaxation in macromolecules and membranes. Biophysical
Journal, 1980. 30(3): p. 489-506.

69.

Kinosita Jr, K., A. Ikegami, and S. Kawato, On the wobbling-in-cone analysis of
fluorescence anisotropy decay. Biophysical Journal, 1982. 37(2): p. 461-464.

70.

Szabo, A., Theory of fluorescence depolarization in macromolecules and membranes.
The Journal of Chemical Physics, 1984. 81(1): p. 150-167.

71.

Fery-Forgues, S., J.-P. Fayet, and A. Lopez, Drastic changes in the fluorescence
properties of NBD probes with the polarity of the medium: involvement of a TICT state?
Journal of Photochemistry and Photobiology A: Chemistry, 1993. 70(3): p. 229-243.

72.

Mazères, S., et al., 7-nitrobenz-2-oxa-1,3-diazole-4-yl-labeled phospholipids in lipid
membranes: differences in fluorescence behavior. Biophysical Journal, 1996. 71(1): p.
327-335.

73.

Greenough, K.P. and G.J. Blanchard, Evaluating the Role of Chromophore Side Group
Identity in Mediating Solution-Phase Rotational Motion. The Journal of Physical
Chemistry A, 2006. 111(4): p. 558-566.

74.

Starr, T.E. and N.L. Thompson, Formation and characterization of planar phospholipid
bilayers supported on TiO2 and SrTiO3 single crystals. Langmuir, 2000. 16(26): p.
10301-10308.

75.

Starr, T.E. and N.L. Thompson, Fluorescence pattern photobleaching recovery for
samples with multi-component diffusion. Biophysical Chemistry, 2002. 97(1): p. 29-44.

76.

Starr, T.E. and N.L. Thompson, Fluorescence pattern photobleaching recovery for
samples with multi-component diffusion. Biophys Chem, 2002. 97(1): p. 29-44.

77.

Kasotakis, E., et al., Design of metal-binding sites onto self-assembled peptide fibrils.
Biopolymers, 2009. 92(3): p. 164-72.

78.

Kumara, M.T., S. Muralidharan, and B.C. Tripp, Generation and characterization of
inorganic and organic nanotubes on bioengineered flagella of mesophilic bacteria. J
Nanosci Nanotechnol, 2007. 7(7): p. 2260-72.

126

79.

Audette, G.F., R.T. Irvin, and B. Hazes, Purification, crystallization and preliminary
diffraction studies of the Pseudomonas aeruginosa strain K122-4 monomeric pilin. Acta
Crystallogr D Biol Crystallogr, 2003. 59(Pt 9): p. 1665-7.

80.

Audette, G.F. and B. Hazes, Development of protein nanotubes from a multi-purpose
biological structure. J Nanosci Nanotechnol, 2007. 7(7): p. 2222-9.

81.

Menzella, H.G., E.A. Ceccarelli, and H.C. Gramajo, Novel escherichia coli strain allows
efficient recombinant protein production using lactose as inducer. Biotechnology and
Bioengineering, 2003. 82(7): p. 809-817.

82.

Li, K., et al., Fed-batch fermentor synthesis of 3-dehydroshikimic acid using recombinant
Escherichia coli. Biotechnology and Bioengineering, 1999. 64(1): p. 61-73.

83.

Wallace, B.A. and R.W. Janes, eds. Modern technices for circular dichroism and
synchrotron radiation circular dichroism spectroscopy. Advances in Biomedical
Spectroscopy, ed. P.I. H. 2009, IOS Press Amsterdam. 256.

84.

Greenfield, N.J., Using circular dichroism spectra to estimate protein secondary
structure. Nat. Protocols, 2007. 1(6): p. 2876-2890.

85.

Mihalyi, E., Numerical values of the absorbances of the aromatic amino acids in acid,
neutral, and alkaline solutions. Journal of Chemical & Engineering Data, 1968. 13(2): p.
179-182.

86.

Veazey, J.P., Scanning probe studies of the pilus nanowires in Geobacter sulfurreducens,
in Physics. 2011, Michigan State University: East Lansing. p. 149.

87.

Kyte, J. and R.F. Doolittle, A simple method for displaying the hydropathic character of
a protein. Journal of Molecular Biology, 1982. 157(1): p. 105-132.

88.

Malhotra, A., R.B. Richard, and P.D. Murray, Chapter 16 Tagging for Protein
Expression, in Methods in Enzymology. 2009, Academic Press. p. 239-258.

89.

Gentz, R., et al., Association of degradation and secretion of three chimeric polypeptides
in Escherichia coli. J. Bacteriol., 1988. 170(5): p. 2212-2220.

127

90.

Fong, B.A., W.-Y. Wu, and D.W. Wood, The potential role of self-cleaving purification
tags in commercial-scale processes. Trends in Biotechnology, 2010. 28(5): p. 272-279.

91.

Xu, M.-Q. and T.C. Evans Jr., Purification of Recombinant Proteins from E. coli by
Engineered Inteins, in E. coli Gene Expression Protocols, P.E. Vaillancourt, Editor.
2003, Humana Press: Totowa, New Jersey. p. 43-68.

92.

Switzer, R.L. and L.F. Garrity, Experimental Biochemistry. 3rd edition ed. 1999, New
York, NY: W. H. Freeman and Company. 450.

93.

Matsuo, K., H. Watanabe, and K. Gekko, Improved sequence-based prediction of protein
secondary structures by combining vacuum-ultraviolet circular dichroism spectroscopy
with neural network. Proteins: Structure, Function, and Bioinformatics, 2008. 73(1): p.
104-112.

94.

Davis, A.M., S.A. St-Gallay, and G.J. Kleywegt, Limitations and lessons in the use of Xray structural information in drug design. Drug Discovery Today, 2008. 13(19-20): p.
831-841.

95.

Montserret, R., et al., Involvement of Electrostatic Interactions in the Mechanism of
Peptide Folding Induced by Sodium Dodecyl Sulfate Binding. Biochemistry, 2000.
39(29): p. 8362-8373.

96.

Watts, T.H., D.G. Scraba, and W. Paranchych, Formation of 9-nm filaments from pilin
monomers obtained by octyl-glucoside dissociation of Pseudomonas aeruginosa pili.
Journal of Bacteriology, 1982. 151(3): p. 1508-1513.

97.

Hjelmeland, L.M., A. Chrambach, and B.J. William, [16] Solubilization of functional
membrane proteins, in Methods in Enzymology. 1984, Academic Press. p. 305-318.

98.

le Maire, M., P. Champeil, and J.V. Møller, Interaction of membrane proteins and lipids
with solubilizing detergents. Biochimica et Biophysica Acta (BBA) - Biomembranes,
2000. 1508(1-2): p. 86-111.

99.

Whitmore, L. and B.A. Wallace, DICHROWEB, an online server for protein secondary
structure analyses from circular dichroism spectroscopic data. Nucleic Acids Research,
2004. 32(suppl 2): p. W668-W673.

128

100.

Whitmore, L. and B.A. Wallace, Protein secondary structure analyses from circular
dichroism spectroscopy: Methods and reference databases. Biopolymers, 2008. 89(5): p.
392-400.

101.

Provencher, S.W. and J. Gloeckner, Estimation of globular protein secondary structure
from circular dichroism. Biochemistry, 1981. 20(1): p. 33-37.

102.

van Stokkum, I.H.M., et al., Estimation of protein secondary structure and error analysis
from circular dichroism spectra. Analytical Biochemistry, 1990. 191(1): p. 110-118.

103.

Sreerama, N. and R.W. Woody, Estimation of Protein Secondary Structure from Circular
Dichroism Spectra: Comparison of CONTIN, SELCON, and CDSSTR Methods with an
Expanded Reference Set. Analytical Biochemistry, 2000. 287(2): p. 252-260.

104.

Scanning tunneling microscopy and spectroscopy: theory, techniques and applications,
ed. D.A. Bonell. 2001, New Yoork: Wiley-VCH.

105.

Chen, C.J., Introduction to scanning tunneling microscopy. 2nd ed. Oxford series in
optical and imaging sciences. 2008, New York: Oxford University Press.

106.

Howorka, S., Rationally engineering natural protein assemblies in nanobiotechnology.
Current Opinion in Biotechnology, 2011. 22(4): p. 485-491.

107.

Honda, M. and et al., Metal-molecular interface of sulfur-containing amino acid and
thiophene on gold surface. Journal of Physics: Conference Series, 2008. 100(5): p.
052071.

108.

Bard, A.J. and L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications.
1st ed. 1980, New York: John Wiley & Sons, Inc.

109.

Johannsmann, D., I. Reviakine, and R.P. Richter, Dissipation in Films of Adsorbed
Nanospheres Studied by Quartz Crystal Microbalance (QCM). Analytical Chemistry,
2009. 81(19): p. 8167-8176.

110.

Kotarek, J.A., K.C. Johnson, and M.A. Moss, Quartz crystal microbalance analysis of
growth kinetics for aggregation intermediates of the amyloid-beta protein. Analytical
Biochemistry, 2008. 378(1): p. 15-24.

129

111.

Bradford, D.C., et al., Surface-Enhanced Infrared Ellipsometry of Self-Assembled
Undecanethiol and Dodecanethiol Monolayers on Disordered Gold Nanoisland
Substrates. The Journal of Physical Chemistry B, 2005. 109(44): p. 20914-20922.

112.

Manne, S., et al., Direct Visualization of Surfactant Hemimicelles by Force Microscopy
of the Electrical Double Layer. Langmuir, 1994. 10(12): p. 4409-4413.

113.

Levchenko, A.A., et al., Kinetics of Sodium Dodecyl Sulfate Adsorption on and
Desorption from Self-Assembled Monolayers Measured by Surface Plasmon Resonance.
Langmuir, 2002. 18(22): p. 8464-8471.

114.

Sek, S., K. Swiatek, and A. Misicka, Electrical Behavior of Molecular Junctions
Incorporating Alpha-Helical Peptide. The Journal of Physical Chemistry B, 2005.
109(49): p. 23121-23124.

115.

Wu, J.-C., et al., Controlled growth of aligned alpha-helical-polypeptide brushes for
tunable electrical conductivity. Applied Physics Letters, 2011. 98(13): p. 133304.

130