3. .

A .wwfiafi
.. 9.2,

3.15.,

H...» .. ..., . ,3
i 5,. , .

1w..." 3%..“ n5.

Lhasa“... .... #1.}.

A an: ".5“le g1 .

.2...- . Ifl‘llciusmtv . Bur vfimufi.‘
. “3...... n. ad.

.n..

0!? 51...)!
tit:: 10....

3.1.
Rail".
WA»: 7.-

.. ,
V . 3.:
$31.. a
. s .... m9:
43km
.. 12
v.

.. . .
.flmfififififl?
.. I I ”a

fit
.mw.

“Havana“... at .
. . . 1:1..th

It’l.\:lv.t 5.....«sllatsvbz

1;! tintvkal 9:...niivaln 410.

1.2.3.3.! 30.3%:
it: .63.: n ..

 

J. .1 Eli
.- qui..ial
: 5134...:

$.31
a?

££txhfi|flhflil§§ .
Z :1Ilivvl-‘l4lI‘
211:5). i1 1

5,001.81:
.lxlizzll
5131:»..‘3‘

35-9..

14!: ’3':
$11!. .

is. méwuffiiufi E17,.

2

a

543%.;
. .

kin: P23
£53.,
€1,152

L!
r ... . rt... 3

:4 .t.

¢ K

. V. . . .a
A. L1 : .

 

In... ..'J
:3. L: .
. 1..

 

 

LIBPARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

Formation and Characterization of Air Stable
Phospholipid Adlayers on Modified Substrates

presented by

Benjamin Oberts

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistq

 

% gflrz

( VMajor Professor’s Signature
‘3/ ’%W

Date

MSU is an Affinnative Action/Equal Opportunity Employer

 

 

is checkout from your record.
T0 AVOlD FINES return on or before date due.

PLACE IN RETURN BOX [0 remove [h

MAY BE RECALLED with earlier due date if requested.

DATE DUE

 
  
 

  
  

     

DATE DUE

  

DATE DUE

 

   

FORMATION AND CHARACTERIZATION OF AIR STABLE
PHOSPHOLIPID ADLAYERS ON MODIFIED SUBSTRATES

By

Benjamin Oberts

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Chemistry

2009

 

 

 

 

 

ABSTRACT

FORMATION AND CHARACTERIZATION OF AIR STABLE PHOS-
PHOLIPID ADLAYERS ON MODIFIED SUBSTRATES

By

Benjamin Oberts

Plasma membranes are essential to the function of cellular systems. As a con-
sequence of this fact many proteins and other biological species are not active outside
of their native environment. It would be ideal to create an artificial structure that could
house selected biomolecules in their active forms, enabling their use in applications such
as bio-sensing, for example. As the understanding of these complex and dynamic bilayer
structures increases, it has become clear that their stability in air is not sufficient for use
in many chemical sensing applications. For this reason, the work presented here was
undertaken to find ways to optimize interactions between lipids and planar substrates,
enabling the formation of air-stable lipid adlayers. Several phospholipids were explored;
1,2-dimyristoyl-sn-glycer0-3-phosphatidic acid (DMPA), l,2—dimyristoyl-sn—glycero-3-
phosphocholine (DMPC), 1,2-dimyristoyl-sn-glycero-3—phosphoethanolamine (DMPE),
l,2-dimyristoyl-sn-glycero-3-[phospho-rac-( 1 -glycerol)] (DMPG), and 1,2-dimyristoyl-
sn-glycero—3-[phospho-L—serine] (DMPS). These lipids, in vesicle form, were exposed
to chemically modified substrates and underwent vesicle fusion to create lipid adlay-
ers. The Au substrates on which the adlayers were deposited were modified to interact
with the phospholipid headgroups. The resulting lipid adlayer was stable with respect to

transport across the water/air boundary. The chemical modification of the interface was

A

 

 

arson
actcn.
reacts
funcli
mlh [
into [1’
lipid a
3M
sim in
lechm
Self-(l:

blill}' l

accomplished using 6—mercapto-1-hexanol to form a self-assembled monolayer char-
acterized by a polar, hydrophilic interface. The 6—mercapto-l-hexanol monolayer was
reacted with POCl3 and water to create a phosphate-terminated interface. The phosphate
functionalities were populated with Zr4+ ions, rendering them capable of complexation
with phospholipid phosphate moieties. Other metal salts were also used to gain insight
into the effect of metal ion identity on the binding of phospholipids. To characterize the
lipid adlayers, time correlated ellipsometry, water contact angle, cyclic voltametery, XPS,
31P-NMR and F TIR measurements where used. All of the analyses performed were ex-
situ from the lipid deposition vessel, demonstrating air-stable adlayer formation. Each
technique interrogated a different adlayer property. The discovery of a novel family of
self-assembling adlayer comprised of biologically important molecules opens the possi-

bility of future success in the creation of robust biomimetic interfacial structures.

 

wan: 7 "mm
mi“? »- -‘

 

 

Copyright by
Benjamin Oberts
2009

 

 

 

To my wife Carrie
and our cat Lucky

Acknowledgements

I would like to first thank my advisor Gary Blanchard for his support and guid-
ance through out my time in graduate school. He has been a guide through these tough
years and has always been there to point me in the right direction or supply a new idea.

I have learned a lot from his teachings and will carry with me the experience and con-
fidence that he has instilled in me. As he once said to me “the best way to learn how
something works it to take it apart and put it back together.” This holds true and I will
apply this mentality were ever I end up and toward any problem I face. I have been very
fortunate to have known him and work with him.

I would also like to thank my friends and fellow group mates as they have always
been there for me through failed experiments and new discoveries. I will miss our week-
ly lunches out as well as the time we spent together in the lab. Being friends with and
working with such talented and knowledgeable people made it a joy to get in the lab and
made it easier to work the long hours that some experiments required.

I would like to say thanks to my parents Chuck and Leslie Oberts, who have
always been there to support me through all my endeavours. They have always gone far
beyond what was necessary to help me. Much of what I am today I owe to them and their
continued support.

Finally, my greatest thanks goes to my wife Carrie. From moving to Michigan
to constantly reassuring me that I will figure out the next problem she has been my rock.
Her patience and constant belief in me, even when I did not believe in myself, has been
the driving force behind my success. I truly would not have gotten my degree with out
her. Out of all the things in my life I am most proud and thankful to be her husband and

to have her at my side through it all.

vi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TABLE OF CONTENTS
List of Tables viii
List of Figures ix
Chapter 1 - ' “ J 1
1-1: Understanding Lipid Bilayers 2
1-2: Phospholipid Vesicles 3
1-2. 1: F luidity and Impact of Impurities on Lipid Vesicles ....................... 5
1—2.2: Lipid Rafts 7
1—3: Supported Lipid Bilayers 10
1-3. 1: SLB Formatinn 11
l-3.2: Inter-Layer Exchange (T ‘ ‘h-n) 13
1-4: Controlling Vesicle Fusion 14
1-5: Introduction of Performed Work 16
Chapter 2 - Formation of Air-Stable Supported Lipid Monolayers and Bilayers ............. 19
2-1: Introduction to Air-Stable SLRs 19
2-2: Materials and Instrumentation Utili7ed 21
2-3: Experimental Setup 21
2—4; Results 25
241: Air Stable Supported Bilayer F ormatinn 25
2-4.2: Air-Stable Lipid Monolayer F ormminn ' 30
2-4.3: NMR Analysis 34
2-5: I." ' 36
2-6: How Mono- and Bilayers Form on the Modified Au Substrates .................. 37
2-7: Kinetics of Lipid Adlayer F nrmatinn 38
2—8: C ‘ ' 39
Chapter 3 - Phospholipid Headgroup Dependent Assembly of Lipid Adlayers on Zirco-
nium Phosphate-Terminated Interfaces 42
3-1: Introduction to Surface Modification 42
3-2: Experimental Set up and Materials Used 45
3-2.1: Sample F r " 46
3-3: Individual Lipid Results 47
3-3.1: DMPA 48
3-3.2: DMPF 50
3—3.3: DMPF 52
3—3.4: DMPG 53
3—3.5: DMPS 55
3—4: Overall Lipid Observations 56
3-4.1: Lipids with Strong Headgroup ' ‘ " 58
3-4.2: Lipids with Weak Headgroup ' ‘ " 59

 

3-5; Diemieeinn 59
3-6; (‘nnr‘lncinns 62

Chapter 4 - Ionic Binding of Phospholipids to Interfaces: Dependence on Metal Ion

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Identity 64

4-1: ' ‘“ J " 64

4—2: Metals for Modification of Au Substrate-e 66

4-3: Experimental Setup and Instrumentation Utili7ed 67

4-3.1: Substrate Preparation 68

4-3.2: DMPA Vesicle P r " 69

4-3.3: Adlayer Formation 69

4-42 Results Of DMPA Fansnre 70

4-4.1: 7 ' ' 70

4-4.2: Iron ll] 73

4-3.3: Nickel 74

4-4.4: 7 im‘ 76

4—4.5: Calcium 77

4-4.6: “ a ' 78

4—4.7: Copper 1 80

4-4.8: Copper 11 82

4-5: Comparison of Metal-ion Modified Interfaces 83

4-6: F ' ' 85
Chapter 5 - F ‘ ' 87
Literature Cited 93

 

viii

 

Table 2-1.

Table 3-1.

Table 4-1:

 

LIST OF TABLES

Calculated activation energy for lipid flipping as a function of Arrhenius pref-
actor

Electrochemical data for Ru(NH ) 3+l2+ and Fe(CN) ”4' probes for interfaces
studied here. The SAM indicated in the Table is 6-mercapto—1-hexanol. End-
point values are reported vs. Ag/AgCl reference electrode ........................... 60

Data for the interfaces examined in this work. In the first column, metal ion-
surface coverage reported by the ratio of the XPS M+n to Au4f signal inten-
sities; second column, ellipsometric thickness in A; Water contact angle of
DMPA-terminated interfaces, in degrees; water contact angle hysteresis in de-
grees; Ru+3/+4 CV peak splitting for metal ion-terminated in-terface; Ru+
CV peak splitting for DMPA-terminated interfaces; F e+2/ +3 CV ratio of non-

F aradaic current for metal-terminated interface to DMPA-terminated interface.
All CV data has an associated 0.5 mV error 85

 

Fig

Fig

Fig

F
igi

FigL

LIST OF FIGURES

Figure 1-1: 1,2-Distearoyl-sn-Glycero-3-Phosphocholine (DPPC) ................................... 2

 

Figure 1-2: Model vesicle structure 3

Figure 1-3: Fluorescence micrographs of GUVs at 25°C and compositions of 30% choles-
terol mixed with 2:1 DOPC/DPPC, 1:1 DOPC/DPPC, and 1:2 DOPC/DPPC.
The scale bars indicate 20 um. Figure adapted from Veatch and coworkers.4. 8

Figure 1-4: a) Langmuir-Blodgett and Langmuir-Schaefer lipid deposition. b) Vesicle
fireinn 12

Figure 2-1. Schematic of the Teflon® flow-cell used for the deposition of DMPC adlay-
ers by vesicle fusion. The depth of the reservoir is ca. 2 mm ...................... 23

Figure 2-2: F TIR spectra of (a) a 6-mercapto-l-hexanol self-assembled monolayer de-
posited on a gold surface. (b) a DMPC adlayer deposited on the 6-rnercapto-
l-hexanol-modified gold inter-fem 26

 

Figure 2-3: Cyclic voltammograms of a 6-mercapto-1-hexanol self-assembled mono-
layer deposited on a gold surface (solid line) and of the same interface with
an adlayer of DMPC deposited (dashed line), recorded using (a) 1 mM
K4[Fe(CN)6] in 0.1 M LiClO4 and (b) 1 mM [Ru(NH3)6]Cl3 in 0.1 M KCl as
electrochemical probes 27

 

Figure 2-4: a) Ellipsometric thickness measurements for a 6-mercapto-l-hexanol-mod-
ified Au surface, acquired at a series of reaction times to follow DMPC depo-
sition. b) Water contact angle measurements acquired at a series of reaction
times to follow DMPC deposition the same inter-face. For both panels,
initial data points correspond to t=0 29

Figure 2-5: (a) Ellipsometric thickness measurements for a zirconated 6-mercapto-1-
hexanol/Au self-assembled monolayer, acquired at a series of reaction times
to follow DMPC deposition. (b) Water contact angle measurements acquired
at a series of reaction times for the same interface. For both panels, initial
data points correspond to t=0 30

Figure 2-6. F TIR spectra of (a) a zirconated 6-mercapto-l-hexylphosphate self-assem-
bled mono-layer deposited on a gold surface. (b) a DMPC adlayer deposited
on the zirconated 6-mercapto-l-hexylphosphate-modified gold interface... 31

 

'6‘
—.In'h

Figure 2-7:

Figure 2-8:

Figure 3-1:

Figure 3-2:

Figure 3-3:

Figure 3-4:

Figure 3-5:

 

Cyclic voltammograms of a 6— --mercapto l -hexanol/Au self-assembled mono-
layer reacted with POCl3 and ZrOCl to zirconate the interface (solid line)
and of the same interface with an adlzayer of DMPC deposited (dashed line),
recorded using (A) 1 mM K [Fe(CN)6 ] in 0.1 M LiClO4 and (B) 1 mM
[Ru(NH3 )6]C13 in 0. l M KCl as electrochemical probes4 ............................... 32

3 IP MAS NMR spectra. (A) Silica gel coated with POC13, (B) surface shown
in panel a exposed to ZrOC12(aq), (C) surface shown in panel b exposed to
DMPC 35

Phospholipid headgroups used as well as the overall acyl chain. Base acyl
chain does not change for each lipid with the R group being the different

 

D r

DMPA results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the
dashed line is advancing, and the dotted line is receding. C) CV of DMPA
monolayer (solid line) and blank substrate (dashed line) when exposed to
Ru(NH )6 C13. D) CV of DMPA monolayer (solid line) and blank substrate
(dashed36 lme)3 when exposed to K3Fe(CN)b 49

DMPC results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the
dashed line is advancing, and the dotted line is receding. C) CV of DMPC
monolayer (solid line) and blank substrate (dashed line) when exposed to
Ru(NH )6 Cl. D) CV of DMPC monolayer (solid line) and blank substrate
(dashedaé lrne)3 when exposed to K3Fe(CN)0 51

DMPE results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the
dashed line is advancing, and the dotted line is receding. C) CV of DMPE
monolayer (solid line) and blank substrate (dashed line) when exposed to
Ru(NH )6 Cl. D) CV of DMPE monolayer (solid line) and blank substrate
(dashed36 lrne)3 when exposed to K3Fe(CN)O 53

DMPG results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the
dashed line is advancing, and the dotted line is receding. C) CV of DMPG
monolayer (solid line) and blank substrate (dashed line) when exposed to
Ru(NH )6 C13. D) CV of DMPG monolayer (solid line) and blank substrate
(dashed36 11ne)3 when exposed to K 3OFe(CN) 54

xi

Figure 3-6: DMPS results. A) Ellipsometric thickness measurement in time. B) Water

Figure 3-7:

Figure 4-1:

Figure 4-2:

Figure 4-3:

Figure 4-4:

contact angle in time with the solid line being the initial drop angle, the
dashed line is advancing, and the dotted line is receding. C) CV of DMPS
monolayer (solid line) and blank substrate (dashed line) when exposed to
Ru(NH )6 Cl. D) CV of DMPS monolayer (solid line) and blank substrate
(dashed,6 lrne) when exposed to K3Fe(CN)O 56

A) Overlay of all thickness measurements in time with DMPA (solid line),
DMPC (dashed line), DMPE (dash-dot-dot-dash line), DMPG (dotted line),
DMPS (dash-dot-dash line). B) Overlay of initial drop water contact angle
measurements with DMPA (solid line), DMPC (dashed line), DMPE (dash-
dot- dot-dash line), DMPG (dotted line), DMPS (dash-dot-dash line). C)
Overlay of Ru(N H 36) Cl probe CVs with Blank (solid line), DMPA (dashed
line), DMPC (bold 3solidaline), DMPE (dotted line), DMPG (dash-dot-dot—dash
line), DMPS (dash- dot-dash line). D) Overlay of K F e(CN) probe CVs with
Blank (solid line), DMPA (dashed line), DMPC (bo d solid lrne), DMPE (dot-
ted line), DMPG (dash-dot-dot-dash line), DMPS (dash-dot—dash line) ......... 58

a) XPS spectrum of Zr+4-modified thiol/gold substrate. b) CV of
Ru(NH3)6C13 for a DMPA monolayer on Zr+4-terminated interface (dashed
line) and for the Zr +4-terminated inter-face with no adlayer (solid line). 0)
CV of K F e(CN) for a DMPA monolayer on Zr+4 -terrninated interface

(dashed line) and for the Zr+4-terminated interface with no adlayer (solid
line) 71

a) XPS spectrum of Fe+3 -mod1fied thiol/gold substrate. b) CV of
Ru(NH3)6 Cl3 for a DMPA monolayer on Fe -terminated interface (dashed
line) and for 3the Fe+ 3’-terminated inter-face with no adlayer (solid line). c)
CV of K 36Fe(CN) for a DMPA monolayer on Fe+ 3-terminated interface

(dashed line) and for the F e 3 -term1nated interface with no adlayer (solid
line) 73

3) XPS spectrum of Ni+2-modified throl/ gold substrate. b) CV of
Ru(NH3)6Cl3 for a DMPA monolayer on Ni +2-terrninated interface (dashed
line) and for the Ni 2-terminated inter-face with no adlayer (solid line). c)
CV of K F e(CN) for a DMPA monolayer on Ni+2- terminated interface

(dashed Fine) and6 for the Ni 2-terrninated interface with no adlayer (solid
line) 75

a) XPS spectrum of Zn+2-modified thiol/gold2 substrate. b) CV of

Ru(NH3 )6C1 for a DMPA monolayer on ZnJr 2-terrninated interface (dashed
line) and for the Zn 2-terminated inter-face with no adlayer (solid line). c)
CV of K F e(CN) for a DMPZA monolayer on Zn+ 2-terminated interface
(dashed Fine) and 6for the ZnJr 2-terminated interface with no adlayer (solid
line) 76

 

 

 

ll
1

hi!
.1;

i

Fist
5

f'a-JUF:

f‘m‘ ..
ru‘
: R

 

 

Figure 4—5:

Figure 4—6:

Figure 4—7:

Figure 4—8:

 

a) XPS spectrum of Ca+2-modified thiol/gold substrate. b) CV of

Ru(NH3)6 C13 for a DMPA monolayer on Ca 2-terminated interface (dashed
line) and for the Ca 2-terminated inter-face with no adlayer (solid line). c)
CV of K F e(CN) for a DMPA monolayer on Ca+ 2-terminated interface
(dashed Fine) and6 for the Ca +2-terrninated interface with no adlayer (solid
line) _ 78

a) XPS spectrum of Mg+2—modified thiol/gold2 substrate. b) CV of

Ru(NH3 ) 6Cl3 for a DMPA monolayer on Mg+ 2-terminated interface (dashed
line) and 6for the Mg 2-terminated inter-face with no adlayer (solid line). 0)
CV of K Fe(CN)6 for a DMPA monolayer on Mg+ 2-terminated interface
(dashed fine) and 6for the Mg 2-terminated interface with no adlayer (solid
line) 79

a) XPS spectrum of Cu+ -modified thiol/gold substrate. Inset shows Cu2P
spectral region. b) CV of Ru(NH3 )6 C13 for a DMPA monolayer on Cu+ -ter-
minated interface (dashed line) and 63for the Cu +-terminated interface with no
adlayer (solid line). c) CV of K F e(CN) for a DMPA monolayer on Cu+ -
terminated interface (dashed line) and for the Cu+-terminated interface with
no adlayer (solid line)

a) XPS spectrum of Cu 2-modified thiol/gold substrate. Inset shows Cu2P
spectral region. b) CV of Ru(NH3 )6Cl3 for a DMPA monolayer on Cu+ 2-ter-
minated interface (dashed line) an3d for the Cu +2-terminated interface with no
adlayer (solid line). 0) CV of Fe(CN) for a DMPA monolayer on Cu +2-
terminated interface (dashed line) and for the Cu 2-terrninated inter—face with
no adlayer (solid line) 83

xiii

 

 

Chapter 1

Introduction

Biological systems are complex and dynamic, and consequently it is difficult to
gain molecular scale insight into their behavior. The cellular level is the typical start-
ing point for the evaluation of biological function, but progress in our ability to measure
complex molecular structures at low concentrations has lead to further questions on how
such molecular complexity serves to produce functional biological systems. Mammalian
cells contain components such as mitochondria, a nucleus (except for erythrocytes), and
a cell membrane, among others. Each of the sub-structures within a cell is character-
ized by significant molecular complexity and there is much current research activity in
the area of detailed characterization of the component parts of cellular systems. Plasma
membranes serve as an excellent example of this complexity. The lipid bilayer structure
that defines the plasma membrane serves a critical role in supporting trans-membrane
proteins, which regulate cellular functions such as transport of ions and other species
into and out of the cell. Perhaps the most basic function of the plasma membrane is to
maintain cellular integrity and serve as an impediment to pathogens. Mammalian plasma
membranes are composed of more than 100 different components,1 including glycero-
phospholipids, sphingolipids and cholesterol, which make up ~50% of the cell wall. The
balance of the plasma membrane is comprised of integral and peripheral proteins, and a
detailed inventory of plasma membrane components varies according to cell fimction.2’3
The complexity of the plasma membrane poses a substantial challenge to achieving a
detailed understanding of membrane form and function. It is thought that this composi-
tional complexity plays a role in stabilizing the folding of transmembrane proteins and

thus mediating their function. It is clear that the relationship between molecular-scale

 

membrane composition and dynamics and transmembrane protein function is a prerequi-

site to the broad-based use of synthetic lipid bilayers in applications ranging from cellular
function to chemical and/or biological sensing. It is not the purpose of this dissertation to
achieve a connection between bilayer composition and transmembrane protein function.
The focus of this work is on the more fundamental issue of how bilayer composition and

interactions with interfaces can serve to mediate lipid dynamics and organization.

1-1: Understanding Lipid Bilayers

A prerequisite for the study of lipid bilayer structures is the ability to form such
structures in a reproducible manner. Given the complexity of plasma membranes and the
desire to construct biomimetic structures that are compositionally simpler, the primary
focus of this work will be on phospholipids. The choice of phospholipids has been made
because this family of molecules comprises the largest fraction of plasma membranes and
because phospholipids can form bilayer structures without the addition of other compo-
nents, such as sterols, for example. Phospholipids are amphipathic, containing two
hydrophobic carbon tails that are bound to any of several hydrophilic head group func—

tionalities through a phosphate linkage. Fig. 1-1 shows a phosphocholine as an example

Fig. 1-1: 1,2-Distearoyl-sn-Glycero-3-Phosphocholine (DSPC)

phospholipid. The head group (choline is shown in Fig. 1.1) can vary in size, polarity,
and charge while the acyl chains vary in length and the presence of unsaturation(s). The
acyl chains can either be identical or different, both in length and in the presence of

unsaturations. This structural variability allows for complex mixtures of phospholipids in

 

 

 

l‘ill
org;
izcc
Sign
diw
are i
158 l
llOlCL
int-err
1‘13";

these

l-I. P'

 

 

 

 

 

bilayer structures, yielding a commensurate complexity in terms of understanding bilayer

organization and stability. Lipid bilayers are quasi-two dimensional structures character-
ized by compositional heterogeneity and fluidity, and it is this latter property that adds
significant complexity to the study of these systems.4'l6 The work presented in this
dissertation focuses on comparatively simple lipid mono- and bilayer structures, which
are intended as starting points in developing supported biomimetic structures for potential
use in areas such as chemical sensing and selective lipid detection.17 It should also be
noted that while this dissertation focuses on planer supported lipid bilayers on modified
interfaces, it is expected that the dynamics and molecular interactions of non-supported
bilayers are going to carry over to supported bilayers. It is then important to first discuss

these interactions.

1—2: Phospholipid Vesicles

Phospholipid vesicles where devised as a possible means of simulating a plasma
membrane. In an aqueous medium, phospholipids will self-assemble to form a bilayer
structure called a vesicle (Fig. 1-2). The key structural features of the vesicle are that the
lipid tail region for each lipid layer is directed inward, toward the center of the bilayer,
with the hydrophilic head groups oriented outward
to be in contact with the (aqueous) solution, and
that the bilayer structure closes on itself to preclude
exposure of the acyl chain region to an aqueous
environment. The presence of the inner lipid layer

differentiates a vesicle from a micelle structure.

 

Through the study of bilayer structures,

Fig 1-2: Model vesicle structure mostly in the form of vesicles, it has been found

Kim "
Ls we

 

 

 

 

that even for binary and ternary systems, phase separation between the constituents can

occurg’15 '18'20 The phase separation of different lipids and/or lipids and sterols plays a
major role in determining the organization and fluidity of the vesicles. For such compara-
tively simple systems to provide meaningful insight into organization in the substantially
more complex plasma membranes, it is important to understand first the organizational
behavior of the model system. It is thought that the phase separation and fluidity that is
characteristic of plasma membranes mediate the function of transmembrane protein.2l’22
The issue of bilayer composition is inherently related to bilayer fluidity. Bilay-
ers exhibit phase transitions between a variety of phases, depending on the composition
of the bilayer and the temperature. It has been observed that several phase transitions
occur with increasing temperature. At a low temperature for a given lipid, a bilayer will
exist in a crystalline gel phase (LC), which undergoes a transition to lamellar gel phase
(LB)’ which can undergo a further transition to a rippled gel phase (PB) with increasing
temperature.23 The gel—to—fluid phase transition, which occurs with a further increase in
temperature, is labeled the gel-to-fluid phase transition temperature (Tm), and is thought
to proceed because of the acyl chains undergoing a structural change from predominantly
all-trans to a conformation with a significant contribution fiom trans-gauche conform-
ers.24 The rationale for phase transitions in bilayer systems is the balance between
thermal energy and the attractive inter-chain interactions that operate in the acyl chain
region of the bilayer. It is clear on physical grounds that the value of Tm, or any of the
preceding transitions will be influenced by the length and degree of unsaturation of the
lipid acyl chains. The measurement of Tm is a well established means of characterizing
bilayer structures.21’22’25'27 The relative ease of measuring Tm compared to other phase
transitions, and the sensitivity of the transition temperature to impurities within the bi-

layer, provides an important gauge of lipid purity and suggests that impurities may play

-—

an important if underestimated role in determining the range of literature values that have

been reported for a given system.

1-2. 1: Fluidity and the Effect of Impurities on Lipid Vesicles
Koan and coworkers28 examined the effect of impurities on the phase transition

temperature of the C14 phosphocholine DMPC using time-resolved fluorescence measure-
ments. These experiments reported on the anisotropy decay dynamics of a chromophore
inserted into the acyl chain region of the vesicles as function of the amount of impurity
(14:1 PC) present in the bilayer. It is useful to review this measurement to understand
how Tm is extracted from the anisotropy decay data. The induced orientational anisotro-
py function, which is a measurement of the orientational relaxation of molecules excited
by polarized light to a random distribution, can be performed using time-correlated single
photon counting (TCSPC) spectroscopy. Data collected from this measurement contains
information pertaining to the motion of a chromophore within the lipid bilayer, which
is present in this case in the form of a vesicle. The goal of anisotropy decay measure-
ments is to acquire molecular reorientation information, specifically the reorientation
time constant(s), which depend on the identity of the chromophore and the environ-
ment in which it resides. 1n the case of this work, temperature-dependent changes in the
reorientation time constant reveal the temperature, Tm, at which the gel-to-fluid phase
transition occurs. The chromophore local environment will affect the functional form
of induced orientational anisotropy decay. Calculation of I”(t) — Ii (t)

r = -l)
the anisotropy decay function with equation 1-1 is accom- Ill“) ‘1' 21 J. (t )
plished by taking the normalized difference between the

r(t) = r(0)exp[t——t] (1-2)

0R
flzL (1-3)
kBTS 6D

fluorescence transients polarized parallel and perpendicular

to the incident vertically polarized excitation pulse. In its t OR =

 

simplest interpretation, the anisotropy decay function is fit to equation 1-2 to extract the
reorientation time, Tor. In equation 1-2, r(0) is the initial anisotropy, which is related to
the angle between the excited and emittng transition dipole moments, and can range from
-0.2 to 0.4. Measurement of the reorientation time allows the calculation of the viscos-
ity h according to the modified Debye-Stokes-Einstein equation (equation 1~3).29'31 In
this equation, Vis the hydrodynamic volume of the chromophore, calculated according
to Edward’s formulation,3 2 k1) is the Boltzmann constant, T is the temperature in K, f is
the “friction” coefficient for the solvent-solute interaction boundary condition, and S is
the shape factor to account for non-spherical shapes, which ranges from 0 to 1. This data
yields a general picture of the system.

Koan et al,28 used perylene as the “probe” chromophore. Perylene has a well
characterized linear response and is a planar polycyclic aromatic hydrocarbon, a structure
useful for the interrogation of the lipid bilayer acyl chain region. A time-correlated single
photon counting apparatus was used and the results from this system allowed detailed
information to be obtained on the chromophore local environment.33'34 The bilayer sys-
tem used was composed of two different lipids; 1,2-dimyri.,:v ,1 5:1 _ ‘ r‘ “I ‘ ' "
(14:0 PC) and an “impurity” of l,2-dimyristoyl-sn-glycero—3-phophocholine (14:1 PC).
Fluorescence lifetime and anisotropy data was collected for perylene in vesicles contain-
ing controlled ratios of 14:0 PC to 14:1 PC at several temperatures. These data exhibited
a discontinuous change in tor at Tnn which is interpreted in the context of a change in the
molecular scale organization of the lipid bilayers.

Results from this work indicate that impurities can significantly influence the
phase transition temperature Tm.28 Vesicles of 14:0 PC exhibited Tm at 24°C, in agree-
ment with the literature.35 Upon addition of 14:1 PC “impurity”, Tm decreased consider-

ably. F or 0.3 mol% of 14:1 PC, Tm decreased by ~15°.28 Increasing the concentration

 

of 14:1 PC further lowered Tm, but the change in Tm with 14:1 PC not as great as it was

for the initial addition. Tm changes with increasing 14:1 PC concentration exhibit a
monotonic decrease, demonstrating that a relatively small amount of impurity gives rise
to large changes in the organization of the bilayer. It is interesting to note that while the
value of Tm changed significantly with the presence of the impurity, the viscosities sensed
by the chromophore did not. For the fluid phase viscosities were in the range of 8.5 :L- 1.5
cP and the gel phase viscosities lied in the range of 14.5 i 2.5 CP.27 While the impurities
may perturb the organization of the bilayers, the interactions responsible for maintaining
bilayer structure do not change.

The overall conclusion to be drawn by this work is that even the simplest systems
pose problems in understanding their intrinsic complexity at the molecular scale, and
further work is necessary to develop an understanding before it will be possible to create
systems that can function as biomimetic plasma membrane structures. While this study
focused on the bilayers in the form of vesicles, the interactions probed are pertinent to

supported bilayer structure described in this work.

l-2.2: Lipid Rafts

Lipid heterogeneity in vesicle systems has lead to further observations and chal-
lenges in regard to the complexity of these systems. Binary mixtures of cholesterol and
phospholipids are found to be miscible however, when cholesterol is mixed with two
phospholipids of varying melting temperature (Tm) to form a ternary mixture, cholesterol
can complex preferably with one phospholipid creating “rafis” or condensed complexes
that phase separate.”39 Rafts have been observed as distinct domains with reduced
diffusion within cell membranes.2’8 These physical properties may suggest that these

organized regions could be locations for specific proteins to interact with the membrane,

 

which could lead to further bilayer satiability. It has also been theorized that some pro-

teins require the presence of raft structures and, as such, pure phospholipid bilayers will
not provide the required biomimetic environments for functional biomolecules.3’8 These
theories suggest that lipid rafts are important due to their putative role in mediating pro-
tein functionality as well as other cellular functionsm’39

It has been proposed that the cholesterol rafts will form between mixtures of
cholesterol and saturated lipids because cholesterol is immiscible in unsaturated lip-
ids.38 Mixtures of cholesterol and two lipids with different melting points have exhibited
two liquid phases, particularly if the Tm’s for the two different lipids are widely differ-
ent.“”’4042 To observe and characterize these lipid-cholesterol domains, Veatch and co-
workers43 utilized lH-NMR techniques and fluorescence microscopy. Giant unilarnellar
vesicles (GUVs) were created that were composed of 30% cholesterol and varying ratios
of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and l,2-dipaplmitoyl-sn-glycero-
3-phosphocholine (DPPC). 1H-NMR data reveal that not only are domains being formed
but it is possible to map the phase boundary of the domains from one relatively homog-
enous domain in the vesicles, to two liquid domains containing lipid and cholesterol or
lipid only. Fluorescence microscopy was utilized to image the GUVs for that investiga-
tion. Fig. 1-3 shows the fluorescence micrographs of the GUVs at different ratios of lip-
ids.43 It is clearly visible from
Fig. 1-3 that lipid domains are

being formed, as well as upon

 

larger concentrations of the

Fig. 1—3: Fluorescence micrographs of GUVs at 25°C miscible lipid DPPC, the do-
and compositions of 30% cholesterol mixed with 2:1

DOPC/DPPC, 1:1 DOPC/DPPC, and 1:2 DOPC/DPPC. mains become larger until they
The scale bars indicate 20 um. Figure adapted from

encom ass the ma‘ '
Veatch and coworkers.42 p jorrty 0f the

 

GUV, leaving DOPC minor domains. Both the 1H-NMR and fluorescence data indicate

the formation of two phase domains. From this work it was found that the domains can
be as small as ca. 80 nm and projected phase diagrams are consistent with other studies.“
To quantify the results reported by Veatch and coworkers,43 a thermodynamic
model was developed by Radhakrishnan and McConnell.36’37 In their model it is as-
sumed that there are four different species in the liquid phase of the bilayer at equilibri-
um. These consist of un-reactive lipids (DOPC), reactive lipids (DPPC), free cholesterol,
and the reactive lipid-cholesterol complex. Assuming all species are present in equimolar

0
amounts, the Gibbs free energy can be express as Equations 1-4.37 In this equation ”1
_ 0 0
G _Zix,(m +kT1nxi)+2kZi<jx,xj]}j (1—4)

is the chemical potential of pure component i and xi is the equilibrium mole fraction. In
the simplest model, all of the chemical potentials would be constant except that of the
complex, which would leave —kT In K, while all of the critical temperatures (Ti)?
are assumed to be below 298 K with the exception of the binary un-reactive/complex pair
labeled Tc0 .37 From these assumptions it is then possible to model the phase behavior
of the DOPC/DPPC/Cholesterol system for varying temperatures and equilibrium con-
stants for the complex (K). This approach allows for a qualitative model to account for
the experimental data.

To firrther understand these multi-component systems, allowing for their poten-
tial future use, significant work to incorporate functioning proteins in to these structures
has been investigated. Evidence that rafts could be the site of important protein lipid

interactions has surfaced.45

One such study focused on the disorder of extracellular
inorganic phosphates. It has been observed that type IIa renal apical brush border mem-

brane (BBM) sodium/phosphate co-transporter (Na/Pi) protein systems play a major role

 

 

unt'
oft
lilll
loc;

Th

Slitl

10

 

CL

 

 

 

in mediation of the renal proximal tubular BBM Na/Pi transport which helps prevent
hypophosphatemia or hyerphosphatemia by regulating renal and gastrointestinal phos-
phate reabsorption.44 Finding the mediating factors of the Na/Pi would allow a better
understanding of how to address the reabsorption of the renal phosphate. While the bulk
of this work is not pertinent to the discussion at hand, it was found that samples taken
from control rats indicated that the majority of the Na/Pi in the samples was found to be
located by highly enriched areas of cholesterol, sphingomyelin, and glycosphingolipids.45
This is indicative of the impact of these lipid rafts on the function of cellular systems and
shows that further work is still necessary for a further understanding to be had with this
multicomponent lipid rafis.

What these studies show is that lipid heterogeneity in plasma membranes is a
crucial factor that leads to bilayer stability as well as bilayer interactions with other spe-
cies. While the supported bilayers discussed in this dissertation focus on much simpler
systems, future work on supported bilayers will explore interactions of analogous com—
plexity in vesicle systems, and having an understanding of these systems is thus germane

to the design of future expemiments.

1-3: Supported Lipid Bilayers

With a working understanding of lipid vesicle systems it is possible to move on
to supported lipid bilayers. Vesicles are limited to some extent because of the variabil-
ity associated with the curvature intrinsic of vesicles, which can differ widely from the
curvature of a typical plasma membrane. Vesicles are also suspended in solution, and for
applications such as chemical or biological sensing, a planar, supported bilayer structure
is of more immediate interest.

Supported lipid bilayers (SLBs) are interesting not only because they provide a

10

 

structural motif that may be useful for chemical sensing, but also because the connection

of the bilayer to a planar substrate can pose a significant chemical and physical challenge,
and the properties of the resulting bilayer structure may be perturbed by the presence of
the substrate. Supported lipid bilayers are typically deposited on a planar substrate which
removes issues arising from the radius of curvature that lipid vesicles give rise to. Ad-
ditionally, these same characteristics have been thought to lend the SLBs to be a more
ideal model systems for plasma membranes opposed to vesicles.46'50 It should be noted
that real cells are curved so while vesicles may mimic this aspect of plasma membranes.
Since the lipids are supported in the SLBs, they are easier to work with and lead to
several possible materials applications, including the ability to be utilized as biochips or
biosensors utilizing naturally existing biomolecules as well as for multiple uses including
medical diagnosis and drug discovery.”54 For one possible example of a fiiture appli—
cation, it has been observed that many biomolecules lose their activity away from their
native environment, one means of getting around this is to fabricate an artificial environ-
ment that mimics the native environment such as a SLB.55'57 Therefore it is beneficial to
have a quick and efficient way to develop SLBs that can then be used for further analysis.

One down side to SLBs though is that they are harder to fabricate.

1-3.l: SLB Formation

There are two primary ways that SLBs are formed. The first is Langmuir—
Blodgett (LB) deposition followed by Langmuir-Schaefer (LS) deposition.57 LB-LS lipid
deposition (Fig l-4a) has been utilized to create SLBs.16’57 This approach has several
advantages, including the ability to create monolayers or bilayers depending on which
deposition method(s) are applied. As is shown in Fig 1-4a, withdrawal of the vertically

mounted substrate to effect LB deposition produces a monolayer on each side of the

11

 

substrate. With LS deposition, horizontal
contact with a surface adlayer occurs. The
substrate is suspended and contacts a sur-
face layer held at a specific surface pres-
sure. Performing this horizontal dip on a
substrate that already contains a lipid layer
deposited using LB methodology produces
a lipid bilayer structure.57 This approach
allows for the formation of comparatively

well organized bilayers.57 SLBs can

 

form on hydrophilic substrates of various
Fig. 1-4: A) Langmuir-Blodgett and Lang-

composition16 because lipid headgroups muir-Schaefer lipid deposition. B) Vesicle
fusion

will be attracted to such surfaces, facilitat-

ing lipid adlayer organization. Though LB/LS deposition methodology does have many

advantages, it is time consuming and labor intensive, and from a chemical standpoint it

is sensitive to chemical contaminants.

An alternative to the step-wise construction of SLBs is to deposit lipid bilayers
on hydrophilic surfaces using vesicle fusion. This technique is useful because it allows
for the solution phase creation of the lipid bilyers with wide control over vesicle
composition, followed by the single-step deposition of the bilayer onto the substrate of
interest. Vesicle fusion is a self assembly technique that starts with the lipids in vesicle
format and, upon contact with the appropriate hydrophilic surface, the vesicles open and
bind, either by physisorption or chemisorption, to the surface. Experimentally, vesicle

fusion is a relatively quick and straightforward means to deposit SLBs and has been

utilized in a variety of studies.58’59 The details of vesicle fusion are not understood fully

12

 

at this point. It is believed to proceed with an intact vesicle approaching and adsorbing

to a hydrophilic substrate surface, followed by vesicle deformation caused by increased
lipid interaction with the substrate, to the point of vesicle breakage onto the surface,
producing a region covered with a bilayer. The resulting bilayer can then diffuse laterally
and merge with other bilayer domains on the surface to form a large bilayer domain.
Schonherr and coworkers59 utilized atomic force microscopy (AF M) to follow this
process and develop this mechanism for vesicle fusion. The diffusion of the lipids across
a substrate surface has been found to be fast, with a rate of 1-8x10'8 cm2/s for DMPC
and DOPC bilayers above their phase transition temperatures. '6 This finding may imply
relatively fast lipid dynamics within a single lipid leaflet. Vesicle firsion has been shown
to be adsorption limited; higher concentrations of vesicles in solution do not produce

faster vesicle fusion.58

l-3.2: Inter-Layer Exchange (Translocation)

In addition to the substantial effort dedicated to the examination of dynamics
and composition within a single lipid leaflet, there has also been an effort to understand
inter-leaflet dynamics in bilayer structures.28’60’61 Inter-layer exchange, termed trans-
location, is the movement of a lipid from one layer or “leaflet” to the other layer in a
bilayer system. Translocation is important for several reasons, including lipid renewal,62
lipid organization and packing,63 protein sorting, and even rafi formation.64 The litera-
ture extant provides apparently contradictory information, with some reports suggesting

65-68

that a “flippase” protein is required to mediate the translocation process, and other

reports using probe molecules, indicating that spontaneous translocation can occur. It is

63,64,69

known that plasma membranes are asymmetric, a situation that could argue either

for or against lipid translocation. There are a variety of ways in which translocation can

13

proceed, with protein mediation being the most widely accepted means.70 There are two
types of mediating proteins: ATP-independent and ATP-dependent. ATP-independent
proteins are thought to move lipids to/from either bilayer leaflet, but not against a com-
positional gradient.67 ATP-dependent proteins can overcome compositional gradients,
allowing lipids to be moved across a bilayer regardless of compositional differences
between the layers.“69 There is also a body of work that indicates spontaneous translo-
cation, but in cases where “probe” lipids were used, the structural issues associated with
the probes called into question the time constants derived from such work.67’71 The most
elegant investigation of lipid translocation involved the use of nonlinear sum-frequency
vibrational spectroscopy.72 In that work, the “probes” were perdeuterated phosphocho-
lines, which could differ in mass from the proteolipids by as much as 10%,72 depending
on acyl chain length. From this there is still room for understanding in terms of SLB

lipid dynamics.

1-4: Controlling Vesicle Fusion

With the dynamics of supported bilayers in mind, in order to make usefirl struc-
tures it is necessary to have some from of control over the vesicle fusion process. Vesicle
fusion on unmodified polar supporting substrates does not yield significant control over
lipid organization or structure of the bilayer. To address this limitation, several tech-
niques have been developed that range from physical manipulation to chemical modifica-
tion of the substrate. Initial work in this area utilized a physical barrier that was grafted
onto the substrates to prevent lipid spreading and covering all of the substrate.48 That
study focused on creating barriers of aluminum oxide, ITO, chromium, and gold on a
silica substrate which would then be exposed to vesicles. Varying the identity of the bar-

riers produced a range of results upon lipid exposure. Aluminum oxide inhibits vesicle

14

 

fusion to areas with accessible silica and would act as a barrier where no lipid could

pass. The remaining barriers (ITO, chromium, and gold) can adsorb the lipids from the
vesicle however, the lipids become tacked to those barriers and lateral diffusion becomes
blocked.48 These barriers allow control over vesicle firsion to the substrate by control-
ling the locations where fusion can proceed, allowing in turn patterns of lipid bilayer to
be formed on the silica substrate. This work does suffer from the fact that while one can
achieve control the overall physical structure of the SLB, there remains a lack of control
over the type or quality of SLB formed, thus limiting their utility.

It is also of interest to gain control over the ability of the surface to support the
formation of a monolayer or bilayer, or to control the quality of the adlayer formed.
Chemical modification of the substrate is required to achieve this level of control. As the
work discussed above has shown, forming bilayers on hydrophilic substrates by vesicle
fusion is relatively straightforward. However, modifying the surface of a 8102 substrate
with either self-assembled monolayers (SAMs) or by controlling the oxidation level of
the 8102 surface, it is possible to control lipid adlayer deposition and improve the quality
of the adlayers formed-’3’74 Hydrophilic lipid headgroups interact with the SiO2 surface
causing the lipids to adsorb onto the surface via the lipid head group, causing vesicle
rupture and the subsequent deposition of a bilayer. This approch is usefiil to form bilayer
structures however as mentioned modification of the surface can allow for control of the
structure formed as well which could lead to further possibilities that must be explored.
The resulting hydrophobic SAM modified surface does not interact strongly with lipid
headgroups from other vesicles but can interact with lipid acyl tail regions yielding sig-
nificantly different results form the bare SiO2 surface discussed preveously.73’74 If the
hydrophobic region of the vesicle is exposed to the SAM modified surface, through either

the fusion of the vesicle on to a clear SiO2 area on the surface and lateral diffusion over

15

the SAM region or free lipid in solution absorbing to the SAM, the resulting Van Der
Waal’s interactions of the SAMs and lipid acyl tails causes the formation of a lipid mono-
layer over top the SAM.73 This approach now allows for structural control over lipid
adlayer formation.

To facilitate better bilayer formation it is necessary to make the surface as hy-
drophilic as possible. Thorough oxidation of the surface and minimal opportunity for
contamination are important factors in producing a hydrophilic surface.72 In the study
conducted by Isono and coworkers, it was found that the extent of SiO2 surface oxida-
tion can influence the quality of the bilayers formed upon vesicle exposure.74 Bilayer
formation in this study was followed by fluorescence imaging with NBD-PC and, upon
chemical oxidation, showed a significant increase in bilayer quality compared to that of
a hydrogen terminated silica surface which served to demonstrate the existence of unrup-
tured vesicles. Thermal oxidation over a long period of time fiirther improved upon that
bilayer quality significantly.74 What this indicates is the use of SAMs and other forms
of chemical surface modification allows for some level of control over the lipid adlayers

formed as well as the quality of those layers.

1-5: Introduction of Work Performed

The ability to form SLBs has allowed for opportunities to investigate some of the
details of lipid mono- and bilayer organization. SLBs can be made to be stable in air and
can be formed in multiple ways and the supported substrate can be modified to allow for
more selective control over the lipid layers formed. These studies have provided some
level of insight regarding the dynamics of membrane adlayer constituents. A limitation
of this area of work is that many of the known supported bilayer systems are fragile and

cannot pass through the water/air interface without degradation. This degradation in-

16

cl

01:

U.
iii
in
lo
LL

I“

 

 

 

 

cludes the formation of defects in the bilayer to complete collapse of the the bilayer struc-
ture. A lack of stability in air makes further examination of supported limited bilayers
significantly more difficult. All analysis of the bilayers formed has to be in situ, limiting
several spectroscopic techniques that could be useful in probing the details of lipid/lipid
interactions within the bilayer structure.

The work reported here is intended to achieve a better understanding of the for—
mation of selected lipid adlayers. Specifically, the formation of air stable bilayers was
a primary focus of this work based on the hypothesis that the strength of lipid-substrate
interactions was the primary factor that limited the formation of stable lipid adlayers. To
create robust interactions between the lipid adlayer and the substrate, a well organized,
hydrophilic substrate is required, and hydroxythiol SAMs on Au are one such surface.
It is also a comparatively simple matter to react the terminal SAM hydroxyl groups to
form a phosphate group, followed by zirconation of the phosphate group. The strategy
to creating strong lipid-substrate interactions is to utilize robust interaction chemistry. A
goal of this dissertation is to establish how well zirconium phosphate/bisphosphate (ZP)
chemistry can be used to bind phospholipid head groups. The acyl tails of the lipids are
composed of carbon-carbon and carbon-hydrogen bonds, leaving little interaction other
that Van der Waals forces to mediate organization and stability. As previous studies have
shown, these forces are central to the phase transitions in lipid bilayers and are important
in the development of self-assembling monolayer systems. The lipid headgroups allow
for the use of polar and/or ionic interactions to develop energetically favorable structures
that are capable of maintaining their structural integrity under a wider range of environ-
mental conditions that are hostile to physisorbed lipid bilayer assemblies.

Results from this work could point the way toward several possible opportunities

in interface-mediated science, such as chemical/biological sensing and chemical sepa-

l7

 

rations. It is also necessary to develop an understanding of the interactions that could

give rise to lipid-specific surfaces, which could be utilized in lipid separations. Utilizing
hydroxythiol SAMs and ZP chemistry to modify Au substrates allowed probing of inter-
actions that lead to stable lipid adlayer formation. The initial aspect of this work were to

investigate these surface/lipid interactions and move towards development of reproduc-

ible, air stable lipid adlayers.

 

Chapter 2

Formation of Air—Stable Supported Lipid Monolayers and Bilayers

2-1: Introduction to Air-Stable SLBs

The creation of planar supported lipid bilayers (SLBs) has been pursued because
of their potential to function as biomimetic structures to facilitate studies aimed at under-
standing the function of transmembrane proteins. Such supported bilayers can be used to
help understand the structural complexity and compositional heterogeneity of multicom-
ponent lipid bilayer systems. Lipid bilayers, as discused earlier, are quasi-two dimen-
sional structures characterized by both structural heterogeneity and fluidity, and it is this
latter property that adds significant complexity to the study of these systems.4'l6

While a great deal of work has been done in creating biomimetic bilayer films,
the current understanding in the field is that such films cannot maintain their structural
integrity upon exposure to air.75 The instability of the bilayers in air is thought to origi-
nate at the point when the bilayer crosses the water/air interface following deposition,
giving rise to bilayer disorganization and/or decomposition that is ultimately determined
by the limited strength of interaction(s) between the bilayer and the supporting substrate.
Several techniques have been explored to synthesize air-stable SLBs, including polym-
erization of the lipids that comprise the bilayers, the addition of stabilizing components
such as cholesterol to the lipid mixture, or even the creation of physical boundaries on the
solid support to prevent uncontrolled lipid spreading on the support surface.75'77
Substantial effort has been dedicated to examining the dynamics and composition of
single lipid leaflets in-situ, and there have also been efforts to understand inter-leaflet

dynamics in bilayer structures.28’60"51 However, the intrinsic structural and stability

limitations of SLBs have prevented analogous experiments from being performed on that

19

 

 

 

fami

subs
imm
plies
Show
be po

sition

lopmr
groups
strong].
sunhcc
exposu}
action c
Sll'0ng l‘
letfagc‘
36a. 30
USlng 0p

1
“mmcni
character

lace; Th

 

family of structures.

Among the issues of concern is the extent to which the bilayer leaflets act coop-
eratively to stabilize the overall structure. It would be useful to have a facile means of
creating either a monolayer or bilayer structure by simply controlling the identity of the
substrate on which the layer is deposited. The use of Zr-bisphosphate (ZP) chemistry to
immobilize lipid headgroups is an inviting possibility. This chemistry is known to bind

phosphates and phosphonates strongly,78'86

and if phosphocholine headgroups can be
shown to bind to a zirconated surface, the creation of supported lipid monolayers would
be possible without resorting to Langmuir-Blodgett (LB) or Langmuir-Shafer (LS) depo-
sition methods.

In the work presented in this chapter, monolayer interfacial chemistry is used
to prepare two surfaces; a polar interface capable of interacting with lipid bilayer head-
groups noncovalently, and a second interface capable of binding the lipid headgroups
strongly. Bilayers form readily and are stable for extended periods of time on an Au
surface modified with 6-mercapto-l-hexanol, even following removal from solution and
exposure to air. For Au substrates modified with 6-mercapto-1-hexanol, followed by re-
action of the terminal —OH groups with POCI3, H20 and Zr”, it is observed that there are
strong interactions between Zr+4 and the phosphocholine lipid headgroup.87 For this in-
terface, lipids originally present as a partial bilayer evolve to a monolayer structure over
a ca. 20 minute time period. This partial bilayer-to-monolayer conversion was monitered
using optical ellipsometry and contact angle measurements.

To develop a full understanding of the interactions taking place FTIR, cyclic vol-
tammetry, time-resolved ellipsometry and contact angle data are utilized, allowing for the
characterization of the formation of lipid adlayers on the two chemically modified inter-

faces. The coupling of phospholipids to the Zr-terminated surface was demonstrate using

20

 

3‘P NMR measurements.

2-2: Materials and Instrumentation Utilized

The lipid used for these experiments was 1,2-dimyristoyl-sn-glycero-3-phos-
phocholine (DMPC) dissolved in chloroform obtained from Avanti Polar Lipids, Inc.
The solvents acetonitrile, ethanol (100%) and ethyl acetate, as well as the reagents
6-mercapto-1-hexanol, phosphorus oxychloride (POCl3), zirconyl chloride octahydrate
(ZrOC12-8 H20), 2,4,6-collidine, potassium ferrocyanide trihydrate, lithium perchlorate,
hexaminerutheniumflll) chloride, and potassium chloride were all obtained from Sigma—
Aldrich in the highest purity grade available. 18 M9 Water was used for all experiments
and all reagents were used as received, without further purification.

The specific instrumentation used is as follows: Reflectance F TIR spectra of
modified gold substrates were collected with a Nicolet Magna-IR 560 spectrometer in
reflection mode with 2 cm'1 spectral resolution. The data are presented as normalized
absorbance spectra. Electrochemical data were acquired using a CH Instruments 650
Electrochemical Analyzer. All optical ellipsometry measurements were recorded with a
J. A. Woollam Co., Inc. spectroscopic ellipsometer with a wavelength range of 185-1100
nm utilizing 44 wavelengths at a time. Water contact angle measurements were recorded
on an ACT Products Inc. VCA 200 Video Contact Angle System. Except where noted

experiments were performed at 20°C.

2-3: Experimental Setup
For the electrochemical measurements two electrochemically active probes were
used to characterize the interfaces. These were K3Fe(CN)6-3 H20 (1.32 mM) in 0.1

M LiClO4 and Ru(NH3)6Cl3 (1.00 mM) in 0.1 M KCl. These two probes were chosen

21

 

 

because of their different electron transfer kinetics.88 Cyclic voltammetry was performed
with each probe being cycled three times at a scan rate of 0.1 V/s. The F e(CN) 63 '/ 4'
probe was scanned from -0.1 to 0.5 V and the Ru(NH3)63+/2+ probe was scanned from
-0.4 to 0.1 V vs. Ag/AgCl using a Pt counter electrode.

Each substrate was prepared in a similar fashion. Gold substrates were prepared
using a procedure described previously.89 Briefly, the substrates were rinsed with water
and ethanol, cleaned in a UV—cleaner for 15 min., then exposed to 10 mM 6-mercapto-
l-hexanol in ethanol for 6 hrs. The resulting interface was rinsed with ethyl acetate and
ethanol, then dried under a stream of N2(g). For Zr-modified interfaces, the mercapto-
hexanol monolayer was reacted with POCl3 (0.4 mL) in dry acetonitrile (10 mL), and
catalyzed with 2,4,6-collidine (0.4 mL) for 3 hrs. The phosphate-modified monolayer
was rinsed with ethanol and water, dried with N2(g), and exposed to 5 mM ZrOCl2 in
a 60:40 ethanol/water solution for 12 hrs. The zirconated monolayer was dried under
N2(g), then exposed to the solution containing DMPC unilamellar vesicles.

To prep the DMPC for adlayer formation, unilamellar vesicles of DMPC were
prepared first as described previously.60 The vesicles used are comprised solely of
DMPC. Chloroform was first evaporated from the DMPC solution using a N2 stream.
The lipids were then exposed to vacuum to remove any remaining chloroform. The solid
DMPC was next dissolved in a 10 mM Tris(hydroxymethyl)-aminomethane hydrochlo-
ride (Tris®, Aldrich) pH 7.5 buffer solution to a final concentration of 1 mg/mL. The
solution was treated with five freeze-thaw-vortex cycles to ensure thorough suspension of
the lipids prior to extrusion.90 A syringe-based mini-extruder was used to form uniform
unilamellar vesicles (Avanti Polar Lipids, Inc.).9l'93 The lipid suspension was passed
eleven times through a polycarbonate filter with an average pore diameter of 400 nm to

produce unilamellar vesicles of that diameter.

22

, d-‘r'w m“-

~-:,.‘ . m
. ““3: a.

 

 

 

With the unilamellar vesicles of DMPC formed, mono- and bilayer formation can
then proceed. Planar DMPC adlayers were formed by spontaneous fusion of the unila-
mellar vesicles formed.75 The clean gold substrates were placed in a custom-made Tef-
lon® flow cell (Fig. 2-1) with
an approximate volume of 1
mL. A flow cell was employed

to ensure that the lipid suspen-

 

sion was in full contact with the

substrate during bilayer forma-

 

 

 

 

tion. Tris® buffer was flowed

 

 

 

over the substrate at approxi-

 

 

 

 

 

mately 5 mL/min prior to lipid

Fig. 2-1. Schematic of the Teflon® flow-cell used for
the deposition of DMPC adlayers by vesicle fusion. The

lipid suspension, flowed at the depth 0f the reservoir iS ca. 2 mm.

deposition, followed by the

same rate until the buffer solution is displaced. The lipid-containing solution is then held
in place to allow exposure to the substrate for a given time. Following exposure, the sub-
strate was washed with water at the same flow rate, the water is then removed fiom the
cell at a rate of 1 mL/min, and the substrate is removed from the flow cell and allowed to
air dry by hanging the substrate by one end.

In order to confirm the coordination of the lipid headgroup with the Zr substrate
3 lP-NMR was performed. Solid-state NMR samples were prepared as described previ-
ously.86 One gram of silica gel 60 (40-60 pm, 470-530 m2/ g surface area, EMD) was
dried and placed under an Ar atmosphere. Silica gel was chosen as a substrate as it has
the necessary —OH groups to perform the phosphonation and zirconation reaction; POCl3

(2 mL) and dry acetonitrile (10 mL) were added to the silica gel and mixed. 2,4,6-Colli-

23

dine (2 mL) was added to this silica gel suspension and the solution was reacted for 3 hrs.
The resulting suspension was dried and washed with acetonitrile to rinse away excess
POCl3 or other phosphorus oxides. A 5 mM ZrOCl2 solution in 60:40 ethanol/water
solvent was added to the silica gel, and the suspension was mixed for 12 hrs. The result-
ing suspension was dried and washed with ethanol and water. A 1 M DMPC solution in
Tris® buffer was prepared and exposed to the silica gel for several hours. The resulting
suspension was dried by filtration and washed with water. A sample of the functional-
ized silica gel was then taken at each of the drying steps in this process. The solution
phase NMR sample was prepared using a l M DMPC solution. The chloroform was first
evaporated from an aliquot of the stock DMPC solution and the resulting solid phase
lipids were placed under vacuum to remove any remaining chloroform. The lipids were
then dissolved in chloroform-d for NMR analysis.

Two different NMR instruments were utilized to collect the NMR measurements.
31P MAS NMR measurements were made using a 400 MHz Varian Infinity Plus NMR
spectrometer. The 31P nuclei resonance frequency was 161.82 MHz for all spectra ac-
quired with this instrument. Each sample was packed in a 6 mm diameter zirconium ro-
tor and spinning speed was set to 4 kHz. MAS-NMR spectra were not proton-decoupled,
and the nuclei were excited using a 90° pulse for a duration of 15 us and an internal pulse
delay of 1 s. 100 Hz of line-broadening was applied prior to processing the spectra to
increase the S/N of the spectra.

Solution phase 3 lP-NMR measurements were performed using a Varian UNITY-
plus 500 MHz NMR spectrometer with a 31F resonance frequency of 202.38 MHz. The
spectra were proton-decoupled and the nuclei were excited with a 60° pulse for 7 11s and
a recycle delay of 2 s. 1 Hz of line-broadening was applied prior to Fourier transform to

increase the S/N of the spectra.

24

 

2-4: Results

The primary purposes of this work was to demonstrate a supported lipid bilayer
that is stable in air, and to explore the consequences of a phosphocholine binding to a zir-
conated surface. Discussed first is the formation of an air-stable supported lipid bilayer,
and then the formation of a supported lipid monolayer. With these interfacial structures
characterized, the energetics of transition of an initially-deposited DMPC bilayer to

monolayer on the zirconated surface is considered.

2-4. 1: Air Stable Supported Bilayer Formation.

The creation of a supported lipid bilayer structure that is stable in air has been
sought widely, with many attempts yielding interfacial lipid organization that is charac-
terized by defects and possibly void regions.75 There have been several methods devised
for creating bilayers capable of surviving the passage through the air-water interface

following their forrnation,75'77’94

with methods for their formation ranging from the use
of complex, multi-component systems to the growth of the bilayers on substrates such as
PDMS.94 In the work presented here, lipid bilayers comprised of DMPC are used, and
these bilayers have been deposited via vesicle fusion onto substrates that are character-
ized by comparatively high densities of polar terminal functionalities. It is believed that
the ability to form these air-stable supported lipid layer structures is dependent on the
interfaces use.

DMPC has been deposited on a gold surface that has been modified with a self-
assembled monolayer (SAM) of 6-mercapto-l-hexanol to produce a hydroxyl-terminated
interface yielding a lipid bilayer. Present in Fig. 2-2a the F TIR spectrum of the interface

prior to deposition of the bilayer. It should be noted that the background correction ap-

plied to these data has produced an artefactual sloping background that has no physical

25

or chemical significance.

It is the band positions

he
I

Oil-i
b
m

that provide the relevant

r chemical informa-

9
4;
I

tion. These data reveal

absorption (arb. units)
O
O‘
l

.O
N
l i

aliphatic resonances

 

 

 

0-0 — r L r r 1 r
l l l l LLI 1 l l 1 l I l l l l Ll_l L l 1 L l l -l
4000 3500 3000 2500 2000 1500 1000 “2854 cm and 2928
-1
frequency (cm ) cm“, characteristic of a
”g 1’0 h b comparatively disordered
E 0 8
g ‘ SAM aliphatic region.
2” 0-6 This is not a surpris-
o
E. 0-4 n ing result, based on the
m
‘3 0-2 knowledge that aliphatic
0.0~ ,,,H,,,,+,U,,.,H,, ,Lg chainsshorterthan

 

 

4000 3500 3000 2500 2000 1500 1000

frequency (cm'1) C9 produce analogous

Fig. 2-2: F TIR spectra of (a) a 6-mercapto-1-hexanol self- TCSUItS in 2111010611110l
assembled monolayer deposited on a gold surface. (b) a DMPC
adlayer deposited on the 6-mercapto-l-hexanol-modified gold self-assembled monolay-

inter-face.
ers.95 Also noted are the

presences of a broad band centered at ca. 3300 cm], indicative of hydrogen-bonded —OH
functionality.

To gauge the quality of this film, cyclic voltammograms of these interfaces were
measured in the presence of the electrochemical probes Fe(CN)63'/4' (Fig. 2-3a, solid line)
and Ru(NH3)63+/2+ (Fig. 2-3b, solid line). Krysinski and Brzostowska-Smolska88 have
demonstrated these two probes produce different results at monolayer interfaces, owing

to the modest difference in electron transfer kinetics that characterize the two probes, as

26

- - “"0!vvxvuzsvweeruzawr 9:!!‘!tfl“t*!‘~lh

 

well as their formal charges.88

 

 

 

 

 

 

500 '—
Using both probes thus provides 400 L
2 300 r
a gauge on the quality of the 3 200 f
‘E 100 _—
interfaces. The data report here g 0 _-
-100 _-
are typical for thiol-modified -200 :
-300 _-
interfaceswheresom m 400 ' ‘ ‘4 ‘ ' ‘ ‘ ‘ ‘ ' ‘ 1 ‘
e a ount -0.1 0.0 0.1 0.2 0.3 0.4 0.5
of disorder is present. Specifi- po‘em‘al vs“ Ag/AgCl (V)
60 -
cally, for the F e(CN) 63 '/ 4' probe 40 L
and the 6-mercapto-l-hexanol i 20 T
2' 0 ~
monolayer, it is observed that 3'20 '_
”—40 b
no discernible redox waves ,—
-60 —
(Fig. 2-3a, solid line), but for _80 i r m r L r . r z r . r . r
-0.4 -0.3 —0.2 -0.1 0.0 0.l 0.2
the Ru(NH3)63+/2+ redox couple potential vs. Ag/AgCl (V)

(Fig. 2'3b’ solid line), there is a Fig. 2-3: Cyclic voltammograms of a 6-mercapto-1-

hexanol self-assembled monolayer deposited on a gold

surface (solid line) and of the same interface with an

_154 mV, with a peak separation adlayer of DMPC deposited (dashed line), recorded
using (a) 1 mM K 4[F e(CN)6] in 0.1 M LiClO4 and (b)

of 120 mV. These findings un- 1 mM [Ru(NH3)6]Cl3 in 0.1 M KCl as electrochemical
probes.

derscore the higher sensitivity of

measured midpoint potential of

3+/2+

the Ru(N H redox couple to SAM quality and indicate that the adlayer is mediating

3)6
electron transfer kinetics.

Using substrates prepared in this manner, bilayers have been formed on them
by means of vesicle fusion (vide infra). This procedure is well established, but what is
unique about these results is that once the bilayers are formed in an aqueous environ-

ment, they have been successfully been removed from that environment and the bilayer is

exposed to air. The resulting bilayer has been characterized using F TIR, cyclic voltam-

27

metry, optical ellipsometry and water contact angle measurements, all ex-situ from the
vessel in which they were formed. The FTIR spectrum of the resulting bilayer is present-
ed in Fig. 2-2b. These data show C-H stretching resonances at 2854 cm'1 and 2925 cm'l,
suggesting the bilayer acyl chains are somewhat disordered. This is not surprising given
the gel-to-fluid transition temperature for DMPC is 24°C, and the internal temperature

of the FTIR spectrometer is at or above 24°C. It is also important to note the presence

of phosphate peaks at 1265, ~1100, and 820 cm], further supporting the presence of the
lipid bilayer.

The cyclic voltammetry data for the bilayer, using both the F e(CN)63 44‘ and
Ru(NH3)62+/3+ probes are presented as the dashed lines in Fig. 2-3. These data indicate
that the presence of the bilayer provides some additional screening for the Fe(CN)63'/4’
probe, but without the presence of redox peaks, it is not possible to quantitate these
differences. For the Ru(NH3)62+/3+ probe, there is both a diminution in current and a
substantial increase in peak separation, with a midpoint potential of -200 mV and a peak
separation of 255 mV. This is not surprising and indicates that the resulting bilayer con-
tains some imperfections and further impedes the electron transfer kinetics. Recognizing
that it would be preferable to acquire CV data for adlayers both before and after removal
from the flow cell in which they were deposited, due to design restrictions of the flow cell
used, it is not possible to acquire CV data in situ at the present time.

Optical ellipsometry is used to measure the thickness of the DMPC bilayer as a
function of exposure time to a solution of vesicles. This approach has been used before
to measure bilayer formation kinetics.58 Each surface was exposed to DMPC for a spe-
cific period of time before being removed from the solution, then washed, dried and mea-
sured. Initial measurements of the substrates after reaction with 6-mercapto-l-hexanol

showed thicknesses of ca. 12 A, and thicknesses of ca. 76 A were observed following de-

28

 

 

position of DMPC. Fig. 2-4a shows

00
O

the change in ellipsometric thickness

\l ‘1
CU.
rTFT1
-i

t—O—l

 

 

 

 

 

<
:3 65 _- of the interface as a function of time
:33 6°: '
t; 55 p i of exposure to a vesicle-containing
g 50 -
a 45 f . solution. The data are characterized
§_ 40 - i
:6 35' r‘ . r . . . r .4 . r . . . r . r byarapidbuild-up,withatimecon-
-50 0 50 mo ISO 200 250 300 350 400
: (1)3 _- time (min) stant of 67 i 5 sec to a thickness that
3 :3 : ‘r is consistent with that of a DMPC bi-
s 70 :-
2 60 f i layer (64 :1: 2 A) based on molecular
E0 50 :- 1
g :3 T iii ii I mechanics calculations and experi-
g 20 i I I I 96 - ~
0 10 L mental data. While there rs uncer-
0 b . . r . r r r . A r r . r r 1 r r _
~50 0 50 100 150 200 250 300 350 400 tainty in each data point, resulting

time (min)
fiom multiple measurements on a
Fig. 2-4: (a) Ellipsometric thickness measurements

for a 6-mercapto-l-hexanol-modified Au surface, given sample and the use of multiple
acquired at a series of reaction times to follow

DMPC deposition. (b) Water contact angle mea- samples, the plateau value is CODSiS'
surements acquired at a series of reaction times to _ _ . _

follow DMPC deposition the same inter-face. For tent wrth a DMPC lipid bilayer.

both panels, initial data points correspond to t=0.
Contact angle measurements

have been used extensively for both the study of lipid interactions with surfaces as well
as for the determination of the degree of surface coverage.72’97'99 Water contact angle
measurements are performed to further evaluate the polarity and uniformity of the bilay-
ers. The measured water contact angle of the bilayer-containing interface was found to
have a value of ca. 30° (Fig. 2-4b). The contact angle of an Au surface modified with
6-mercapto-l-hexanol was measured to be ca. 54°. These data are thus consistent with
a polar interface being formed by the phosphocholine head groups.77 The contact angle

variation across the substrate is ca. 2° for all measurements, and the contact angle hys-

29

 

 

 

 

 

 

mien
clccu
Ol‘lht
chart:
the hi

to air.

24.2:

layer P

Zircon;

fundam-

ness of
angle 0
DMPC
Were WI
13391 lh;
from lhr
CVaIUaIc

Fig 2~

(J:

t

thicknes‘

 

teresis (advancing — receding) is ca. 5°, indicating a comparatively uniform bilayer, both

microscopically and macroscopically. These findings are consistent with the F TIR and

electrochemical data, and indicate that a DMPC lipid bilayer forms rapidly upon exposure

of the hydroxythiol-terminated surface to a vesicle containing solution. The fact that the

characterization steps were performed following removal from the initial vessel in which

the bilayer was deposited demonstrates that the resulting bilayer is stable upon exposure

to air.

2-4.2: Air-Stable Lipid Mono-
layer Formation

Vesicle exposure to a
zirconated surface results in a
fundamentally different inter-
face. Initial values for the ZP
treated surface showed a thick-
ness of ca. 20 A with a contact
angle of ca. 72°. Following
DMPC exposure, measurements
were repeated and the ZP-ad-
layer thickness was subtracted
from the total thickness to
evaluate the adlayer thickness.
Fig. 2-5a shows the temporal
evolution of the ellipsometric

thickness of the interface fol-

ellrpsometnc thickness (A)
N w DJ «5
U! C U. G
I—H /

I
I
I
I
I
I
:
I
l
I
I
I
I
l—LH
r—o—r

A A A l 1 I A A n l A l A l A l A I

-50 0 50 100 150 200 250 300 350 400

time (min)

1 r
90 iii: i i

w

l
T

contact angle (deg.)
A»
c

 

0 . AllllllllnAlLIA'

-50 0 50 100 150 200 250 300 350 400

time (min)

Fig. 2-5: (a) Ellipsometric thickness measurements for
a zirconated 6-mercapto-l-hexanol/Au self-assembled
monolayer, acquired at a series of reaction times to
follow DMPC deposition. (b) Water contact angle
measurements acquired at a series of reaction times for
the same interface. For both panels, initial data points
correspond to t=0.

30

 

 

 

 

lowing exposure to DMPC vesicles. By fitting these data, an initial growth to 53 i 8 A

was obtained, followed by a decay in thickness to a value of 28 i 3 A. The time constant
of the decay is 1200 :1: 720 s. This uncertainty may seem large but, due to the nature of
the measurements, such uncertainty is unavoidable. Water contact angle data reveal a

nonpolar, homogeneous in-

 

 

A 1.0 a
If) .
g 0 8 ' terface, With a contact angle
3 0 6 of ca. 90° (Fig. 2-51)), with a
g 0 4 ’ variation across the substrate
g 0.2 _ of :t 2° and a hysteresis of
h A ca. 5°. These data, taken
O-Oh. r.1414.H.r...IrI...r....r

 

4000 3500 3000 2500 2000 1500 1000 collectively, are consistent

frequency (cm")
with the resulting interface

 

 

1;: 1.0 — b
g 0 8 ' being comprised of a mono-
? 0 6 ' layer of DMPC, where the
,§ 0 4 polar headgroup of the lipid
e . o n
g 0 2 complexes essentially 1r-

0 0 _ reversibly with the Zr-phos-

' n l n n 1 r l n L n n I n n l 1 I l 1 14 I n :

 

4000 3500 3000 2500 2000 1500 1000 phate surface and the nonpg-

frequency (ch)
lar acyl tails are exposed.

Fig. 2-6. F TIR spectra of (a) a zirconated 6-mercapto-1-
hexylphosphate self-assembled mono-layer deposited on a The FTIR spectra
gold surface. (b) a DMPC adlayer deposited on the zircon-

ated 6-mercapto-l-hexylphosphate-modified gold interface. 0f the zirconated interfaces

are shown in Fig. 2-6a. The
POCl3 reacts essentially completely with the 6-mercapto-l-hexanol terminal -OH groups,
and after zirconation only the phosphate peaks at 1265 cm", ~1100 cm"1 and 820 cm'1

are observed. Upon exposure to DMPC, the monolayer that is formed shows the same

31

 

 

 

phosphate peaks with the addition of acyl chain peaks (Fig. 6b). These data also show

C-H stretching resonances at 2854

 

 

 

 

 

 

cm"1 and 2926 cm'l, suggesting the 400 F
300 -
monolayer acyl chains are some- A -
3200-
what disordered. Interestingly how- E '00 __
5 .
ever, these peaks, when compared 0 0 -

to the bilayer acyl chain peaks, are "00 ' , . 1 . J . J g, _ , . , ,

_ -0.I 0.0 DJ 0.2 0.3 0.4 0.5
slightly different and show more potential vs. Ag/AgCl (V)

60 -

disorder in the monolayer than in 40 ’_ b
the bilayer. This expected result i 20 f
f 0 -
is likely due to the monolayer acyl 2-20 I.
. . 0.40 t.
chams berng exposed to the hydro- .
.60 ..

philic environment were the bilayer -80 1 . 4 J L L k . ' . ‘ . '

-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2
has the hydrophilic head groups 90mm vs' Ag/Agc‘ (v)

Fig. 2-7: Cyclic voltammograms of a 6-mercapto-
l-hexanol/Au self-assembled monolayer reacted
with POCl and ZrOCl2 to zirconate the interface
(solid line) and of the same interface with an ad-
making the bilayer slightly more layer of DMPC deposited (dashed line), recorded

using (a) 1 mM K 4[Fe(CN)6] in 0.1 M LiClO4 and
crystalline. (b) 1 mM [Ru(NH3)6]Cl3 in 0.1 M KCl as electro-
chemical probes.

exposed protecting the acyl chains

from the hydrophilic environment

While contact angle and
ellipsometry measurements are useful in providing a mesoscopic gauge on the properties
of the interfacial layer(s), electrochemical measurements are critical to evaluating their
molecular-scale organization. Cyclic voltammetric measurements were performed on
each lipid layer to probe the mono- or bilayer structures for defects. Figs. 2-3 show the
cyclic voltammograms of the hydroxythiol-terminated interfaces with a DMPC bilayer

using F e(CN)63’/ 4' (Fig. 2-3a, dashed line) and Ru(NH3)63+/2+ (Fig. 2-3b, dashed line)

32

as probes. Figs. 2-7 shows the cyclic voltammograms of the zirconated interfaces with

DMPC using F e(CN) 63 '/ 4' (Fig. 2-7a) and Ru(NH 3+/2+ (Fig. 2-7b) as the probes. The

3%
peak splittings and positions for both electrochemical probes are consonant with literature
data,57 and it is possible to gain some insight into the interface through a detailed exami-
nation of the CV data.

First, it should be noted that the capacitance for these interfaces was found to be
0.0074 pF/sz for the bilayer on the 6-mercapto-l-hexanol SAM interface and 0.0087
pF/sz for the monolayer on the zirconated SAM interface. This finding would sug-
gest that the interfaces are close to the same thickness, provided their defect densities are
similar. The ellipsometric data (vide infra) indicate that there is a factor of two difference
in the interface thickness, implying that the adlayer capacitance we measure is defect-
mediated for both systems. The F e(CN) 63 '/ 4‘ CV data for the 6-mercapto-l-hexanol SAM
indicate better blocking of the probe than the same interface with the lipid adlayer depos-
ited (Fig. 2-3a). There is a large splitting (> 400 mV) of a weak wave seen for the lipid
adlayer, suggesting very slow electron transfer kinetics at this interface. The presence of
the wave following the addition of the lipid adlayer, which is not present with the SAM
alone, implies that the adlayer is perturbing the organization of the SAM to some extent
upon deposition.

The corresponding data for the Ru(NH3)63+/2+ probe reveal the presence of a
redox wave for both adlayers as well as for the SAMs on which the adlayers are de-
posited (Figs. 2-3b, 2-7b). It is not surprising to observe this electrochemical behav-
ior because lipid adlayers are likely to be characterized by some defect sites, and the
Ru(NH3)63+/4+ probe is more sensitive to interface defect sites. It is clear for both in-
terfaces that the electron transfer kinetics are slowed significantly by the presence of the

lipid adlayer. As noted above, for the 6-mercapto-l-hexanol SAM, the peak separation

33

 

 

inter

same

ltise

With 1
lipid

zircoi

acid
lion 2

gm ‘tt

for the Ru(NH3)63+/2+ waves is ca. 120 mV with a midpoint potential of -154 mV, and
this separation increases to ca. 255 mV (and the midpoint potential shifts to -200 mV)
with the addition of the lipid bilayer (Fig. 2-3b). For the zirconated interface (Fig. 2-7),
the redox peak splitting is ca. 117 mV (midpoint potential -153 mV) and the addition of
the lipid adlayer increases the splitting to ca. 143 mV (midpoint potential -l79 mV). For
a reaction that is reversible on the time scale of the potential scan, the expected (Epa-Epc)
splitting should be 59/n mV at 25° C. The values obtained for these interfaces are all in
excess of 59 mV, indicating that the kinetics of the electron transfer are mediated by the
interfacial adlayers in all cases. The splitting and midpoint potentials are essentially the
same for both the 6-mercapto-l-hexanol and zirconated SAMs, which is not surprising.
It is also not surprising that the lipid bilayer impedes the electron transfer kinetics to a
greater extent than the lipid monolayer, as manifested by a comparison of the peak split-
ting and midpoint potential data. The electrochemical data are in qualitative agreement
with the ellipsometric and contact angle data, and are consistent with the formation of a
lipid bilayer structure on the 6-mercapto-1-hexanol SAM and a lipid monolayer on the

zirconated SAM.

2—4.3: NMR Analysis

The ZP-lipid interactions that are central to understanding the formation of the
monolayer data can be characterized using 31P NMR measurements. 31P NMR measure-
ments have been used in the past to characterize ZP multilayer structures, and this work
follows those same procedures.87 All NMR spectra were referenced to 85% phosphoric
acid (5=0ppm). Fig. 2-8 shows the 31P NMR spectrum for each step of interface prepara-
tion and DMPC exposure. For this work silica gel was used as the substrate for interface

growth because of the surface area advantage. As seen in Fig. 2-8a, after exposure of

34

 

I
~—-‘\ v ~ --~-_——

 

"r- alv'u'm‘ mwmarv‘ - 3".

’
\—,I 's. - - “
--fi ,1" v “4",

 

10 0 O - 80 PP“
140 100 60
140 100 '60
.r—_--.__‘_r.- affirm-Jr. ’1‘. “,‘I r.“ 1. I
140 100 so ' 20

20'1'

.‘ x
, ,t,‘

I
o”350"
!
t
i
0'lid"w'

pp-

| .
‘L”‘1}VI".""

r . r
~-- ‘- ‘.-'.- k—r
, Lu ‘—'.,_.-.4‘

-60 -100' bps

Fig. 2-8: 3 IP MAS NMR spectra. (a) Silica gel coated

with POCl ,
ZrOCl (aqi,
DMP .

(b) surface shown in panel a exposed to
(c) surface shown in panel b exposed to

the silica gel to POCI3 a strong
resonance is observed at 8=-l .1
ppm, and a smaller resonance
is seen at 6=-14.9 ppm, cor-
responding to physisorbed and

chemically bound phosphate at

' .-60.....-._ 2136“”pr the silica gel surface, respective-

ly. The resonance at -1.1 ppm

is extremely strong because the
physisorbed phosphate cannot
be removed easily. The remain-
ing bands are due to the spectra
not being proton decoupled,

and the presence of spinning
side bands.87 On exposure to
ZrOClz, the spectra exhibit a
characteristic shift (Fig. 2-8b).87
Poorly resolved peaks at 8=-4.7,
-12.0, and -l9.7 ppm result from
zirconium coordination with a
phosphate. The prominent peak
at -l .1 ppm is eliminated because

the Zr+4 in solution complexes

the physisorbed phosphate, removing it fi'om the silica gel surface. The spectrum shown

in Fig. 2-8c was recorded following exposure of the zirconated surface to DMPC. The

35

v --u- p3..qetn{sv_1(“.1".MAI'lifleig-ii'fl‘“ ’ " “ ‘ ‘

 

broad peaks at 5=-12.1 and -l9.4 ppm correspond to the zirconium-phosphate complex,
and the sharp resonance at -O.5 ppm is due to zirconium complexation by the phospho-
choline headgroup of DMPC. This assignment is made by comparing these data to a
solution phase DMPC spectrum, which has a single peak at 6=-0.32 ppm (not shown).
Washing the silica gel with water does not remove the resonance at -0.5 ppm, indicating
that the DMPC phosphocholine moiety complexes with the Zr+4 present at the interface.
While this is not a surprising result when viewed in the context of the literature on ZP

78,79,81-86,100-103

chemistry, it is one of the first reports that is available showing a direct

complexation between a phospholipid and a zirconated substrate.

2-5: Discussion

The FTIR, ellipsometry, electrochemistry, contact angle and 31P NMR data pro-
vide a self-consistent picture of these interfaces. A stable DMPC lipid bilayer is formed
by vesicle fusion on the 6-mercapto-1-hexanol-coated Au substrate and a DMPC lipid
monolayer is ultimately formed on the zirconated substrate, with the phospholipid head-
group interacting strongly with the Zr+4. The F TIR data confirm the presence of phos-
pholipids on the surfaces through the carbonyl stretches, which could be there only as a
result of phospholipid deposition. Ellipsometric thickness data are consistent with the
presence of a lipid bilayer on the mercaptohexanol-terminated surface and a lipid mono-
layer on the zirconated phosphate surface. These data however, do not provide explicit
information on the organization of the phospholipids. The F TIR data suggest that the
acyl chains of DMPC are not in a fully crystalline state, as gauged by the band positions
of the asymmetric and symmetric CH2 stretches. The cyclic voltammetry data, using two
different electrochemical probes, demonstrate that the DMPC mono- and bilayers are not

free of defects, but they do cover the interface to a significant extent. Water contact angle

36

data are complementary to the ellipsometric and electrochemical data and show that the
bilayer is hydrophilic while the monolayer is hydrophobic. Taken collectively, the F TIR,
ellipsometry, contact angle, and electrochemical data confirms that a DMPC bilayer is
being formed on the 6-mercapto-l-hexanol coated substrate and a monolayer is formed
on the zirconated substrate and that these interfacial structures remain even after being
removed from the aqueous environment in which they are formed. The 31P NMR data
demonstrate that there is a measurable interaction between the DMPC head group and the
zirconated interface, consistent with the known chemistry of Zr—bisphosphonate systems
and the ellipsometric and contact angle data presented here.

There is limited precedent for the formation of air-stable bilayers. The approach
to creating air-stable bilayers described here relies on the formation of a well organized
interface on which the phospholipid mono- or bilayer can form. The mercaptohexanol
base layer is comparatively well organized and tightly packed, based on the CV data
(Figs. 2-3). Such an interface allows for substantial interaction of the terminal —OH
fimctionalities with the phospholipid head group. For the case of the zirconated interface,
the dominant chemical interaction is the formation of a ZP-like complex with the phos-
pholipid headgroup, as confirmed by 3 1P NMR measurements. For both interfaces, the
interactions between the terminal chemical functionality and the phospholipid headgroups
are sufficiently strong to allow the formation of structures that are stable in a range of

environments.

2-6: How Mono- and Bilayers Form on the Modified Au Substrates
The interpretation of this data in the context of a lipid bilayer forming on a 6-mer-
capto-l-hexanol SAM and a lipid monolayer ultimately forming on a zirconated SAM

requires a consideration of how each of these structures can form and has implications

37

in terms of the mass of phospholipid that ultimately deposits at the interface. For lipid
adlayer formation on the 6-mercapto-l-hexanol SAM, vesicle fusion gives rise to the for-
mation of a bilayer structure. When a vesicle contacts the interface, then spreads on that
interface, there is the opportunity for lateral mobility of the bilayer constituents owing to
the nature of the interface-lipid interactions. Because of the translational mobility of the
planar bilayer on the 6-mercapto-l-hexanol SAM, vesicle fiision can proceed on open
regions of the SAM until an essentially complete bilayer forms.

It is asserted that the physical “picture” for the zirconated interface is fimda—
mentally different. While vesicle fusion proceeds on the zirconated interface, once the
phospholipid headgroups of the bottom leaflet of the planar bilayer contact the zirconated
interface, they bind and are not free to execute translational motion. Thus the interface
coverage is heterogeneous initially, and the interstices between covered regions are, of
necessity, left open once they become smaller than an area that can accommodate vesicle
deposition and fusion. Over time, mediated by translational motion of the top lipid leaflet
and translocation of the top leaflet constituents once they reach the edges of the bottom-
leaflet “islands”, the open interstitial regions are filled in with phospholipid molecules
and a lipid monolayer results.

The coverage of a planar surface achieved by vesicle fusion is difficult to model,
but if it is assumed that the spherical vesicle fuses to form a circular bilayer which is not
free to translate on the surface, one can estimate the maximum achievable surface cover-
age. F or a hexagonal close-packed array of circles, the maximum coverage would be
94% and for a face-centered cubic arrangement, coverage would be ca. 78%, based on
simple geometric models. Because of the nominally random deposition of the vesicles on
the zirconated surface, it is improbable that such high initial surface coverage is achieved.

There may ultimately be a slight excess of lipid molecules present by the end of the par-

38

. v .- r"o.|1:!.VDQP[4{ "A". hfl‘t!" "01".“

 

tial bilayer deposition and monolayer formation process. Any excess “top leaflet” mole-
cules will likely dissolve into the bulk solvent and/or reincorporate into lipid vesicles that

remain in solution.

2-7: Kinetics of Lipid Adlayer Formation

With this model of the interface in mind, one can consider the energetics and
kinetics of lipid monolayer formation on the zirconated interface. From the data shown
in Fig. 2-5a, there appears to be a time constant associated with the formation of the lipid
monolayer. Initially it appears that a partial lipid bilayer structure forms on the surface,
and over time the partial bilayer converts to a monolayer. Through the use of a zircon-
ated interface, a condition has been established where a lipid initially in the top leaflet
will bind to the zirconated surface essentially irreversibly once it migrates to an edge of
the lipid island and executes a translocation to the bottom leaflet. The energetics associ-
ated with this process based on the time constant observed for the evolution of the mono-
layer structure can be estimated. Using the ansatz that, for the interactions of lipids with

zirconated surfaces, the relevant unimolecular reactionlo’65’72 is

k

N

top

lop—bottom > Nbonom (2_ 1 )

 

While it is possible that there is some dissociation from the lipid-ZP complex, the rate
of dissociation is expected to be slow. The energy of formation for a Zr-bisphosphonate
complex, which is similar to the lipid-ZP complex, is known to be 2250 kJ/mol, cor-
responding to a dissociation constant at room temperature of ~3x10'44 for 1:1 lipid:Zr
stoichiometry.80 The decay of interface thickness shown in Fig. 2-2a corresponds to the

time constant for lipid top-to-bottom flipping, k = 1:"1 = 8.3 (+125, «3. 1) x10'4

top-bottom

39

s'1 , a value that is similar to literature reports for lipid translocation.10 By treating this

interface evolution reaction as an activated process, the Arrhenius prefactor associated
with such a reaction will lie within the range of 1010 — 1015 Hz. Because temperature-

dependent data for the formation of these interfaces has not been examined, it is not

 

 

 

 

 

 

 

 

A (5") Ba (kJ/mol)
1010 73.3 Table 2-1: Calculated activation en-
” ergy for lipid flipping as a fimctron of
10 79'0 Arrhenius prefactor.
1012 84.6
1013 90.2 ktlolp-bo om zt-le'texPGEa/Rfll}
w ere t 4,0“ = 8.3 x10-4 s ,
10l4 95'8 T = 300 ie, ananit = 8.314 J/mol-K.
1015 101.4

 

 

 

possible to determine the prefactor experimentally. One can, however, determine the
range of activation energies consistent with the unimolecular prefactor range (Table 2-1).
Calculating the activation energies for DMPC lipid trans-leaflet migration using an Arrhe-
nius prefactor in this range and with the experimental time constant measured, activation
energies in the range of 73 — 101 kJ/mol are obtained. It is noted that the values for Ba
are in relatively close correspondence with those reported by Kornberg and McConnell
for translocation of a tagged lipid,10 while Liu and Conboy72 report Ea values on the
order of 200 kJ/mol or more for lipid translocation.

The fact that phosphocholines interact strongly with a zirconated interface sug-
gests the ability to scavenge certain phospholipids from multi-component solutions using
the appropriate interface chemistry. At the present time it is not clear that all phospholip-
ids will interact equally well with the zirconated interface, and the process of understand-
ing the role of phospholipid headgroup identity on the strength of lipid binding to zircon-
ated interfaces is discussed in the following chapter. The phospholipid headgroup is not
all that is anticipated that regulates the formation of lipid monolayer structures, and this

points the way toward evaluating the strength of interaction of phospholipids with other

40

metal ions as well.

2-8: Conclusion

Here was observed a means of forming lipid mono- and bilayer structures that can
cross the air/water interface and remain intact. The interfaces formed have been charac-
terized using F TIR spectroscopy, cyclic voltammetry, optical ellipsometry, contact angle
measurements and 31P NMR data. Reported here is the formation of interfacial lipid
layers on substrates that are either polar (hydroxythiol on Au) or are capable of binding
the phospholipid headgroup (phosphated and zirconated interface). Complexation of the
DMPC phosphocholine headgroup to the zirconated interface was confirmed by 31P NMR
data, demonstrating for the first time, at the time of this writing, knowledge of the com-
plexation of a phospholipid in such a manner. It hasbeen found that for a DMPC bilayer
on the zirconated surface, the time constant for partial bilayer-to-monolayer conversion
is ca. 20 minutes, and for a unimolecular reaction,lo’65’72 this time constant is consistent
with an activation energy between 75 and 100 kJ/mol. Given the interaction between
phospholipids and ZP-terminated interfaces, it is important to understand how the interac-

tion varies with the identity of the phospholipid headgroup, which is discussed next.

41

 

Chapter 3

Phospholipid Headgroup Dependent Assembly of Lipid Adlayers on Zirco-

nium Phosphate-Terminated Interfaces

3-1: Introduction to Surface Modification

The basis for the formation of a lipid bilayer structure is the balance of intermo-
lecular interactions between the lipid nonpolar acyl chain regions and the polar head-
group interactions with the (aqueous) medium with which the bilayers are in contact.
Lipid bilayers that are present in biological systems are comprised of many constituents
and are structurally complex. It is thought that this complexity plays a role in stabilizing
the folding of transmembrane proteins and thus mediating their function.

There is a significant research effort involved with chemical sensing based on the
use of biomolecules as the chemically selective elements. To succeed in using certain
biomolecules as chemical sensing elements, an interface is required that can stabilize the
structure of the biomolecule and at the same time fimction as part of a transduction sys-
tem to relay the chemical signal of interest to instrumentation. Supported lipid bilayers
are an appropriate choice for such purposes. The bilayer composition and the manner in
which the bilayer interacts with the interface to which it is bound need to be investigated
as an initial step in this effort. It is therefore of interest to bind selected phospholipids to
chemically modified interfaces, and one way to perform this binding is through interac-
tions between the phospholipid headgroup moieties and the supporting surface similarly
to that discussed in chapter 2. The work discuss here however, is focused on the depo-
sition and characterization of lipid monolayers, not bilayers of several phospholipids.
Investigation in to lipid monolayers was pursued because they provide an interesting

opportunity to develop an understanding of the interaction(s) between the substrate and

42

the lipid headgroups. This is valuable because it is these interactions that ultimately will
govern the bilayer integrity and physical properties. Once the substrate-lipid interactions
are understood, it is possible to add an outer phospholipid leaflet by Langmuir-Schaefer
deposition,m4'108 for example.

Zr-bisphosphonate and Zr—bisphosphate (ZP) chemistry is again suitable for this
work as it is a type of self-assembly that has been used in the formation of interfacial ad-
layers for some time.79’8]'84’8738'109'I 15 The primary motivation for the use of ZP chem-
istry is that the Zr—phosphate/phosphonate association is energetically very favorable,80
resulting in an essentially irreversible complexation that is characterized by fast reaction
kinetics. It has been demonstrated recently that ZP complexation chemistry can be used
to form phosphocholine lipid adlayers on surfaces terminated with a zirconium phosphate
moiety. 1 16 In that work, 1,2-dimyristoyl-sn-glyceroé3'-phosphocholine (DMPC) com-
plexed with Zr+4 bound to surface phosphate functionalities to produce a self-assembling
lipid adlayer.116 31P NMR data demonstrated that the lipid-interface interaction was
through the phospholipid headgroup phosphate moiety. The resulting adlayer remained
intact after removal from the solution in which it was deposited, and was stable in air,
suggesting its use as a foundation for biomimetic films. The only phospholipid inves-
tigated in that work was DMPC. Because of the compositional complexity of plasma
membranes,117 it is of interest to understand how the structure of the phospholipid head-
group influences the self-assembly of lipid-ZP complexes at planar interfaces.

In addition to the importance of binding phospholipids to interfaces as a step in
the creation of biomimetic interfaces, understanding such interactions may have immedi-
ate utility in characterizing lipid profiles in selected biological systems. Specifically, if
an interface can be identified that binds one or more types of phospholipids selectively,

its use would facilitate the rapid characterization of plasma membranes. The initial step

43

 

in this endeavor is to determine what intrinsic, chemical selectivity for phospholipids is

manifested by the ZP interface. Phospholipid headgroups vary significantly in size, po-

larity, and charge. These factors can and will affect the binding affinity of the phosphate

moiety to D“, and the chemical headgroup-dependent binding efficiency of selected

lipids is discussed in this work.

The use of gold-thiol self-assembled monolayer chemistry

95,118-125 to build a

monolayer on a gold surface that can be modified subsequently to bind selected phos-

pholipids was utilized again in this study. Au substrates are first exposed to 6-mercapto-

l-hexanol to form a
self-assembled monolayer
(SAM), followed by the
reaction of the SAM
terminal -OH group
with POCl3, H20 and
Zr+4.80’87’89 The result-
ing zirconated surface
has been shown to bind
DMPC,l l6 and the focus
of this work is on under-
standing the chemical
and steric factors that are
important to this lipid-
binding process. Exam-
ined here are the affinity

of lipids possessing the

R: —H DMPA
R:

\/\,!.0/ DMPC
R:

m3», DMPE

O

DMPS

 

R
Fig. 3-1: Phospholipid headgroups used as well as the overall
acyl chain. Base acyl chain does not change for each lipid
with the R group being the different headgroups.

44

 

selected headgroups, including l,2-dimyristoyl-sn-glycero-3-phosphatidic acid (DMPA),

l,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC), l,2-dimyristoyl-sn-glycero-3-
phosphoethanolamine (DMPE), l ,2-dimyristoyl-sn-glycero-3-[phospho-rac-( 1 - glycero 1)]
(DMPG), and l,2-dimyristoyl-sn-glycero-3-[phospho-L-serine] (DMPS). The structures
of these lipids are shown in Fig. 3-1. Acyl chains for all of the lipids were C14 with no
unsaturations, allowing evaluation of the role of lipid headgroup structure on the surface-
binding process.

The results for lipid-binding to a zirconated surface are presented here, including
cyclic voltammetry, time resolved ellipsometry, and water contact angle data to elucidate
the formation and to a limited extent, the organization of the monolayers. The data are
consistent with structurally-based expectations, where steric factors and intermolecular
interactions such as hydrogen-bonding can play a significant role in mediating surface-

binding phenomena.

3-2: Experimental Set up and Materials Used

Materials utilized in these studies include the following and are similar to those
used in the work discussed in chapter 2. They include: 1,2-dimyristoyl-sn-glycero-3-
phosphatidic acid (DMPA, monosodium salt), l,2-dimyristoyl-sn-glycero-3-phospho-
choline (DMPC) dissolved in chloroform, l,2-dimyristoyl-sn-glycero-3-phosphoetha-
nolamine (DMPE), l,2-dimyristoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (sodium
salt) (DMPG), and l,2-dimyristoyl-sn-glycero-3-[phospho-L-serine] (sodium salt)
(DMPS), all in a mixture of chloroform, methanol and water, and are obtained from
Avanti Polar Lipids, Inc. The solvents acetonitrile, ethanol (100%) and ethyl acetate, as
well as the reagents 6-mercapto-l-hexanol, phosphorus oxychloride (POC13), zirconyl

chloride octahydrate (ZrOC1208 H20), 2,4,6-collidine, potassium ferrocyanide trihydrate,

45

 

lithium perchlorate, hexaminerutheniumflll) chloride, and potassium chloride, are all ob-

tained from Sigma-Aldrich in the highest purity grade available. 18 M0 Water was used
for all of the experiments and all reagents were used as received, without fiuther pmifica-
tion.

The instumentaion again was the same as the previous study as well and includes:
a CH Instruments 650 electrochemical analyzer for electrochemical analysis, a J. A.
Woollam Co., Inc. spectroscopic ellipsometer model EC110 with a wavelength range of
185-1100 nm, utilizing 44 wavelengths simultaneously for optical ellipsometry measure-
ments, and an ACT Products Inc. VCA 200 video contact angle system for all water con-
tact angle measurements. Unless noted otherwise, experiments were performed at 20°C .

Two electrochemically active probes were used to characterize the interfaces;
K3Fe(CN)6-3 H20 (1.32 mM) in 0.1 M LiClO4 and 'Ru(NH_,))6Cl3 (1.00 mM) in 0.1 M
KCl. These two probes were chosen because of their different electron transfer kinet-
ics and ionic charges.88 Cyclic voltammetry (CV) was performed with each probe being
cycled three times at a scan rate of 0.1 V/s. The F e(CN) 63 44' probe was scanned from

-0.1 to 0.5 V and the Ru(NH 3H?“ probe was scanned from -0.4 to 0.1 V vs. Ag/AgCl

3%
using a Pt counter electrode. It should be noted that due to calculation error, 10 times the

iron probe concentration was used initially leading to larger then expected currents.

3—2. 1: Sample Preparation

Gold substrates were prepared using a procedure described previously.89 Briefly,
the substrates are rinsed with water and ethanol, cleaned in a UV—cleaner for 15 min., then
exposed to 10 mM 6-mercapto-l-hexanol in ethanol for 6 hrs. The resulting interface is
then rinsed with ethanol and ethyl acetate, then dried under a stream of N2(g). For Zr-

modified interfaces, the 6-mercapto-l-hexanol monolayer is reacted with POCl3 (0.4 mL)

46

 

 

Uidr}

phosr

N,(gl

zncor

soluni

VCSlCIt
r0f0rn'
sohnio
dncdli
ndclTi
11111th 1':
lipids p;
$011579.
lSpassc

lfincsro

 

in dry acetonitrile (10 mL), and catalyzed with 2,4,6-collidine (0.4 mL) for 3 hrs. The

phosphate-modified monolayer formed is then rinsed with ethanol and water, dried with
N2(g), and exposed to 5 mM ZrOCl2 in a 60:40 ethanol/water solution for 12 hrs. The
zirconated monolayer is then finaly dried under N2(g), then is ready for exposure to the
solution containing lipid unilamellar vesicles.

Unilamellar vesicles of each lipid are prepared as described previously.60 The
vesicles used here are comprised of phospholipid only, with no other constituents. Chlo-
roform or the chloroformzmethanolzwater ternary system is first evaporated from the lipid
solution using a N2 stream, and any remaining solvent was removed under vacuum. The
dried lipid is next dissolved in a 10 mM tris(hydroxymethyl)-aminomethane hydrochlo-
ride (Tris®, Aldrich) pH 7.5 buffer solution to a final concentration of 1 mg/mL. The so-
lution is treated with five freeze-thaw-vortex cycles to ensure thorough suspension of the
lipids prior to extrusion.90 To form unilamellar vesicles with a narrow size distribution, a
syringe-based mini-extruder is used (Avanti Polar Lipids, Inc.).9l'93 The lipid suspension
is passed through a polycarbonate filter with an average pore diameter of 400 nm eleven
times to produce unilamellar vesicles of that diameter.

Planar lipid adlayers are formed by spontaneous fusion of unilamellar vesicles.75
The modified gold substrates are placed in a custom-made Teflon® flow cell (Fig. 2-1)
with an approximate volume of 1 mL that has been described in detail in the previous

chapter. 1 16

The flow cell is used to ensure the lipid vesicle suspension was in full contact
with the substrate during adlayer formation. Tris® buffer is flowed over the substrate at
ca. 5 mL/min. prior to lipid deposition, followed by the lipid suspension, flowed through
the cell at the same rate, until the buffer solution is displaced. The vesicle-containing

solution is then allowed to remain in contact with the substrate for a fixed period of time.

For the electrochemical experiments, the vesicle-containing solution is in contact with the

47

substrate for at least six hours to allow for the maximum adlayer formation. After expo-
sure to vesicle solution, the substrates are washed with water at the same flow rate. The
water is then aspirated from the cell at 1 mL/min, and the substrate is removed from the

flow cell and allowed to air dry by hanging vertically.

3-3: Individual Lipid Results

The primary purpose of this work is to evaluate the affinities of selected phospho-
lipid headgroups for a zirconated surface and thus gauge the extent to which lipid adlayer
self-assembly occurs. Considered first is the experimental data for the different phospho-
lipids individually, then by comparing these results, it is possible to assess which factors

are of primary importance in determining lipid-interface interactions.

3-3.1: DMPA

The planar substrate used in this work are Au that is first reacted with a 6-mercap-
to-l-hexanol to form a hydroxyl-terminated SAM. The resulting interface is subsequent-
ly reacted with ZrOCl2 to produce a zirconated surface. The zirconated substrate is then
exposed to a solution containing DMPA vesicles, and optical ellipsometry is used to mea-
sure the thickness of the resulting adlayer ex situ as a function of vesicle exposure time.
DMPA exhibits a rapid build-up to a thickness of 3012 A (Fig. 3-2a), consistent with the
formation of a lipid monolayer.96 Once the lipid adlayer formed, the thickness remained
constant as a function of vesicle exposure time. Water contact angle measurements are
performed on these same interfaces, providing some insight into their polarity and homo-
geneity.72’97'99 The water contact angle for a DMPA adlayer is 104°, with a hysteresis
(the difference between advancing and receding contact angles) of ca. 7° (Fig. 3-2b). The

value of 104° indicates that the chemical functionality of the adlayer in contact with the

48

 

 

 

 

 

 

40 a 300 C
35 ' T “A 200
. ' E ’
30 _ Vixi 8 [00 _ /,# < R 3 <_‘
A i 1].;// 1 /
.4: 25 - v 0— / , / ’
v b /' /
. / , / .
a 20 - a .100 - / _.
D E / / .'
r) ”T / .
§ 15 - 'c -.00 » \\ ,1
O 0— ix: .
.... c .. \ _,
J: 10 — o -300 -
...
I:
5 - 8 400 -
0 I L (w I l A I
”5 50 mo 150 200 250 300 .04 03 02 —Ol 00 01 0.2
IIO - f3 b n" '00 d
_ . E _
i05 - 4“; s s , g - , a -- 7 Q
A £9 1’ ,,,,,, " ...... < 75
ch '00 ' Y . Y 1
o ‘f ’ v
'0 95 .1 _.T J...
v .».Lr I I 3:? 50 _
a) 90 f3 w
30 85 5
g '0 25 »
,_, so ‘5
o 0 5 , ,4”
N 75 > I: 0 79,; «2 ’:v ’
H f ::_::.:.—3/"“
S 70 - a r! 7 .
° 65 - .25
60 I I l I J 1 A I A l I I I I
0 50 I00 i50 200 250 300 or 0.0 0.1 0.2 0.3 0.4 0.5
time (min.) potential vs. Ag/AgCl (V)

Fig. 3-2: DMPA results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the dashed line is
advancing, and the dotted line is receding. C) CV of DMPA monolayer (solid line) and
blank substrate (dashed line) when exposed to Ru(NH ) Cl . D) CV of DMPA mono-
layer (solid linc) and blank substrate (dashed line) when exposed to K3Fe(CN)6.

water droplet is nonpolar, suggesting the lipid acyl chains are the outermost component of
the adlayer. By comparison, the blank ZP surface is characterized by a contact angle of
ca. 72°. The hysteresis seen for the DMPA adlayer suggests modest spatial heterogeneity
in the organization of the adlayer.

Cyclic voltammetry of electroactive probes in solution over the adlayers is also
performed to evaluate interface uniformity. Two electrochemical probes are utilized, with
CV data for Ru(NH3)6Cl3 (Ru probe) shown in Fig. 3-2c and K3Fe(CN)6 (Fe probe) in
Fig. 3-2d. For adlayers and probes where redox waves were detectable, the reactions are

found to be reversible, redox waves for both probes, with peak splitting for both probes

49

 

 

 

 

 

 

being consistent with literature reports.57 For the zirconated adlayer, probe access to

the electrode is limited for the Ru probe and is heavily attenuated for the Fe probe. The
measured peak splitting is found to be 128 mV with a midpoint potential of 190 mV vs.
Ag/AgCl for the Fe probe with the ZP treated 6-mercapto-1-hexanol SAM surface. For
the DMPA-terminated interface, it is observed that little to no Fe probe electrochemi-

cal response is found. The Ru probe yields a peak splitting of 164 mV and a mid-point
potential of -1 70 mV vs. Ag/AgCl for the ZP treated 6-mercapto-1-hexanol SAM. The
Ru probe electrochemical signal is attenuated slightly for the DMPA-terminated interface,
yielding a splitting of 212 mV and a midpoint potential of -l67 mV vs. Ag/AgCl. These
results are compiled for both probes and all adlayers studied in Table 3-1 which is dis-
cussed later in this chapter. The splitting data for all measurements suggests interfacial
adlayer mediation of the probe electron transfer kinetics. A peak splitting of 59 mV is
expected for a fiilly reversible reaction with fast electron transfer kinetics. This increase
in peak splitting data could also contain a contribution from hindered diffusion, but this is
believe to be less likely to account for these reported findings than mediation of electron
transfer kinetics by the adlayer. This statement is based on the fact that the diffusional
properties of both probes are similar, and the electron transfer kinetics for the Ru probe
are somewhat faster than for the Fe probe. If hindered diffusion accounted for these data,
greater similarity in the peak splitting for both probes, and a dependence of the peak split-
ting on adlayer identity which follows the ellipsometry and/or contact angle data would
be expected. This is not observe for either of these trends (Table 3-1). For the Ru probe,
it is clear that the lipid adlayer is slowing the electron transfer kinetics to a greater extent
than for the 6-mercapto-1-hexanol SAM-terminated interface alone. The electrochemi-
cal and water contact angle data point collectively to the DMPA monolayer containing a

measurable quantity of defects. There is precedent for the formation of a lipid monolayer

50

 

mm -."..._-. '

 

at a zirconated interface. Recent work on DMPC interactions with a zirconated interface
show that a monolayer does form, with the dominant chemical interaction being shown

by 31P NMR to be the complexation of the Zr+5 by the lipid phosphocholine group. 1 16

3-3.2: DMPC

DMPC forms a stable adlayer on the zirconated Au surface, following a ca. 20
minute equilibration time. 1 '6 The ellipsometric thickness of the DMPC adlayer was
measured to be 28i3 A,116 (Fig. 3-3a) similar to that found for DMPA and consistent with

a lipid monolayer being deposited at the interface. The water contact angle for the result-

 

 

 

 

 

 

 

50 » a 300 C
l .r‘ l
45‘ it 5 200+
40 \ 4 -
[T < 100- / \ \g,
35 i 1 _, ,7,
A l, ‘ V .7 / / ’
o< 30 , ‘ k I// ” } b 0» / /
v ,x \ \ Y ”v.1.
a 25 i i . 5 -mo- 5. ,
o x A
.5 20- , “U ,,00_ /
O—l a \ ' /
O : -‘\\ /
.— I5 - 0 \ 74/
'5 I: .300~
l0 - 8
5 _ .400 ~
0 A I l I I I l l I Sm L I I I A 1 l
5 0 so 100 ISO 200 250 300 350 400 .0.4 03 .02 -0.I 0.0 0.1 0.2
iio m" 100 d
. E _ I
105 Q 1"
2:. we 2%
.. .. . t v
3 ‘ i .33 50
Q) 90‘ K m
'3“) 35 "i 5
5 ‘ l! i ‘ . 1: 25-
80 - H
5 E
«r 75- l‘: 0- . 4mg a" "
"‘ _. .7. 7,, r:;~‘;—‘—"":: , 'r’
8 7o— 3 ,
° 65- 25
(,0 1 I L I 1 r r I J J r I I L I
—50 0 50 100 150 200 250 300 350 400 -0.l 0.0 0.: 0.2 0.3 0.4 0.5
time(min.) potential vs. Ag/‘AgCl (V)

Fig. 3-3: DMPC results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the dashed line is
advancing, and the dotted line is receding. C) CV of DMPC monolayer (solid line) and
blank substrate (dashed line) when exposed to Ru(NH3) 6C13. D) CV of DMPC mono-
layer (solid line) and blank substrate (dashed line) when exposed to K3Fe(CN)6.

51

 

ing interface is ca. 90° (Fig. 3-3b) with a hysteresis of 5°. While the water contact angle

is slightly less than that seen for the DMPA adlayer, it is consistent with the lipid acyl
chains being the moiety which defines the outer portion of the adlayer. The comparative-
ly low hysteresis is consistent with the existence of a slightly more homogeneous adlayer
than was seen for DMPA, in keeping with the electrochemical data. Cyclic voltammetry
data for the DMPC interface using the Ru and Fe electrochemical probes (Figs. 3-3c and
3-3d, respectively) indicate that the DMPC adlayer effectively blocks access of the Fe
probe to the Au surface, while allowing the Ru probe access. For the Ru probe, the ZP
treated 6-mercapto-l-hexanol interface yields a splitting of 166 mV with a midpoint po-
tential of -164 mV vs. Ag/AgCl (Fig. 3-3c). For the DMPC-terminated interface, we ob-
serve 167 mV of peak splitting and a midpoint of -1 70 mV vs. Ag/AgCl. Both interfaces
mediate the electron transfer kinetics of the probe to the electrode, and given the similar-
ity of the data, it appears that DMPC does not influence the organization of the underly-
ing 6-mercapto-1-hexanol SAM adversely, as is seen for the DMPE-terminated interface
(vide infra). These electrochemical findings are also consistent with a well organized
adlayer, because the Ru probe is energetically favored over the Fe probe in terms of being

able to penetrate the interfacial adlayer.88

3-3.3: DMPE

Ellipsometry and water contact angle measurements for the zirconated substrate
exposed to DMPE unilamellar vesicles yielded data consistent with a somewhat less
organized adlayer than that of either DMPA or DMPC. The ellipsometric thickness
was measured to be l6il A (Fig. 3-4a) suggesting either sub-monolayer coverage or a
relatively uniform lipid adlayer exhibiting a ca. 45° tilt angle with respect to the surface

normal. Water contact angle data for the DMPE adlayer reveals a hydrophobic interface

52

 

 

 

 

 

 

40 a A 300 C
35 - g 200 -
\ /./’ \._
30 — i 100 . / ;
, /
'3; 25 E‘ 0 " / .-"/
w 20 E .100 - ,/ .7/
8 o /' .' /
E "U , /’ ' r /
. .3 , _ x ,1
g 15 T t: 200 “ \ 77/
- ‘ a) \ .'
5 l0 — . E .300 — f
5 ° 400
0 I r I r_ L . I 1 5“) I r I I J L J
0 so l00 150 200 250 300 .04 —0.3 -0.2 -0.l 0.0 0.1 0.2
NS -
b A d
110 - NE 100 -
T
105 - ‘
at r»? 2
’5 ”A A V, i A 75 _
no '00 v7? . iv "' J? J; 3
.8 95 3 .Y T : , y . t
v dI *1 I Q 50 _
a) 90 ‘ o . m
'3'” 85 - ' ‘ { ° 1 ’i’ 5
g 1 . ’ l -o 25 — ,.
so . I 0-0 _ .7/
e E
S 75 ' I: 0 - ,4 A». ,L 7:153”
5 70 - g 2'" 7' ’
o .
65 - 25 ’
w I I I I I I l I l I I I I
0 so 100 150 200 250 300 —0.l 0.0 0.: 0.2 0.3 0.4 0.5
time (min.) potential vs. Ag/AgCl (V)

Fig. 3-4. DMPE result. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the dashed line is
advancing, and the dotted line is receding. C) CV of DMPE monolayer (solid line) and
blank substrate (dashed line) when exposed to Ru(N H3) 6C13. D) CV of DMPE mono-
layer (solid line) and blank substrate (dashed line) when exposed to K3Fe(CN)6.

with a contact angle of ca. 103° (Fig. 3-4b). The contact angle hysteresis for the DMPE
interface is >10°, a value consistent with a heterogeneous surface and arguing for sub-
monolayer coverage. The ellipsometry and water contact angle data are both consistent
with a DMPE interfacial adlayer that is measurably less well organized than either the
DMPA or DMPC adlayers, and likely present as a partial adlayer. Cyclic voltammetry
data for the DMPE interface (Figs. 3-4c and d) reveal a more prominent Ru redox wave
than was seen for DMPA or DMPC, but for the Fe probe, no signal is seen for either the
ZP treated SAM or the DMPE-terminated interface. The Ru probe data exhibit a peak

splitting of 164 mV and a midpoint potential of -l70 mV vs. Ag/AgCl for the ZP treated

53

 

 

 

 

 

SAM, and a peak splitting of 151 mV with a midpoint potential of -124 mV vs. Ag/AgCl

for the DMPE-terminated interface. The addition of the DMPE adlayer apparently facili-

tates the electron transfer kinetics compared to the 6-mercapto-l-hexanol SAM, implying

that the DMPE disrupts the organization of the underlying SAM. These data point col-

lectively to the existence of a heterogeneous interface, consistent with partial monolayer

coverage by DMPE.

3-3.4: DMPG

Adlayers formed using DMPG exhibited properties that are significantly differ-

ent from those observed for DMPA and DMPC adlayers, and more akin to that seen for

 

 

 

 

40 a
”A
35 E
30 i
V
A
‘35 25 *3."
m 20 '
8 8
13
5 I5 *" '
p s
S w 5
5 O
0 I L l A l L ‘ L l J
0 50 mo I50 200 250 300
IIS b A
(‘0
no
-, 5
I05 -rw-I'” 2
ab I00 WY .................................... g e
8 95 ‘i Q 3‘
V
2 m i a—
DD 85 E
5 so 1::
‘5 E
3 75 o
g 70 5
O
65
60. 4 L l A l e + l ‘ l ‘ l ‘
0 50 I00 I50 200 250 300
time (min.)

300

200 ~

 

 

_ c
I
I A 4
OJ 0.2
* d

 

 

flat-z” ‘* "r“
I . r A J _ n A 1 1 + J #
-0.| 0.0 or 0.2 0.3 0.4 05

potential vs. Ag/AgCl (V)

Fig. 3-5: DMPG results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the dashed line is
advancing, and the dotted line is receding. C) CV of DMPG monolayer (solid line) and
blank substrate (dashed line) when exposed to Ru(NH3)6Cl3. D) CV of DMPG mono-
layer (solid line) and blank substrate (dashed line) when exposed to K3Fe(CN)6.

54

 

 

Matt “7'— ”I
m —

 

 

DMPE. The ellipsometric thickness of the DMPG adlayers is 1628 A, requiring an hour

to form (Fig. 3-5a). This thickness value is similar to that seen for DMPE and is consis-
tent either with a partial, spatially heterogeneous adlayer, or an adlayer that displays a ca.
45° tilt angle relative to the interface normal for its acyl chains. The water contact angle
value for the DMPG adlayer is ca. 101° (Fig. 3-5b), with a hysteresis of ca. 10°. The
interface is clearly nonpolar but the magnitude of the hysteresis indicates the presence of
spatial heterogeneity in the adlayer, consistent with fractional coverage. Cyclic voltam-
metry data for the Ru and Fe probes indicate that the addition of the lipid adlayer can
diminish the organization of the thiol SAM (Fig. 3-5d). It is found that for the Ru probe
there is a peak splitting of 166 mV for the ZP treated SAM, with a midpoint potential of
-l69 mV vs. Ag/AgCl. With the addition of the DMPG adlayer, an increase in splitting
to 254 mV and a midpoint potential of -l43 mV vs. Ag/AgCl are observed. It is clear
that the presence of the DMPG adlayer is mediating the electron transfer kinetics at this
interface, but perhaps more important is the observation that the magnitude of the cur-
rent is the same for both interfaces. In all cases, the Ru probe has significant access to
the electrode. For the Fe probe, it is found that the thiol adlayer yields a small Fe redox
wave, with a peak splitting of 152 mV and a midpoint potential of 225 mV vs. Ag/AgCl.
In this case, the presence of the DMPG adlayer hinders access of the Fe probe to the
electrode surface. These data indicate, collectively, that there can be significant access of
the electrochemical probe to the electrode, implying a disordered, incomplete adlayer that

can influence the organization of the 6-mercapto-1-hexanol SAM is formed by DMPG.

3-3.5: DMPS
DMPS is also used to evaluate its propensity for adlayer formation on a zirconated

interface. Ellipsometric thickness measurements yielded erratic results with data ranging

55

 

 

thickness (A)

contact angle (deg)

 

 

 

 

 

a
50 100 150 200 250 300
. t
l
1
50 I00 I50 200 250 300
time (min.)

current density (uA/cmz)

§

§

current density (uA/cmz)

at"

N ~l
VI LII

I
N
La

3%.

O
I

as:

 

 

 

g r

O
.,1

 

 

-0.4 -0.3 -0.2 0 I 0.0 0 I 0.2
. (’7’
-0. I 0.0 0. I 0.2 0.3 0.4 0.5

potential vs. Ag/AgCl (V)

Fig. 3-6: DMPS results. A) Ellipsometric thickness measurement in time. B) Water
contact angle in time with the solid line being the initial drop angle, the dashed line is
advancing, and the dotted line is receding. C) CV of DMPS monolayer (solid line) and
blank substrate (dashed line) when exposed to Ru(N H3) 6C13. D) CV of DMPS mono-
layer (solid line) and blank substrate (dashed line) when exposed to K3Fe(CN)6.

from 4 A to ca. 10 A (Fig. 3-6a). These data indicate the formation of what could opti-

mistically be termed a partial adlayer. Water contact angle data on these same interfaces

indicated the existence of a hydrophobic adlayer being formed (Fig. 3-6b), with a contact

angle of ca. 100° and a hysteresis of ca. 15°. Such a large hysteresis indicates a highly

nonuniform interface, consistent with the ellipsometric thickness data. The interaction of

DMPS with a zirconated interface is not sufficiently favorable to produce an identifiable

adlayer, and what does adsorb onto the zirconated interface does so only afler extended

exposure time. Cyclic voltammetry data using the Ru and Fe probes indicate an adlayer

with fractional interface coverage. Redox waves are seen for the Ru probe but not for

Fe probe (Fig. 3-6c,d). For the ZP treated SAM, the Ru probe produces a peak splitting

56

 

of 166 mV and a midpoint potential of -l7l mV vs. Ag/AgCl. With the addition of the

DMPS adlayer, the splitting is 162 mV and the midpoint potential shifts to -128 mV vs.
Ag/AgCl. Given the comparative magnitudes of the voltammograms for the SAM and
DMPS interfaces and peak splitting data, it is possible that lipid deposition disrupts the

organization of the zirconated SAM.

3-4: Overall Lipid Observations
With an overview of the ellipsometry, water contact angle and CV data for each of

the lipid adlayers that have been studied, the next consideration is how these data com-

 

pare to one another. Specifically, of concern is how the phospholipid headgroup identity
mediates lipid adlayer formation. All data collected for each phospholipid headgroup

is overlaid in Fig. 3-7 in order to gain further insights on the variations between head-
groups. The data point to the division of the lipids into two broad categories; those phos-
pholipids characterized by comparatively strong headgroup interactions with the zirc0n-
ated interface (DMPA, DMPC), and those phospholipids that interact only to a limited
extent with the zirconated interface (DMPE, DMPG, DMPS). This draws on the asser-
tion that there is both a steric component and an “interaction” component (e. g. hydrogen
bonding) that account for these findings.

A word is in order at this point on the formation of lipid monolayers rather than
bilayers in this study. The physical picture for monolayer formation as discused in chap-
ter 2, is believed to proceed as follows: initially, vesicle fusion proceeds on the zircon-
ated interface, and once the phospholipid headgroups of the bottom leaflet of the planar
bilayer contact the zirconated interface, they bind in an essentially irreversible manner,
provided the headgroup structure allows for this interaction to occur. The bottom leaflet

is thus not free to execute translational motion. It is believed that the interface coverage

57

 

 

 

 

 

 

 

 

 

 

fl) - 60 —
9.9 ‘0-
” ~ I '.
. ‘ A 20—
% If M - 5 a ..
: ,.. E 20_
3 _0 g _ :
g m . 2 _, f 0 Mb
.0 . “ 1:3} - IIIIII ‘ .
g A.“ (to-
10 - Ll: ..... ..
. 5 .
0 —so.
-IOO ‘
0.2
d
20-
j
3 A '5 //
3 3 ,
v «- IO 1/
'3 I /
a g 5 ' l/ / .1
0 / ‘,/.. -_'_‘_'..
..///. ..... ,»._,,,
0 . .. .. ;§:—.":‘-";‘v,;;-—-,
5133:5557: -~
0 -s. v/
0 50 100 I50 21” 250 300 4).] 0.0 0.1 0.2 03 04 0‘5
- ( I ) potential“. Ag/Agcl (v)

Fig. 3-7: A) Overlay of all thickness measurements in time with DMPA (solid line),
DMPC (dashed line), DMPE (dash-dot-dot-dash line), DMPG (dotted line), DMPS (dash-
dot-dash line). B) Overlay of initial drop water contact angle measurements with DMPA
(solid line), DMPC (dashed line), DMPE (dash-dot-dot-dash line), DMPG (dotted line),
DMPS (dash-dot-dash line). C) Overlay of Ru(NH3) 6C1, probe CVs with Blank (solid
line), DMPA (dashed line), DMPC (bold solid line), DMPE (dotted line), DMPG (dash-
dot-dot-dash line), DMPS (dash- dot-dash line). D) Overlay of K F e(CN) probe CVs
with Blank (solid line), DMPA (dashed line), DMPC (bold solid line), DMPE (dotted
line), DMPG (dash-dot-dot-dash line), DMPS (dash-dot-dash line).

of lipids is heterogeneous initially, and the interstices between covered regions remain
open if they are smaller than the area required for vesicle adsorption and fusion. Subse-
quent to initial deposition, translational motion of the top lipid leaflet, and translocation
of those lipids, once they reach the edges of the bottom-leaflet islands, serves to fill in the
open interstitial regions with phospholipid molecules, resulting in a lipid monolayer. In
cases where the phospholipid headgroup interaction is weak (vide infra), the formation of

a partial adlayer is achieved.

58

3-4. 1: Lipids with Strong Headgroup Interactions

The phospholipids DMPA and DMPC both exhibit strong interactions with the
zirconated interface. Both adlayers produce ellipsometric thicknesses of ca. 30 A, con-
sistent with a monolayer of C14 phospholipid based on molecular mechanics calculations
and experimental data.96 While all phospholipids yielded a water contact angle of 2 90°,
implying a very hydrophobic interface, contact angle hysteresis has proven to be more
informative in terms of adlayer quality. The magnitude of the contact angle hysteresis
scales with interface heterogeneity, with a hysteresis of 5° or less implying a compara-
tively homogeneous interface, and a hysteresis of 10° or more, implying significant struc-
turalheterogeneity.126 The DMPA and DMPC adlayers yield the lowest contact angle
hysteresis of all the lipid adlayers studied, and these findings are in qualitative agreement
with the electrochemical data, which indicate that these two adlayers are the least perme-

able to the Ru(NH3)6Cl3 and K3Fe(CN)6 probes.

3-4.2: Lipids with Weak Headgroup Interactions

The second group of lipids, DMPE, DMPG and DMPS, are all characterized by
comparatively thin (ca. 15 A) adlayer thicknesses, and contact angle hysteresis of 10° or
more. The CV data point to a measurable ability of the electrochemical probes to un-
dergo electron transfer with the Au electrode. In some cases (e.g. DMPE), the addition of
the lipid adlayer appeared to diminish the integrity of the underlying SAM. Taken col-
lectively, the data point to a heterogeneous interface with sub-monolayer coverage of the

lipids. Considered next is the physical and chemical basis for these findings.

3-5: Discussion

There is an important feature of the electrochemical data which can be understood

59

in terms of access of the electrochemical probes to the zirconated layer (Table 3-1). It is
noted that the midpoint potential for all of the zirconated interfaces is located at ca. -170
mV vs. Ag/AgCl. For a zirconated interface, it is expected that the Zr+4 will be coordi-
nated by either OH- or Cl- ions in solution, and it is also likely that non-stoichiometric

127"” These ligands must be

water will be associated with the interface terminal group.
displaced upon complexation of the phospholipid headgroup to the Zr”. It has been es-
tablished previously that the presence of Zr-phosphate at an electrode interface can shift

the midpoint potential of a redox-active species to more positive values if the redox-

 

 

 

 

 

 

 

 

Ru(N l13)63 ”2+ probe F e(CN) 63 ”l 4' probe

Lipid SAM SAM+lipid SAM SAM+lipid

Splitting E _ Splitting E _ Splitting E . Splitting E .

(mV) "’3' (mV) m." (mV) mi' (mV) m."

(FRI? 3m 2%? (€30
DMPA 164 -l70 212 -167 128 190 -- --
DMPC 166 -164 167 -170 -- -— -- --
DMPE 164 -l70 151 -124 -- -- -- _-
DMPG 166 -169 254 -143 152 225 -- --
DMPS 166 -171 162 -128 -- -- -- --

 

 

 

 

 

 

 

 

 

 

Table 3-1. Electrochemical data for Ru(NH )63+/2+ and F e(CN)63'/ 4' probes for interfaces
studied here. The SAM indicated in the Tab e is 6-mercapto-l-hexanol. Endpoint values
are reported vs. Ag/AgCl reference electrode.

active species can gain direct access to the ZP moiety. 129 These experimental midpoint
potential data show no change for DMPA and DMPC interfaces and a ca. 40 mV positive
shift for DMPE, DMPG and DMPS interfaces. For structurally heterogeneous, partial
lipid adlayers, such a shift is expected, and this finding is consistent with the ellipsometry
and contact angle results.

When possible, the Zr—bisphosphate (ZP) structure will form because of the
substantial thermodynamic driving force for this reaction.130 The formation of the ZP
structure can be precluded by either steric interference or by competition from other

intermolecular interactions. As mentioned above, it is known that ZP interfaces with Zr+4

60

 

" Kim

I-

_-

I. _S

 

 

E5.

 

as the topmost layer will attract non-stoichiometric water, presumably due to the hydro-

.127’128 When a ligand complexes with the Zr+4 ion, it must

philic nature of the Zr+4 ion
first displace the water surrounding the metal ion. If the ligand also has a propensity for
interaction with water, the water may not be displaced or may be competitively bound

by the ligand, leading to diminished complexation between the metal and the phosphate
moiety. This situation could lead to structural disruption of the ZP interface. Consider-
ing the phOSpholipids in the context of their propensity for H-bonding interactions, it is
anticipated that the phospholipids DMPA, DMPE, DMPG and DMPS can all undergo
extensive H-bonding with water due to the presence of phosphate, amine, hydroxyl and
carboxylate moieties, respectively. In the case of DMPC, analogous H-bonding with wa-
ter is hindered by the terminal trimethylamine moiety, and experimentally it is found that
the existence of the choline substituent on the lipid phosphate group does not sterically
preclude formation of the ZP complex. For DMPA, the formation of a ZP complex with a
sterically unhindered phosphate can occur, displacing H20 in the process. For the sub-
stituted phospholipids DMPE, DMPG and DMPS, the side groups are capable of H-bond
formation with any water in the vicinity of the ZP group, and can thus maintain the phos-

phonate moieties physically separate from the Zr+4

. For these lipids, it is possible that the
formation of a partial adlayer disrupts the organization of the underlying SAM, leading

to enhanced exposure of the ZP moieties to the electrochemical probes, as manifested by
the observed ca. 40 mV positive shifi in the midpoint potential. For phospholipids that
either do not participate substantially in aqueous H-bonding (DMPC) or for those where
there is no steric issue with respect to access to the phosphate moiety (DMPA), we ob-
serve comparatively strong chemical interactions with the zirconated surface, resulting in

the formation of a structure that resembles a monolayer of lipid. Under these conditions,

electrochemical probe access to the ZP moieties and underlying Au electrode surface is

61

precluded by the presence of the lipid and the midpoint potential seen for these inter-
faces resembles that seen for the (presumably well organized) ZP-terminated interface,
where the ZP moieties are likely coordinated to OH' and/or Cl'. For phospholipids with
phosphate pendant functionalities capable of significant H-bonding, less well organized
adlayers are observed. The data for the phospholipids that fall into this group (DMPE,
DMPG, DMPS) suggest the formation of a spatially heterogeneous partial adlayer on the
zirconated interface.

The results obtained for the DMPS adlayer is somewhat surprising in light of
the fact that Zr-phosphate-carboxylate (ZPC) complexation is known to occur.89’l3 1'13 2
While this complexation is not as energetically favorable as ZP complexation, the pres-
ence of two functionalities in the DMPS headgroup offers the opportunity for multiple
types of complexation, neither of which are found to contribute significantly, based on

our experimental results.

3-6: Conclusions

In this study the time-dependent ellipsometric thickness and contact angle of a se-
ries of phospholipid adlayers bound to a zirconated interface have been measured. These
data, in concert with cyclic voltammetry data using two electrochemical probes of the
resulting interfaces, serve to characterize the comparative binding efficiency and quality
of the lipid adlayers. The data reveal that DMPA and DMPC form the most organized
adlayers with DMPE, DMPG, and DMPS producing adlayers characterized by spatial
heterogeneity and sub-monolayer coverage. The adlayer properties are then related to
the ability of the phospholipid substituted phosphate moiety to interact with the zircon-
ated interface. Both steric issues and the propensity of the phospholipid headgroups to

H-bond with water and possibly other lipids in the proximity of the interface likely play

62

   

- . $02“!!! amt-wtnfiflkflnflv'r~

 

roles in determining the quality of the phospholipid adlayers. One issue that remains

to be investigated is whether or not the means of lipid adlayer deposition influences the
observed interfacial film thickness and uniformity. Lipid fusion is used to effect the
self-assembly process. It is possible that the pre-assembly of lipid adlayers as Langmuir-
Blodgett films could influence the properties of films deposited on the same substrates
used here, and an effort is underway to explore this possibility.

It is clear that the understanding of this novel class of self-assembling adlayers
would benefit from imaging measurements, and this is an effort that is ongoing. It is
hoped that these adlayers will find use in the formation of supported lipid bilayer struc-
tures where the extent of interaction between the lipid adlayer and the support can be
adjusted chemically through the composition of the lipids used.

It is important however, just like with the varying lipid headgroups, to explore
varying substrates. By changing the metal on the surface from Zr to other various met-
als, different lipid-metal interactions could be observed. This could then translate to the
ability to have multi-metal surfaces that have tailored effects on mixed lipid solutions.
Initially however, it is important to gain an understanding of how different metals interact

with a lipid headgroup, which is discussed in the following chapter.

63

 

 

 

Chapter 4
Ionic Binding of Phospholipids to Interfaces: Dependence on Metal Ion
Identity

4-1 : Introduction

The goal of creating bilayer systems is to utilize them to simulate the plasma
membrane in a manner that allows for the presence of trans-membrane proteins in their
active form(s).57 Success in this area requires that the lipid bilayer and the immediate
environment on both sides of the bilayer be sufficiently hydrophilic to mimic a cellular
system, and this issue has led to the design of bilayer structures that reside on hydrophilic
underlayers, for example.59’74’77 In addition to these structural requirements, there are
the issues of lipid bilayer fluidity and the manner in which the bilayer is bound to the
underlayer many of these issues already have been addressed. Simple physisorption of
bilayers onto most substrates yields an interface that is not sufficiently robust to main-
tain its structural integrity in the long term. It is thus important to identify ways to make
more robust the lipid bilayer interaction(s) with the support on which they reside. In the
previous chapters this has been explored, with this work showing that Zr+4 can interact

”6"33 and there is anecdotal evidence that Ca+2

with phospholipid phosphate headgroups,
is required to achieve a high quality lipid bilayer under conditions where the bilayer is
physisorbed to the interface. For these reasons it is also of interest to explore the strength
of interactions between other interfacial metal ions and selected phospholipids, and these
findings are reported here.

To reiterate, mammalian plasma membranes are complex systems that are com-

prised of more than 100 different components.1 This compositional complexity is

thought to be essential for housing transmembrane proteins as well as making the bilayer

64

 

structure sufficiently robust that it is capable of maintaining its structural integrity upon

exposure to air.75'77’1 16 It is found in the earlier work presented that it is possible to cre-
ate a hydrophilic interfacial adlayer with a high density of surface hydroxyl groups, and
this interface can support a physisorbed phospholipid bilayer. The resulting interface is
important for the creation of a well organized and robust biomimetic interface.116 By
modifying the hydroxylated interface to create a Zr-phosphate (ZP) fimctionality, it was
determined that structurally robust lipid monolayers were formed because of the interac-
tion of phosphocholine head groups with the surface-bound ZP functionality. The forma-
tion of a Zr-bisphosphate complex was verified by 31P NMR measurements. 1 16 The use
of Zr+4-phosphate/phosphonate complex formation to create organized mono- and mul-
tilayer interfacial structures is well known, and the success of this approach to controlled
adlayer formation is based on the essentially irreversible Zr+4 interaction with the phos-
phate moieties.89’134
With the establishment of phospholipid binding to zirconated interfaces, 3 key
issue to evaluate was the role of phospholipid headgroup identity in mediating the com;-
plexation process. Utilizing the gold-thiol self-assembled monolayer chemistry used in

earlier studiesgs’1 18"”

to build a monolayer on a gold surface. This can be modified sub-
sequently to bind selected phospholipids. Au substrates are first exposed to 6-mercapto-
l-hexanol to form a self-assembled monolayer (SAM), followed by reaction of the SAM
terminal —OH group with POCl3, H20, and then ZrOClz. It was found that phospho-
choline and phosphatidic acid lipids complexed with Zr+4 ions strongly, while phospho-
ethanolamine, phosphoglycerol and phosphoserine lipids did not form organized lipid
adlayers.133 Contact angle and optical ellipsometry data indicate that the adlayers formed

using these lipids were incomplete and spatially heterogeneous. These findings are un-

derstood in the context of the propensity of the lipid headgroups to hydrogen bond with

65

water in the vicinity of the Zr-phosphate interface. Lipids with H-bonding headgroups
do not complex substantially with the surface-bound Zr+4 because of competitive interac-
tions with nonstoichiometric water in the vicinity of the interface. In this next group of
experiments, phosphatidic acid was chosen to eliminate issues that could be related to

steric contributions to ZP complex formation.

4-2: Metals for Modification of Au Substrates

It is clear that, for reasons of the phospholipid headgroup structure, there is intrin-
sic chemical selectivity associated with the formation of lipid adlayers in this manner. It
is also important to consider whether the metal ion used in the formation of the interfacial
metal-phosphate structure will play a role in mediating the interface-lipid interactions.
Metal-phosphates are known for a range of metal iOns.78’135'142 Most of the metal ions
tested have been divalent transition metals, and Mg2 and Ca+2 have also been used.136'
139 Following the same methods used in the modification of Au substrates with Zr”,116
the examination of the ability of several metal ions, some with biological significance,
to form interfacial complexes with phosphatidic acid has been chosen. Metal ions Ca”,
Mg”, Zn”, Ni+2 and Cu+2 are chosen based on their known propensity for interactions
with phosphates. F e+3 was chosen because iron coordinates phosphate strongly.136 Be-
cause of the oxidative instability of F e+2, it was necessary to work with F e+3 due to the
fact that experiments are performed in air or in an aqueous medium where no effort had
been made to deoxygenate the solution. To gain firrther insight the phosphate interactions
with Cu+ was examined in an attempt to understand whether or not metal ionic charge
played a significant role in the formation of the supported lipid adlayer. Since lipid inter-
actions with Zr+4 have been characterized,133 the data reported here are compared to the

Zr+4-modified interface.

66

 

 

pl

llll

 

 

The results for lipid binding to a metal modified surface include X-ray photo-
electron spectroscopy (XPS), cyclic voltammetry, optical ellipsometry and water contact
angle data, to elucidate the formation and, to a limited extent, the organization of the lipid
adlayers. The data indicate that lipid-metal coordination is a complex process that is me-
diated by the identity and loading density of the metal ion that coordinates to the surface-

bound phosphate groups.

4-3: Experimental Setup and Instrumentation Utilized

The lipid chosen for these experiments was l,2-dimyristoyl-sn—glycero-3-phos-
phatidic acid (DMPA, monosodium salt) and was obtained fi'om Avanti Polar Lipids,
Inc. The solvents and reactants: Acetonitrile, ethanol (100%), ethyl acetate, 6-mercapto-
l-hexanol, phosphorus oxychloride (POCl3), zirconyl chloride octahydrate (2100; 8
H20), 2,4,6—collidine, lithium perchlorate, potassium chloride, as well as the electro—
chemical probes potassium ferrocyanide trihydrate and hexamineruthenium(III) chloride
were obtained from Sigma-Aldrich in the highest purity grade available. The metal salts
calcium chloride (CaClz), zinc chloride (ZnClZ), nickel chloride (NiClz), magnesium
chloride (MgClz) and ferric chloride (F eCl3) were obtained fi'om Spectrum Chemicals.
While cupric chloride (CuClz) was obtained from J .T. Baker Inc. and cuprous chloride
(CuCl) obtained from Mallinckrodt. All metal salts were purchased in the highest purity
grade available and used as received. 18 M9 Water was obtained from an in-house Bam-
stead system and used for all experiments.

The instrumentation utilized is as follows: all electrochemical data were acquired
using a CH Instruments 650 electrochemical bench. Optical ellipsometry measurements
were performed using a J. A. Woollam Co., Inc. model EC] 10 spectroscopic ellipsometer

with a wavelength range of 185-1100 nm, utilizing 44 wavelengths simultaneously. The

67

 

 

water contact angle measurements were performed on an ACT Products Inc. VCA 200

video contact angle system. XPS measurements were performed on a Perkin Elmer Phi
5400 instrument equipped with a Mg-Ka X-ray source. Samples were analyzed at pres-
sures between 10'9 and 10'8 Torr with a pass energy of 29.35 eV and a take-off angle of
45°. The spot size is ca. 250 m2. Atomic concentrations were determined using known
sensitivity factors. All peaks were referenced to the C 1 5 peak associated with adventi-
tious C at 284.6 eV. Unless noted otherwise, experiments were performed at 20°C.

The electrochemical measurements were performed similarly to the experiments
in chapter 2 and 13.116’13 3 Two electrochemically active probes were used to character-
ize the interfaces that were studied; K3Fe(CN)6- 3 H20 (1.32 mM) in 0.1 M LiClO4 and
Ru(NH3)6Cl3 (1.00 mM) in 0.1 M KCl. These two probes were chosen because of their
different electron transfer kinetics across alkanethiol SAMs and their different ionic
charges.88 Cyclic voltammetry (CV) was performed with each probe being cycled three
times at a scan rate of 0.1 V/s. The Fe(CN)63'/4’ probe was scanned from -0.1 V to +0.5

V vs. Ag/AgCl and the Ru(NH 3+/2+ probe was scanned from -0.4 V to +0.1 V vs. Ag/

3)6

AgCl, using a Pt counter electrode.

4-3. 1: Substrate Preparation

Gold substrates were prepared using a procedure described previously.89 Briefly,
the substrates were rinsed with water and ethanol, cleaned in a UV-cleaner for 15 min.,
then exposed to 10 mM 6-mercapto-1-hexanol in ethanol for 6 hrs. The resulting inter-
face was rinsed with ethanol and ethyl acetate, then dried under a stream of N2(g). For
metal modified interfaces, the 6-mercapto-l-hexanol monolayer was reacted with POCI3
(0.4 mL) in dry acetonitrile (10 mL), and catalyzed with 2,4,6-collidine (0.4 mL) for 3

hrs. The phosphate-modified monolayer was rinsed with ethanol and water, dried with

68

 

 

N2(g), and exposed to 5 mM concentrations of one metal salt in a 60:40 ethanol/water
solution for 12 hrs. For each metal ion used, the substrate was prepared in the same man-
ner utilizing metal chloride salts (except for Zr”, where ZrOCl2 was used). The resulting
metal ion-containing monolayer was dried under N2(g), then exposed to a solution con-

taining a DMPA unilamellar lipid vesicles.

4-3.2: DMPA Vesicle Preparation.

Unilamellar vesicles of DMPA were prepared as described previously.60 The
vesicles were comprised of the phospholipid only, with no other constituents. The
chlorofonnzmethanolzwater solvent system was first evaporated from the lipid solution
using a N2 stream. The lipid was then exposed to vacuum to remove any remaining
solvent. The dried lipid was dissolved in a 10 mM‘ tris(hydroxymethyl)-aminomethane
hydrochloride (Tris®, Aldrich) pH 7.5 buffer solution to a final concentration of 1 mg/
mL. The solution was then mixed using five freeze-thaw-vortex cycles to ensure suspen-
sion of the lipids prior to extrusion.90 A syringe-based mini-extruder was used to form
unilamellar vesicles with a narrow size distribution (Avanti Polar Lipids, Inc)”93 The
lipid suspension was then passed through a polycarbonate filter (average pore diameter

400 nm) eleven times to produce unilamellar vesicles of that diameter.

4-3.3: Adlayer Formation

Planar DMPA adlayers were formed by spontaneous fusion of unilamellar ves-
icles.75 The modified gold substrates were placed in a custom-made Teflon® flow cell
that has been described in chapter 2.116 The flow cell was used to ensure the lipid ves-
icle-containing solution was in fill] contact with the substrate during bilayer formation.

Tris® buffer was flowed over the substrate at ca. 5 mL/min. prior to DMPA deposition,

69

 

‘ .-_ '2"'!t'£~1t-'¢?' 931'...“ u tiflx°n*-tvf‘ul§ewcmfi_

 

then the vesicle-containing solution was flowed through the cell at the same rate until the
buffer solution was displaced, and this solution remained in contact with the substrate for
two hours. After exposure to the vesicle-containing solution, the substrate was washed

with water. Following washing, the water was aspirated fiom the cell. The substrate was

then removed from the flow cell and allowed to dry in air while being held vertically.

4-4: Results of DMPA Exposure

The primary purpose of this work is to evaluate the interactions between the
DMPA headgroup and selected metal ions bound to surfaces through a phosphate group,
and thus gauge the extent to which lipid adlayer self-assembly proceeds. First the ex-
perimental data for the metal ions individually is discussed, then the comparison of these
results to assess which metal ions give rise to phospholipid self-assembly, and which
factors are of primary importance in determining the lipid-interface interaction is looked

into.

4-4.1 : Zirconium

As noted above, phosphate-terminated SAMs as the substrate for vesicle deposi-
tion were used. By reacting the phosphate-terminated interface with ZrOClz, it is pos-
sible to produce a Zr+4-terminated surface. XPS was used to determine Zr-surface cover-
age. Analysis of the ratio of ZrzAu4f concentrations yields a value of 0.34 (Fig. 4-la),
which we take to indicate substantially complete surface coverage based on the known
strongly favored complex formation behavior of Zr+4 with ROPO3'2.80’134 It should be
recognized that this concentration ratio is not quantitative due to the fact that the signal
from a monolayer (or less) of metal ions is compared to the signal from a comparatively

thick Au layer, but these ratio data for the different metal ions serve as a useful com-

70

 

 

 

 

parison. The zirconated sub-

strate was exposed to a solution
containing DMPA vesicles, and
optical ellipsometry was used

to measure the thickness of the
resulting lipid adlayer ex situ
(30i2 A), consistent with the
formation of a lipid monolayer.96
Water contact angle measure-
ments were performed on these
same interfaces, providing insight
into their polarity and homoge-
neity.72'97'99 The water contact
angle for a DMPA adlayer is
104°, with a hysteresis (the dif-
ference between advancing and
receding contact angles) of ca.

7°. The value of 104° indicates
that the chemical functionality
of the adlayer in contact with the
water droplet is nonpolar, con-
sistent with the lipid acyl chains
being the outermost component
of the adlayer.133 The hysteresis

seen for the DMPA adlayer sug-

 

 

. xl 0A5 a
3 v
-Au4f7.
ZSi {>15 -Zr3pl
l 1035 ‘-Zcr3p
2 F
‘ O—KLL
1S—Zr3d
-CKLL
01. 5V"!

W
._. 1

gm

 

 

 

 

60 _ M“
40— /'\\
.. 2o- z. -- ------ /
3 °' / ’/
a -2o- ;
0 LZ><f
‘60? &\\ /
-so - \/
_100 I I I I I I I
-0.4 -0.3 -0.2 -0 l 0.0 0.1 0.2
201 c
A 15 — //
15,. ' //
IO — / /
a . / /
5 _ /
g . fl
0 _ ‘W .4,“ ‘‘‘‘‘ #Mf;;__: ::.
-5 _ ”a.” — -
-0.1 0.0 0.1 0.2 0.3 0.4 0.5
WNW-As/MCICV)
Fig. 4-1: a) XPS spectrum of Zr+4-modified thiol/

gold substrate. b) CV of Ru(NH ) CI for a DMPA
monolayer on Zr+4-terminated interface (dashed line)
and for the Zr+4-terminated inter-face with no adlayer
(solid line) 4c) CV of K F e(CN) for a DMPA mono-
layer on Zr 4-terminate<i interfacée (dashed line) and
for the Zr 4-terrninated interface with no adlayer (solid
line).

71

 

   

gests modest spatial heterogeneity in the organization of the adlayer. Typically, hysteresis
of ca. 2-3° is taken to indicate a homogeneous interface, and hysteresis of 10° or more
indicates a structurally heterogeneous interface.

Cyclic voltammetry of electroactive probes in solution was also used to gauge
the presence of defects in the lipid interface. Two electrochemical probes were utilized
because of their different ionic charges and consequent abilities to penetrate nonpolar
adlayers,88 with CV data for Ru(NH3)6Cl3 (Ru probe) shown in Fig. 4-1b and K3Fe(CN)6
(Fe probe) in Fig. 4-lc. The data reveal reversible redox waves for both probes, with
peak splitting for both probes being consistent with literature reports.57 Probe access to
the electrode is limited for the Ru probe and is essentially blocked for the Fe probe with
the Zr-lipid adlayer. For the Ru probe, it was found to have a measured peak splitting
of 164 mV for the ZP treated 6-mercapto-l-hexanol SAM. The Ru probe electrochemi-
cal signal is attenuated slightly for the DMPA-terminated interface, and characterized
by a splitting of 212 mV. It is noted that the splitting data for all measurements indicate
interfacial adlayer mediation of the probe electron transfer kinetics. A peak splitting
of 59 mV would be the expected peak splitting for a fully reversible reaction with fast
electron transfer kinetics. For the Ru probe, it is clear that the lipid adlayer is influenc-
ing the electron transfer kinetics to a greater extent than for the 6-mercapto-l-hexanol
SAM-terrninated interface alone. For the DMPA-terminated interface, the ratio of non-
F aradaic current to the non-Faradaic current of the blank was found to be 0.54. The lipid
adlayer that is formed on the surface is analogous to the dielectric medium in a parallel
plate capacitor, with the gold substrate and the water/lipid interface as the plates. The
spacing between the plates determines the capacitance of the system. The ratio of the
capacitive current measured for the lipid-adlayer terminated interface to that of the blank

(Zr+4-terminated) interface allows for an estimate of the lipid layer thickness, in the limit

72

I1..\OII|!-l\. 11..\4 I

 

 

 

 

 

V

3'1"—

 

 

‘_-

 

x10A 5 a of an interface characterized by very

 

-F LMM few defects. The capacitive current
l -o KLL e -Au4f7

-015 -Cls ‘ ratio of the non-Faradaic currents for

 

the DMPA-terminated interface to

the Zr+4-terminated substrate indi-

c/I
_—_—-‘_

cates adlayer deposition. The cyclic

 

05L 1 4 voltammetry and water contact angle

data point collectively to the DMPA

 

 

 

o l J l J A L l I l A
Iioomoosooaoomqaoosoowoaoozooroo 0 . .
WW) b monolayer contarnrng a measurable

3,,-"'"'_'.',‘_'.:'- quantity of defects. There is prec-
edent for the formation of a lipid

monolayer at a zirconated inter-

 

face.116 Recent work on DMPC in-

' 4,; _o_3 4,2 ‘ -o.1 ‘ o.o ‘ 0.1 ' 02 teractions with a zirconated interface

c showed that a monolayer does form,
with the dominant chemical interac-
tion being shown by 31P NMR to be
the complexation of the Zr +4 by the

lipid phosphocholine group.116

 

 

-0.1 0.0 0.1 0.2 0.3 0.4 0.5
pow VI. Ag/Aga (V)

. +3 _ . 4-4.2: Iron III
Fig. 4- 2: 3) XPS spectrum of Fe -mod1fied thIol/

g01d substrate. b)3 CV Of RU(NH3 )6 C13 for a DMPA Fe+3 is an important metal
monolayer on Fe 3-t3erminated interface (dashed
line) and for the F e 3-terminated inter-face with ion from a biological perspective,
no adlayer (solid line). c)+ CV of I(3Fe(CN)6 for
a DMPA monolayer on F e -3terminated interface and it is known to form metal bis-

(dashed line) and for the Fe 3-terminated interface 143
with no adlayer (solid line) phosphonate Structures. XPS

73

 

measurements of the Fe modified surface revealed a ratio of F e:Au of 0.46 (Fig. 4-2a),
similar to the surface coverage seen for Zr+4 (vide infra). Complexation of F e+3 with
DMPA is also observed by the formation of a lipid adlayer with a thickness of 262t2 A.
Water contact angle measurements of the lipid-tenninated F e+3 interface yield a contact
angle of 102i2°, indicating a hydrophobic lipid monolayer oriented such that the acyl
chains form the outer surface. The contact angle hysteresis for this substrate was found
to be 7°, the same as was found for the Zr+4 interface and indicative of a macroscopically
moderately well organized lipid monolayer. Cyclic voltammetry measurements for the
Fe+3 interface were performed with the two redox probes to gain a insight on the micro-
scopic organization. The Ru probe data reveal that the Fe modified interface is charac-
terized by a peak splitting of 175 mV (Fig. 4-2b) and the DMPA-terminated interface
produces a 277 mV peak splitting. The non-Faradaic current ratio calculated for the Fe
probe was found to be 0.57 (Fig. 4-2c). The electrochemical data are in good agreement
with the contact angle and ellipsometric data. These measurements point collectively

to the formation of a DMPA monolayer that is characterized by a modest defect density.
The F e+3 modified surface is similar to the Zr+4 modified surface, indicating the forma-
tion of a substantially complete lipid monolayer. It should be noted that the examination
of the Fe+3-modified substrate is potentially complicated by the use of the Fe(CN)6'3"4
probe. During potential cycling, the surface-bound F e‘L3 will be reduced, but the current
from this species will be small relative to that of the solution-phase probe, and because
Fe+3 is present on both the blank and the DMPA-terminated interfaces, it will represent a

constant contribution to any interface-dependence seen in the data.

4-3.3: Nickel

XPS analysis of nickel coordination with the modified gold substrate found a

74

 

 

surface coverage ratio of 0.08 N izAu
(Fig 4-3a). There appears to be mini-
mal N i+2 bound to the phosphate-ter-
minated the interface. In contrast to
the behavior of the Cu+2-terminated
interface (vide infra), the Ni+2-ter-
minated interface is seen to form a
complex with the DMPA headgroup.
The thickness of the DMPA adlayer
is 27:53 A, with a water contact angle
of 50:1:3°. By comparison, the Ni+2-
terminated interface yields a contact
angle of 62d:4°. The contact angle
hysteresis for this system is ca. 2°,
suggestive of a uniform interface.
These data are apparently contra—
dictory, with an adlayer thickness
consistent with a lipid monolayer and
contact angle data pointing to a polar
interface. For a heterogeneous inter-
face characterized by bare or coated
domain sizes that are large (ca. um
scale), contact angle hysteresis is

not an accurate gauge of surface

heterogeneity. It is possible to have

 

 

 

a
A

5 mos , . -

45'

r -Au4f5

3.5“

 

 

 

J

kl

 

 

I I 'TTi' I

r11r'

 

b

 

 

 

 

 

75

- 0.3
potential vs. Alt/Asa (V)

0.4

0.5

Fig. 4 3: a) xps spectrum of Ni+2-modified
thiol/gold substrate. b) CV 2of Ru(NH 3)6Cl3 for
a DMPA monolayer on Ni 2-2terminated interface
(dashed line) and for the Ni 2-terminated inter-face
with no adlayer (solid line). 0) CV of 1(3Fe(CN)6
for a DMPA monolayer on Ni+2-terminated In-
terface (dashed line) and for the Ni 2-terminated
interface with no adlayer (solid line).

 

a
x1045
3.5 fl t r
3 - +
-Zn LMM3 'Au4f5.

'AU493 -Au4d3

    

 

 

 

.....

 

 

-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2

mm

 

 

_25 1 . I I l . I . I . J . I .
4).] 0.0 0.1 0.2 03 0.4 0.5

Mal vs. AgIAgCl (V)

 

Fig. 4- 4. 3) XPS spectrum of Zn+2 -modified
thiol/gold substrate. b) CV of Ru(NH3 )6 C13 for
a DMPA monolayer on Zn 2-term2inated6 inter-
face (dashed line) and for the Zn 2-terminated
inter-face with no adlayer (solid line). c) CV of
F e(CN)6 for a DMPA monolayer on Zn 2-ter2-
minated interface (dashed line) and for the Zn+2
terminated interface with no adlayer (solid line).

76

a macroscopically heterogeneous
interface that produces such results,
and electrochemical data are use-
ful in resolving whether or not this
is the case for the Ni+2-terminated
interface. Using the Ru probe, the
Ni+2-terminated interface produces
a splitting of 209 mV (Fig. 4-3b)
and upon exposure to DMPA, a
peak splitting of 171 mV is ob-
served. Such a significant decrease
in peak splitting would suggest
that, upon the initial exposure of
the Ni to the substrate, the Ni co-
ordinates with multiple interfacial
phosphate groups, limiting access
of the Ru probe to the electrode
surface. The addition of a DMPA
adlayer perturbs this organization
and enhances access of the probe to
the electrode surface. The ratio of
the DMPA interface-to-Ni+2-terrni-
nated interface capacitance is 1.28
for interfaces examined with the

Fe probe (Fig. 4-30), a value indi-

 

eating a thickness intermediate between that of a monolayer and a bilayer based on data

obtained for both types of adlayers, suggesting that the DMPA adlayer formed is spatially

heterogeneous with a relatively large characteristic domain size.

4—4.4: Zinc

The behavior of Zn+2 is similar to that of Ni”, suggesting sub-monolayer cover-
age of the phosphate-terminated interface, and yielding a XPS anAu concentration ratio
of 0.08 (Fig. 4-4a). Ellipsometry data point to a DMPA adlayer thickness of 3 1i2 A.
Water contact angle measurements show a contact angle of 66il° with a hysteresis of
ca. 8°, suggesting some amount of surface heterogeneity. These data, taken collectively,
point to a heterogeneous interface with a thickness slightly greater than that expected for
a uniform monolayer and a water contact angle similar to that of the unmodified Zn+2-
terminated interface, which is characterized by a contact angle of 58i4°. Electrochemical
data for the Zn+2-terminated interface, collected with the Ru probe, show a splitting of
166 mV (F i g. 4-4b) and the DMPA-terminated interface exhibits a 214 mV peak splitting.
Again, the presence of the lipid adlayer mediates electron transport at the interface, and
the ratio of capacitance for the DMPA-terminated interface to the Zn+2-terminated inter-
face was found to be 4.72 using the Fe probe (Fig. 4-4c). The electrochemical capaci-
tance data point to the DMPA adlayer enhancing the organization of the Zn+2-terminated
supporting SAM. The capacitance data, by themselves, suggest a thick DMPA adlayer.
Based on all of the data collected for this system, it is more likely that a heterogeneous
surface comprised of regions of DMPA bilayer is formed, and that defect areas as well as
possibly poorly organized regions of the lipid adlayer contribute to these findings. In any

case, Zn+2 appears to produce only modest interactions with the DMPA headgroup.

77

 

:r-r~

 

uiskr

phosp
onrfl:
unflar
hcklt
atan

Uncra

sonal

adhn

 

 

 

4-4.5: Calcium

Ca+2 is of interest because
it is known that calcium facilitates
phospholipid bilayer formation
on planar substrates formed from
unilamellar vesicles.”I44 It is
held that the presence of Ca+2 ions
at an interface somehow mediates
interactions of the phospholipid
headgroups and produces a rea-
sonably uniform interfacial lipid
adlayer. The surface coverage of
Ca+2 was measured by XPS to be
the same as that seen for Ni+2 and
Zn”, with a metal-to-Au ratio of
0.08, indicating low coverage of
metal ion (Fig. 4-5a). Ellipsometry
measurements performed follow-
ing DMPA exposure yielded a
thickness of 39:1:2 A, significantly
higher than that seen for a DMPA
monolayer, and approaching that
of a bilayer. The presence of a
partial bilayer is further indicated

by the contact angle data, with a

 

 

 

x1045 a
45 [ v u r , fir r v v 1 1
4 .
3.5’ ~Au4f ‘
3.
01 -Au4d3
.. 2.5 * ' 5 ~Ca2p
b -Au4p3 -Ca2$

 

 

0.5”

 

 

 

mmnamaaaanao
Ioo BIndIng Fnetgy (0V) b
r

t

 

 

 

-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2

 

 

or 0.0 0.1 0.2 03 0.4 0.5
Mann/A30 (V)

Fig. 4- 5: a) XPS spectrum of Ca+2-modified thiol/
gold substrate. b) CV of RuO‘lH3 ) 6Cl3 for a DMPA
monolayer on Ca +2-t2erminated interface (dashed
line) and for the Ca+ 2-terminated inter-face with

no adlayer (solid line). c) CV of K3 F e(CN)6 for

a DMPA monolayer on Ca z-zterminated interface
(dashed line) and for the Ca 2-terminated interface
with no adlayer (solid line).

78

 

5[x10’:5 . g f a
4.5L 4
4* -Au4f I
3.5 F
3t
52.5 _ -Au4d3

 

 

 

 

 

 

 

 

 

-M92p
o I I I 41 L I 1 l I I
11001000900 mo 7W.“ 5W 4W 3W 2W 1W 0
m0”) 1)
-0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2
30 - C
20 l-
A h 13/
3 lo -
g 0 r
-10 ..
.20 p
l l l l l I l

 

 

-0.l 0.0 0.1 0.2 0.3 0.4

WNW-Adm (V)

Fig. 4- 6: a) XPS spectrum of Mg+2 -modified thiol/
gold substrate. b) CV of Ru(NH3 )6 C13 for a DMPA

monolayer on Mg+ 2-terminated interface (dashed
line) and for the Mg 2-terminated inter-face with
no adlayer (solid line). c) CV of l(3Fe(CN)6 for a
DMPA monolayer on Mg 2-terminated interface

(dashed line) and for the Mg 2-terminated interface

with no adlayer (solid line).

79

water contact angle of 372i:6° and a
hysteresis of 7°. Electrochemical
data for the Ru probe at the Ca”-
modified interface shows a peak
splitting of 237 mV (Fig. 4-5b).
The same interface, modified with a
DMPA adlayer, exhibits a 224 mV
peak splitting. The similarity of
these splitting values argues for the
unimportance of the DMPA adlayer
in mediating the electron transfer
process at this interface. Using the
Fe probe, an observed capacitance
ratio of 4.7 is found for the DMPA-
terminated interface relative to the
Ca+2-terminated interface (Fig.
4-50). These data point to compar-
atively strong interactions between
the Ca+2 ions and the phosphates of
both the underlying SAM and the
lipid headgroup. Ca+2 appears to
interact with the lipid headgroups
but does not form a tightly bound
interfacial structure. This point is

considered below.

4-4.6: Magnesium

Mg+2 is a metal that is found widely in biological systems. XPS measurements of
phosphate-terminated SAMs that had been exposed to Mg+2 were found to yield a Mg:Au
concentration ratio of 0.08 from XPS data (Fig. 4-6a). This finding is consistent with
measurements of the other divalent metals investigated in this work. An ellipsometric
thickness of 38il A, water contact angle of 6li10°, with a hysteresis of 8° was found
for the DMPA-terminated Mg+2 interfaces. These results point to a lipid adlayer similar
to that formed on the Ca+2-modified substrates. Cyclic voltammetry data further indicate
a similar surface being formed as that found with Zn”, with the Ru probe data for the
Mg+2-terminated interface yielding a peak splitting of 206 mV (Fig. 4-6b). The DMPA-
terminated interface exhibited a 157 mV peak splitting, less than that of the Mg+2-termi-
nated interface. Capacitance measurements, performed with the Fe probe, show a ratio of
the Mg+2-terminated surface-to-DMPA-terminated surface to be 1.77 (Fig. 4-6c). These
data point to the disruption of the supporting SAM structure upon exposure to of DMPA.
This result is not surprising, because a decrease in the organization of the support mono-
layer is also seen for other divalent metals ions (vide infra). This is attributed to the loss
of organization upon addition of the lipid adlayer in all of these cases (Ni, Zn, and Mg)
to the charge carried by the metal ions and the consequent inability to accommodate the
presence of two divalent ligands (ROPO3'2) while maintaining interfacial integrity. Ini-
tial exposure of the phosphate-terminated interface to Mg+2 produces coordination with
the bound phosphates and a consequent increase in the organization of the interface by
virtue of the metal ion coordinating with more than one phosphate moiety. Upon expo-
sure to DMPA, the Mg+2-terminated interface rearranges to accommodate the presence of
the DMPA ligands, resulting in a decreased interaction with the phosphate moieties that

are bound to the interface. The result is a decrease in the organization of the interface

80

with a consequent increased acces-
sibility of the solution phase elec-
trochemical probes to the electrode
surface. Capacitance measurements
performed with the Fe probe support
the ellipsometry and contact data col-
lected, indicating an interfacial thick-
ness intermediate between that of the
Cu+2-containing DMPA adlayer (vide
infra) and the Zr+4-containing DMPA

adlayer.

4-4.7: Copper 1

Cu+ was found to bind to the
phosphate-terminated SAM. XPS
data showed that the ratio of Cu+zAu
was 0.31 indicating a substantially
complete coverage of the interface
with CuJr (Fig. 4-7a). Following
exposure of the Cu+-tenninated
interface to DMPA, ellipsometry
measurements yielded a thickness of
l4i5 A and the water contact angle
for this interface was found to be

52:1:3° with a hysteresis of 6°. In

ctn'rent (ILA)

§§sssscssss

cumnt (HA)

 

      

   

a
A
4 x10 5 ' .
3.5 *
-Au3f7
3 ~ .
-Cu LMMI J
2.5 i - CU LMM3 _o1s
5 { KLL -o KLL ’
2 ~ -Cu2p3

 

 

975 965 955 945 935 925

 

   

-Cls

M

 

LI

 

 

o I I I 4 L L I I I I
IIoolooosooaoo'loqogosoo 420ng 200 100 0
We

9533
I'I'I

.10 l-

 

~0.l 0.0 0.1 0.2

 

 

0.3 0.4 0.5

muunmwwuw

81

Fig. 4-7: a) XPS spectrum of Cu+-modified thiol/
gold substrate. Inset shows Cu2P spectral region.
b) CV of Ru(NH3)6Cl3 for a DMPA monolayer on
Cu+-terminated interface (dashed line) and for the
Cu+-terminated interface with no adlayer (solid
line). c) CV of K3 F e(CN) 6 for a DMPA monolayer
on Cu+ -terminated interface (dashed line) and for
the Cu+ -terrninated interface with no adlayer (solid
line).

comparison, the water contact angle of the Cu+-terrninated interface was measured to be
57d:4°. The ellipsometry data point to fractional coverage of the interface with DMPA,
giving rise to an interface that is dominated by unreacted sites. The contact angle data
support this interpretation because there is not a measurable change in contact angle of
the interface upon exposure to DMPA. Cyclic voltammetry data for the DMPA-termi-
nated interface using the Ru and Fe electrochemical probes (Figs. 4-7b and 4-7c, respec-
tively) indicate that the DMPA adlayer disrupts the underlying SAM. For the Ru probe,
the Cu+-te11ninated interface yields a splitting of 155 mV (Fig. 4-7b) and the DMPA-
terrninated interface exhibits a 256 mV peak splitting. These data suggest that Ru probe
access to the electrode surface is hindered by the addition of the DMPA adlayer, but, as
can be seen in Fig. 4-7b, the Cu+-terminated interface produces two oxidative peaks. It is
found that exposure of the phosphate-terminated interface to Cu+ leaves a noticeable film
on the interface, suggesting the presence of excess physisorbed Cu+. Upon electrochemi-
cal cycling, the film desorbs, producing a second oxidative peak in addition to allowing
the electrochemical probe access to the electrode. Examination of the surface with the
Fe probe also indicates that the addition of DMPA disrupts interfacial organization, also
allowing access of the probe to the electrode surface. Comparison of the non-Faradaic
current of the DMPA terminated interface to that of the Cu+-terminated interface yields

a value of 2.2 l , suggesting a larger than expected DMPA-terminated adlayer thickness.
This finding is attributed to the presence of non-stoichiometric Cu+ at the interface.
Upon exposure of the interface to DMPA, the phospholipid will coordinate to the non-
stoichiometric Cu+, which desorbs into solution, leaving hydrophilic open regions in the
interface. Subsequently, DMPA in solution can interact with the hydrophilic interface,
allowing the formation of bilayers.116 It is these resulting bilayers that contribute to the

cyclic voltammetry capacitance measurements.

82

 

a 4-4.8: Copper ll

 

 

6 x10¢5 ‘ r . . fi fi ‘
Modification of the phos-

5 h I o o

phate-termmated Interface wrth
'Au4f +2
‘ ’ ‘ Cu produces a fimdamen—
975 965 953 945 935 925
53. -Au4d3 . tally different interaction upon
-Au4d5

DMPA exposure than is seen for

 

 

the Cu+-terminated interface.

H
T

Ad XPS measurements indicate

 

 

 

0 . +2 . .
IIooIooosooooo7ooooosoo4ooaoozooIoo o
3-1.1, (w) b very lIttle Cu being deposued
" on the phosphate-terminated
interface, with a concentration

ratio of Cu+21Au of 0.03 (Fig.

4-8a). This finding indicates

 

 

 

. - - - . - ' little 0 ortuni for adla er
-o.4 -o.3 -o.2 -o.1 0.0 0.1 02 pp ty y

 

 

 

 

4o _ 0 formation in a manner similar
30 f to that seen for other divalent
'5; 2° '
lo 1 metal ions. Without the deposi-
‘g o _ , ,. . +2
. tIon of Cu at the phosphate-
-10 _
-20 . terminated interface, the inter-
-3o - - . - - - . - - - . - 1 - , .
'0-1 0-0 0-1 0-2 03 0-4 0-5 face that Will be available for

interaction with DMPA will
Fig. 4 8: a) xps spectrum of Cu+2 modified thiol/gold
substrate. Inset shows Cu2P spectral region. 2b) CV of be (101111113th by the presence
Ru(NH3)6Cl3 for a DMPA monolayer on Cu: 2-termi- . .
nated interface (dashed line) and for the Cu+2 -termi- 0f phosphate, resulting 1“ the
nated interface with no adlayer (solid line). 2c) CV of . . . . 116
F e(CN) for a DMPA monolayer on Cu+2 -terminat- formation ofa hpld bilayer.
interface (dashed line) and for the Cu 2-terminated

inter-face with no adlayer (solid line). The ellipsometry and contact

83

angle confirm this assertion with the DMPA adlayer thickness being 633:5 A and a water
contact angle of 23i2°. Due to the significantly hydrophilic surface formed by this ad-
layer, hysteresis measurements could not be performed with any degree of accuracy. The
CV measurements are consistent with the presence of a lipid bilayer at the interface. For
the Ru probe, the Cu+2-terminated interface produces a splitting of 178 mV (F ig. 4-8b)
and the DMPA-terminated interface exhibits a 185 mV peak splitting. The absence of a
measurable difference in the peak splitting for the two interfaces suggests a bilayer with
a high defect density. Using the Fe probe, a capacitance ratio of 2.92 (Fig. 4-8c) was

obtained, indicating a comparatively thick adlayer forms upon DMPA deposition.

4-5: Comparison of Metal-ion Modified Interfaces

With the data reported for the different metal ions, now it is time to con-
sider how these results compare to one another. Table 4-1 shows the results for each
interface. From these data it emerges that there are three classes of interface that form.
Some metal ions (F e+3 , Zr”) form a relatively well organized monolayer structure upon
complexation with DMPA, some metal ions (Cu+2) yield a bilayer structure, and some
metal ions (Zn+2, Ni+2, Cu+, Ca”, Mg”) produce a partial adlayer characterized by
limited organization. These different characteristic interfaces are attributed to the result
of the ability of the metal ions examined to form ionic complexes with phosphates, and
this complex-forming ability is correlated for the most part with thecharge of the metal
ion. For some metal ions (Ni+2, Ca”, Mg”) the addition of the DMPA overlayer appears
to produce a disruption of the organization of the supporting SAM. This disruption is
attributed to a stoichiometric deficiency of the metal ion based on its formal charge and

consequent inability to form a regular structure when sandwiched between two phosphate

planes. Specifically, the issue of charge compensation likely plays a central role in deter

84

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

XPS Ellipso- Water Contact 81:33:; ;?;:i/;; FELZ?
Me :2: 2:: new one
ratio (A) ( deg.) sis ( de g.) for M” DMPA current
surface surface ratio
Cu+ 0.21 l4:t5 52:1:3 6 155 256 2.21
Cu+2 0.03 63i5 23i5 -- 178 185 2.92
Ni+2 0.08 27:1:3 50i3 2 209 171 1.28
Zn” 0.08 31i2 66i2 8 166 214 4.72
Ca+2 0.08 39i2 37i6 7 237 224 0.19
Mg+2 0.08 38i1 61:1:10 8 206 157 1.77
Fe” 0.46 26:1:2 1023c2 7 175 277 0.57
Zr+4 0.34 30i2 104i] 5 163 212 0.54
Table 4-1: Data for tie interfaces examinec. in this work. In the first column, metal ion-

surface coverage reported by the ratio of the XPS M“ to Au4f signal intensities; second
column, ellipsometric thickness in A; Water contact angle of DMPA-terminated inter-
faces, in degrees; water contact angle h steresis in degrees; Ru+3/+4 CV peak splitting for
metal ion-terminated in-terface; Ru+3/ CV peak splitting for DMPA-terminated inter-
faces; Fe+2/+3 CV ratio of non-Faradaic current for metal-terminated interface to DMPA—
terminated interface. All CV data has an associated 0.5 mV error.

mining the organization of the interfaces. The divalent metal ions are characterized by
comparatively low surface loading densities and, in order to achieve surface charge com-
pensation upon reaction with metal ions, there must also be residual H+ present. Upon
addition of the phospholipid, charge compensation becomes a more significant issue,
with Na+ and/or tn's(hydroxymethyl)-aminomethane cation from the buffer in the vesicle-
containing solution being the only additional cationic species present that are capable
of charge compensation. It is the structural accommodation of these additional cationic
species that must be responsible at some level for the changes in organization of the self-
assembled monolayer upon addition of DMPA, and this effect can be seen in certain of
the CV results.

F e+3 and Zr+4 are known to bind strongly to phosphate and have sufficiently high

formal charge to accommodate the presence of the DMPA functionality without disrup-

85

tion of the underlying phosphate layer. Copper is unique because Cu+2 is found to bind to
the phosphate-terminated SAM more weakly the Cu+. The limited extent to which Cu+2
bonds to phosphate gives rise to the formation of a DMPA bilayer rather than a mono-
layer, consistent with very weak Cu+2 — phosphate interactions. This finding implies an
interface that is closer to a hydroxyl-terminated interface in terms of the strength of inter-
action with the lipid adlayer.116 Cu+ gives rise to the formation of a more monolayer-like
lipid adlayer, albeit with extensive heterogeneity and likely regions of little or no DMPA
deposition. It is apparent fiom this data that a prerequisite for the formation of a reason-
ably well organized lipid adlayer is the presence of sufficient metal ion loading density at
the interface. XPS data shows comparatively high interface coverage for F e+3 and Zr”.
Metal ions that are deposited at the phosphate interface at lower loading density give rise
to a lower quality lipid adlayer, as can be seen from contact angle data and especially

electrochemical data.

4-6: Conclusions

In this study the interactions between selected metal ions, a phosphate-terminated
interface and the phospholipid DMPA have been investigated. The chemical interactions
investigated in this work are akin to those examined for metal bisphosphonate multilayer
structures with selected metal ions, with similar results.145 The goal of this work, how-
ever, is aimed at understanding the limits on the ability to form strongly bound supported
lipid mono- and bilayer structures. This data point to the use of metal ions with high
ionic charge and small ionic radius as being the most useful for forming lipid adlayer
structures. In general, divalent metal ions give rise to partial, spatially heterogeneous
structures that may or may not be useful in the creation of biomimetic interfaces, depend-

ing on the application.

86

' _'
W

V‘D 4‘,

 

 

 

Chapter 5

Conclusions

In this work the formation of air stable lipid monolayers and bilayers has been
explored. The goal in forming these various adlayers is to create a one step approach
utilizing vesicle fusion for the formation of lipid adlayer structures that can be removed
fi'om the aqueous environment in which they are formed. To act as a first step in creating
an interface that could be usefill in simulating a plasma membrane structure in a sensor
format, these interfaces need to be stable in air and be comparatively free of structural
defects. The imposition of a robust physical structure on a system that must perform as a
two-dimensional fluid is a challenging prospect.

The initial step in this work was to explore how to create an air-stable lipid adlay-
er. To succeed in this attempt, (Chapter 2) it is necessary to maximize the intermolecu-
lar interactions between the phospholipid lipid headgroup and the supporting substrate.
Two distinct but related interfaces were utilized for this purpose; 6-mercaptohexanol and
zirconium phosphate modified gold substrates. These modifications allowed for different
molecular interactions to be explored. The mercaptohexanol -OH terminal functional-
ity allowed for polar interactions with the lipid headgroup, allowing for the deposition
of a lipid bilayer structure. DMPC does not covalently bond to this surface, which then
allows for lateral diffusion of the lipids to produce a bilayer structure. The Zr+4 modi-
fied surface utilized the known strong interaction of Zr ion to phosphates/bisphosphates
(ZP chemistry). The phosphate group on DMPC coordinates to the bound Zr+4 prevent-
ing fiirther lateral diffusion or translocation leaving the lipids in place. This molecular
interaction was verified by 31P-NMR spectroscopy. What is seen is that a partial bilayer

initially formed would gradually decay over a period of 20 min. to a DMPC monolayer.

87

Monolayer formation is attributed to the irreversibility of the Zr+4llipid headgroup coor-
dination. This work showed that, by maximizing strong intermolecular or ionic forces, it
is possible to stabilize the lipid adlayer that is exposed to the substrate, making the corre-
sponding adlayer air stable. This result is desirable for the formation of an air-stable lipid
adlayer, and it is the use of ZP chemistry that allows this structural motif to be realized.
The initial step of this work lead to the ability to create air stable lipid bilayers
and monolayers. What also resulted from this work was the observation that DMPC, a
phosphocholine, would coordinate with the Zr+4 modified surface. To further understand
the interactions at the Zr+4/headgroup point of interaction, other lipid headgroups interac-
tions must be explored (Chapter 3). It was desirable to examine lipids with headgroup
functionalities that varied in both size and polarity. The operating hypothesis is that each
headgroup would have a distinctly different interaction with the Zr+4 on the surface based
on these lipid properties. It should be first noted that the bound Zr+4 will have excess
coordinated water surrounding it because no attempt was made to make the substrates an-
hydrous. Therefore, if the phospholipid headgroup is to bind to the Zr”, the water must
be displaced. This becomes problematic with headgroups that can hydrogen bond as they
will interact with the surrounding water opposed to the Zr leaving weaker coordination
and in turn poor adlayers. Steric and polar interactions will have a minor effect as well
but are secondary to the H-bonding effects. As a result it was observed that PA and PC
showed strong interaction with the surface, where PE, PG, and PS showed poor adlayer
formation, consistent with less energetically favorable interactions with the surface-
bound Zr. This is an expected result because PA is a small headgroup which eliminates
steric factors contributing to it’s interaction with Zr”, and while it is polar and can
hydrogen bond with the excess water around the Zr+4, the lack of steric hindrance allows

for the PA to displace the coordinated water and coordinate strongly to the bound Zr. PC,

88

though being a larger headgroup, does not have the chemical functionality required to hy-
drogen bond substantially because there are three methyl groups on the amine portion of
the choline headgroup. Phospholipid headgroups that do not undergo strong H-bonding
interactions allow for displacement of the water and good coordination with the bound
Zr”. PE, PG, and PS are all larger headgroups which will limit access of the headgroup
organophosphate moiety to the Zr+4, while polar interactions and to a greater extent the
ability of these headgroups to form H-bonds with any surrounding water, minimizes in-
teractions with the bound Zr, consequently producing poor adlayers. The comparison of
the data illustrated the wide range of interactions that are possible with the various lipid
headgroups and the Zr+4 surface. Given the headgroup-dependent nature of lipid inter-
actions with the surface-bound Zr+4, it is also likely that the interactions between lipid
headgroups and selected metal ions will differ significantly.

The goal of the work presented in Chapter 4 was to investigate the effect the metal
has on the lipid headgroups interaction and adlayer formation. As discussed above, previ-
ous studies have shown that metal ions other than Zr+4 can bind strongly with phosphates.
The work presented in Chapter 4 was designed to explore the interactions between a
phosphatidic acid and selected metal ions bound to a phosphate-terminated interface.
Phosphatidic acids (PA) were found to have the strongest interaction with Zr+4 from the
work presented in Chapter 3 and, therefore, PA was chosen for the metal ion-dependence
study. The goal was to minimize steric contributions to the metal-phosphate interactions.
The substrates were modified with biologically relevant metals in an effort to investigate
the effect of the metal on the metal/phosphate interaction with the lipids and the surface
bound metal. The interactions between DMPA and the several surface-bound metal ions
that the strength of lipid-interface interaction depended significantly on the identity of

the metal ion. The thickness and uniformity of the metal ion-bound DMPA adlayer was

89

 

 

 

found to depend on the formal charge of the metal ion used. The strongest metal ion-lipid
interactions were found for metal ions with comparatively high (+3, +4) ionic charge
and small ionic radius. F e+3 and Zr+4 formed the strongest complexes with DMPA and
thus the most uniform adlayers resulted. It is also found that the range of interactions of
DMPA with the bound metal ions is limited by the quality of the surface—bound metal ion
layer formed on the substrate. To form a surface-bound complex, the metal ions must
interact with a bound phosphate group on the substrate and the phosphate functional-

ity of DMPA. Owing to the requirement of macroscopic charge neutrality, divalent and
monovalent metal ions would be more likely to compete with other ions (e. g. H+, Na+)
and thus an interface with a comparatively high density of defects would result. If there
is not enough charge to balance the PA anions, then poor adlayer formation is observed.
This situation was found to obtain for Ca”, Mg”, Cu+, Cu”, Zn”, and Ni”. Without a
balance of charge that can be satisfied only by the metal ion and the phosphate moieties
in the layer where complexation occurs, it is difficult to form high quality adlayers. This
phenomenon is consistent with what is known about ZP layered materials and suggests
that much of the knowledge relating to ZP layered materials can be applied to the forma-
tion of lipid adlayers by these means.

One issue that was not explored through these studies and would be a logical next
step, is the effect of acyl chain length on adlayer formation. While it is most probable
that the dominant molecular interaction in the deposition of metal ion-bound lipid adlay-
ers is the interaction between the lipid headgroup and the metal ion, it is possible that the
structure and conformation of the acyl chains could play a role. The gel-to-fiuid phase
transition for phospholipids is thought to be dominated by acyl chain length and extent
of unsaturation, and the propensity of the acyl chains to organize in concert with the

headgroup-metal ion complexation process could, in principle, play a role in the forma-

90

tion of lipid adlayers. An experiment that could be performed to evaluate the importance
of lipid acyl chain length and structure would utilize selected phosphatidic acids and
complexation would proceed with Zr-modified substrates. Choice of the PA to use would
be important because the acyl chains impact several physical properties of the lipids that
could potentially influence adlayer formation. First, the issue of the lipids solubility in
water must be considered. Lipids exhibit limited solubility in water, and solubility will
depend on acyl chain length. A decrease in lipid solubility could result in less eflicient
vesicle fusion. Additional factors that would factor into lipid choice would be the sec-
ondary organizational role that the acyl chain chains play in mediating adlayer organiza-
tion. These considerations are reflected in the gel-to-fluid phase transition temperature
Tm of the lipids. The work discussed in this dissertation has been performed with lipids
chosen to have the same chain length (C14) and‘thus similar transition temperatures Tm,
near 24° C. While the use of a lipid with a higher Tm could interfere with adlayer deposi-
tion for reasons of solubility, the increased chain length of such a lipid may also serve to
introduce organization to the adlayers, once they form. The careful choice of lipid acyl
chain length and degree of unsaturation will likely allow these issues to be resolved in
future experiments.

The work presented in this disertation provides insight into the factors that al-
low for the formation of air-stable, surface-bound lipid bilayers and monolayers. From
this foundation, it will be possible to explore potential applications that utilize air stable
bilayers and/or metal ion coordination with phospholipid headgroups. The availability
of air-stable lipid adlayers may form the foundation for novel multilayer structures, and a
key issue will be the extent to which the surface-bound adlayer can mimic the fluid nature
of a plasma membrane. If the lipid adlayers can be organized and/or modified to become

biomimetic, an effort that is presently underway in the Blanchard group, the next step

91

 

will be to explore the incorporation of biologically active molecules (e. g. transmembrane

proteins) groups into the bilayer structure. Owing to the structural and compositional
complexity of plasma membranes, the incorporation of multiple bilayer constituents,
including cholesterol for example, will be required to produce a biomimetic interface.
While this goal will likely not be realized for several years, initial studies to evaluate the
stabilization of the bilayer with the addition of multiple components as well as the exami-
nation of any phase segregation which may take place, will be necessary.

It is also possible that further exploration of the interactions between interface-
bound metal ions and phospholipids will lead to the creation of interfaces that adsorb
lipids with different headgroups selectively. Since it has been established that various
lipid headgroups interact differently to the presence of Zr4+, it is possible that differ-
ent interface structural motifs or the use of different metal ion complexing agents could
produce higher lipid selectivity than has been seen to date. The effect of lipid acyl chain
length and saturation could also influence the efficiency of interfacial lipid adsorption, an
issue that bears investigation in the future (vide infia).

The work reported in this dissertation points to the benefits and possibilities that
could result from creating modified substrates as support structures for lipid bilayers or
monolayers. Through careful chemical control of the modified substrate, it is possible to
create air stable bilayers and lipid monolayers using vesicle fusion. These findings point

the way toward possible applications in chemical sensing and the design of biomimetic

interfaces.

92

(1)
(2)

(3)
(4)
(5)
(6)

(7)

(8)
(9)

(10)
(11)
(12)

(13)

(14)

(15)

(16)
(17)

Literature Cited

Edidin, M. Nature Reviews. Molecular Cell Biology 2003, 4, 414.

Ikeda, M.; Kihara, A.; Igarashi, Y. Biological & Pharmaceutical Bulletin
2006, 29, 1542.

Lee, A. Current Biology 2001, 11, R811.
Anderson, T. G.; McConnell, H. M. Biophysical Journal 2001, 81, 2774.
Anderson, T. G.; McConnell, H. M. Biophysical Journal 2002, 83, 2039.

Benvegnu, D. J .; McConnell, H. M. Journal of Physical Chemistry 1993,
97, 6686.

Brian, A. A.; McConnell, H. M. Proceedings of the National Academy of
Sciences 1984, 81, 6159.

Keller, S. L.; McConnell, H. M. Physical Review Letters 1999, 82, 1602.

Keller, S. L.; Radhakrishnan, A.; McConnell, H. M. Journal of Physical
Chemistry B 2000, 104, 7522.

Kornberg, R. D.; McConnell, H. M. Biochemistry 1971, 10, 1111.
Lee, K. Y. C.; Klingler, J. F.; McConnell, H. M. Nature 1994, 263, 655.

Radhakrishnan, A.; McConnell, H. M. Journal of the American Chemical
Society 1999, 121, 486.

Radhakrishnan, A.; McConnell, H. M. Biophysical Journal 1999, 77,
1507.

Radhakrishnan, A.; McConnell, H. M. Journal of Physical Chemistry B
2002, 106, 4755.

Subramaniam, S.; McConnell, H. M. Journal of Physical Chemistry 1987,
91, 1715.

Tamm, L. K.; McConnell, H. M. Biophysical Journal 1985, 47, 105.

Xu, Y.; Shen, Z. Z.; Wiper, D. W.; Wu, M. Z.; Morton, R. E.; Elson, P.;
Kennedy, A. W.; Belinson, J .; Markman, M.; Casey, G. Jama-Journal of

93

(18)

(19)

(20)

(21)
(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)
(30)
(31)
(32)
(33)
(34)
(35)

(36)

the American Medical Association 1998, 280, 719.

Discher, B. M.; Maloney, K. M.; Grainger, D. W.; Sousa, C. A.; Hall, S. B.
Biochemistry 1999, 38, 374.

Samsonov, A. V.; Mihalyov, 1.; Cohen, F. S. Biophysical Journal 2001, 81,
1486.

Yechiel, E.; Edidin, M. Journal of Cell Biology 1987, 105, 755.
Epand, R. M. Biochemical Society Transactions 1997, 25, 1073.

Tocanne, J. F.; Cezanne, L.; Lopez, A.; Piknova, B.; Schram, V.; Toumier,
J. F.; Welby, M. Chemistry and Physics of Lipids 1994, 73, 139.

McMullen, T. P. W.; McElhaney, R. N. Biochimica Et Biophysica Acta-
Biomembranes 1995, 1234, 90.

Pabst, G.; Amenitsch, H.; Kharakoz, D. P.; Laggner, P.; Rappolt, M. Physi-
cal Review E 2004, 70.

Stevens, B. C.; Ha, T. Journal of Chemical Physics 2004, 120, 3030.

Prenner, E. J .; Lewis, R.; Kondejewski, L. H.; Hodges, R. S.; McElhaney,
R. N. Biochimica Et Biophysica Acta-Biomembranes 1999, 1417, 211.

Lewis, R.; McElhaney, R. N. Biophysical Journal 2000, 79, 2043.

Koan, M. M.; Blanchard, G. J. Journal of Physical Chemistry B 2006, 110,
16584.

Perrin, F. Journal De Physique Et Le Radium 1936, 7, 1.

Debye, P. Polar Molecules; Chemical Catalog Co.: New York, 1929.
Zwanzig, R.; Harrison, A. K. Journal of Chemical Physics 1985, 83, 5861.
Edward, J. T. Journal of Chemical Education 1970, 47, 261.

Khan, T. K.; Chong, P. L. G. Biophysical Journal 2000, 78, 1390.
Lakowicz, J. R.; Knutson, J. R. Biochemistry 1980, 19, 905.

Marsh, D. Handbook of Lipid Bilayers; CRC Press: Boca Raton, F1, 1990.

McConnell, H. Biophysical Journal 2005, 88, L23.

94

(37)

(38)

(39)
(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)
(49)
(50)

(51)

(52)

(53)

Radhakrishnan, A.; McConnell, H. Proceedings of the National Academy
of Sciences of the United States of America 2005, 102, 12662.

McConnell, H. M.; Radhakrishnan, A. Biochimica Et Biophysica Acta-
Biomembranes 2003, 1610, 159.

Simons, K.; lkonen, E. Nature 1997, 387, 569.
Veatch, S. L.; Keller, S. L. Biophysical Journal 2003, 85, 3074.
Veatch, S. L.; Keller, S. L. Physical Review Letters 2005, 94.

Dietrich, C.; Bagatolli, L. A.; Volovyk, Z. N.; Thompson, N. L.; Levi, M.;
Jacobson, K.; Gratton, E. Biophysical Journal 2001, 80, 1417.

Veatch, S. L.; Polozov, I. V.; Gawrisch, K.; Keller, S. L. Biophysical Jour-
na12004, 86, 2910.

Tenenhouse, H. S.; Murer, H. Journal of the American Society of Nephrol-
ogy 2003, 14, 240.

Inoue, M.; Digman, M. A.; Cheng, M.; Breusegem, S. Y.; Halaihel, N.;
Sorribas, V.; Mantulin, W. W.; Gratton, E.; Barry, N. P.; Levi, M. Journal
of Biological Chemistry 2004, 279, 49160.

McConnell, H. M.; Watts, T. H.; Weis, R. M.; Brian, A. A. Biochimica Et
Biophysica Acta 1986, 864, 95.

Grakoui, A.; Bromley, S. K.; Sumen, C.; Davis, M. M.; Shaw, A. S.; Allen,
P. M.; Dustin, M. L. Science 1999, 285, 221.

Groves, J. T.; Boxer, S. G. Accounts of Chemical Research 2002, 35, 149.
Plant, A. L. Langmuir 1993, 9, 2764.
Sackmann, E. Science 1996, 271, 43.

Cornell, B. A.; BraachMaksvytis, V. L. B.; King, L. G.; Osman, P. D. J .;
Raguse, B.; Wieczorek, L.; Pace, R. J. Nature 1997, 387, 580.

Groves, J. T.; Ulman, N.; Boxer, S. G. Science 1997, 275, 651.
Jenkins, A. T. A.; Boden, N.; Bushby, R. J.; Evans, S. D.; Knowles, P. F.;

Miles, R. E.; Ogier, S. D.; Schonherr, H.; Vancso, G. J. Journal of
the American Chemical Society 1999, 121, 5274.

95

(54)

(55)

(56)

(57)
(58)

(59)

(60)

(61)

(62)

(63)

(64)

(65)

(66)

(67)

(68)

(69)

Cremer, P. S.; Yang, T. L. Journal of the American Chemical Society 1999,
121, 8130.

Ataka, K.; Giess, F.; Knoll, W.; Naumann, R.; Haber-Pohlmeier, S.; Rich-
ter, B.; Heberle, J. Journal of the American Chemical Society 2004,
126, 16199.

Giess, F.; Friedrich, M. G.; Heberle, J .; Naumann, R. L.; Knoll, W. Bio
physical Journal 2004, 87, 3213.

Dominska, M.; Krysinski, P.; Blanchard, G. J. Langmuir 2008, 24, 8785.
Stroumpoulis, D.; Pan'a, A.; Tirrell, M. Aiche Journal 2006, 52, 2931.

Schonherr, H.; Johnson, J. M.; Lenz, R; Frank, C. W.; Boxer, S. G. Lang-
muir 2004, 20, 11600.

Greenough, K. P.; Blanchard, G. J. Journal of Physical Chemistry B 2006,
110, 6351.

Lapinski, M. M.; Blanchard, G. J. Chemistry and Physics of Lipids 2007,
150, 12. .

Mohandas, N.; Wyatt, J.; Mel, S. F.; Rossi, M. E.; Shohet, S. B. Journal of
Biological Chemistry 1982, 257, 6537.

Pomorski, T.; Holthuis, J. C. M.; Hemnann, A.; van Meer, G. Journal of
Cell Science 2004, 117, 805.

Lin, W. C.; Blanchette, C. D.; Ratto, T. V.; Longo, M. L. Biophysical Jour-
na12006, 90, 228.

K01, M. A.; de Kruijff, B.; de Kroon, A. Seminars in Cell & Developmen-
tal Biology 2002, 13, 163.

Menon, A. K.; Watkins, W. E.; Hrafnsdottir, S. Current Biology 2000, 10,
241.

Raggers, R. J.; Pomorski, T.; Holthuis, J. C. M.; Kalin, N.; van Meer, G.
Traffic 2000, 1, 226.

Menon, A. Trends in Cell Biology 1995, 5, 355.

Devaux, P. F.; Lopez-Montero, 1.; Bryde, S. Chemistry and Physics of
Lipids 2006, 141, 119.

96

(70)

(71)

(72)

(73)

(74)

(75)

(76)

(77)

(78)

(79)

(80)

(81)

(82)

(83)

(84)

(85)

Anglin, T. C.; Liu, J.; Conboy, J. C. Biophysical Journal 2007, 92, L1.
Liu, J.; Conboy, J. C. Biophysical Journal 2005, 89, 2522.

Ross, E. E.; Rozanski, L. J.; Spratt, T.; Liu, S. C.; O’Brien, D. F.; Saave-
dra, S. S. Langmuir 2003, 19, 1752.

Tero, R.; Takizawa, M.; Li, Y. J .; Yamazaki, M.; Urisu, T. Langmuir 2004,
20, 7526.

Isono, T. T., H.; Ogino, T. Journal of Surface Science and Nanotechnology
2007, 5, 99.

Groves, J. T.; Ulman, N.; Cremer, P. S.; Boxer, S. G. Langmuir 1998, 14,
3347.

Deng, Y.; Wang, Y.; Holtz, B.; Li, J. Y.; Traaseth, N.; Veglia, G.; Stottrup,
B. J.; Elde, R.; Pei, D. Q.; Guo, A.; Zhu, X. Y. Journal of the American
Chemical Society 2008, 130, 6267.

Ross, E. E.; Bondurant, B.; Spratt, T.; Conboy, J. C.; O’Brien, D. F.;
Saavedra, S. S. Langmuir 2001, 17, 2305.

Cao, G.; Hong, H. G.; Mallouk, T. E. Accounts of Chemical Research
1992, 25, 420.

Katz, H. E.; Scheller, G.; Putvinski, T. M.; Schilling, M. L.; Wilson, W. L.;
Chidsey, C. E. D. Science 1991, 254, 1485.

Kohli, P.; Rini, M. C.; Major, J. S.; Blanchard, G. J. Journal of Materials
Chemistry 2001, 11, 2996. .

Lee, H.; Kepley, L. J .; Hong, H. G.; Akhter, S.; Mallouk, T. E. Joumal of
Physical Chemistry 1988, 92, 2597.

Putvinski, T. M.; Schilling, M. L.; Katz, H. E.; Chidsey, C. E. D.; Mujsce,
A. M.; Emerson, A. B. Langmuir 1990, 6, 1567.

Schilling, M. L.; Katz, H. E.; Stein, S. M.; Shane, S. E; Wilson, W. L.;
Buratto, S.; Ungashe, S. B.; Taylor, G. N.; Putvinski, T. M.; Chidsey, C. E.
D. Langmuir 1993, 9, 2156.

Venneulen, L. A.; Snover, J. L.; Sapochak, L. S.; Thompson, M. E. Journal
of the American Chemical Society 1993, 115, 11767.

Vermeulen, L. A.; Thompson, M. E. Nature 1992, 358, 656.

97

(86)
(87)

(88)

(89)

(90)

(91)

(92)

(93)

(94)

(95)

(96)

(97)

(98)

(99)

(100)

(101)

(102)

(103)

Venneulen, L. A.; Thompson, M. E. Chemistry of Materials 1994, 6, 77.
Kohli, P.; Blanchard, G. J. Langmuir 2000, 16, 695.

Krysinski, P.; BrzostowskaSmolska, M. Journal of Electroanalytical
Chemistry 1997, 424, 61.

Mazur, M.; Krysinski, P.; Blanchard, G. J. Langmuir 2005, 21, 8802.

Mayer, L. D.; Hope, M. J .; Cullis, P. R. Biochimica Et Biophysica Acta
1986, 858, 161.

Hunter, D. G.; Frisken, B. J. Biophysical Journal 1998, 74, 2996.
Moscho, A.; Orwar, O.; Chiu, D. T.; Modi, B. P.; Zare, R. N. Proceedings
of the National Academy of Sciences of the United States of America
1996, 93, 11443.

Patty, P. J.; Frisken, B. J. Biophysical Journal 2003, 85, 996.

Holden, M. A.; Jung, S.-Y.; Yang, T.; Castellana, E. T.; Cremer, P. S. Jour-
nal of the American Chemical Society 2004, 126, 6512.

Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. Journal of the
American Chemical Society 1987, 109, 3559.

Brechling, A.; Sundermann, M.; Kleineberg, U.; Heinzmann, U. Thin
Solid Films 2003, 433, 281.

Foss, B. J.; Sliwka, H. R.; Partali, V.; Naess, S. N.; Elgsaeter, A.; Melo, T.
B.; Naqvi, K. R. Chemistry and Physics of Lipids 2005, 134, 85.

Munro, J. C.; Frank, C. W. Langmuir 2004, 20, 3339.

Toca-Herrera, J. L.; Krustev, R.; Muller, H. J .; Mohwald, H. Journal of
Physical Chemistry B 2000, 104, 5486.

Burwell, D. A.; Valentine, K. G.; Thompson, M. E. Journal of Magnetic
Resonance 1992, 97, 498.

Colon, J. L.; Yang, C. Y.; Clearfield, A.; Martin, C. R. Journal of Physical
Chemistry 1988, 92, 5777.

Hanken, D. G.; Corn, R. M. Analytical Chemistry 1997, 69, 3665.

Home, J. C.; Blanchard, G. J. Journal of the American Chemical Society

98

(104)

(105)

(106)

(107)
(108)
(109)
(110)

(111)

(112)

(113)

(114)

(115)

(116)
(117)

(118)

(119)

(120)

(121)

1998, 120, 6336.

Conboy, J. C.; Liu, S.; O’Brien, D. F.; Saavedra, S. S. Biomacromolecules
2003, 4, 841.

Liu, J .; Conboy, J. C. Langmuir 2005, 21, 9091.

McConnell, H. M.; Watts, T. H.; Weis, R. M.; Brian, A. A. Biochim. Bio-
phys. Acta 1986, 864, 95.

Plant, A. L. Langmuir 1999, 15, 5128.

Starr, T. E.; Thompson, N. L. Langmuir 2000, 16, 10301.

Burwell, D. A.; Thompson, M. E. Chemistry of Materials 1991, 3, 730.
Burwell, D. A.; Thompson, M. E. Acs Symposium Series 1992, 499, 166.
Frey, B. L.; Hanken, D. G.; Corn, R. M. Langmuir 1993, 9, 1815.

Hong, H. G.; Sackett, D. D.; Mallouk, T. E. Chemistry of Materials 1991,
3, 521. '

Katz, H. E.; Schilling, M. L. Chemistry of Materials 1993, 5, 1162.

Ungashe, S. B.; Wilson, W. L.; Katz, H. E.; Scheller, G. R.; Putvinski, T.
M. Journal of the American Chemical Society 1992, 114, 8717.

Newman, J. D. S.; Roberts, J. M.; Blanchard, G. J. Analytica Chimica Acta
2007, 602, 101.

Oberts, B. P., Blanchard, G. J. Langmuir 2009, 25, 2962.
Edidin, M. Nature Reviews Molecular Cell Biology 2003, 4, 414.

Chidsey, C. E. D.; Porter, M. D.; Allara, D. L. Journal of the Electrochemi-
cal Society 1986, 133, C130.

Dubois, L. H.; Zegarski, B. R.; Nuzzo, R. G. Journal of Vacuum Science
& Technology a-Vacuum Surfaces and Films 1987, 5, 634.

Nuzzo, R. G.; Allara, D. L. Journal of the American Chemical Society
1983, 105, 4481.

Nuzzo, R. G.; Fusco, F. A.; Allara, D. L. Journal of the American Chemi-
cal Society 1987, 109, 2358.

99

(122)

(123)

(124)

(125)

(126)
(127)

(128)

(129)

(130)

(131)

(132)

(133)

(134)

(135)

(136)

(137)

(138)

Nuzzo, R. G.; Zegarski, B. R.; Dubois, L. H. Journal of the American
Chemical Society 1987, 109, 733.

Ulman, A. Chemical Reviews 1996, 96, 1533.

Whitesides, G. M.; Troughton, E. B.; Bain, C.; Holmesfarley, S. R.; Was-
serman, S. R.; Strong, L. H. Journal of the Electrochemical Society 1987,
134, C110.

Xia, Y. N.; Whitesides, G. M. Annual Review of Materials Science 1998,
28, 153.

Fang, C. P.; Drelich, J. Langmuir 2004, 20, 6679.
Major, J. S.; Blanchard, G. J. Langmuir 2001, 17, 1163.

Tulock, J. J .; Blanchard, G. J. Journal of Physical Chemistry B 2002, 106,
3568.

Malinauskas, A.; Ruzgas, T.; Gorton, L. Bioelectrochemistry and Bioener-
getics 1999, 49, 21.

Pet’kov, V. 1.; Asabina, E. A.; Markin, A. V.; Smirnova, N. N.; Kitaev, D.
B. Journal of Thermal Analysis and Calorimetry 2005, 80, 695.

Bakiamoh, S. B.; Blanchard, G. J. Langmuir 1999, 15, 6379.

Bakiamoh, S. B.; Blanchard, G. J. Langmuir 2001, 17, 3438.

Oberts, B. P., Blanchard, G. J. Langmuir 2009, XX, XXXX.

Akhter, S.; Lee, H.; Hong, H. G.; Mallouk, T. E.; White, J. M. Journal of
‘lVggtglum Science & Technology a-Vacuum Surfaces and Films 1989, 7,

Bujoli, B.; Courilleau, A.; Palvadeau, P.; Rouxel, J. European Journal of
Solid State and Inorganic Chemistry 1992, 29, 171.

Bujoli, B.; Pena, O.; Palvadeau, P.; Lebideau, J .; Payen, C.; Rouxel, J.
Chemistry of Materials 1993, 5, 583.

Cao, G.; Lee, H.; Lynch, V. M.; Mallouk, T. E. Inorganic Chemistry 1988,
27, 2781.

Cao, G.; Lynch, V. M.; Swinnea, J. S.; Mallouk, T. E. Inorganic Chemistry
1990, 29, 2112.

100

(139)
' (140)

(141)
(142)
(143)
(144)

(145)

Cao, G.; Mallouk, T. B. Journal of Solid State Chemistry 1991, 94, 59.
Cao, G.; Mallouk, T. E. Inorganic Chemistry 1991, 30, 1434.

Cao, G.; Rabenberg, L. K.; Nunn, C. M.; Mallouk, T. E. Chemistry of Ma-
terials 1991, 3, 149.

Cao, H.; Tokutake, N.; Regen, S. L. Journal of the American Chemical
Society 2003, 125, 16182.

Feldheim, D. L.; Mallouk, T. E. Chemical Communications 1996,
2591.

Kim, Y. H.; Rahman, M. M.; Zhang, Z. L.; Misawa, N.; Tero, R.; Urisu, T.
Chemical Physics Letters 2006, 420, 569.

Yang, H. C.; Aoki, K.; Hong, H. G.; Sackett, D. D.; Arendt, M. F.; Yau,

S. L.; Bell, C. M.; Mallouk, T. E. Journal of the American Chemi-
cal Society 1993, 115, 11855.

101