.3.

.l .1 Amman-as...
.. .5
. tan-rm

.y {VIVI 1.5

$.-

11. .nr... :
29.5., a . Liv»...
.p ...
,1
.s .1 3.1.3:;
.1. impact???
#32

 

 

 

 

 

 

 

. _(x¢QI->3.Amv . ‘-.M§-x=..n , c nay;
. I: x. . A _or 1.). I 3
A». . I A 3 rt. 57 .w .
. Pk Aggiififiafiuv.c...az.i.mm

. gang; 7 . . V

a... .

..

‘ flaw, : K“

:1 . . (wa391 I
my? éfimfirfiku

 

"kg"? _
hum

.~-
"()1

"3

This is to certify that the
dissertation entitled

Dynamic Solvation of Zn(ll)-Cytochrome c in the Native Fold
and the Unfolding Transition State

presented by

Sanela Lampa-Pastirk

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Chemistry

 

 

W ffiQ/e

 

Major Professor’s Signature

08/1 9/2005

 

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State

 

-.-.-.-.-.-.-.-.-

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

 

DATE DUE

DATE DUE

DATE DUE

 

5m 12.932909

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:IClRC/Date0ue.indd-p.1

 

 

Dynamic Solvation of Zn(II)-Cytochrome c in the Native Fold

and the Unfolding Transition State

by

Sanela Lampa-Pastirk

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2005

ABSTRACT

Dynamic Solvation of Zn(II)-Cytochrome c in the Native State and at the
Unfolding Transition State

by
Sanela Lampa-Pastirk

This dissertation introduces a new approach for the study of protein motions at the
native structure and near the transition state for protein unfolding or misfolding reactions.
An embedded intrinsic electronic chromophore is used as a solvatochromic probe of the
motions on the energy landscape of the surrounding protein and associated solvent and as
a trigger for the unfolding reaction. This approach is demonstrated with
Zn(II)-cytochrome c, a small globular protein containing a porphyrin macrocycle as an

intrinsic chromophore.

Picosecond time-resolved measurements of the dynamic Stokesshifi of the
fluorescence from Zn(II)-cytochrome c with Q-band excitation show that the fastest
protein motions occur in the hydrophobic core; much slower motions occur in the
solvent-contact layer. The two classes of motion are apparently not strongly coupled. The
nonpolar response that accompanies the excited-state photodissociation of the

Zn(II) ion’s protein derived axial ligands occurs on a different timescale.

With Soret-band excitation, the fluorescence spectrum from Zn(II)-cytochrome c
exhibits a prompt red shift that reports an entry of water molecules into the hydrophobic
core. The initial red shift is followed by an unusual blue shift that reports an extrusion of
water from the hydrophobic core. These results suggest a new approach for optically
triggering protein unfolding reactions from the native state with complete control over the
surrounding solution conditions and temperature. We call this method the optical G jump
because it provides an effectively instantaneous transition from the equilibrated native
state to an excited Gibbs-free energy level. The protein system is then allowed to
propagate on the energy landscape using whatever trajectories are available under the
solution conditions and temperature. This method opens exciting new perspectives for

studying disease related protein misfolding reactions.

Igoru

iv

Acknowledgments

I thank Professor Jane Vanderkooi and Wayne Wright from the School of
Medicine, University of Pennsylvania for the generous gift of the Zn(II)—substituted
cytochrome c sample that was used in this work. I am also grateful for the indispensable
assistance of Professor Gary Blanchard during the picosecond TCSPC fluorescence
measurements. I would like to thank past and present members of the research group,

Elizabeth Carson, Katherine Shelly and Kevin Dillman.

I would particularly like to thank my advisor Warren Beck, for help, guidance and
invaluable discuSsions through my graduate studies and supporting me both

professionally and personally.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. x
LIST OF FIGURES ........................................................................................................... xi
Introduction ......................................................................................................................... l

1. Polypeptide and Solvent Dynamics in Native and Misfolded States of Proteins 4

1.1.

1.2.

1.3.

1.4.

1.5.

Protein Folding ................................................................................................ 7
1.1.1. Energetics .................................................................................................. 7
1.1.2. Mechanisms ............................................................................................. 11
1.1.3. Energy Landscape and Funnels ............................................................... 13
1.1.4. Thermodynamic and kinetic control ........................................................ 16
1.1.5. Experimental and Theoretical Approach ................................................. 17
1.1.6. Role of Intrinsic and Surface Water ........................................................ 18

Structural Fluctuations and Time-Correlation Function ................................ 20
1.2.1. Molecular Liquids .................................................................................... 24
1.2.2. Polypeptide and Folded Proteins ............................................................. 25

Structure, Dynamics and Folding Reactions of Cytochrome c ...................... 27
1.3.1. Structure ................................................................................................... 27
1.3.2. Folding Intermediates .............................................................................. 28
1.3.3. Role of Heme Ligation in the Folding Reaction ...................................... 29

Proposed work ............................................................................................... 30

References ...................................................................................................... 33

vi

2.

Experimental Procedure ........................................................................................ 41

2. 1 . Introduction .................................................................................................... 41

2.2. Sample Preparation ........................................................................................ 42
2.3. Modeling of Continuous Wave Absorption and Fluorescence Emission

Spectra of ZnCytc .......................................................................................... 44

2.4. The TCSPC Method ...................................................................................... 47

2.4.1. The TCSPC Apparatus ............................................................................ 48

2.4.2. Data Analysis ........................................................................................... 51

2.5. References ...................................................................................................... 55

Excited-State Axial-Ligand Photodissociation and Nonpolar Protein-Matrix

Reorganization in Zn(II)-Substituted Cytochrome c ............................................ 56
3.1. Introduction .................................................................................................... 56
3.2. Experimental .................................................................................................. 58

3.2.1. Zn(II)-Substituted Cytochrome c ............................................................. 58

3.2.2. Continuous-Wave Absorption and Fluorescence Instrumentation .......... 59

3.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy ......................... 59

3.3. Results ............................................................................................................ 60

3.3.1. Continuous-Wave Absorption and Fluorescence Emission Spectra ....... 60

3.3.2. Time-Resolved Fluorescence Spectra ...................................................... 60

3.4. Discussion ...................................................................................................... 66

3.4.1. Interpretation of Vibronic Structure in Zn(II)-Porphyrin Complexes ..... 67
3.4.2. Origin of the Biexponential Axial-Ligand Photodissociation Response in

ZnCytc ..................................................................................................... 68

3.4.3. Protein-Matrix and Solvent Reorganization Dynamics ........................... 70

3.5. References ...................................................................................................... 73

vii

4.

Polar Solvation Dynamics in Zn(II)-Substituted Cytochrome c: Diffusive
Sampling of the Energy Landscape in the Hydrophobic Core and Solvent-Contact

Layer ..................................................................................................................... 77
4.1. Introduction .................................................................................................... 77
4.2. Experimental .................................................................................................. 80

4.2.1. Zn(II)-Substituted Cytochrome c ............................................................. 80

4.2.2. Continuous Wave Absorption and Fluorescence Instrumentation .......... 80

4.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy ......................... 81

4.3. Results ............................................................................................................ 82

4.3.1. Continuous-Wave Absorption and Fluorescence Emission Spectra ....... 82

4.3.2. Dynamic Stokes Shift .............................................................................. 84

4.4. Discussion ....................... p ............................................................................... 87

4.4.1. Solvent Response Functions in Proteins .................................................. 87

4.4.2. Dynamic Solvation in ZnCytc. ................................................................ 88

4.4.3. Coupling of Polar and Nonpolar Solvation in ZnCytc ............................ 91
4.4.4. Estimation of the Magnitude of the Internal Solvation Response in

ZnCytc ..................................................................................................... 92

4.5. References ...................................................................................................... 96

Light-Induced Protein Unfolding of Zn(II)—substituted Cytochrome c: Dynamics
of Entry and Exit of Water Molecules to and from the Hydrophobic Core During

Unfolding and Refolding .................................................................................... 100
5.1. Introduction .................................................................................................. 100
5.2. Experimental ................................................................................................ 102

5.2.1. Zn(II)-Substituted Cytochrome c .......................................................... 102
5.2.2. Continuous-Wave Absorption and Fluorescence Instrumentation ........ 103
5.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy ....................... 103
5.3. Results .......................................................................................................... 104

viii

5.3.1. Time-Resolved Fluorescence Spectra .................................................... 104
5.3.2. Dynamic Stokes Shift ............................................................................ 107
5.4. Discussion .................................................................................................... 111

5.4.1. Nature of Dynamical Stokes Shifi in ZnCytc with Soret-Band
excitation ................................................................................................ 1 l 1

5.4.2. The Optical G Jump in ZnCytc: Dynamics of Unfolding and Refolding

following Soret-Band excitation ............................................................ 116 '
5.5. References .................................................................................................... 122
Future Directions ................................................................................................ 125
6.1. References .................................................................................................... 128

LIST OF TABLES

Table 1. The time evolution of Q01 / Q00 for ZnCytc in two solvent systems at 22 °C. .. 64

Table 2. Fit parameters for the time evolution of v00 for Zn(II)-substituted cytochrome c
in three solvents at 295 K ...................................................................................... 86

LIST OF FIGURESQ

Figure

Figure

Figure

1.1 Top panel. Hypothetical funnel-shaped energy landscape. During the folding
process protein moves down the funnel from the unfolded structure to the folded
at the bottom of the funnel which is accompanied by a decrease in energy E and
entropy, and with increase in native contacts Q. On its way to the native state
protein goes through the region of molten globule MG and overcomes the
transition state TS barrier before it finally reaches the native state. After Onuchic
and Wolynes34 Bottom panel. Enlarged view of near vicinity of a native state
position in the funnel depicting the thermodynamic (a) and kinetic (b) paths.38 . 15

1.2 Solvation dynamics and the dynamic Stokes shift. Following electronic
excitation of a solute chromophore, the solvent molecules still reside in the
configuration that was reached at equilibrium with the ground state solute’s dipole
moment. Over time, the solute molecules reorganize in response to the change in
the solute’s dipole moment; the fluorescence spectrum shifts to the red as the
system approaches equilibrium. After Stratt and Maroncelli.72 ............................ 22

1.3 The structure of horse-heart90 ferricytochrome c, rendered using PyMOL.9' . 28

Figure 2.1. Continuous-wave absorption spectrum from iron-free cytochrome c, in water

Figure

Figure

at room temperature. The spectrum shows the characteristic four bands in the Q
region. ................................................................................................................... 42

2.2 Continous-wave absorption spectrum from ZnCytc, in water at room
temperature. The spectrum shows characteristic two bands in the Q region. ....... 43

2.3. Continous-wave absorption spectrum (dots) from ZnCytc, in water at room
temperature. The spectrum is plotted as relative dipole strength and normalized to
unit area. Superimposed with the absorption spectrum is a model (line) described
with equation 2.1 ................................................................................................... 45

Figure 2.4. Continous-wave fluorescence spectrum (dots) from ZnCytc, in water at room

Figure

temperature. The fluorescence spectrum was obtained with excitation at 585 nm.
The spectrum is plotted as relative dipole strength and normalized to unit area.
Superimposed with the fluorescence spectrum is a model (line) described with
equation 2.2. .......................................................................................................... 46

2.5 Schematic for the time-correlated single-photon-counting apparatus: CFD,
constant fraction discriminator; TAC, time-to-amplitude converter; MCP PMT,
multichannel plate photomultiplier; S, sample. .................................................... 50

 

‘8 Images in this dissertation are presented in color.

xi

Figure 2.6 Typical single-wavelength transient and instrument response function obtained

Figure

in TCSPC experiment. .......................................................................................... 51

2.7 Typical time-wavelength-intensity surface constructed from the single
wavelength transients collected in TCSPC experiment. ....................................... 52

Figure 2.8 Typical time-resolved fluorescence spectrum obtained from time-wavelength-

intensity surface (data points). Superimposed with the time-resolved fluorescence
spectrum is a model (line) described with equation 2.2 ........................................ 54

Figure 3.1. Structure of the heme and axial ligands from horse-heart ferricytochrome c, as

Figure

Figure

rendered using PyMOL22 from the pdb file lHRC.21 ........................................... 58

3.2. Continuous—wave absorption (right) and fluorescence (lefi) spectra from
ZnCytc at 22°C. The spectra are plotted as dipole strengths, A(v)/v and F(v)/v3,
respectively, and are normalized to unit area. The spectra are superimposed with
fitted curves employing eqs 1 and 2 as models. The fluorescence spectrum was
obtained with excitation tuned to the 0—0 transition energy, where the absorption
and fluorescence spectra cross. ............................................................................. 61

3.3. Time-resolved fluorescence spectra from ZnCytc in water at 22°C (data
points) and fitted spectra (smooth curves, see eq 2) at three delays. .................... 62

Figure 3.4 Time evolution of the dipole strengths for the 0—1 (Q01) and 0-0 (Q00) Vibronic

transitions in time-resolved fluorescence spectra from ZnCytc in water at 22°C.
Top: As normalized by the total dipole strength SjQOj. Bottom: As the ratio
Q01 /Q00 . The parameters for the fit to the ratio are listed in Table 1. ................ 65

Figure 3.5. Time evolution of the ratio of the dipole strengths for the 0-1 (Q01) and 0-0

Figure

(Q00) Vibronic transitions in time-resolved fluorescence spectra from ZnCytc in
50% glycerol at 22°C. The fit parameters are listed in Table l. ........................... 66

4.1 Structure of the horse-heart ferricytochrome c, as rendered using PyMOLl7
from pdb file lHRC.18 Top: Ribbon representation showing the polypeptide’s
secondary and folded structure, with heme and axial ligands to the Fe(III) ion,
Met 80 and His 18, rendered as tube structures. Bottom: Solvent-contact surface,
with the heme as a space-filling (CPK) representation. ........................................ 79

Figure 4.2. Continuous-wave absorption and fluorescence emission spectra from ZnCytc

in water at room temperature. The spectra are plotted as relative dipole strengths
and are normalized to unit area. The fluorescence spectrum was obtained with
excitation tuned to the origin (O-O transition energy), where the two spectra cross.
The two spectra are displaced by 2)»: 290 cm", where It represents the
reorganization energy ............................................................................................ 82

Figure 4.3. Continuous-wave absorption and fluorescence emission spectra from ZnCytc

in 50% (v/v) glycerol at room temperature. The spectra are plotted as relative

xii

dipole strengths and are normalized to unit area. The fluorescence spectrum was
obtained with excitation tuned to the origin (0-0 transition energy), where the two
spectra cross. The two spectra are displaced by 2}» = 290 cm", where X represents
the reorganization energy ...................................................................................... 83

Figure 4.4. The time evolution v00 (see eq 1) for ZnCytc in water. The fit parameters are

Figure

Figure

listed in Table 2 ..................................................................................................... 84

4.5. Time evolution v00 (see eq 1) for ZnCytc in 50% (v/v) glycerol. The fit
parameters are listed in Table 2 ............................................................................ 85

4.6. Model for the early time evolution v00 for ZnCytc in water. The ordinate

corresponds to the remaining Stokes shift at the indicated delay following
excitation. The thick solid line shows the fitted trend obtained from Figures 4 and
5 from fits to the time-resolved fluorescence spectra observed over the lOO-ps to
12-ns time window. The thin solid line extrapolates the fitted trend to zero time.
The dashed line shows a hypothetical Gaussian trend that connects the
reorganization energy (see Figures 4.2 and 4.3) with the observed shift of v00 at

the IOO-ps delay. ................................................................................................... 95

Figure 5.1. Time-resolved fluorescence spectra from ZnCytc in aqueous solution at 22 °C

at six delays. The dashed line depicts the initial red shift of the Q00 Vibronic band

of the fluorescence spectrum followed by the blue shift at time-delays longer than
250 ps. ................................................................................................................. 106

Figure 5.2. Early time response of the frequency of the center of the Q00 Vibronic line of

the fluorescence spectrum from ZnCytc in aqueous solution at 22 °C. The
excitation laser was tuned to 420 nm, the peak of the Soret absorption band. 107

Figure 5.3. Time evolution of the frequency of the Q00 intensity maximum of the ZnCytc

Figure

fluorescence spectrum in aqueous solution at 22 °C. The top curve was obtained
with the excitation laser tuned to 420 nm, the peak of the Soret absorption band.
The early time portion of this trace is shown on Figure 2. The bottom curve is that
obtained with Q-band excitation (chapter 4). Each data trace is superimposed with

an exponential fit of the form y = C + 2i Ale-W" . ........................................... 109

5.4. Comparison of the time evolution of the frequency of the Q00 intensity
maximum of the fluorescence spectrum from ZnCytc in three solvent systems at
22 °C. The excitation laser was tuned to 420 nm, the peak of the Soret absorption
band. The data is superimposed with exponential fit of the form

y=C+ziAie-’/ri. ........................................................................................... 110

Figure 5.5. Comparison of the time evolution of the linewidth of the Q00 Vibronic line of

the fluorescence spectrum from ZnCytc in three solvent systems at 22 °C. The

xiii

excitation laser was tuned to 420 nm, the peak of the Soret absorption band. The
data is superimposed with exponential fit of the form y = C + 2i Ate-Wi . ..... 112

Figure 5.6. Energy level diagram for Zn(II) porphyrins, showing an absorption (A)
transition in the Soret band to the Sn manifold of states followed by internal
conversion (IC) to the 81 state and recovery of the ground state So by emission of
a fluorescence photon (F) .................................................................................... 113

Figure 5.7. Sequence of events following the excitation of the electronic probe in the
hydrophobic core of a target protein in the optical G jump experiment: 1. the
initial vibrational relaxation of the probe and associated vibrational energy
transfer to the protein, 2. structural evolution of the protein at an excited
vibrational level, 3. transfer of enthalpy from the protein to the surrounding
solvent, and the associated vibrational equilibration of the protein at a displaced
position on the energy landscape, 4. initiation of a refolding (4a) or misfolding
(4b) trajectory, 5. hydrophobic collapse of the unfolded protein and ejection of
water, 6. slow refolding into the final, native state. ............................................ 118

xiv

Introduction

Knowledge of the protein folding process has developed remarkably in recent
years from an understanding that proteins fold in a linear multi-state kinetic pathway to
the idea of a multidimensional energy landscape that involves many possible folding
trajectories. An important concept in studying the mechanism of protein folding was that
protein structures can exist in a large number of conformations and actively switch
between them. Some of these conformations are crucial for the function of the protein;
however, some appear to be misfolded and toxic species appear that can be responsible
for a significant number of diseases. The work presented in this dissertation is mainly
motivated by the wish to gain an understanding of the molecular mechanism that
produces misfolded proteins from their native folded structures. The difficulty in
determining the role played by misfolded proteins in the molecular mechanism of the
diseases is that no methods have been available that permit synchronous triggering of the
unfolding reactions under physiological conditions that do not involve the manipulation
of the bulk solution composition or temperature. This work will propose a new

experimental approach to address this problem.

The dissertation is organized as follows:

The role of protein folding and misfolding in disease is discussed in chapter 1. We
review the main concepts of protein folding, the energetics, the models and the methods

of investigation of folding reactions.

Chapter 2 describes preparation of the ZnCytc and its spectroscopic
characteristics. Information about the time-correlated single-photon counting method is
included with the detailed description of the experimental set up used in this work and the

method of data analysis.

In Chapter 3 we examine nonpolar solvation dynamics of the native
Zn(II)-substituted cytochrome c with Q—band excitation. We show that Zn(II) porphyrin
exhibits excited-state photodissociation of the protein-derived axial ligands. We discuss
the protein-matrix response to the axial-ligand photodissociation reaction in ZnCytc in
terms of nonpolar solvation dynamics, a reorganization of the protein in response to a

change in size of an imbedded structure.

The main focus of Chapter 4 is polar solvation dynamics of the native
Zn(II)-substituted cytochrome c with Q-band excitation of Zn(II) porphyrin in
cytochrome c. The solvent-response function obtained from the dynamic Stokes shift
measurements of the time-resolved fluorescence spectrum exhibits biexponential decay
with two distinct time constants. We discuss these two components in terms of the
motions of the hydrophobic core and the motions in the solvent-contact layer. We show
that the polar solvation dynamics is not strongly coupled to the nonpolar dynamics

discussed in Chapter 3.

In chapter 5 we present the solvation dynamics of Zn(II)-substituted cytochrome c
with Soret-band excitation. We introduce the optical G jump as a new method for optical

triggering of the protein unfolding reaction. When irradiated with light in a blue region of

the spectrum, the Zn(II) porphyrin as an internal chromophore can serve as a trigger for
unfolding the protein and as a sensor for the subsequent motions of the protein and water
molecules. We show the direct experimental observation of the fast dynamics of entrance

and exit of water molecules Item the hydrophobic core.

Chapter 6 concludes the dissertation by proposing future investigations of the
protein unfolding and misfolding reactions based on the methods developed in the work

presented here.

1. Polypeptide and Solvent Dynamics in Native and
Misfolded States of Proteins

For a long time the main motivation for understanding the protein folding
mechanism was the remarkable efficiency and accuracy with which these large structures
fold into highly specific and biologically active structures that are essential for the
fimction of living organisms. However, the motivation for studying protein folding
changed in the 19903 with the discovery that proteins on their own can be disease-causing

species.

Today, it is well established that the inability of proteins to fold properly or to
maintain the folded state is a direct cause for a number of diseases. Markedly associated
with protein misfolding are the transmissible spongiform encephalopathies that include
Creutzfeldt—Jakob disease in humans, scrapie in sheep, and bovine spongiform
encephalopathy (“mad cow disease”). Also involving misfolding are the amyloidoses, the
most famous of which are Alzheimer’s disease and Parkinson’s disease."3 All of these
diseases are thought to be associated with the accumulation of amyloid fibrils, but the

critical step seems to be the decay of the native folded form of the protein in question.4

Native proteins are soluble molecules with distinct structures that perform a
variety of functions in living organisms through their selective interaction with other
molecules in the surrounding .solvent. The protection of the hydrophobic core of the

protein from the surrounding solvent is critical to the stability of the native fold.5 Unlike

the native forms, the amyloid fibrils derived from them are insoluble generic forms of
protein.6 The first step that leads to the formation of amyloid fibrils is the failure of the
protein to assume the proper native fold. Misfolded or partially unfolded proteins form
small soluble aggregates, which associate into protofibrils and then finally into insoluble
mature fibrils. The aggregation always occurs from a misfolded or partially unfolded
state of the protein."7 The fully formed fibrils exhibit nearly identical structures
regardless of the protein of origin; they all have a characteristic cross-beta-sheet
configurationL7 despite having essentially identical structures. The rates at which the
misfolded proteins aggregate depend strongly on the physicochemical properties of the
proteins, such as the spatial organization of charges, a preference for certain secondary

structural features, and hydrophobicity.8

While it is known that amyloidic polyneuropathy arises directly from a single-
point mutation of the protein transthyretin that results in the formation of amyloid
structures,9'” the ability to form amyloid fibrils is apparently a general property of
polypeptides and is not necessarily dependent on specific amino-acid sequences or
native-fold topologies.‘2 This idea is supported by the finding that proteins not related to
the amyloid diseases can unfold and aggregate in vitro and create amyloid fibers of
almost identical structure to those found in amyloidoses.l3 Dobson and co-workers
showed that even myoglobin and cytochrome c can form amyloid fibrils under the
conditions that favor unfolding of the native structures.W5 Further, the SH3 domain from
bovine phosphatidyl-inositol-3’-kinase and the N-terminal domain of E. coli Hpr

protein can form fibers in vitro.'6’17 It is notable that the native structure of these globular

proteins mostly contains alpha helices as the secondary structural element; in order to
form amyloid fibrils, some of these alpha helices have to unfold and refold to form

b-sheet-rich structures. ”"5

The role of misfolded or aggregated protein structures in the progress of disease is
the subject of current vigorous debate. The key questions under consideration are which
of the various forms of a misfolded or aggregated protein is the species that initiates the

disease, and how this species takes part in the molecular mechanism for the disease.””"”'9

The current trend in this discussion seems to indicate that small aggregate
structures formed from the misfolded proteins are the disease-causing species. It was
found that structurally disorganized pre-fibrilar aggregates are highly cytotoxic while
structurally well-defined mature fibrilar forms of the same protein are generally
harmless.”23 The prefibrilar forms are thought to be toxic because they expose
hydrophobic regions of the protein to the surrounding solvent.21 An important example of
this principle is the protein alpha-synuclein, mutant forms of which are associated with
Parkinson’s disease. This protein can be present in several different aggregation states:
the unfolded monomer, an oligomeric form that is rich in beta-sheet-containing
structures, a protofibrilar form, and the mature amyloid fibril. The protofibrilar form is
likely to be the toxic agent, but the possibility that the unfolded monomeric form also is

toxic cannot be excluded.24

The long-term goal of understanding the molecular mechanism behind diseases

associated with protein misfolding and aggregation will require a combination of

computational and experimental approaches. In this chapter we will provide background
for protein folding mechanisms and possible approaches for studying mechanisms of

folding and misfolding.

1.1. Protein Folding

1.1.1. Energetics

Folding of the protein is a process during which a one-dimensional, almost linear
structure of polypeptide with a certain amino acid sequence forms a characteristic
three-dimensional structure of a biologically firnctional native protein. The characteristic
features of this structure are stability, compact and unique internal organization and
presence of the nonpolar core. The stability of native protein can be expressed in terms of
Gibbs free energy, which is minimized during the folding process as a result of interplay
of enthalpic and entropic contributions. The folded structure is only marginally more
stable than unfolded. For a small protein (<200 residues) under physiological conditions
the folded state is typically only 4 to 20 kcal/mol more stable than the unfolded. The
energetically small difference in stability is a result of large contributions from several
major forces. These forces are involved in the protein folding process by impacting
enthalpic or entropic contributions to Gibbs free energy.25 This section underlines the

dominant energies and forces that guide the protein from the unfolded to the folded state.

5

The Hydrophobic Effect. Almost 50 years ago Kauzrnann recognized the

hydrophobic effect as a principal driving force in protein folding. The hydrophobic effect

can be described as the tendency of nonpolar groups to stick together when surrounded
with polar solvent. The major factor in protein stability is the lack of a hydrogen bond
between a nonpolar residue and water, rather than an interaction between nonpolar
molecules. Accordingly, a hydrophobic amino acid residue creates a hole in the
surrounding water, forcing it to reorganize and become more ordered. As a result, the
entropy of water decreases, which is a thermodynamically unfavorable process. To
minimize the surface tension created by the reorganization of water and to release free
energy confined in the hydrogen bond network, hydrophobic amino acid residues tend to
pack together in the interior of the molecule. A number of experimental results and
theoretical approaches confirm that the hydrophobic effect is a major force in protein

folding that provides an ample amount of energy to the folding reaction.

Hydrogen Bonds. The hydrogen bond is a weak but very abundant interaction in
proteins. On average there are 1.1 intramolecular hydrogen bonds per amino acid residue
in a protein. The energy of a hydrogen bond in a protein varies between 3 and 9 kcal/mol,
depending on the type. Initially, it was believed that hydrogen bonding is the most
important force responsible for folding stability. Kauzmann5 argued that the hydrogen
bond can not be the leading force in protein folding since there was no proof that buried
intramolecular hydrogen bonds in proteins have lower fiee energy than those of the
unfolded chain to water. Hydrogen exchange experiments showed that the energy of the
solvated hydrogen bond is almost equal to the energy of the buried hydrogen bond.
Conversely, hydrogen exchange studies also showed that the rate-limiting step in

unfolding can be associated with the disruption of the hydrogen bond network. A number

of experiments that involved site-directed mutations indicated that hydrogen bonds could
have an overall positive contribution to the protein stability. Although the energetic role
of hydrogen bonding in protein stability is still a matter of debate, it is clear that once the
internal hydrogen bonds are formed, they cannot be broken easily. It would be
energetically very expensive to bury the unsatisfied polar group inside the hydrophobic
core. Hydrogen bonds can guide protein folding by making it energetically unfavorable to
convert already built structures rich in hydrogen bonding into other, structurally less-

organized conformations.”27

Conformational Entropy. The major force that opposes the hydrophobic effect is
conformational entropy. The number of conformational states accessible to the unfolded
protein is enormous. The polypeptide chain of the unfolded protein is very flexible due to
the large number of degrees of freedom. Of particular interest in the unfolded state are the
degrees of freedom related to the side chain rotation around torsion angles. As the protein
folds, parts of its structure are confined to the hydrophobic core, where free rotation
around the torsional angles is hindered due to steric effects and formation of a large
number of hydrogen bonds. This results in a limited number of conformations, and as a
consequence, a substantial loss of entropy that accompanies the process of folding. The
folding is favored only if the hydrophobic effect exceeds the loss of conformational

entropy.25

Electrostatic Interactions. The electrostatic interactions that take place between
charged particles are characterized by an energy change that is distance dependent, as

described by Coulomb’s law. The energy of electrostatic interaction depends on the

9

nature of the charges, the distance between them and the dielectric constant of the
medium. The dielectric constant of the protein exhibits large variations due to the
diversity of protein structure and the presence of the protein-water interface. The
dielectric constant inside the protein is believed to be significantly lower than water (2-4,
in comparison to 80, respectively). The dielectric constant at the water-protein interface is
thought to be similar to the one at the lipid-water interface, which is less than 40. The fact
that the exact values for the dielectric constants of the protein are not known makes the
determination of the electrostatic interactions and their influence on protein stability very

challenging.5’26’27

In general, it is believed that the electrostatic interactions can contribute to the
protein stability in two ways, through classical effects or as specific charge interactions.
The classical effects are related to the nonspecific repulsions that occur between the
protein and the surrounding solvent. These are conditions of high pH and/or ionic
strength when protein is highly charged. The increase in the total charge of the protein
results in the increased repulsion between the protein and solvent and, consequently, the
destabilization of the native protein. The acid or base denaturation of the protein results

in a state with lower electrostatic free energy.

Specific charge interactions, such as the formation of ion pairs, are the result of
attractive forces between two oppositely charged amino acids. Unlike the electrostatic
interactions in the classical sense, these interactions stabilize the protein structure. The
largest contribution to the stability of the native state comes from the ion pairing on the

surface of the protein. 26‘”

10

Van der Waals Interactions. Van der Waals interactions are nonpolar interactions
between two dipoles, fixed or induced. The energy of these interactions is usually
described with the Lennard-Jones potential, where the distance between atoms is defined
as the sum of the Van der Waals radii. These interactions are generally very weak and
characterized by low energies, so the resulting effect on protein stability is small but

additive.5'26’27

All the interactions described above play a part in the stabilization of the native
structure. How these interactions balance one another to create a stable native structure
remains a puzzle. In the following we review theoretical models and experimental

methods that have been applied to address this question.

1.1.2. Mechanisms

Almost 40 years ago, Anfinsen showed that proteins can fold reversibly without
any help from other biological molecules. Anfinsen suggested that the amino acid
sequence of a protein fully determines the three-dimensional structure. He proposed that
the native state is the conformation corresponding to the lowest possible free energy. This
idea is known as the thermodynamic model of protein folding.28 However, Levinthal
showed that the number of accessible conformations that a protein would have to search
through in order to find the native state is endless. He showed that it would be impossible
even for a very small protein to fold in a reasonable time frame if it had to sample them
all.29 Yet, many proteins fold on the millisecond to second timescale. This disagreement

in actual and calculated time required for protein folding is known in literature as

11

Levinthal’s paradox. As a solution, Levinthal proposed that the protein folds into the
native state that does not necessarily correspond to the global free energy minimum. It
does so by following a limited number of well-defined pathways. This is in contradiction
to Anfinsen’s thermodynamical model, and implies that protein folding is a kinetically
controlled process. Levinthal’s work initiated the development of several models for

protein folding that involved a limited number of pathways.

The nucleation model of protein folding was based on the hypothesis that the
amino acid sequence of each protein contains several residues that serve as markers.
These residues are capable of forming initial contacts from which parts of the native
secondary structure are built. They can act as a nucleus from which the structure can
propagate further, in a stepwise manner, towards the native state. This model implies that
there are no intermediates in the folding, and was abandoned with the first experimentally

detected intermediates.25

The updated version of this model is called the
search nucleation.30 In this model protein folding is envisioned as the search for the
minimal number of contacts needed to achieve the native state. The search is performed
through a trial-and-error process that traps critical, structure-formative hydrogen bonds.

This leads to the formation of a nucleus with native-like topology which rapidly proceeds

to the fully folded state.3 "32

The diflusion-collision model, also known as the framework model, proposes that
the formation of the secondary and tertiary structures occurs independently. In this model
the conformational search is reduced by the formation of early-time secondary structures.

These structures diffuse until they collide and reorganize into the tertiary structure.

12

The main hypothesis of the hydrophobic-collapse model was that the rapid
collapse around the hydrophobic side chains of the protein creates an intermediate with
restricted conformational space and far-off tertiary contacts. From this intermediate state

the structure can rearrange further, governed by the native tertiary interactions.25

The experimental observation of intermediates in protein folding raised questions
about their role in the process. One opinion is that their presence increases the rate of
folding but it is clearly inconsistent with the experimental finding that many small
globular proteins are fast-folders that follow two-state kinetics. The other opinion is that
folding intermediates correspond to traps, which slow down the folding process. Even
though the models discussed above provide an explanation for some of the experimental
results, none of them provide a unified explanation for very complex experimental

findings that went in favor of both thermodynamic and kinetic control of folding.

1.1.3. Energy Landscape and Funnels

In the new view advanced especially by Wolynes, Onuchic, Brooks and their
coworkers, folding proteins are steered towards the native state as they fold by the
funnel-shaped topology of the potential-energy hypersurface, also called the energy
landscape.”37 This view requires consideration of the dynamics of an ensemble of
folding trajectories, rather than the standard single-path reaction-coordinate picture with
few well-defined intermediates that is suggested by most biochemical and spectroscopic
experiments.34 The hypothetical funnel shaped energy landscape for a small protein is

shown in Figure 1.1 in a two-dimensional representation. The reader should keep in mind

13

that the real energy landscape is a complex multi-dimensional surface, with as many
independent coordinates as there are internal degrees of freedom.” The vertical
coordinate on the firnnel-shaped surface is internal energy, E, defined as the total free
energy without the entropy contribution. The entropy is represented by the width of the
funnel. One additional parameter is proposed in this representation of the protein folding
energy surface, and that is the measure of the native-like contacts, Q. The position of the
ensemble of states within the funnel is described by the E and Q parameters. At the very
top of the funnel is an ensemble of conformations for an unfolded protein. The unfolded
state of the protein can be characterized as a state with large conformational freedom
where individual groups and parts of the polypeptide chain can freely rotate relative to
one another. As a consequence, the number of accessible conformational states for an
unfolded protein is very large and has high entropy. As the protein folds into an ordered
native structure, the number of different conformational states decreases and,
consequently, entropy decreases, too. The decrease in the number of different
conformational states is a direct consequence of the formation of the larger number of
correct native contacts. When the number of correct native contacts becomes significant
(about 25% or more), a partially folded protein reaches the ensemble of states called

molten globule. Structural characteristics of the molten globule state are the presence of

14

entropy

 

energy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1.1 Top panel. Hypothetical fimnel-shaped energy landscape. During the folding
process protein moves down the funnel from the unfolded structure to the folded at the
bottom of the funnel which is accompanied by a decrease in energy E and entropy, and
with increase in native contacts Q. On its way to the native state protein goes through the
region of molten globule MG and overcomes the transition state TS barrier before it
finally reaches the native state. After Onuchic and Wolynes34 Bottom panel. Enlarged
view of near vicinity of a native state position in the funnel depicting the thermodynamic
(a) and kinetic (b) paths.”

15

some secondary structure and some tertiary interactions and also the presence of the
hydrophobic core. This state is very similar in structure to the native state, but less
compact and energetically less stable. After the molten globule state, the folding process
reaches the transition state region or so-called energy bottleneck on the potential energy

landscape, after which the structure collapses into the compact and stable native state.25

The idea of the funnel-shaped potential energy landscape overcomes Levinthal’s
paradox by seeing the folding process as an ensemble average of a process that is in
reality very heterogeneous. In this representation every molecule that undergoes folding
can choose a different trajectory, but all molecules in the process will reach the same
final state. Unlike the other models of protein folding, the energy landscape model

explains folding as a process that is thermodynamically and/or kinetically controlled. 35

1.1.4. Thermodynamic and kinetic control

The thermodynamically controlled folding process is the one defined by the
energy of the conformational states and guided by the tendency to reach the state that
corresponds to the global minimum on the energy landscape. In contrast, the kinetically
controlled process is guided by the rate of the transformation from some random
conformation to the native state, where the fastest route would be preferable.”39 In the
potential energy landscape, the shape of the folding funnel and presence of the final state
signify the thermodynamical control of protein folding. The final state is defined as the
conformation with lowest energy. The ruggedness of the potential energy funnel implies

kinetic control of protein folding by introducing a number of local minima where the

16

protein can temporarily or finitely reside depending on the height of the barrier.”

The comparison of two types of control can be seen at the bottom of Figure 1.1,
which represents an enlarged view of the vicinity of the native state in the folding funnel.
State X is obviously thermodynamically preferred, since it is energetically closer to the
native state. However, state Y is kinetically preferred since it implies a smaller barrier to
cross on the way to the native state. Depending on whether the system chooses path a or
path b, the folding can be either thermodynamically or kinetically controlled.”40 The
proteins that have smooth funnels show a preference towards the folding process which is
mostly thermodynamically controlled, while the proteins with rugged folding funnels
could fold via the kinetically controlled process. This implies that the native state does
not necessarily have to be the state corresponding to the global energy minimum; it can

be the state close to the global energy minimum that is kinetically preferred.

1.1.5. Experimental and Theoretical Approach

Protein folding reactions have been examined using NMR linewidth
measurements from which structural information can be obtained.‘”’42 The protein
stability and structural dynamics are extensively studied with NMR methods coupled to
hydrogen exchange.”45 Temperature jump methods were applied to studies of small

peptides, and protein unfolding and refolding dynamics/’6’49

Protein folding free energy surfaces can be sampled with molecular dynamics
(MD) simulations. The MD simulations are based on Newton’s laws of motion that are

applied to every atom in the protein and surrounding solvent.25’50 This requires

17

calculations of the large number of conformational states over a long period of time. As a
result the calculations of folding reactions are limited to relatively short timescales. MD
studies are suitable for the characterization of unfolding reactions since the initial
parameters can be obtained from X-ray and NMR structural studies. Furthermore,
simulating denaturants or increasing temperature can accelerate the unfolding reaction
and reduce calculation time. The folding reaction can be examined by running the MD
simulations for the unfolding reaction in reverse. MD simulations were successfully used
for characterization of transition states and intermediate and denatured states in a variety

of proteins including chymotrypsin inhibitor 2,51 bovine pancreatic trypsin inhibitor,52

53 54

barnase and myoglobin. Recently MD simulations were applied for studying
conformational transition of amyloid b-peptide, a small protein related to Alzheimer’s

disease55 and prion protein related to transmissible spongiform encephalopathies}6

1.1.6. Role of Intrinsic and Surface Water

Water plays an important role in the dynamics, function, structure and stability of
proteins. Several different types of biological water can be recognized, depending on
their position with respect to the protein molecule: structural or buried water, hydration
water located in the vicinity of the protein surface and bulk water. Structural water is an
integrated part of the protein structure; it interacts with the protein side chains and can
organize in clusters in the hydrophobic regions inside of the protein. Hydration water
interacts with the protein surface, resulting in the coupling of fluctuations of hydration
water to the protein dynamics. This interaction stabilizes the folded structure but can also

have a large effect on protein dynamics, increasing the complexity.

18

The most obvious reason why water has to be involved in studies of the protein
folding process is its role in the hydrophobic effect, one of the main effects that govern
folding.57 One approach to studying water involvement in protein folding is to examine
desolvation during the folding process starting with a fully solvated unfolded polypeptide
chain. Onuchic and coworkers58 used the minimalist model to investigate the desolvation
of the SH3 protein during the folding process by including the effects of expulsion of
water in the energy landscape theory. They found the presence of an intermediate with a
partially solvated hydrophobic core that is structurally very close to the collapsed
structure of the native state. These results are in agreement with Zhang et a159'6' and Mok

1.,62 who confirmed experimentally the existence of an intermediate SH3 structure

et a
with a hydrated hydrophobic core. The formation of this near-native structure precedes
the final step in the folding of the protein, which according to Onuchic and coworkers is
the desolvation of the hydrophobic core. In another contribution,63 authors performed a
detailed theoretical and experimental study of folding rates using mutations in the
hydrophobic core on an a-spectrin SH3 protein. These studies suggested that the

desolvation of the hydrophobic core in this protein is the most important step in

overcoming the transition state barrier for folding.

Wolynes and coworkers suggested that the smooth regions of the potential energy
fimnel result from long-range water-mediated interactions. Hence, the water is directly
involved in the folding process by insuring proper steering through the funnel and
increasing the efficiency of folding.64 Similar results were obtained for proteins A and G

from atomistic simulation studies.“ 66

19

The experimental studies of the dynamics of biological water report wide range of
time scales depending on the method used, ranging from the several hundreds of

R,“68 to several nanosecond as determined by dielectric

picoseconds determined by NM
relaxation experiments. Zewail and coworkers used surface tryptophan as a natural probe
in solvation dynamics studies of surface water.”70 They observed a bimodal hydration
response with time constants on the order of ps. The slower component of the response
was attributed to the surface-bound water; the fast component was attributed to the free
water. In the unfolded protein both time constants significantly lengthened. The authors

proposed that the longest time constant (~60 ps) observed in the study with unfolded

protein involves motions of the protein backbone.

1.2. Structural Fluctuations and Time-Correlation Function

Proteins are characterized by a high level of order but this does not mean that they
are static structures. Even in the native state proteins exhibit motions over a wide range of
timescales, amplitudes and energies. The motions can span from the global motions of
the entire protein to local motions related to the atomic fluctuations and rotation of the
amino-acid side chains. The high level of dynamics in proteins implies that in the native
state protein fluctuates around some average structure, so even at room temperature it can
sample a large number of conformations. Water also has a large impact on the motions of

1.71 argued that motions in the protein can be

proteins and their dynamics. Fenimore et a
coupled to fluctuations in surrounding water. The protein motions are affected by the

surrounding solvent mainly because of the difference in the dielectric coefficients of

protein and water. The desolvated protein has a significantly smaller dielectric coefficient

20

than water. The fluctuations of the water dipoles surrounding the protein couple to the
fluctuations of the water dipoles in the hydration shell and with charged residues at the
protein surface, affecting the dynamics of the protein. This suggests that it could be
possible to monitor solvent-protein interaction by detecting the dipole moment
fluctuations of a carefully chosen probe molecule attached to the protein. The work
presented in this dissertation utilizes that idea by experimentally observing protein
dynamics in the folded and unfolded state through measurement of a solvent response

function.

Upon excitation with light, the chromophore experiences an instantaneous
transition from the equilibrium in the ground state to a higher electronic state (see
Figure 1.2). This transition is accompanied by a fast change in the dipole moment of the
chromophore. The solvent molecules that surround the chromophore cannot reorganize
instantaneously in order to adapt to the new dipole moment of the chromophore, which as
a consequence has a different chromophore-solvent interaction in the excited state than in
the ground state. The energy released by the solvent molecules as they equilibrate within
the excited state after F ranck-Condon transition is called solvent reorganization energy, I.
As the system equilibrates in the excited state, the energy gap between the two states will

decrease in time and transition frequency (Deg red-shifts.

In the condensed phase, the ground-to-excited-state transition frequency was of an

electronic chromophore fluctuates with respect to time, owing to collisions and to the

electrostatic interactions with molecules in motion outside the first solvation shell.

21

 

 

Energy

 

 

 

 

 

Solvation Coordinate

Figure 1.2 Solvation dynamics and the dynamic Stokes shift. Following electronic
excitation of a solute chromophore, the solvent molecules still reside in the configuration
that was reached at equilibrium with the ground state solute’s dipole moment. Over time,
the solute molecules reorganize in response to the change in the solute’s dipole moment;
the fluorescence spectrum shifts to the red as the system approaches equilibrium. After
Stratt and Maroncelli.72

22

The fluctuation can be expressed in terms of the difference dwega) between the

transition frequency for a certain molecule at a given instant in time and that frequency

averaged over the ensemble, (meg) :73“77

weg(t)=<a)eg>+6weg(t)+e (1.1)

The parameter a is that part of the difference from the average for a given chromophore—

solvent site that does not vary with time. The time correlation firnction M (t) describes the

loss of memory of the transition energy owing to the solvent-induced fluctuations:

M(t) = (dweg(0)6a)eg(t)>
<§weg2>

 

(1.2)

The time evolution of the fluorescence spectrum defines a normalized solvent-

response function, SV (t) , defined usually in terms of the frequency at the fluorescence
intensity maximum, n(t).72

v(t) - v(oo)

S" (t) = v(0) — v(oo)

(1.3)

M (t) and SV (t) are equivalent in the high-temperature limit, where the breadth of the

energy fluctuations is much less than kBT, and in the linear-response regime, where the

perturbation caused by the ground-to-excited-state electronic excitation of the probe is

23

not so large that the motions of the solvent are different in the two electronic

7 8
813168.72' 5’76'7

1 .2.1 . Molecular Liquids

Extensive studies performed on molecular liquids showed that it is possible to
extract information about the time evolution of the solvent-solute interaction energy by
detecting the time-resolved emission spectrum from an electronically excited
chromophore/9'81 The obtained experimental data suggests that the solvation dynamics in
liquids are biphasic in nature. Fleming and coworkers82 observed the fast inertial phase of
solvation dynamics that occurred on 100 fs timescale. Maroncelli and Stratt attributed
this fast phase of solvation dynamics to the librational motions of the solvent molecules
surrounding the probe molecules"83 Fleming and Barbara associated the slower, diffusive
component with reorientational motions74’79’84 The experimental results of Maroncelli8|
and theoretical work of Ladanyi and Stratt85 suggest that solvation dynamics in both polar
and nonpolar media result from the coupled rotational and/or translational dynamics
chromophore-solvent system. While the polar solvation response is the result of a change
in the probe’s dipole moment, the non-polar solvation response is the result of a change
in the probe’s size or structure. The details of solvation dynamics in two media can vary
due to the differences in the intermolecular interactions between the chromophore and
solvent and solvent and solvent, but the solvation response exhibits a fast decay due to
the inertial dynamics, followed by the slow diffusive dynamics in both polar and

nonpolar media.

24

1.2.2. Polypeptide and Folded Proteins
The fluctuations that contribute to the decay of M (t) and S, (t) in proteins might

involve collision-like interactions between the probe and the surrounding polypeptide
host and any intrinsic water molecules. Owing to the low effective dielectric constant in
the interior of a protein,86 however, the fluctuations may also involve more distant
interactions of the chromophore’s dipole moment with dipoles in the protein structure or,
perhaps, on the surface of the protein than are possible in a molecular solvent with a high
dielectric constant, where the interactions with the distant dipoles are attenuated. A
consensus has not been reached on how a protein’s structure and the included solvent

contribute to M (t) and S, (t).

In previous work in this laboratory, an intrinsic extended tetrapyrrole
chromophore was used in transient-hole burning and stimulated photon-echo peak-shift
experiments to sense the dynamics of a small globular protein, the alpha subunit of C-
phycocyanin. A large fraction (90 %) of the decay of M (t) in C-phycocyanin was
observed on the <100 fs scale.”88 In molecular liquids, a probe’s energy gap exhibits
fluctuations on this time scale owing to the inertial (free-steaming, free-rotor or
librational) motions of the surrounding solvent molecules/37’88 The diflusive portion of the
decay of M (t) in C-phycocyanin extends, perhaps, even to the us time scale, as it might
in a slow-relaxing, viscous solvent. This part of the time-correlation function arises from
random reorientational motions in a molecular solvent. Both the inertial and diffusive

parts of M (t) were assigned to motions of the surrounding protein structure and intrinsic

water moleculesgg’g'9 A large inertial solvation component was also observed by Fleming

25

and co-workers in a chicken egg-white lysozyme, which was extrinsically probed with
the fluorescent dye eosyn. The results indicated that the protein may make a contribution

to the diffusive part of M (t), but the large inertial component was strongly correlated

with motions of the bulk aqueous solvent.74 The solvation response observed by the same
group in bacteriorhodopsin using the intrinsic retinal chromophore as the probe consists
solely of an inertial component, with the bulk of the decay confined to the sub-50 fs
regime; no diffusive decay was observed over the 1 ps-window in which stimulated

photon echoes could be observed.73

Hochstrasser and co-workers78 probed the solvation dynamics of calmodulin in
pump—probe studies of the dynamic Stokes shift of the stimulated-emission spectrum of
coumarin 343, which was used to label the protein. The inertial component that they
observed on the sub-IOO-fs time scale was followed by a smaller diffusive component on
the ps time scale. Both solvation components were assigned to motions of the

surrounding protein rather than to the bulk solvent molecules.

Cohen et al.86 observed a generally biphasic solvation response in the B1 domain
of protein G using the fluorescent synthetic amino acid aladan incorporated at specific
sites in the amino-acid sequence. The femtosecond fluorescence upconversion technique
was employed in this work. The dynamic Stokes shift response that was detected both at
the buried and solvent-exposed sites has a prominent sub-ps component followed by an
eXponential or multiexponential diffusive response that reached to the end of the reported
150-ps decay window. The magnitude of the solvation response was dependent on the
degree of solvent exposure, with the exposed sites exhibiting a larger reorganizational

26

energy, but a substantial fraction of the reorganization was observed in the sub-ps portion
at all of the probed sites. An important conclusion from this work is that the solvation
response should be expected to be heterogeneous and strongly dependent on the location
of the probe in the folded protein. For that reason it might be suitable to study solvation

response with proteins that have naturally present chromophore, such as heme proteins.

1.3. Structure, Dynamics and Folding Reactions of Cytochrome c

1.3.1. Structure

Cytochrome c is an electron-transfer protein present in the mitochondrial
membrane of all aerobic organisms. The role of cytochrome c is to carry electron
between the cytochrome c reductase and cytochrome c oxidase in the final steps of the
respiratory chain. It is a small globular protein with a single polypeptide chain of 103 to
111 amino acid residues, depending on the species. Cytochrome c contains a prosthetic

group, an iron protoporphyrin IX ring (heme) (see Figure 1.3).

Unlike in some other heme proteins, myoglobin and hemoglobin for example,
heme in cytochrome c is covalently bonded to the polypeptide chain via thioether bonds
to the invariant Cys 14 and Cys 17. The Fe(II) in the heme is octahedrally liganded and
can be in either the Fe(II) or Fe(IlI) oxidation state. Four octahedral ligands to the iron
atom are supplied by the porphyrin ring. Axial ligands to the iron atom are contributed by
the nitrogen of the His 18 side chain and the sulfur of Met 80. Hydrophobic residues

occur in clusters, mostly in the heme region. The heme-binding region, hydrophobic

27

residues in its vicinity and two water molecules are highly conserved in cytochrome c
molecules from different species. The small size of this protein and the presence of the
natural chromophore, which is well-enveloped by the protein matrix and stays bonded to
the polypeptide chain even when the protein is fully unfolded, makes cytochrome c a

very attractive target for protein folding studies.

 

Figure 1.3 The structure of horse-heart90 ferricytochrome c, rendered using PyMOL.91

1.3.2. Folding Intermediates

The unifying picture of the cytochrome c folding process is a result of
experimental studies performed with a variety of methods. The results suggest that the
bottleneck for the folding reaction of cytochrome c has its origin in the final steps in the

search for the native structure and not in the initial condensation of the polypeptide chain.

28

Cytochrome c folding starts with formation of an early intermediate species characterized

by a more compact Polypeptide chain.92'93

When initiated from the acid-denatured state, folding of cytochrome c is rapid and
occurs on two distinct time scales of 500 ms and about 15 ms. Biphasic folding in this
case is attributed to the formation of helices, which occurs stepwise},4 The same authors
found two distinct intermediates that occur on-pathway during refolding of cytochrome c
(at 160 ms and 500 ms). The faster time constant is related to the rapidly formed
intermediate, which has some helix content and is almost two times larger in radius than
the native state. This structure evolves into the second intermediate at 500 ms, which is
more compact and in which the majority of helical content is formed. The presence of
these two distinct intermediate forms was also confirmed by small angle X-ray
scattering.95 The native state is formed at a much slower rate (IO-100 ms), which strongly
depends on the refolding conditions and the formation of intermediates that involve the

improper coordination of heme.

1.3.3. Role of Heme Ligation in the Folding Reaction

Denaturation of cytochrome c causes the displacement of the Met 80 ligand while
His 18 stays coordinated to the iron in the heme. Spectral data showed that Met 80 is
most commonly replaced with another His residue.96 The non-natively ligated His
residues in horse heart cytochrome c are typically His 26 or His 33, and their ligation to
the heme iron can effect the overall rate of the folding of cytochrome c. Roder and

coworkers97 showed that folding initiated from the base-denatured state of cytochrome c

29

is slowed to several hundred milliseconds. The slow folding is attributed to the formation
of an off-pathway intermediate due to the mis-coordination of iron in a heme group by
His 26 or His 33. They98 proposed the kinetic mechanism for the folding/unfolding of
oxidized cytochrome c, which involves heme ligand exchange coupled to structural
changes. In the early stages of folding several intermediate species are present that
involve bounding of the non-native histidine. The misligation of the histidine represents a
trap on the folding pathway and is the rate-limiting step. This mechanism is supported by
experiments performed by a variety of methods: resonance Raman spectroscopy,93
stopped—flow measurements monitored by fluorescence, heme absorbance, and CD over a
range of conditions.99"°2 In contrast, the reduced form of cytochrome c seems to exhibit a
two-state folding mechanism. The ligand exchange reactions were detected in the folding
mechanism of reduced cytochrome c by triggering refolding with laser-induced
photolysis of the CO ligand. These reactions involved alternative His and Met ligands
and occurred on time scales much faster than in the oxidized form of cytochrome c, and

long-lived intermediate species were not detected.

1.4. Proposed work

The investigation of the molecular mechanism involved in creation of the disease
related species along the folding trajectory needs to start with a detailed evaluation of the
native state. The protein in the native state is a highly dynamic structure that exhibits a
wide range of motions while switching from one low energy conformation to another.
Some of these motions might be responsible for trapping the protein in undesirable

conformations which could ultimately result in a misfolded, toxic form. We will

30

investigate motions that protein exhibits in the native state by employing picosecond
time-resolved fluorescence studies of solvent response fiom a Zn(II) substituted

cytochrome c.

The native cytochrome c is not suitable for studies of the dynamics of a solvated
protein with time resolved fluorescence, since the naturally present chromophore with the
F e(II) ion has a very short fluorescence lifetime. However, substitution of the F e(II) ion
in the heme with the Zn(II) ion provides the chromophore with long-lived fluorescence103
and makes it a suitable probe for solvation dynamics. The studies of the Zn-substituted
cytochrome c (ZnCytc) with continuous absorption and fluorescence spectra and
2D-NMR showed that replacement of the Fe(II) ion with Zn(II) in the heme does not
result in a different structurelm'105 Vanderkooi and coworkers applied fluorescence line
narrowing spectroscopy to study ZnCytc in the native and denatured states.'°6 Winkler
and coworkers probed the environment of the heme chromophore in the native and

denatured states of ZnCytc by monitoring triplet state decay kinetics.107

We will use this spectroscopically well characterized protein to probe not only
solvation response in the native state but also near the transition for unfolding, since the
chromophore stays bound to the protein even when it is fully unfolded. In the studies of
the various diseases it has been indicated that toxicity of the misfolded species might be
caused with increased exposure of the inside of the protein to the surrounding solvent and
inappropriate hydrophobic interactions. The importance of water in the folding
mechanism is well established, but the influence of water on the dynamics of proteins is

not fully understood. Of particular interest is entrance or exit of water into and from

31

hydrophobic core since it is directly involved in the hydrophobic collapse of the protein
structure. While the dynamics of the surface water is studied to the certain extent, the
direct observation of the water entrance or exit into the hydrophobic core, its dynamics

and its effect on the protein motions are not known.

9 So far, the studies of protein folding, unfolding and misfolding reactions were
limited to the applications of methods that involve significant manipulation of the protein
solution conditions or the temperature. We intend to address the above questions by
studying the protein folding, unfolding and misfolding reactions triggered without
imposing external manipulation of the solvent, or without the direct change of the

temperature of the protein solution.

32

1.5.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

References

Prusiner, S. B. Proc. Natl. Acad. Sci. USA 1998, 95, 13363—13383.

Dobson, C. M. Trends BiochemSci. 1999, 24, 329—332.

Caughey, B.; Lansbury, P. T. Annu. Rev. Neurosci. 2003, 26, 267—298.

Dobson, C. M. Nature 2003, 426, 884—890.

Kauzmann, W. Adv. Protein. Chem. 1959, 14, 1—63.

Dobson, C. M. Philos. Trans. R. Soc. Lond. Ser. B-Biol. Sci. 2001, 356, 133—145.

Sunde, M.; Blake, C. The structure of amyloid fibrils by electron microscopy and
X-ray diffraction. In Adv. Prot. Chem, 1997; Vol. 50; pp 123—159.

Chiti, F.; Stefani, M.; Taddei, N.; Ramponi, G.; Dobson, C. M. Nature 2003, 424,
805.

Blake, C. C. F.; Geisow, M. J.; Oatley, S. J.; Rerat, B.; Rerat, C. J. Mol. Biol.
1978, 121, 339—356.

Coelho, T. Curr. Opin. Neurol. 1996, 9, 355—359.

Plante-Bordeneuve, V.; Said, G. Curr. Opin. Neurol. 2000, 13, 569—573.

Chiti, F.; Webster, P.; Taddei, N.; Clark, A.; Stefani, M.; Ramponi, G.; Dobson,
C. M. Proc. Natl. Acad. Sci. USA 1999, 96, 3590—3594.

Selkoe, D. .1. Nature 2003, 426, 900—904.

33

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

Pertinhez, T. A.; Bouchard, M. L.; Tomlinson, E. J.; Wain, R.; Ferguson, S. J .;
Dobson, C. M.; Smith, L. J. Febs Lett. 2001, 495, 184—186.

Fandrich, M.; Forge, V.; Buder, K.; Kittler, M.; Dobson, C. M.; Diekmann, S.
Proc. Natl. Acad. Sci. USA 2003, 100, 15463-15468.

Guijarro, J. 1.; Sunde, M.; Jones, J. A.; Campbell, I. D.; Dobson, C. M. Proc. Natl.
Acad. Sci. USA 1998, 95, 422M228.

Chiti, F .; Bucciantini, M.; Capanni, C.; Taddei, N.; Dobson, C. M.; Stefani, M.
Prot. Sci. 2001, 10, 2541—2547.

Lambert, M. P.; Barlow, A. K.; Chromy, B. A.; Edwards, C.; Freed, R.; Liosatos,
M.; Morgan, T. E.; Rozovsky, I.; Trommer, B.; Viola, K. L.; Wals, F.; Zhang, C.;
Finch, C. E.; Krafft, G. A.; Klein, W. L. Proc. Natl. Acad. Sci. USA 1998, 95,
6448—6453.

Walsh, D. M.; Townsend, M.; Podlisny, M. B.; Shankar, G. M.; Fadeeva, J. V.; El
Agnaf, O.; Hartley, D. M.; Selkoe, D. J. J. Neurosci. 2005, 25, 2455—2462.

Hartley, D. M.; Walsh, D. M.; Ye, C. P. P.; Diehl, T.; Vasquez, S.; Vassilev, P.
M.; Teplow, D. B.; Selkoe, D. J. J. Neurosci. 1999, 19, 8876—8884.

Bucciantini, M.; Giannoni, E.; Chiti, F.; Baroni, F.; Formigli, L.; Zurdo, J. S.;
Taddei, N.; Ramponi, G.; Dobson, C. M.; Stefani, M. Nature 2002, 416, 507-511.

Bucciantini, M.; Calloni, G.; Chiti, F.; Formigli, L.; Nosi, D.; Dobson, C. M.;
Stefani, M. J. Biol. Chem. 2004, 279, 31374—31382.

Cleary, J. P.; Walsh, D. M.; Hofmeister, J. J .; Shankar, G. M.; Kuskowski, M. A.;
Selkoe, D. J.; Ashe, K. H. Nat. Neurosci. 2005, 8, 79—84.

Volles, M. J .; Lansbury, P. T. Biochemistry 2003, 42, 7871—7878.

Fersht, A. Structure and mechanism in protein science: aguide to enzyme catalysis
and protein folding; W. H. Freeman and Company: New York, 1997.

34

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

Dill, K. A. Biochemistry 1990, 29, 7133.

Pace, C. N.; Shirley, B. A.; McNutt, M.; Gajiwala, K. Faseb Journal 1996, 10,
75.

Anfinsen, C. B. Science 1973, I81, 223.

Levinthal C. Journal De Chimie Physique Et De Physico-Chimie Biologique
1968, 65, 44.

Sosnick, T. R.; Mayne, L.; Englander, S. W. Proteins 1996, 24, 413—426.

Plaxco, K. W.; Simons, K. T.; Baker, D. J. Mol. Biol. 1998, 27 7, 985—994.

Krantz, B. A.; Sosnick, T. R. Biochemistry 2000, 39, 11696-11701.

Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598—1603.

Onuchic, J. N.; Luthey-Schulten, Z.; Wolynes, P. G. Annu. Rev. Phys. Chem.
1997, 48, 545—600.

Chan, H. S.; Dill, K. A. Proteins: Struct. F unct. Genet. 1998, 30, 2—33.

Brooks III, C. L.; Onuchic, J. N.; Wales, D. J. Science 2001, 293, 612.

Shea, J .-E.; Brooks III, C. L. Annu. Rev. Phys. Chem. 2001, 52, 499—535.

Dill, K. A. Protein Sci. 1999, 8, 1 166—1180.

Baker, D.; Agard, D. A. Biochemistry 1994, 33, 7505—7509.

Dill, K. A.; Chan, H. S. Nature Struct. Biol. 1997, 4, 10—19.

35

(41)

(42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

(50)

(51)

(52)

(53)

(54)

(55)

Burton, R. E.; Huang, G. S.; Daugherty, M. A.; Fullbright, P. W.; Oas, T. G. J.
Mol. Biol. 1996, 263, 311.

Huang, G. S.; Oas, T. G. Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 6878.

Englander, S. W.; Sosnick, T. R.; Mayne, L. C.; Shtilerman, M.; Qi, P. X.; Bai, Y.
W. Acc. Chem. Res. 1998, 31, 737.

Englander, S. W.; Mayne, L. Annu. Rev. Biophys. Biol. Struc. 1992, 21, 243—265.
Englander, S. W. Annu. Rev. Biophys. Biol. Struc. 2000, 29, 213-238.

Gruebele, M. Annu. Rev. Phys. Chem. 1999, 50, 485—516.

Callender, R.; Dyer, R. B. Curr. Opin. Struc. Biol. 2002, 12, 628.

Ballew, R. M.; Sabelko, J.; Gruebele, M. Proc. Natl. Acad. Sci. U. S. A. 1996, 93,
5759—5764.

Williams, S.; Causgrove, T. P.; Gilmanshin, R.; Fang, K. S.; Callender, R. H.;
Woodruff, W. H.; Dyer, R. B. Biochemistry 1996, 35, 691—697.

Daggett, V.; Levitt, M. J. Mol. Biol. 1993, 232, 600.

Day, R.; Daggett, V. Protein Science 2005, 14, 1242.

Daggett, V.; Levitt, M. Proc. Natl. Acad. Sci. U. S. A. 1992, 89, 5142.
Caflisch, A.; Karplus, M. Proc. Natl. Acad. Sci. U. S. A. 1994, 91, 1746.
Onufriev, A.; Case, D. A.; Bashford, D. J. Mol. Biol. 2003, 325, 555.

Xu, Y. C.; Shen, J. J.; Luo, X. M.; Zhu, W. L.; Chen, K. X.; Ma, J. P.; Jiang, H. L.
Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 5403.

36

(56)

(57)

(58)

(59)

(60)

(61)

(62)

(63)

(64)

(65)

(66)

(67)

(68)

(69)

Barducci, A.; Chelli, R.; Procacci, P.; Schettino, V. Biophys. J. 2005, 88, 1334.

Harano, Y.; Kinoshita, M. Chem. Phys. Lett. 2004, 399, 342.

Cheung, M. S.; Garcia, A. E.; Onuchic, J. N. Proc. Natl. Acad. Sci. USA 2002, 99,
685—690.

Zhang, 0. W.; F ormanKay, J. D. Biochemistry 1997, 36, 3959.

Farrow, N. A.; Zhang, 0. W.; FormanKay, J. D.; Kay, L. E. Biochemistry 1997,
36, 2390.

Zhang, 0.; Formankay, J. D. Biochemistry 1995, 34, 6784.

Mok, Y. K.; Kay, C. M.; Kay, L. E.; Forman-Kay, J. J. Mol. Biol. 1999, 289, 619.

Femandez-Escamilla, A. M.; Cheung, M. 8.; Vega, M. C.; Wilmanns, M.;
Onuchic, J. N.; Serrano, L. Proc. Natl. Acad. Sci. USA 2004, 101, 2834.

Papoian, G. A.; Ulander, J .; Eastwood, M. P.; Luthey-Schulten, Z.; Wolynes, P.
G. Proc. Natl. Acad. Sci. USA 2004, 101, 3352.

Sheinerman, F. 8.; Brooks, C. L. J. Mol. Biol. 1998, 278, 439.

Garcia, A. E.; Onuchic, J. N. Proc. Natl. Acad. Sci. USA 2003, 100, 13898.

Otting, G.; Liepinsh, E.; Wuthrich, K. Science 1991, 254, 974.

Halle, B.; Andersson, T.; F orsen, S.; Lindman, B. Journal of the American
Chemical Society 1981, 103, 500.

Pal, S. K.; Peon, J .; Bagchi, B.; Zewail, A. H. J. Phys. Chem. B 2002, 106,
12376—12395.

37

(70)

(71)

(72)

(73)

(74)

(75)

(76)

(77)

(78)

(79)

(80)

(81)

(82)

(83)

Peon, J.; Pal, S. K.; Zewail, A. H. Proceedings of the National Academy of
Sciences of the United States of America 2002, 99, 10964.

Fenimore, P. W.; Frauenfelder, H.; McMahon, B. H.; Parak, F. G. Proc. Natl.
Acad. Sci. USA 2002, 99, 16047.

Stratt, R. M.; Maroncelli, M. J. Phys. Chem. 1996, 100, 12981—12996.

Kennis, J. T. M.; Larsen, D. S.; Ohta, K.; Facciotti, M. T.; Glaeser, R. M.;
Fleming, G. R. .1. Phys. Chem. B 2002, 106, 6067—6080.

Jordanides, X. J .; Lang, M. J.; Song, X.; Fleming, G. R. J. Phys. Chem. B 1999,
103, 7995—8005.

Cho, M.; Yu, J.-Y.; Joo, T.; Nagasawa, Y.; Passino, S. A.; Fleming, G. R. J. Phys.
Chem. 1996, 100, 11944—11953.

Fleming, G. R.; Cho, M. Annu. Rev. Phys. Chem. 1996, 47, 109—134.

100, T.; Jia, Y.; Yu, J.-Y.; Lang, M. J.; Fleming, G. R. J. Chem. Phys. 1996, 104,
6089—6108.

Changenet-Barret, P.; Choma, C. T.; Gooding, E. F .; DeGrado, W.; Hochstrasser,
R. M. J. Phys. Chem. B 2000, 104, 9322-9329.

Jarzeba, W.; Walker, G. C.; Johnson, A. E.; Barbara, P. F. Chem. Phys. 1991,
152, 57—68.

Maroncelli, M.; Maclnnis, J.; Fleming, G. R. Science 1989, 243, 1674—1681.

Maroncelli, M. J. Mol. Liq. 1993, 57, 1—37.

Rosenthal, S. J.; Xie, X.; Du, M.; Fleming, G. R. J. Chem. Phys. 1991, 95, 4715-
4718.

Stratt, R. M.; Cho, M. J. Chem. Phys. 1994, 100, 6700—6708.

38

(84)

(85)

(86)

(87)

(88)

(89)

(90)

(91)

(92)

(93)

(94)

(95)

(96)

(97)

Kahlow, M. A.; Jarzeba, W.; Kang, T. J .; Barbara, P. F. J. Chem. Phys. 1989, 90,
151.

Ladanyi, B. M.; Stratt, R. M. J. Phys. Chem. 1995, 99, 2502-2511.

Cohen, B. E.; McAnaney, T. 8.; Park, E. S.; Jan, Y. N.; Boxer, S. G.; Jan, L. Y.
Science 2002, 296, 1700—1703.

Riter, R. E.; Edington, M. D.; Beck, W. F. Inertial protein-matrix solvation of a

light-harvesting chromophore. In Ultrafast Phenomena X; Barbara, P., Knox, W.,
Zinth, W., Fujimoto, J ., Eds.; Springer-Verlag: Berlin, 1996; pp 324-325.

Homoelle, B. J.; Edington, M. D.; Diffey, w. M.; Beck, w. F. J. Phys. Chem. B
1998,102,3044—3052.

Homoelle, B. J .; Beck, W. F. Biochemistry 1997, 36, 12970—12975.
Bushnell, G. W.; Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 585—595.

Delano, W. L. The PyMOL molecular graphics system; DeLano Scientific: San
Carlos, California, 2002.

Shastry, M. C. R.; Roder, H. Nature Struct. Biol. 1998, 5, 385.
Yeh, S.-R.; Han, S.; Rousseau, D. L. Acc. Chem. Res. 1998, 31 , 727-736.

Akiyama, S.; Takahashi, S.; Ishimori, K.; Morishima, 1. Nature Struct. Biol. 2000,
7, 514.

Pollack, L.; Tate, M. W.; Damton, N. C.; Knight, J. B.; Gruner, S. M.; Eaton, W.
A.; Austin, R. H. Proc. Natl. Acad. Sci. USA 1999, 96, 10115.

Babul, J.; Stellwag.E. Biopolymers 1971, 10, 2359.
Colon, W.; Wakem, L. P.; Sherman, F.; Roder, H. Biochemistry 1997, 36, 12535.

39

(98)

(99)

(100)

(101)

(102)

(103)

(104)

(105)

(106)

(107)

Colon, W.; Elove, G. A.; Wakem, L. P.; Sherman, F.; Roder, H. Biochemistry
1996, 35, 5538.

Colon, W.; Roder, H. Nature. Struct. Biol. 1996, 3, 1019.

Elove, G. A.; Chaffotte, A. F.; Roder, H.; Goldberg, M. E. Biochemistry 1992, 31 ,
6876.

Elove, G. A.; Bhuyan, A. K.; Roder, H. Biochemistry 1994, 33, 6925.

Sosnick, T. R.; Mayne, L.; Hiller, R.; Englander, S. W. Nature Struct. Biol. 1994,
1, 149.

Vanderkooi, J. M.; Adar, F .; Erecinska, M. Eur. J. Biochem. 1976, 64, 381.

Anni, H.; Vanderkooi, J. M.; Mayne, L. Biochemistry 1995, 34, 5744-5753.

Quian, C.; Yao, Y.; Tong, Y.; Wang, J.; Tang, W. J. Biol. Inorg. Chem. 2003, 8,
394—400.

Logovinsky, V.; Kaposi, A. D.; Vanderkooi, J. M. Biochim. Biophys. Acta 1993,
1161, 149-160.

Kim, J. E; Pribisko, M. A.; Gray, H. B.; Winkler, J. R. Inorg. Chem. 2004, 43,
7953.

40

2. Experimental Procedure

2.1. Introduction

In order to study protein and solvent dynamics on the picosecond—nanosecond
timescale by time resolved fluorescence spectroscopy, we have employed the time-
correlated, single-photon counting (TCSPC) method. This method has been used for
many years with a wide span of applications, including fluorescence lifetime
measurements and anisotropies,l fluorescence resonant energy transfer (FRET),
fluorescence correlation spectroscopy (F CS),2’3 and determination of solvation dynamics
in polar and nonpolar liquids.4 Today the TCSPC method is widely used in combination

with scanning microscopes for cell imaging.5

With the development of femtosecond laser techniques such as photon echo
peak-shift spectroscopy, solvation dynamics measurements employing the picosecond
TCSPC method have appeared in the literature less frequently because the focus in
liquids has been on the very short timescales. In this dissertation we show that the
TCSPC method can be of exceptional value for sensing the protein and solvent dynamics
of diffusive character on the 100 ps to 50 ns time scale, when used with an intrinsic
electronic chromophore in a globular protein. This chapter provides information about the
preparation of ZnCytc, the protein of interest in this dissertation. We also outline the

basics of TCSPC and provide details of the experimental apparatus and data analysis.

41

2.2. Sample Preparation

ZnCytc samples were prepared in the Vanderkooi laboratory at the University of
Pennsylvania School of Medicine in Philadelphia. The procedure of Vanderkooi and
co-workers6 was employed to remove the native Fe(II) ion and replace it with a Zn(II)
ion. A horse-heart cytochrome c (Sigma) was treated with anhydrous hydrogen fluoride
in order to extract the Fe(II) ion from the porphyrin ring. The change in color of the
protein solution from red to purple indicates completion of this reaction. The protein
solution contained free porphyrin and intact protein, which were separated on the
Sephadex G50 gel-filtration column. The absorption spectrum of an iron-free cytochrome

c sample obtained in the Vanderkooi laboratory is presented in Figure 2.1.

 

 

Soret

2.5

'5

8 Trp 59

E

-e

O.)

.‘E

E

o

2 Q

0 —
l l I I l
300 400 500 600 700
Wavelength (nm)

Figure 2.1. Continuous-wave absorption spectrum from iron-free cytochrome c, in water
at room temperature. The spectrum shows the characteristic four bands in the Q region.

42

The band arising from Trp 59 is observed at 280 nm. The very intense Soret (or
B) band occurs at 404 nm. The Soret band arises from porphyrin It —> If transitions to

the manifold of higher excited states, Sn (n22). The four Q bands, arising from
porphyrin tr —> if transitions to the SI state, occur further to the red, in the 500—630 nm

region of the spectrum. The ZnCytc was prepared by adding zinc acetate to a solution of
iron-free cytochrome c followed by heating in water bath at 50° C. Measuring the
absorption spectrum of the solution in short time intervals allows monitoring of the
progress of reaction. The four Q bands, characteristic for absorption spectrum of iron-free
cytochrome c, collapse into two bands upon the introduction of the Zn(II) ion into the
porphyrin ring (see Figure 2.2). The Soret band red shifts in comparison to iron-free

species, and occurs at 420 nm. The lower energy Q band in the absorption spectrum of

 

 

 

 

Soret
Z:
'5
r:
8
5
"e
<1)
.E
E
e
Z Q
0—
l I I l l
300 400 500 . 600 700
Wavelength (nm)

Figure 2.2 Continous-wave absorption spectrum from ZnCytc, in water at room
temperature. The spectrum shows characteristic two bands in the Q region.

43

ZnCytc arises from the So(v = 0) —) S1(v = O) Vibronic transition and occurs at 580 nm.

The higher energy Q band arises from S0(v=0)—>S1(v=1) transition and occurs at

547 nm.

The purity of the Zn(II)-substituted species following the reaction of the
metal-free form with Zn(II) was determined from a characterization of the absorption and
fluorescence spectra. The sample was dialyzed to remove ammonium acetate,

concentrated, lyophilized and stored in the dark at -—10° C prior to use.

2.3. Modeling of Continuous Wave Absorption and Fluorescence
Emission Spectra of ZnCytc

The representative absorption and fluorescence spectra of ZnCytc (see Figure
2.3) are plotted with respect to the wavenumber v as A(v)/v and F (v)/ v3 , respectively,
with normalization to the unit area. The integrals of these quantities report the dipole
strength, the square of the transition-dipole moment.3’7 The spectral region shown in

Figure 2.3 features the Q bands, which arise from if —> rr‘ transitions between the ground
and first-excited singlet states. The Q bands exhibit a resolved Vibronic structure arising
from a progression in a methane-stretching mode that is polarized in the plane of the

porphyrin macrocycle.8

The absorption dipole-strength spectra D A (v) from ZnCytc can be effectively

described as a Vibronic progression in one mode of frequency VA :

DA (V) = 40040004 V00) + A01401041’00 + VA ) + 40240206 V00 + 2VA) (2- 1)

The scaling parameters 210,- represent the dipole strengths for the O—i Vibronic

transitions, each of which is described by a normalized log-normal (asymmetric

Gaussian)9 line-shape function L0i(v,vo), with v0 defining the center of the line.
Similarly, a good model for the fluorescence dipole-strength spectrum DF(v) from

ZnCytc employs a Vibronic progression in one mode of frequency up, with the normalized

log-normal line shape L0i(v, v0) oriented so that the broader side points to lower energy:

Dr (V) = 1“0040004 V00) + F01401061’00 - VF) + Fozloz (V, V00 - ZVF) (23)

Relative [Dipole Strength

<3
1

 

 

l l l T l
18 20 22 24 26

Energy (103 cm’l)

Figure 2.3. Continous-wave absorption spectrum (dots) from ZnCytc, in water at room
temperature. The spectrum is plotted as relative dipole strength and normalized to unit
area. Superimposed with the absorption spectrum is a model (line) described with

equation 2.1.

45

The scaling parameters FOi are the integrated dipole strengths for the Vibronic lines in the

fluorescence spectrum. Figure 2.4 shows the normalized fluorescence spectrum of

ZnCytc.

The spectra shown in Figures 2.3 and 2.4 are superimposed with fitted curves that
were obtained by a nonlinear least-squares regression to the models defined by equations
2.1 and 2.2. We find that an adequate description of the absorption or fluorescence
spectra from ZnCytc can be obtained by using the same line-width and asymmetry

parameters for LOi in all three Vibronic components. Good fits are obtained with the

asymmetry parameter set to 1, which corresponds to the Gaussian line shape. As will be
discussed in the next section, we use the model for the fluorescence spectrum as a starting

point for a modeling of the time-resolved fluorescence spectra.

Relative Dipole Strength

 

l l l l l
l3 14 15 l6 17

Energy (103 cm")

Figure 2.4. Continous-wave fluorescence spectrum (dots) from ZnCytc, in water at room
temperature. The fluorescence spectrum was obtained with excitation at 585 nm. The
spectrum is plotted as relative dipole strength and normalized to unit area. Superimposed
with the fluorescence spectrum is a model (line) described with equation 2.2.

46

2.4. The TCSPC Method

TCSPC is a photon counting method that uses a pulsed light source, such as a
mode-locked laser, to excite an ensemble of molecules in the sample. One of the basic
prerequisites for this method is that it operates under the condition of low excitation
power to insure only single photon absorption and detection of single emission events.
The excited molecules relax by emitting fluorescence photons with an exponential
distribution of time delays. The resulting fluorescence response function is commonly

described with the equation
Ip<t)=1e"/’, (2.3)
2'

where 1F (t) is the fluorescence intensity at some time t, and z' is fluorescence lifetime.‘0

In the TCSPC method the fluorescence response function is obtained by
measuring the time delays of the fluorescence photons with respect to the excitation
pulse. The excitation pulse is detected by a fast photodiode and serves as a start signal for
counting photons that are time-correlated to the excitation pulse. This start signal triggers
the voltage ramp in the time to amplitude converter (TAC), which is an essential part of
the setup. It transforms the arrival time between a start and a stop pulse into a voltage.
The ramping stops when the photomultiplier tube (PMT) registers the first fluorescence
photon from the sample. The constant fraction discriminator (CF D) is used to minimize
the noise and guarantee that the timing between the start and stop pulses is independent

of the intensity of the signal pulse. The analog-to-digital converter (ADC) converts

47

output voltage from the TAC into a time channel in a multichannel analyzer (MCA). This
process continues until the MCA builds up a probability histogram of the number of
counts as a function of the time delay. For high repetition rate excitation sources, the
TAC does not have enough time to reset between the two excitation pulses and the
resulting fluorescence decay curve may be distorted. To avoid this, the TAC is usually
operated in the reverse mode, where the first detected fluorescence photon serves as the

start signal, and the subsequent excitation pulse as the stop signal.

With an infinitely narrow excitation pulse and infinitely fast detection electronics,
the measured and real fluorescence response functions would be the same. In reality, the
molecules in the sample that are excited at early times are still emitting fluorescence
while the rest of the molecules are excited with the tail of the excitation pulse. This
results in distortions of the measured fluorescence response at early times. To extract the
early-time fluorescence response, it is necessary to deconvolute the instrument response
function from the measurement. The instrument response fimction is collected under the
same conditions as an actual measurement, but with a scattering solution instead of the
sample. It represents the shortest time profile that can be measured with a given
instrument. The TCSPC method is characterized by high resolution and sensitivity, and

excellent signal-to-noise ratio.

2.4.1. The TCSPC Apparatus

The single-wavelength fluorescence transients were acquired with a time-

correlated, single-photon-counting system operated in the reverse-triggered mode. The

48

schematic of the experimental apparatus is shown in Figure 2.5. The excitation source
was a cavity-dumped rhodamine-6G dye laser (Coherent 702-1), which was
synchronously pumped by the 532-nm second harmonic of a mode-locked Nd31-YAG
laser (Coherent Antares 76-8). The saturable—absorber jet of the dye laser was not used.
The cavity dumper was adjusted to produce a 4-MHz pulse-repetition rate; the zero-
background autocorrelation width of the emitted pulses was 10 ps, as measured using an

Inrad 5-14A autocorrelator.

The excitation laser light was passed through an attenuator composed of an
achromatic half-wave retarder and a New Focus 5521 calcite/fused-silica polarizer; a
BK7 lens (5-cm focal length) was used to focus the beam onto the sample position. The
fluorescence emission was collected by another BK7 lens (5-cm focal length), passed
through a Glan-Thompson calcite polarizer (CV1 Laser CPLG-10.0), dispersed with a
0.125-m double-subtractive monochromator (CVI CM112), and detected with a
microchannel-plate photomultiplier tube (Hamamatsu R3809U). The spectral band pass
of the monochromator was 4 nm. To obtain dichroism-free signals, the plane of
polarization of the excitation laser beam was rotated to the magic angle, 54.7°, with

respect to the plane analyzed by the polarizer in the fluorescence beam.3

For timing of the fluorescence and excitation photons, reference pulses were split
from the dye laser output, passed through a fused-silica neutral-density filter and an 80-m
fiber optic delay line, and then focused onto a silicon avalanche photodiode
(RCA C30902E). The arrival times of the reference and fluorescence photons were

determined with a constant fraction discriminator (CFD, Oxford TC 454). The CFD

49

 

/ NdIYAG

 

 

 

 

Dye—laser I—___a\

 

| PC } l
[WWI—[CFDH TAC HCFDJ—x

MONOCHROMATOR
-]

 

 

 

 

 

 

 

 

1 AA
is 1.!

Figure 2.5 Schematic for the time-correlated single-photon-counting apparatus: CFD,
constant fraction discriminator; TAC, time-to-amplitude converter; MCP PMT,
multichannel plate photomultiplier; S, sample.

output pulses corresponding to the reference photons were sent to a time-to-amplitude
converter (TAC, Canberra 2145). The CFD output pulses corresponding to the
fluorescence photons were delayed (Tennelec 412A) and then sent to the TAC. The TAC

output pulses were pulse-height analyzed by a multichannel analyzer (MCA, Canberra

50

MPT-16E). A LabVIEW (National Instruments) program controlled the MCA and the
monochromator, allowing the unattended collection of single-wavelength fluorescence
transients at different emission wavelengths. A total of 106 fluorescence photons were

accumulated for each transient at an effective counting rate of 10 kHz.

2.4.2. Data Analysis

The time evolution of the fluorescence spectra from ZnCytc was characterized by
acquiring a fluorescence time—wavelength—intensity surface. For a given sample, the
surface was obtained by recording a set of the TCSPC fluorescence transients at emission
wavelengths spanning most of the fluorescence spectrum. An example of a fluorescence

transient with a corresponding instrument response function is shown in Figure 2.6.

Counts

 

 

 

 

 

o-l

Time (ns)

Figure 2.6 Typical single-wavelength transient and instrument response function obtained
in TCSPC experiment.

51

Relative Dipole Strength

 

 

l 0
clay (ns)

20
h (Hm) 700 40 3o TimeD

Figure 2.7 Typical time-wavelength-intensity surface constructed item the single
wavelength transients collected in TCSPC experiment.

For each experimentally acquired data set, a given transient recorded at a certain
emission wavelength contributes one column to a matrix of intensities that composes the

I A'— ' A .L

, surface (see Figure 2.7). The grid spacing along the

“mm
“In" v — '-'-°_

wavelength direction corresponds to the band pass of the emission monochromator; the

grid spacing along the time direction is 12 ps, the effective dwell time for the MCA. The

52

relative intensity of each transient is scaled so that its integral with respect to time is
proportional to the relative intensity at the observation wavelength in the continuous-
wave fluorescence spectrum.ll The transients were used without deconvolution of the
instrument-response function (60 ps fwhm) because our interest here is in the response in
the >100—ps regime. Time-resolved spectra are obtained directly as the rows of the

intensity matrix.

Figure 2.8 shows a representative time-resolved fluorescence spectrum at a
selected time delay from ZnCytc in water at pH 7.5. To characterize any changes in the
time resolved fluorescence spectra, we fit the entire matrix of intensities as time-resolved
spectra to the model described in equation 2.2. To do so we wrote procedure in the
IgorPro program that allows us to scan through the entire time—wavelength—intensity
surface modeling the sum of lognorrnal emission lineshapes to the time-resolved spectra
obtained from slices of the surface. For most experiments shown in this dissertation, the

mode frequency (VF) and line-shape parameters were fixed to those obtained from the

analysis of the continuous-wave spectrum. Allowing these parameters to vary during the
analysis did not result in improved fits, nor were trends observed in the floating
parameters with respect to time, so we conclude that the shape and separation of the
Vibronic components did not evolve significantly during the time course from 100 ps
onward. Use of fixed line-shape and mode-frequency parameters in the modeling of the

time-resolved spectra permits a relatively precise estimation of F00, even when the

spectral range does not include the entire width of the 0—0 Vibronic feature. For the

spectra obtained with Q-band excitation the analyzed spectral range begins one spectra]

53

band pass to the red of the excitation wavelength. Examples of the fitted spectra are

shown in Figure 2.8 superimposed with the data points of the three time-resolved spectra.

The analysis of the time-resolved spectra allows for tracking of the intensity and
the position of the Vibronic lines with respect to time. The change in the intensities of the
Vibronic lines with respect to time is discussed in detail in next chapter. The change of
the position of the 0—0 fluorescence transition with respect to time, which reports the
time-scales for the reorganization of the surrounding solvent in terms of Dynamic Stokes

shift, is discussed in chapters 4 and 5.

Relative Intensity
.8 .8

o
l

 

0-1

 

l l l l I I
14.0 14.5 15.0 15.5 16.0 16.5 17.0
Frequency (103 cm'l)

Figure 2.8 Typical time-resolved fluorescence spectrum obtained from time-wavelength-
intensity surface (data points). Superimposed with the time-resolved fluorescence
spectrum is a model (line) described with equation 2.2.

54

2.5.

(1)

(2)

(3)

(4)
(5)
(6)

(7)

(8)

(9)
(10)

(11)

References

Chen, L. X.-Q.; Petrich, J. W.; Fleming, G. R.; Perico, A. Chem. Phys. Lett. 1987,
139, 55—61.

Fleming, G. R. Chemical Applications of Ultrafast Spectroscopy; Oxford
University Press: New York, 1986.

Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Second ed.; Kluwer
Academic/Plenum Publishers: New York, 1999.

Maroncelli, M.; Fleming, G. R. J. Chem. Phys. 1987, 86, 6221—6239.

Berland, K. M.; So, P. T. C.; Gratton, E. Biophys. J. 1995, 68, 694—701.

Vanderkooi, J. M.; Adar, F.; Erecinska, Ml. Eur. J. Biochem. 1976, 64, 381.

Kantor, C. R.; Schimmel, P. R. Biophysical Chemistry, Part II: Techniques for the
study of Biological Structure and Function; Freeman and Co: San Francisco,
1980.

Gouterrnan, M. Optical spectra and electronic structure of porphyrins and related
rings. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978;
Vol. 3; pp 1—159.

Siano, D. B.; Metzler, D. E. J. Chem. Phys. 1969, 51, 1856—1861.

Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum Press, 1983.

Pierce, D. W.; Boxer, S. G. J. Phys. Chem. 1992, 96, 5560.

55

3. Excited-State Axial-Ligand Photodissociation and
Nonpolar Protein-Matrix Reorganization in
Zn(ID-Substituted Cytochrome c'1

3.1. Introduction

Even under conditions that favor the native structure, a protein fluctuates over a
range of thermally accessible states and time scales. The work of Frauenfelder and co-
workers showed early on that an intrinsic chromophore can be used as a probe of these
fluctuations; the motions and structures that a protein accesses on its potential-energy
surface control the probe’s spectroscopic line-shape. The potential-energy surface
(or energy landscape) of a protein resembles that of a glass in exhibiting a rugged
profile,2 at least over short displacements from the vicinity of a given folding state. A
large number of possible paths and intermediate states should contribute to the dynamics
of folding}5 an ensemble view is to be favored over the older concept of a sequential

pathway between a small number of intermediate states.6

Proteins also exhibit some of the properties of liquids, especially at physiological
temperatures.7 Ultrafast spectroscopic techniques have proven effective in the study of
line shape and dynamics in liquids, specifically in studying what kinds of larger-range
motions an ensemble of protein structures makes on its potential-energy surfaces“ The
methods used, femtosecond transient-hole burning spectroscopy and stimulated

photon-echo peak-shift spectroscopy, obtain the ground-excited—state energy-gap time-

 

1 The work in this chapter has been published in other fonnl
56

correlation function, M (t) , from a probe molecule in condensed phases.'2'ls

In this chapter, we begin a discussion of the picosecond—nanosecond time-scale

diffusive part of M (t) in cytochrome c, which contains an intrinsic heme chromophore.

To have access to an intrinsic fluorophore with a well-characterized surrounding protein
structure, we have initially chosen to study the dynamics of the Zn(II)-substituted

”’20 rather than that of the naturally occurring Fe(II,III)-containing molecule”

molecule
The Zn(II)-porphyrin in ZnCytc exhibits a long fluorescence lifetime and a spectrum with

resolved Vibronic structure,‘6 which permits measurement of M (t) in terms of the

dynamic Stokes shift of the time-resolved fluorescence spectrum despite the small

reorganization energy (<200 cm'l).

In performing an initial characterization of the time-resolved fluorescence
spectrum from ZnCytc on the picosecond time scale, however, we were surprised to
observe a pronounced biexponential time evolution of the Vibronic structure. The
mirror-image time evolution of the 0—0 and 0—1 transitions evidences a dissociation of
the Zn(II)-ion’s axial ligands, which are apparently derived in the native state from the
side chains of the nearby Met 80 and His 18 amino acid residues,20 as in the Fe(II,III)
species (see Figure 3.1). The fast phase of the Vibronic structure response that
accompanies release of the first axial ligand is slowed when the external medium is
changed from water to 50% glycerol; these results suggest a rate limitation involving

reorganization of the protein matrix and surrounding solvent.

57

 

Figure 3.1. Structure of22 the heme and axial ligands from horse-heart ferricytochrome c, as
rendered using PyMOL22 from the pdb file II-IRC.”

3.2. Experimental

3.2.1. Zn(II)-Substituted Cytochrome c

ZnCytc samples were prepared by the method of Vanderkooi and coworkers, as
described in detail in chapter 2. The ZnCytc samples were dissolved in a 50-mM N-(2-
hydroxyethyl)piperazine-N’-2-ethanesulfonic acid (HEPES, Research Organics) buffer

solution at pH 7.5 or in a 50% (v/v) glycerol/HEPES buffer solution mixture. The
58

samples were filtered and then diluted in the chosen solution so as to obtain a maximum
absorption of 0.1 in the 0—0 transition of the Q-band for a l-cm path length. The samples
were held in quartz cuvettes with l-cm path lengths and were purged with dry nitrogen

prior to use.

3.2.2. Continuous-Wave Absorption and Fluorescence Instrumentation

Steady-state absorption spectra were acquired with 2-nm spectral band pass with
Hitachi U-2000 spectrophotometer or an ATI Unicam UV2 spectrophotometer. Steady-
state fluorescence spectra were acquired with a Hitachi F-4500 fluorescence spectrometer
(5-nm excitation and emission band pass). The fluorescence intensities were corrected by
the spectrofluorometer’s data-acquisition program using a calibration curve obtained
from a standard lamp. The intensities were recorded in terms of photons per nanometer
band pass. When presented as a function of wavenumber, the fluorescence intensities are
accordingly multiplied by the square of the wavelength to compensate properly for the
fixed (in wavelength units) spectral band pass of the emission spectrometer.23 Scattered
excitation light was eliminated from the fluorescence spectra by subtraction of a blank,

solvent-only sample held in the same cuvette.

3.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy

Single-wavelength fluorescence transients were acquired with a TCSPC apparatus
described in detail in chapter 2. The excitation wavelength used in these experiments was

set to 585 nm so as to correspond to the 0—0 transition energy for ZnCytc spectra in

59

water. Samples used in time-resolved fluorescence measurements were held in l-cm
quartz cuvettes mounted in a temperature-controlled holder (Quantum Northwest TLC
50). All samples were filtered, kept at 22 °C, purged with dry nitrogen, and prepared at
the same concentrations used in steady-state measurements. The instrument-response

function of the detection system was approximately 60 ps (fwhm) in these experiments.

3.3. Results

3.3.1. Continuous-Wave Absorption and Fluorescence Emission Spectra

Figure 3.2. shows the absorption and fluorescence spectra obtained at 22 °C from
ZnCytc at pH 7.5 in water. The spectra are normalized to unit area and plotted as dipole
strength. The absorption and fluorescence dipole strength spectra can be modeled with
line-shape functions described in detail in chapter 2 (equations 2.1 and 2.2). The spectra
shown in Figure 3.2 are superimposed with fitted curves that were obtained by nonlinear

least-squares regression to the models defined by equations 2.1 and 2.2 in chapter 2.

3.3.2. Time-Resolved Fluorescence Spectra

The time evolution of the fluorescence spectra from ZnCytc (see figures 3.3—3.5)
was characterized by acquiring a fluorescence time—wavelength—intensity surface. For a
given sample, the surface was obtained by recording a set of TCSPC fluorescence
transients at emission wavelengths spanning most of the fluorescence spectrum. The
details of this procedure are discussed in chapter 2. Figure 3.3 shows time-resolved

fluorescence spectra at three selected time delays from ZnCytc in water at pH 7.5. During

60

 

 

g .-. -
c:
g .
m
.92
E. o :.
O)
.3 ' .
8 ° °
04
0— /
I I I I I I I I
l3 14 15 l6 17 18 19 20

Energy (103 cm'l)

Figure 3.2. Continuous-wave absorption (right) and fluorescence (left) spectra from
ZnCytc at 22°C. The spectra are plotted as dipole strengths, A(v)/v and F(v)/v3,
respectively, and are normalized to unit area. The spectra are superimposed with fitted
curves employing eqs 1 and 2 as models. The fluorescence spectrum was obtained with
excitation tuned to the 0—0 transition energy, where the absorption and fluorescence
spectra cross.

the fluorescence decay, the relative dipole strength of the 0—1 transition, Q01, increases at
the expense of that of the 0—0 transition, Q00, which is initially larger. To characterize

these and any other changes in the time-resolved fluorescence spectra, we fit the entire
matrix of intensities as time-resolved spectra to the model described in the equation 2.2

chapter 2. The mode frequency (VF) and line-shape parameters were fixed to those

obtained from the analysis of the continuous-wave spectrum. Allowing these parameters

to vary during the analysis did not result in improved fits, nor were trends observed in the

61

 

Relative Intensity
N
O
l

 

0.0 A

1.2—

0.8 —-

0.4 —

 

 

0.0 -4

 

l l l l l I 1
13.5 14.0 14.5 15.0 15.5 16.0 16.5

Energy (103 cm‘ 1)

Figure 3.3. Time-resolved fluorescence spectra from ZnCytc in water at 22°C (data
points) and fitted spectra (smooth curves, see eq 2) at three delays.

62

floating parameters with respect to time, so we conclude that the shape and separation of
the Vibronic components did not evolve significantly during the time course from 100 ps
onward. Use of fixed line-shape and mode-frequency parameters in the modeling of the
time-resolved spectra permits a relatively precise estimation of Q00 even though the
spectral range does not include the entire width of the 0—0 Vibronic feature. The analyzed
spectral range begins one spectral band pass to the red of the excitation wavelength.

Examples of the fitted spectra are shown in Figure 3.3 superimposed with the data points

of the three time-resolved spectra.

The analysis also allows us to track the position of the 0—0 fluorescence transition,

n00, over the entire time scale. A plot of v00 with respect to time reports the dynamic
Stokes shift.”25 In ZnCytc, v00 evidently reports the polar reorganizational dynamics of

the protein that surrounds the Zn(II) porphyrin. We will discuss these details in the next

chapter.

The time evolution of the intensities of the 0—0 and 0—1 Vibronic components in

the time-resolved fluorescence spectra is described by the dipole strengths Q00 and Q01

with respect to time. These parameters are plotted in Figure 3.4 for ZnCytc in water in

terms of relative dipole strengths, QOi/Zj Q0 j , and as the ratio Q01 /Q00. The former

plot normalizes the intensity of the Vibronic component with respect to the total dipole
strength at a given time point, which decays proportionally to the excited-state population
according to the fluorescence lifetime. The latter plot obtains in a single observable the

relative intensity evolution of the two Vibronic lines. A distinct biexponential response

63

was observed in both types of plots. The fit parameters foer/Q00 obtained from

t
Qi,,(f)/Q00(t)= C—Z Aie ltiare listed in Table l. A similar model can be used to fit

separately the normalized Q00 and Q01 responses.

 

 

Table l.
solvent a, T, . p5 02 1’2 , PS C
water 0.50i0.02 1204;50 1.50.4:0.02 7i] 2.1i0.02
50%
0.30i0.02 400i50 O.90i0.02 7:1 1.66:0.02
glycerol

Table 1. The time evolution of Q01 /Q00 for ZnCytc in two solvent systems at 22 °C.

The plot of Q0, /QOO for ZnCytc in 50% glycerol is shown in Figure 3.5. The
time-resolved fluorescence spectra evidence a biexponential rise in the value of Q0, /Q00

over six-tenths of the range that was observed in water only (see Figure 3.4), and the time
constant for the faster of the two exponentials is slowed by a factor of 4 (see Table 1).
Note that the time constant for the slower exponential component and the relative
magnitude of the two exponential components were not significantly affected by the

presence of glycerol.

Although the time—wavelength—intensity surfaces that are described in
Figures 3.3—3.5 and Table l were obtained with excitation close to 0—0 transition energy,
where the fluorescence and absorption spectra cross each other (see Figure 3.2),
additional experiments were conducted with the pump laser tuned farther to the blue by

as much as 700 cm". No effects on the time-resolved spectra nor trends in the fitted

64

 

 

 

 

0.7-
Q01
‘ .0... o ...'o .
0.6T J.."\.}‘ M
r "1‘" ”of... 0. . “
5, r em . o
O. ' - .
W“ 0.5 ‘ ‘ . r
\5 .‘
0 ~ . . Q
0.4- - ‘.' :' ' ' 00
'°' 9 0 cc
@4599. . . .3: . J
o .. 3. ‘g’fijyfi
0.3 -
l I I l l I l
0 2 4 6 8 10 12
1.8- °-'.‘ 5:235" °-
.' ..' 0‘..°;: ~. ' .a‘
‘0'. .. ”0* 0. .0
1.6"" .....:‘: ”M. 4' 3:40.0r‘00 {c . .
3“... ’5’ $.. . .0 .0
1.4— . .: g ' .0 co
8 ‘. s". o '
g 1.2- _ , " '
S
O 1.0-
0.84 '
0.6-1
0.47
l I I I I l T
0 2 4 6 8 10 12
Delay (ns)

Figure 3.4 Time evolution of the dipole strengths for the 0-1 (Q01) and 0-0 (Q00) Vibronic
transitions in time-resolved fluorescence spectra from ZnCytc in water at 22°C. Top: As
normalized by the total dipole strength SjQOj. Bottom: As the ratio Q01 / Q00. The

parameters for the fit to the ratio are listed in Table 1.

65

Q01 /Q00

 

 

 

I l
0 2 4 6 8 10 12

Delay (ns)

_‘
—t
—

Figure 3.5. Time evolution of the ratio of the dipole strengths for the 0-1 (Q01) and 0-0
(Q00) Vibronic transitions in time-resolved fluorescence spectra from ZnCytc in 50%
glycerol at 22°C. The fit parameters are listed in Table 1.

parameters as a function of time were noted. These results show that the time evolution
observed in the Q-band Vibronic structure is not predominantly associated with
vibrational relaxation or a redistribution of vibrational energy; either of these phenomena

would be expected to occur on the < lOO-ps time scale.

3.4. Discussion
We suggest that the time evolution of the Vibronic structure of the Q-band
fluorescence spectrum of ZnCytc can be explained by an excited-state release of axial

ligands from the Zn(II) ion. This suggestion is based on the known effects of ligand

66

binding on the Q-band Vibronic structure. In the following, we discuss a possible
explanation for the biexponential response and the apparent rate limitation owing to

protein-matrix and solvent reorganization that is implied by the experimental results.

3.4.1. Interpretation of Vibronic Structure in Zn(II)-Porphyrin Complexes

The presence of axial ligands in Zn(II)-porphyrin systems can be monitored using
the partially resolved Vibronic structure of the Q-band absorption and fluorescence
spectra. The effect of binding axial ligands on the absorption and fluorescence spectra of
Zn(II) porphyrins is a redistribution of dipole strength between 0—0 and 0—1 Vibronic
transitions. In equilibrium titration studies in which a ligand is added to a
Zn(II)-porphyrin species dissolved initially in a noncoordinating solvent, the intensity of
the absorption or fluorescence 0—1 transition decreases and the intensity of the 0—0
transition increases in response to an increase in the concentration of the ligand. Both
intensity responses are initially linear with respect to the concentration of the ligand; at
higher ligand concentrations, the responses saturate. The entire response as a function of

ligand concentration defines the equilibrium constant for the ligand-binding reaction.”28

The time-resolved fluorescence spectra from ZnCytc exhibit changes in Vibronic
structure that closely resemble the spectral evolution observed in the Zn(II)-porphyrin
ligand-titration studies as the concentration of ligand decreases. Figure 3.4 shows that

Q00 and Q01 dipole strengths mirror each other, with Q00 decreasing and Q01 increasing

in intensity with respect to time. These results strongly suggest that a net

photodissociation of the axial ligands occurs in the excited state. The spectral response

67

should be interpreted as a shift of the equilibria that link the ligated and deligated
structures. The response shown in Figures 3.4 and 3.5 arises from the transient flow of

population that accompanies relaxation toward the new equilibrium constant.

3.4.2. Origin of the Biexponential Axial-Ligand Photodissociation Response
in ZnCytc

The simplest explanation for the biexponentiality of the kinetics observed for the
time evolution of the Q01 /Q00 ratio involves a dissociation of two axial ligands:

I(I K2
La —Zn(P)—Lb;-‘_La +Zn(P)—-Lb;‘:La +Zn(P)+Lb (3.1)

In equation 3.1, the axial ligands L, and LI) are initially coordinated to the Zn(II)

porphyrin, Zn(P); the two equilibria govern the sequential dissociation of the two axial
ligands. The exponential components observed for the time evolution of the

Q01 /Q00 ratio report the relaxation time constants for the two equilibria.

This suggestion conflicts, however, with the standard picture that Zn(II)
porphyrins can bind only one axial ligand. It is generally thought that Zn(II) porphyrins
prefer a five-coordinate ligand environment in solution, with the Zn(II) ion domed from
the plane of the porphyrin on the side of the single axial 1igand,26‘29'31 but there are some
structures in crystals that coordinate two axial ligands.”34 Crystal packing forces impose
two axial ligands upon the Zn(II) ion, a nonplanar distortion of the porphyrin macrocycle

results, and the Zn(II) ion is positioned closer to the nominal plane of the porphyrin.”33

68

A similar situation is likely to exist in the native conformation of ZnCytc. The
structure of the polypeptide chain in chytc has been characterized by two groups with
2D-NMR methods.'9’20 Anni et a1.20 conclude that the polypeptide in ZnCytc assumes the
same configuration that exists in Fe(II) protein. The proton resonances for the
polypeptide backbone and the side chains of His 18 and Met 80 indicate that these
residues are located as they are in the Fe(II) protein; the resonances of residues elsewhere
in the protein are the same in the Fe(II) and Zn(II) forms, showing that the configuration
of the polypeptide is identical in the two forms. The same conclusion was reached
recently by Quian et a1.19 Thus, the polypeptide projects the His 18 and Met 80 side
chains in the correct configuration to permit them to serve as axial ligands to the Zn(II)
ion, and importantly, no other amino acids are ligated to the Zn(II) ion. Anni et al.
conclude on the basis of their NMR results and their studies of low-temperature
absorption and fluorescence spectra that the Zn(II) ion is ligated by His 18 and Met 80 in

the same geometry that is present in the Fe(II)—containing molecule.20

Resonance Raman measurements of porphyrin core marker bands by Kostié and
co-workers, however, suggest that the porphyrin core is not sufficiently expanded in
ZnCytc to permit an orthodox six-coordinate configuration.18 The same group accounts
for their finding that the Marcus reorganization energy for electron self-exchange
between ZnCytc and ZnCytc+ is lower than that obtained for the Fe(II) species by

suggesting that Met 80 does not act as an axial ligand to the Zn(II) ion.3s

Having considered all of these results, we conclude that it is probable that the

ligand coordination sphere for the Zn(II) ion in ZnCytc is six-coordinate, as imposed by

69

the essentially native structure of the cytochrome c polypeptide. Incorporation of the
Zn(II) ion in the porphyrin is, however, likely to result in axial-ligand bonding
geometries that are somewhat different fi'om those present in the Fe(II)-containing
protein. The resulting porphyrin structure should be strained, as in the distorted six-

coordinate Zn(II) porphyrins that have been observed in crystal structures.

We suggest that our results are consistent with this hypothesis. The two

exponential components detected in the response of the Q01 /Q00 dipole-strength ratio
are widely separated in time scale (r1 = 120 ps, I, = 7 ns; see Figures 3.4 and 3.5 and

Table l). A six-coordinate ground-state species would yield a five-coordinate
intermediate upon breaking the first axial-ligand bond, perhaps that to the Met 80 side
chain. The five-coordinate intermediate would be expected to be less strained and
accordingly less reactive than the six-coordinate starting structure. Owing to motion of
the departing ligand away from the Zn(II) ion, the packing forces that were imposed upon
the porphyrin by the two axial-ligand side chains would be relaxed, the Zn(II) ion would
dome, and the porphyrin might be expected to assume a more stable, planar
configuration. Thus, the relaxation of the second axial-ligand equilibrium in equation 3.1

might well be expected to occur on a much slower time scale.

3.4.3. Protein-Matrix and Solvent Reorganization Dynamics

The suggestion that the surrounding protein structure applies packing force upon
the Zn(II) porphyrin in ZnCytc and establishes two axial ligands also implies that a
reorganization of the protein structure would be required in response to axial-ligand

70

photodissociation events. The finding that the fast phase of the Q01 / Q00 dipole-strength

response in ZnCytc is slowed by the presence of 50% glycerol in the surrounding solvent
is consistent with this idea. As a result of the photodissociation of either of the two
ligands, a displacement of the ligand donor atoms away from the plane of the porphyrin
would be expected to occur only as fast as the surrounding protein and solvent could
accommodate the change in the effective volume. The rate of reorganization of the
protein and the surrounding solvent would be expected to be slower when glycerol is
present, owing to the change in the bulk viscosity. It is interesting that only the 120-ps

component of the Q01 /Q00 dipole-strength response is slowed by the addition of glycerol;

it could be that the breaking of the second axial-ligand bond occurs so slowly from the
more relaxed five-coordinate intermediate that reorganization of the surrounding solvent

medium is not a significant rate limitation.

The effect of glycerol on the total range of change of the Q0| / Q00 ratio suggests

that the overall equilibrium constant for the ligand-dissociation sequence is somewhat
smaller in the presence of glycerol in the external solvent medium than in the presence of
water only. One might expect that the forward, ligand-dissociating and reverse, ligand-
binding reactions would exhibit a comparable dependence on the solution viscosity
owing to microscopic reversibility. An observation of an effect of glycerol on the
equilibrium constant suggests, then, that glycerol has an additional effect on the reaction
that does not involve solvent friction. It is possible that the change in equilibrium
constant reflects a subtle change in the structure of cytochrome c that is associated with

the lowered polarity of the glycerol/water solvent mixture.

71

We suggest that the reorganization of protein and surrounding solvent that
accompanies the excited—state dissociation of axial ligands in ZnCytc is an example of
nonpolar solvation, a response of the medium to a change in size of the imbedded
probe.”43 The effects noted here upon a change of the external solvent viscosity are
probably rather different in character from those discussed by Eaton and co-workers(refs)
and Fayer and coworkers44 in their studies of protein dynamics associated with photolysis
of carbon monoxide in myoglobin. These studies considered the effects of solvent and
internal protein friction on the protein reorganization that accompanies the
conformational relaxation of the Fe(II) porphyrin upon photodissociation of the axially
coordinated carbon monoxide ligand. The axial-ligand response discussed in this chapter
involves a direct displacement of the polypeptide backbone away from the porphyrin by
the side chains of the amino acids that serve as ligands to the Zn(II) ion. An increase in
the external solution viscosity would be expected to retard this process because the
associated internal pressure increase would be expected to result in an expansion of the

protein’s solvated volume.

In addition to the nonpolar response, ZnCytc should also exhibit a polar response

to the excited state change in dipole moment that occurs upon the ground-to-excited state

It —) Ir‘ transition. We will discuss in the next chapter the details of polar salvation in

ZnCytc and its possible coupling to the axial-ligand photodissociation dynamics.

72

3.5. References

(l) Lampa-Pastirk, 8.; Beck, W. F. J. Phys. Chem. B. 2004, 108, 16288.

(2) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598-1603.

(3) Onuchic, J. N.; Luthey-Schulten, Z.; Wolynes, P. G. Annu. Rev. Phys. Chem.
1997, 48, 545—600.

(4) Dill, K. A.; Chan, H. S. Nature Struct. Biol. 1997, 4, 10—19.

(5) Chan, H. S.; Dill, K. A. Proteins: Struct. Funct. Genet. 1998, 30, 2—33.

(6) Dill, K. A. Protein Sci. 1999, 8, 1166—1180.

(7) Vanderkooi, J. M. Biochim. Biophys. Acta 1998, 1386, 241—253.

(8) Riter, R. E.; Edington, M. D.; Beck, W. F. Inertial protein-matrix solvation of a
light-harvesting chromophore. In Ultrafast Phenomena X; Barbara, P., Knox, W.,
Zinth, W., Fujimoto, J ., Eds.; Springer-Verlag: Berlin, 1996; pp 324—325.

(9) Homoelle, B. J.; Edington, M. D.; Diffey, W. M.; Beck, W. F. J. Phys. Chem. B
1998, 102, 3044—3052.

(10) Homoelle, B. J.; Beck, W. F. Biochemistry 1997, 36, 12970—12975.

(11) Jordanides, X. J.; Lang, M. J.; Song, X.; Fleming, G. R. J. Phys. Chem. B 1999,
103, 7995—8005.

(12) Loring, R. F.; Yan, Y. J.; Mukamel, S. J. Chem. Phys. 1987, 87, 5840—5857.

(13) J00, T.; Jia, Y.; Yu, J.-Y.; Lang, M. J.; Fleming, G. R. J. Chem. Phys. 1996, 104,
6089—6108.

73

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

Fleming, G. R.; Cho, M. Annu. Rev. Phys. Chem. 1996, 4 7, 109—134.
Everitt, K. F.; Geva, E.; Skinner, J. L. J. Chem. Phys. 2001, 114, 1326.
Vanderkooi, J. M.; Adar, F.; Erecinska, M. Eur. J. Biochem. 1976, 64, 381.

Vanderkooi, J. M.; Landesberg, R.; Hayden, G. W.; Owen, C. S. Eur. J. Biochem.
1977, 81, 339.

Ye, S.; Shen, C.; Cotton, T. M.; Kostic, N. M. J. Inorg. Biochem. 1997, 65, 219—
226.

Quian, C.; Yao, Y.; Tong, Y.; Wang, J.; Tang, W. J. Biol. Inorg. Chem. 2003, 8,
394—400.

Anni, H.; Vanderkooi, J. M.; Mayne, L. Biochemistry 1995, 34, 5744—5753.
Bushnell, G. W.; Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 585—595.

Delano, W. L. The PyMOL molecular graphics system; DeLano Scientific: San
Carlos, California, 2002.

Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum Press, 1983.
Maroncelli, M. J. Mol. Liq. 1993, 5 7, 1—37.

Stratt, R. M.; Maroncelli, M. J. Phys. Chem. 1996, 100, 12981—12996.
Humphry-Baker, R.; Kalyanasundaram, K. J. Photochem. 1985, 31, 105—112.
Kadish, K. M.; Shiue, L. R. Inorg. Chem. 1982, 21, 3623—3630.

Miller, J. R.; Dorough, G. D. J. Am. Chem. Soc. 1952, 74, 3977—3981.

74

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

(42)

(43)

Nappa, M.; Valentine, J. S. J. Am. Chem. Soc. 1978, 100, 5075—5080.

Vogel, G. C.; Beckmann, B. A. Inorg. Chem. 1976, 15, 483—484.

Scheidt, R. W. Porphyrin stereochemistry. In The Porphyrins; Dolphin, D., Ed.;
Academic Press: New York, 1978; Vol. 3; pp 463—511.

Scheidt, R. W.; Eigenbrot, C. W.; Ogiso, M.; Hatano, K. Bull. Chem. Soc. Jpn.
1987, 60, 3529—3533.

Schauer, C. K.; Anderson, 0. P.; Eaton, S. S.; Eaton, G. R. Inorg. Chem. 1985, 24,
4082—4086.

Fleischer, E. B.; Miller, C. K.; Webb, L. E. J. Am. Chem. Soc. 1964, 86, 2342—
2347.

Cmogorac, M. M.; Kostic, N. M. Inorg. Chem. 2000, 39, 5028-5035.

Berg, M. J. Phys. Chem. A 1998, 102, 17—30.

Berg, M. J. Chem. Phys. 1999, 110, 8577—8588.

Bagchi, B. J. Chem. Phys. 1994, 100, 6658—6664.

Ladanyi, B. M.; Stratt, R. M. J. Phys. Chem. 1996, 100, 1266—1282.

Saven, J. G.; Skinner, .l. L. J. Chem. Phys. 1993, 99, 4391—4402.

Kang, T. J.; Yu, J.; Berg, M. Chem. Phys. Lett. 1990, 174, 476—480.

Fourkas, J. T.; Benigno, A.; Berg, M. J. Chem. Phys. 1993, 99, 8552—8558.

Fourkas, J. T.; Berg, M. J. Chem. Phys. 1993, 98, 7773—7785.

75

(44) Rector, K. D.; Rella, C. W.; Hill, J. R.; Kwok, A. S.; Sligar, S. G.; Chien, E. Y.
T.; Dlott, D. D.; Fayer, M. D. J. Phys. Chem. B 1997, 101, 1468—1475.

76

4. Polar Solvation Dynamics in Zn(ID-Substituted
Cytochrome c: Diffusive Sampling of the Energy
Landscape in the Hydrophobic Core and Solvent-Contact
LayerI

4.1. Introduction

The potential-energy surface (or energy landscape) of a protein is thought to have
a rugged profile like that of a glass,2 at least over short displacements from a given
folding state. But at physiological temperatures, proteins are comparable to molecular
liquids in exhibiting local flexibility and in allowing internal diffusion of small
molecules.3 On a local scale, the amino acids and their side chains perform hindered
internal reorientational motions and torsions with respect to the polypeptide backbone;
over a longer range, the polypeptide undergoes conformational and folding motions. An
additional complexity arises from the interface between the protein and the surrounding
aqueous bulk, which hinders motions of the solvated regions of the polypeptide as it

moves between folding states.

In this chapter, we show how picosecond time-resolved fluorescence spectroscopy
can be used with an intrinsic electronic chromophore embedded in the hydrophobic core
of a globular protein to sense the liquid-like diffusive polypeptide motions that contribute

to the random search of the protein’s potential-energy surface. We consider here the

 

1 The work in this chapter has been published in other form“)
77

polar protein and solvent dynamics in ZnCytc. The Zn(II) porphyrin in ZnCytc is

fluorophore with a long lifetime rF > 2ns ,4'5 which permits us to measure the dynamic

Stokes shift, the red shift of the fluorescence spectrum with respect to time, over the 100-
ps to 12-ns regime. The dynamic Stokes shift describes the reorganization of the protein
and solvent dipoles that occurs in response to the ground-to-excited-state change in
dipole moment of the Zn(II)—porphyrin chromophore. The fluorescence spectrum of
ZnCytc exhibits some resolved Vibronic structure,4 which permits the solvent-response
function to be measured with a good signal/noise ratio despite the small reorganizational
energy (<200 cm"'). The structure of ZnCytc has been studied with 2D-NMR methodsf”7
and continuous absorption and fluorescence spectra from ZnCytc have been extensively
characterized.4’6'8’9 Wiersma and co-workers5 used photon-echo spectroscopy to

characterize the glass-like protein dynamics in ZnCytc that are observed at 2 K.

In the previous chapter we reported the changes in Vibronic structure that are
observed in ZnCytc’s time-resolved fluorescence spectrum owing to excited-state
photodissociation of the Zn(II) ion’s axial ligands, which are provided by the side chains
of Met 80 and His 18 (see Figure 4.1). The release of the ligands was shown to be
sensitive to the external bulk viscosity; in 50% (v/v) glycerol, the time constant for the
faster of two exponential components was slowed from 120 ps to 400 ps. The
axial-ligand reactions are evidently rate limited by the reorganization of the protein and
the surrounding solvent. The motions in question can be considered to be a nonpolar
solvation response to a change in the size of an embedded probe.'°'l6 In this chapter we

report that polar reorganizational dynamics of ZnCytc are not strongly coupled to the

78

 

Figure 4.1 Structure of the horse-heart ferricytochrome c, as rendered using PyMOL”
fiom pdb file ll-IRC.l8 Top: Ribbon representation showing the polypeptide’s secondary
and folded structure, with heme and axial ligands to the Fe(IIl) ion, Met 80 and His 18,
rendered as tube structures. Bottom: Solvent-contact surface, with the heme as a space-
filling (CPK) representation.

79

nonpolar response because they occur on a significantly different time scale. The
dynamic Stokes shift of the time-resolved fluorescence spectrum is biexponential, with
only the slower phase of relaxation showing sensitivity to the presence of glycerol in the
external solvent medium. We discuss the observed response in terms of the motions of

the hydrophobic core and of the solvent-contact layer of the folded protein.

4.2. Experimental

4.2.1. Zn(II)-Substituted Cytochrome c

ZnCytc samples were prepared in the Vanderkooi laboratory at the University of
Pennsylvania School of Medicine in Philadelphia by the method described in detail in
chapter 2. For the experiments in this chapter ZnCytc samples were dissolved in a 50 mM
N—(2-hydroxyethyl)piperazine-N’-2-ethanesulfonic acid (HEPES, Research Organics)
NaOH buffer solution at pH 7.5 or in a 35% or 50% (v/v) glycerol/HEPES buffer solution
mixture. The samples were filtered and diluted in the chosen solution so as to obtain a
maximum absorption of 0.1 in the 0—0 transition of the Q-band for a l-cm path length.
The samples were held in quartz cuvettes with l-cm path lengths and were purged with

dry nitrogen prior to use.

4.2.2. Continuous Wave Absorption and Fluorescence Instrumentation

Steady-state absorption spectra were acquired with a 2-nm spectral band pass with
a Hitachi U-2000 spectrophotometer or an ATI Unicam UV2 spectrophotometer. Steady-

state fluorescence spectra were acquired with a Hitachi F-4500 fluorescence

80

spectrophotometer (5-nm excitation and emission band pass). The fluorescence intensities
were corrected by the spectrofluorimeter’s data acquisition program using a calibration
curve obtained from a standard lamp. The intensities were recorded in terms of
photons/nm band pass. When presented as a function of wavenumber, the fluorescence
intensities are accordingly multiplied by the square of the wavelength in order to
compensate properly for the fixed spectral band pass of the emission spectrometer.
Scattered excitation light was eliminated from the fluorescence spectra by subtraction of

the spectrum from a blank, solvent-only sample held in the same cuvette.

4.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy

Single-wavelength fluorescence transients were acquired with a TCSPC apparatus
operated in the reverse-triggered mode. The details of the method and apparatus were
discussed in chapter 2. For the experiments discussed in this chapter the excitation

wavelength was set to 585 nm , the 0—0 transition energy for ZnCytc spectra in water.

Samples used in time-resolved fluorescence measurements were held in l-cm
quartz cuvettes mounted in a temperature-controlled holder (Quantum Northwest TLC
50). All samples were filtered, kept at 22 °C, purged with dry nitrogen, and prepared at
the same concentrations used in steady-state measurements. The instrument-response

function of the detection system was approximately 60 ps (fwhm) in these experiments.

81

4.3. Results

4.3.1. Continuous-Wave Absorption and Fluorescence Emission Spectra
Figures 4.2 and 4.3 show the absorption and fluorescence spectra obtained at
22 °C from ZnCytc at pH 7.5 in water and in 50% (v/v) glycerol, respectively. The
spectra are plotted with respect to wavenumber v as A/ v and F / v3 , respectively, with
normalization to unit area. The integrals of these quantities report the dipole strength, the

19-

square of the transition-dipole moment. 2‘The spectral regions shown in Figures 4.2 and

4.3 feature the Q bands, which arise from It —) 72" transitions between the ground and

 

 

 

 

29o cm‘I

in
I; \
CD
£2
8‘ n
a l
32’
E "000
0)
Ci

0 —

 

 

14 15 16 17 18 19 20
Energy (103 cm'l)

Figure 4.2. Continuous-wave absorption and fluorescence emission spectra from ZnCytc
in water at room temperature. The spectra are plotted as relative dipole strengths and are
normalized to unit area. The fluorescence spectrum was obtained with excitation tuned to
the origin (0-0 transition energy), where the two spectra cross. The two spectra are
displaced by 2)»: 290 cm], where 1. represents the reorganization energy.

82

first-excited singlet states. The Q bands exhibit a partially resolved Vibronic progression
in methine-stretching mode that is polarized in the plane of the porphyrin macrocycle.22
A comparison of Figures 4.2 and 4.3 shows that the absorption and fluorescence line
shape is discemibly narrower in the presence of glycerol than observed in water alone. In

both solvents, the energy distance AEoo between the 0—0 transitions in the absorption and

fluorescence spectra, 290 cm", allows an estimate of the reorganization energy, A , of
145 cm"; if the ground-state and excited-state potentials are assumed to be harmonic and

have the same normal-mode frequencies, then AEOO = 21.2123 This estimate of the

 

 

 

 

 

 

"X
f t
- t
f r
r \
t i.
a 1’ i
r: ' \
Q)
5 t
3 \
§ ,g
Q
\
E \
Q)
a:
"-.
an...
O —
I I I I I I
14 15 l6 17 18 19 20

Energy (103 cm'l)

Figure 4.3. Continuous-wave absorption and fluorescence emission spectra from ZnCytc
in 50% (v/v) glycerol at room temperature. The spectra are plotted as relative dipole
strengths and are normalized to unit area. The fluorescence spectrum was obtained with
excitation tuned to the origin (O-O transition energy), where the two spectra cross. The
two spectra are displaced by 2% = 290 cm", where l. represents the reorganization energy.

83

reorganization energy is valid in the limit of Gaussian line shapes,24as are essentially

observed for the resolved Vibronic components in the continuous spectra from ZnCytc.25

4.3.2. Dynamic Stokes Shift

Figures 4.4 and 4.5 present the time evolution of the fluorescence 0—0 transition

energy, v00(t), from ZnCytc in water and in 50% (v/v) glycerol. These plots were

obtained from an analysis of a set of time-resolved fluorescence spectra, which were

obtained as

described in the chapter 2. A set of TCSPC transients were recorded at

discrete emission wavelengths spanning most of the fluorescence spectrum and spaced by

 

 

 

 

 

 

17.10-
: 11
E
° 130 ‘1
M
2 17.004 . cm
>‘ O
U -a
s: .
d.) .
8- " ,
2 o. 'p...’ . .... 8..“
LL °'.'}.. .,. ‘30 ‘ .‘g. I . . . .::. .. ..: . .Q .5 a.." 0:2 0.."
. . : '. . warn»; -- ref. 3;,» ‘3." 7.120%. y
. v. ..o Qn'WI .- '.o'~ . .'o ..o
. .z ., “3;, figs“; 4.3.6; ”"931,
° 0 .z. " . ..0:.!. ° '0'.
16.90— "
I F l l l l f
0 2 4 6 8 IO l2
Time Delay (ns)

Figure 4.4. The time evolution v00 (see eq 1) for ZnCytc in water. The fit parameters are
listed in Table 2.

84

4 nm, the band pass of the emission monochromator. The transients were used without
deconvolution of the instrument-response function. Time-resolved fluorescence-emission
spectra were obtained directly as slices across the intensity-wavelength—time surface

described by the set of transients at 12 ps intervals, the dwell time per point in the
recorded transients.
To determine v00 (t) , we fit the time-resolved fluorescence spectra as dipole

strengths DF =F(v)/V3 to a vibronic progression defined by the sum of three log-

normal line shapes26 L01(v,v0) offset by the mode frequency vF as described in detail in

 

 

 

the chapter 2.
ll
l7.l0 '—
IE ° 120 cm'1
0 o
M o
O o
:l . 5. . o . .
. §
5‘ .0 .* .0 o . . ° . .. .
§ .0 '.. .2 :VJ‘. 7.. I . ' ° ‘ . o I .0 .0 .
0‘ J. .0. ' ° ' o 0"... g .0 a... .0
g .‘o . 0' .4 “3" .‘.£. '.{..~"..‘o. I ‘0‘, ..:'.’.. q 0 10:3: 1‘ :
a m.“ tn“... ~° 'v-w . r“ V.
I" 0'. V ‘0 O O. ..... O .i‘ 0‘. ..~. ..
17.00 — c. t . . .3. 3.358;”: {5. f¥1¢:“;:'$.0. 6,-3.6
o . . . I . o . . O... :

 

 

—l
_
—
_
fl
—
-

Time Delay (ns)

Figure 4.5. Time evolution v00 (see eq 1) for ZnCytc in 50% (v/v) glycerol. The fit
parameters are listed in Table 2

85

By fitting this model to the entire emission range covered in the time-resolved

spectra, we obtain the time evolution of v00 with a good signal/noise ratio despite the
very small reorganization energy. The line shapes L0,. did not vary significantly in width

or asymmetry over the lOO-ps to 12-ns time window; Gaussians provide a satisfactory

description of the line shapes. The mode frequency VF was also invariant. The scaling
terms Fm. are the integrated dipole strengths for the vibronic lines in the fluorescence

spectrum; as discussed in the previous chapter, these parameters change as a function of

time owing to an excited-state dissociation of axial ligands from the Zn(II) ion.

 

 

Table 2
Solvent 0. , cm" I, , ps (:2 , cm'1 72 , cm'l v0, , cm'l
water 70:10 250: 20 100:10 1450:100 16940:10
0
35M“) 50:15 250:50 115:15 2000:100 16990:15
glycerol
0
504“”) 60:15 260:60 90:15 2200:100 17000:15
glycerol

 

Table 2. Fit parameters for the time evolution of v00 for Zn(II)-substituted cytochrome c
in three solvents at 295 K.

The time evolution of v00 observed from ZnCytc in Figures 4.4 and 4.5 is well

described by an exponential model. The parameters obtained by fitting the data obtained
in water, 35% (v/v) glycerol, and 50% (v/v) glycerol are compared in Table 2. (The

corresponding plot of v00 (I) obtained with 35% (v/v) glycerol is not shown.) The faster

of the two components exhibits the same amplitude and time constant in the three solvent

86

mixtures, an average of 60 cm'1 and 253 ps, respectively. While the amplitude of the slow
component is also independent of the solvent mixture, about 100 cm'l, the time constant
lengthens significantly from 1450 ps to 2200 ps in moving from water to 50% (v/v)
glycerol. The slow component exhibits a time constant of 2000 ps in the presence of
35% (v/v) glycerol, so the trend for this component is monotonic as a fimction of the

glycerol concentration. The change in the V00 offset parameter that is obtained from the
v00(t) profiles is consistent with the solvatochromic blue shift observed with continuous

fluorescence spectra (see Figures 4.2 and 4.3) upon moving from water to 50% (v/v)
glycerol. We did not observe any trend in the range of response nor was a faster
component detected when the excitation wavelength was tuned as far as 700 cm'I to the
blue from the 0—0 transition, so vibrational relaxation and intramolecular vibrational
redistribution do not make detectable contributions to the observed response functions

over the reported time window.

4.4. Discussion

4.4.1. Solvent Response Functions in Proteins
The solvent-response function, S,(t) , describes the time evolution of the

fluorescence spectrum. It is usually defined in terms of the frequency of the fluorescence

intensity maximum, v(t), as discussed in chapter 1. The decay of S, (t) in proteins can

reflect collision-like interactions between the chromophore and the surrounding

polypeptide and any intrinsic water molecules.27 As a consequence of the significantly

87

lower dielectric constant in the interior of a protein, S, (t) can involve more distant

interactions of the chromophore and the protein structure, such as interactions with the

surface of the protein.

The solvation response observed in proteins is a topic of current discussion. In
most of the reported work in proteins the response is biexponential, exhibiting a large
inertial phase followed by a smaller diffusive component. The large amplitude fast
component is similar to the inertial phase in liquids that is attributed to the librational
motions of the surrounding solvent molecules.24’28'29 The slow component of the solvation
response in liquids arises from random reorientational motions in a molecular solvent.
The assignment of components observed in the solvation response in proteins strongly
depends on the type of the protein and location of the chromophore.27 The studies of the
solvation response in proteins show that both the inertial and diffusive components of
response can be attributed to the motions of the surrounding protein structure and
intrinsic water molecules, or that both solvation components can be correlated to the
motions of the surrounding protein matrix rather than to the bulk solvent molecules.30 In
some studies even the complete lack of a slow diffusive component was observed.23 At
the present time it is not fully understood how a protein’s structure and the included

solvent contribute to solvent response.

4.4.2. Dynamic Solvation in ZnCytc.

In this chapter, we have exploited the long-lived fluorescence of the intrinsic

Zn(II) porphyrin in ZnCytc to focus on the diffusive part of the solvent-response

88

function. The methods we employed are similar to those that were used by Bashkin et
al.31 and by Pierce and Boxer32 with solvatochromic dyes in the heme-binding pocket to
obtain the solvation response in apomyoglobin. The results obtained by two groups were
not in agreement. A single-exponential on the ns timescale was observed by Bashkin et
al.31 whereas a complex multiexponential decay spanning the ps-to-ns timescale was
observed by Pierce and Boxer.32 The most likely explanation for the different results
obtained with myoglobin is that different dye-binding configurations in the heme-binding
pocket are obtained with the two dyes. By exploiting the intrinsic Zn(II) porphyrin as a
probe for ZnCytc, we avoid the various complications that can arise from the use of
extrinsic probes, such as the presence of multiple binding sites and the possibility that
the dye moves in its protein host during the excited-state lifetime. Because of its
considerable size and surface area, however, the Zn(II) porphyrin is certainly not as

spatially selective as the intrinsic aladan probe employed by Cohen et al.27

At delays longer than 100 ps, the dynamic Stokes shift response exhibited by
ZnCytc is well described by a double exponential function, with only the slower of the
two components exhibiting sensitivity to the nature of the external bulk solvent. It is clear
that protein motions are mostly responsible for both phases of the response because the
time constants are several orders of magnitude slower than those of the bulk solvent or of
the solvent in the hydration layer itself. Consider that the diffusive part of the solvent-
response function observed in water occurs on a sub-ps time scale;33’34 the motions of
water molecules in the hydration layer of a protein are thought to be slower than the bulk,

with a prominent component observed on the 40-ps time scale.24’35"38

89

The slower of the two solvation components observed in ZnCytc evidently reports
motion of the outer part of the protein that directly contacts and is frictionally hindered
by the surrounding solvent. At least part of the solvent sensitivity is suggested by
Figure 4.1, which includes rendering fi'om the X-ray crystal structure of horse-heart
ferricytochrome c'8 of the solvent-contact surface and its proximity to the porphyrin. The
edge of the porphyrin macrocycle breaks through the solvent-contact surface; it is
estimated that only four atoms (7.4% of the porphyrin’s surface area) are exposed to the
solvent.39 This portion of the porphyrin would be expected to sense directly the motions
of the external solvent; indeed, a solvatochromic blue shift and a narrowing of the
continuous absorption and fluorescence spectra occur as the proportion of glycerol in the
solvent increases. The ns timescale for the solvent-sensitive part of the dynamic Stokes
shift, however, makes it clear that the porphyrin does not stick out of the protein far

enough to sense water molecules in the bulk.

The majority of the Zn(II) porphyrin in ZnCytc is embedded in the hydrophobic
core, which gives rise to the faster part of the solvent response function. The shorter time
scale implies that the range of motion covered by relevant polypeptide motions is shorter
than those involved in the slower part of the solvent-response fimction. The observation
that a biexponential model describes the results very well is itself a surprise; one might
have anticipated that the dynamic Stokes shift would report a range of time scales given
the porphyrin’s span from the core to the edge of the protein. The exponential character
suggests that each relaxation component is associated with a well-defined class of protein

motion that is mechanically uncoupled from that of the other component.

90

4.4.3. Coupling of Polar and Nonpolar Solvation in ZnCytc

The dynamics discussed so far are associated with the polar solvation response in
ZnCytc to the formation of the Zn(II) porphyrin’s excited-state dipole moment. We
discussed in the previous chapter how there is also a nonpolar solvation response that
arises from a ground-to-excited-state change in size of the Zn(II) porphyrin owing to a

dissociation of the Zn(II) ion’s axial ligands.

The time evolution of the vibronic structure in ZnCytc’s fluorescence spectrum

can be expressed in terms of the intensity ratio for the 0—1 and 0-0 transitions, Q01/Q00-
In water and glycerol/water mixtures, the time evolution of Q01 /Qoo observed from

ZnCytc describes a biexponential increasing trend that reports a net dissociation of the
two axial ligands. The faster of the two time constants lengthens from 120 i: 50 ps to
400 :t 50 ps in moving from water to 50% (v/v) glycerol; the slower of the two time
constants, 7.0 i1.0 ns is unaffected by the addition of glycerol. The 95% confidence
intervals for these two time constants exclude those for the dynamic Stokes shift
(see Table 2.). Thus, the time evolution of the vibronic structure and dynamic Stokes shifi
are apparently uncorrelated in time scale. Further, the effect of solvent on the two
parameters occurs on different time scales, with the presence of glycerol slowing the fast
component of the vibronic structure change and the slow component of the dynamic
Stokes shifi. These results show that the nonpolar and polar solvation dynamics of

ZnCytc are largely uncoupled.

9l

A consideration of the spectroscopic origins of the nonpolar and polar solvation
dynamics in ZnCytc accounts for their uncoupled character. The polar solvation
dynamics involve a response of the surrounding protein to the porphyrin’s change in
dipole moment along the direction of the transition-dipole moment for the Q absorption
band, which is polarized in the plane of the porphyrin.22 The vibronic structure of the Q-
band involves a progression in a totally symmetric methine stretching mode that also is
polarized in the plane of the porphyrin. The intensities of the vibronic transitions change
owing to the presence or absence of a Lewis base at the axial ligation position, an
interaction that couples some out-of-plane character into the Q-band transition. As the
axial ligands attack or withdraw from the Zn(II) ion, the surrounding protein and solvent
have to reorganize to accommodate the new positions of the axial ligands. This motion is
mostly orthogonal to the types of motion that would be directly coupled to the Q-band
transition-dipole moment. Thus, two types of solvation dynamics involve motions that

should be at best weakly coupled.

4.4.4. Estimation of the Magnitude of the Internal Solvation Response in
ZnCytc

Although the acquired data limit us to a direct consideration of the time-resolved
fluorescence spectra obtained from ZnCytc from the lOO-ps time point onward, we have
some knowledge of the sub-lOO-ps portion of the solvation response from a consideration
of the range of the dynamic Stokes shift that is subsequently observed and the
reorganization energy that is obtained from the continuous absorption and fluorescence

spectra.23’24‘4O The directly observed range of the dynamic Stokes shift is obtained from

92

fits to v00 (I) (see Table 2. and Figures 4.4 and 4.5) at the lOO-ps delay and extrapolated

to infinite time. The total shift is 125 cm'I on average in the three solvent mixtures. This
shift is about 85% of the reorganization energy, l=145 cm'l, that is obtained as half of the
difference between the 0—0 transition energies of the absorption and fluorescence dipole
strength spectra (see figures 4.2 and 4.3).24 This quantity is strictly equivalent to the full

range of the dynamic Stokes shift, v(0)-v(oo), when excitation at the spectral origin

(or 0—0 transition energy, where the absorption and fluorescence dipole strength spectra
cross) is employed.“ In comparison, the sum of the amplitudes of the two exponential

components, 160 cm", is significantly larger than the reorganization energy.

Figure 4.6 shows that the summed amplitudes extrapolate the observed
biexponential solvation response to the zero of time. The sub-lOO-ps portion of the

response must connect the total range of the shift, v(0)—v(oo)=/1, with the shift

observed at 100 ps; in Figure 4.6, a Gaussian is drawn to suggest one possibility that is
consistent with the shape of the response observed in liquids.42 The point of the exercise
is to make it clear that no matter what fimctional form is chosen, the magnitude of the
sub-100 ps solvation response in ZnCytc is much smaller than the magnitude of the
observed response from 100 ps onward. It is safe to say that the sub-ps fraction of the
solvent-response function that would include inertial motions is restricted in ZnCytc to
less than 10% of the reorganization energy. This result strongly contrasts with the
previous observations of the sub-ps solvent response function in other protein systems,

where the sub-ps fraction accounted for the majority of the response.

93

The most likely explanation for the present results is that most of the dynamic
Stokes shift response observed in ZnCytc arises from a direct sensing of motions of the
hydrophobic core of the protein. This region contains very few embedded charges, polar
residues, or water molecules that are capable of the free-streaming character of motion
that contributes to the polar solvation response on the sub-ps time scale in molecular
liquids. In contrast, the protein systems that exhibit large sub-ps solvation responses were
probed with intrinsic or extrinsic chromophores that sense motions of polar groups near
to or at the surface of the protein. In the case of bacteriorhodopsin, the retinal
chromophore is found in a polar channel through which the transmembrane ion-pumping
action occurs and mobile protein groups are found. In ZnCytc, the hydrophobic region is
apparently densely packed in the native state, and reorganization of the protein
inresponse to the formation of the Zn(II) porphyrin’s excited-state dipole moment is
limited to slow, small-amplitude motions of a relatively large reduced mass. We infer that
the Zn(II) porphyrin chromophore largely obtains its sensitivity to the motions of the
solvent-sheathed outer parts of the protein indirectly through the polypeptide backbone;
but the part of the reorganization energy that must be reserved for a fast solvation
response is on the same order of the porphyrin’s surface area that breaks through the
solvent contact surface. A direct consideration of this hypothesis will have to wait for

future experiments on ZnCytc employing femtosecond experiments.

The present results raise the interesting possibility that the motions that a globular
protein makes in its hydrophobic core mostly lack the free-streaming or inertial character

that dominates the solvation response of water and other small molecular liquids. That the

94

polypeptide motions in the hydrophobic core and in the polar exterior of a protein are
different in time scale and are associated with different types of motion is a novel
suggestion that deserves additional attention because of its possible impact on the

potentials used in molecular dynamics simulations of folded and partially folded proteins.

160 -

)
3
c
l

-1

_ _
DJ b
O O
l l
1
1
1
1
1
c
1
I
I

Frequency (cm
13
O
1

 

110-

 

 

_
_

l I l I
0 50 100 150 200 250

Time (ps)

Figure 4.6. Model for the early time evolution v00 for ZnCytc in water. The ordinate

corresponds to the remaining Stokes shifl at the indicated delay following excitation. The
thick solid line shows the fitted trend obtained fiom Figures 4 and S from fits to the time-
resolved fluorescence spectra observed over the lOO-ps to l2-ns time window. The thin
solid line extrapolates the fitted trend to zero time. The dashed line shows a hypothetical
Gaussian trend that connects the reorganization energy (see Figures 4.2 and 4.3) with the
observed shift of v00 at the lOO-ps delay.

95

4.5.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

References

Lampa-Pastirk, 8.; Beck, W. F. J. Phys. Chem. B. 2004, 108, 16288.

Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598—1603.

Vanderkooi, J. M. Biochim. Biophys. Acta 1998, 1386, 241—253.

Vanderkooi, J. M.; Adar, F.; Erecinska, M. Eur. J. Biochem. 1976, 64, 381.

Leeson, D. T.; Berg, 0.; Wiersma, D. A. J. Phys. Chem. 1994, 98, 3913—3916.

Anni, H.; Vanderkooi, J. M.; Mayne, L. Biochemistry 1995, 34, 5744—5753.

Quian, C.; Yao, Y.; Tong, Y.; Wang, J .; Tang, W. J. Biol. Inorg. Chem. 2003, 8,
394—400.

Vanderkooi, J. M.; Erecinska, M. Eur. J. Biochem. 1975, 60, 199—207.

Cmogorac, M. M.; Kostic, N. M. Inorg. Chem. 2000, 39, 5028—5035.

Kang, T. J .; Yu, J .; Berg, M. Chem. Phys. Lett. 1990, I 74, 476-480.

Saven, J. G.; Skinner, J. L. J. Chem. Phys. 1993, 99, 4391—4402.

Bagchi, B. J. Chem. Phys. 1994, 100, 6658—6664.

Fourkas, J. T.; Benigno, A.; Berg, M. J. Chem. Phys. 1993, 99, 8552—8558.

Ladanyi, B. M.; Stratt, R. M. J. Phys. Chem. 1995, 99, 2502—2511.

Berg, M. J. Phys. Chem. A 1998, 102, 17—30.

96

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

Berg, M. J. Chem. Phys. 1999, 110, 8577—8588.

Delano, W. L. The PyMOL molecular graphics system; DeLano Scientific: San
Carlos, California, 2002.

Bushnell, G. W.; Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 585—595.

Kantor, C. R.; Schimmel, P. R. Biophysical Chemistry, Part II: Techniques for the

study of Biological Structure and Function; Freeman and Co: San Francisco,
1980.

Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Second ed.; Kluwer
Academic/Plenum Publishers: New York, 1999.

McHale, J. L. Molecular Spectroscopy; Prentice Hall: Upper Saddle River, New
Jersey, 1999.

Gouterman, M. Optical spectra and electronic structure of porphyrins and related
rings. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978;
Vol. 3; pp 1-159.

Kennis, J. T. M.; Larsen, D. S.; Ohta, K.; Facciotti, M. T.; Glaeser, R. M.;
Fleming, G. R. J. Phys. Chem. B 2002, 106, 6067—6080.

Jordanides, X. J .; Lang, M. J .; Song, X.; Fleming, G. R. J. Phys. Chem. B 1999,
103, 7995—8005.

Lampa-Pastirk, S.; Lafuente, R. C.; Beck, W. F. J. Phys. Chem. B 2004, 108,
12602—12607.

Siano, D. B.; Metzler, D. E. J. Chem. Phys. 1969, 51, 1856—1861.

Cohen, B. E.; McAnaney, T. B.; Park, E. 8.; Jan, Y. N.; Boxer, S. G.; Jan, L. Y.
Science 2002, 296, 1700—1703.

Homoelle, B. J .; Beck, W. F. Biochemistry 1997, 36, 12970—12975.

97

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

(39)

(40)

(41)

Homoelle, B. J.; Edington, M. D.; Diffey, W. M.; Beck, W. F. J. Phys. Chem. B
1998, 102, 3044—3052.

Changenet-Barret, P.; Choma, C. T.; Gooding, E. F.; DeGrado, W.; Hochstrasser,
R. M. J. Phys. Chem. B 2000, 104, 9322—9329.

Bashkin, J. S.; McLendon, G.; Mukamel, S.; Marohn, J. J. Phys. Chem. 1990, 94,
4757—4761.

Pierce, D. W.; Boxer, S. G. J. Phys. Chem. 1992, 96, 5560.

Rosenthal, S. J.; Xie, X.; Du, M.; Fleming, G. R. J. Chem. Phys. 1991, 95, 4715-
4718.

Maroncelli, M.; Kumar, P. V.; Papazyan, A.; Homg, M. L.; Rosenthal, S. J .;
Fleming, G. R. Studies of the inertial component of polar solvation dynamics. In
Ultrafast Reaction Dynamics and Solvent Effects; Guaduel, Y., Rossky, P. J .,
Eds.; AIP Press: Woodbury, NY, 1994; pp 310—333.

Nandi, N.; Bagchi, B. J. Phys. Chem. B 1997, 101, 10954—10961.

Nandi, N.; Bhattacharyya, K.; Bagchi, B. Chem. Rev. 2000, 100, 2013—2045.

Bizzari, A. R.; Cannistraro, S. J. Phys. Chem. B 2002, 106, 6617—6633.

Pal, S. K.; Peon, J .; Bagchi, B.; Zewail, A. H. J. Phys. Chem. B 2002, 106,
12376—12395.

Brayer, G. D.; Murphy, M. E. Structural studies of eukaryotic cytochromes c. In
Cytochrome c: a Multidisciplinary Approach; Scott, R. A., Mauk, A. G., Eds.;
University Science Books: Sausalito, California, 1996; pp 103—166.

Fee, R. S.; Maroncelli, M. Chem. Phys. 1994, 183, 235—247.

Mukamel, S. Principles of Nonlinear Optical Spectroscopy; Oxford University
Press: New York, 1995.

98

(42) Stratt, R. M.; Maroncelli, M. J. Phys. Chem. 1996, 100, 12981—12996.

99

5. Light-Induced Protein Unfolding of Zn(ID-substituted
Cytochrome c: Dynamics of Entry and Exit of Water

Molecules to and from the Hydrophobic Core During
Unfolding and Refolding

5.1. Introduction

The role of misfolded proteins and the mechanism that produces them from
correctly folded, native structures are not understood.“ A central problem in this area is
that existing methods for studying the kinetics and mechanism of protein folding,
unfolding and misfolding reactions involve macroscopic manipulation of the solution
conditions or temperature. The most common methods involve stopped-flow or
rapid-flow mixing techniques that involve removal or addition of a chemical denaturant
to the solution as the means to start the reaction.7'9 Temperature-jump methods have been
used for many years; the most recent versions of this class of experiment have involved
laser-induced temperature jumps mediated by the near IR overtone absorption of water
molecules in the solution."”16 While these methods have yielded a wealth of information
that has been used in the shaping of current theories for protein-folding dynamics, none
of these methods can be used under the physiological conditions that promote the
progress of disease nor can they be used in intact cells or in vivo. These methods also
usually suffer from the disadvantage of being one-shot in character; the reaction can be
initiated only once, and recycling to reform the initial state requires either a long

relaxation time to permit the solution temperature to reequilibrate or a chemical

100

separation of the reactant proteins from the supporting solution. Study of a misfolding
mechanisms would be aided tremendously if there were a way to initiate synchronously a
reaction in an ensemble of proteins, rendering it possible to detect intermediate structures,
but in a range of solution conditions and temperatures under control of the experimentor

rather than just as required to initiate folding, unfolding or misfolding reactions.

The approach discussed in this chapter is motivated by the idea that there are
many possible folding trajectories between the unfolded and native folded states on the
folding potential-energy surface, which has been termed an energy landscape (see
chapter 1).”'25 Protein misfolding can be taken as a consequence of the possibility of
alternate Gibbs free-energy minima on the energy landscape that act as kinetic sinks or
traps. This picture suggests immediately the possibility that the native structure is formed
under a combination of kinetic and thermodynamic control.26 Further, misfolded proteins
may arise spontaneously from the native state; folded proteins are only marginally
stable.24 Nevertheless, owing to the topology of the fimnel, it is possible that most
unfolding trajectories subsequently are steered along refolding paths toward the native
structure. The infinitesimal number of trajectories that lead to trapping in alternate

minima may be the origin of the misfolded proteins that are active in disease.

It has been long established that the origin of protein stability lies in the
hydrophobic effect, the ability of a protein to pack its hydrophobic side chains inside the
molecule and away from water.27 The hydrophobic effect results in a decrease in the free
energy of protein and its surrounding solvent. Manipulation of the free energy of the

hydrophobic core could be used as a direct means to initiate an unfolding reaction.

101

We suggest here an approach that uses an electronic chromophore in the
hydrophobic core of a folded protein as an optical trigger for unfolding reactions. We call
this method the optical G jump. With excitation sufficiently above the 0—0 vibronic
transition from the ground state to the lowest singlet excited state (S.), the electronic
chromophore injects the Gibbs free energy required to generate the transition state for
unfolding through vibrational energy transfer from the chromophore to the protein. The
vibrationally equilibrated electronic chromophore’s subsequent recovery of the ground
state by emission of a photon provides a local solvatochromic probe of the unfolding
reaction. We show in this chapter that this approach uniquely provides a means to detect
the entry and exit of water molecules into the hydrophobic core that is probably
associated with the formation and decay of the folding/unfolding transition state and

control of the reaction rate.

5.2. Experimental

5.2.l. Zn(II)-Substituted Cytochrome c

The method for preparation of ZnCytc is described in detail in chapter 2. The
experiments in this chapter employ ZnCytc samples dissolved in a 50 mM HEPES buffer
solution at pH 7.5 or in a 40% (v/v) methanol/HEPES or 50% (v/v) glycerol/HEPES

buffer solution mixture. Sample handling is described in detail in chapter 2.

102

5.2.2. Continuous-Wave Absorption and Fluorescence Instrumentation

Steady-state absorption and fluorescence spectra instrumentation is described in
chapter 3. The temperature dependent continuous-wave fluorescence spectra shown in
this chapter were acquired with a Hitachi F-4500 fluorescence spectrophotometer
employing temperature controlled sample holder. The excitation wavelength was 585 nm.
The samples were equilibrated at each temperature for 20 minutes prior to recording the

fluorescence emission spectrum.

5.2.3. Picosecond Time-Resolved Fluorescence Spectroscopy

For the experiments discussed in this chapter, the 355-nm third-harmonic of the
Antares Nd-YAG laser was used to synchronously pump a cavity dumped dye laser
operating with Stilbene 420 laser dye (Exciton). The details of the experimental apparatus
are provided in chapter 2. ZnCytc samples were excited with 420-nm light corresponding
to the Soret-band region of the spectrum. Fluorescence signals were collected at magic
angle, 54.7° with respect to the vertically polarized excitation pulse. The samples were
held in l-cm quartz cuvettes mounted in a home-built temperature controlled mount.
During the experiments, the samples were kept at 22 °C with Neslab RTE-7 circulating

water bath. The instrument response function was 35 ps fwhm.

103

5.3. Results

5.3.]. Time-Resolved Fluorescence Spectra

Figure 5.1 shows a series of time-resolved fluorescence spectra obtained at six
selected time delays from ZnCytc in water at pH 7.5. The spectra are obtained as slices of
a time-wavelength-intensity surface by the method described in detail in chapter 2. The
surface was obtained by recording a set of TCSPC fluorescence transients following
excitation of the Soret-band at 420 nm. The transients, collected with 5 nm spacing, span

most of the fluorescence spectrum.

The time-resolved spectra shown in Figure 5.1 exhibit time evolution of the
intensities of the 0—0 and 0—1 vibronic components. At time delays < 150 ps the relative

intensity of the Q01 vibronic line increases at the expense of Q00. At time delays near

250 ps the time-resolved fluorescence spectra pass through the first tum-around point
characterized by equal relative intensities of 0—0 and 0—1 vibronic components. After the

first tum-around point the relative intensity of the Q00 vibronic line slowly increases at
the expense of that of the Q01. Second tum-around point is reached at 5 ns. Beyond this

point the very slow increase in intensity of Q01 vibronic line is observed.

In chapter 3 we discussed time evolution of the intensities of the 0—0 and 0—1
vibronic components in the time-resolved fluorescence spectra in the native state of
ZnCytc following Q-band excitation. The observed response was biexponential with a

monotonic increase of the intensity of the 0—1 transition. The time evolution of the

104

vibrational structure of the Q-band fluorescence spectrum from ZnCytc was rate limited
by protein-matrix and solvent reorganization. We attributed this response to the excited-
state release of axial ligands from the Zn(II) ion. Figure 1 indicates that the vibronic
structure following Soret-band excitation exhibits multiphasic behavior with at least two

tum-around points. Increase of the Q01 intensity at the initial time delay indicates the

possibility for the release of one of the axial ligands, but the overall response is far more

complex.

Figure 5.1 depicts that O—l vibronic line exhibits additional structure, with a
partially resolved shoulder emerging at 16 250 cm]. The intensity of this structure decays
with respect to time, and at ~250 ps completely disappears. The type of spectral changes
observed here could result from distortions of the heme caused by perturbations in its
near vicinity.28 The observed response is an indication of the significant structural
changes that occur in the protein upon the Soret-band excitation that were not observed in

the Q-band experiments described in chapter 3.

The time-resolved spectra at early time delays from 100 to 250 ps exhibit a red

shift of the center of the Qooline. At delays beyond 250 ps, the center of the Q00 line

shifts to the blue. The dashed line on Figure 5.1 depicts this behavior.

105

 

20— 500 ps
10‘

Relative Intensity
o 8 '

 

I
I
1.2— '
0.8- 5 “S |
0.4— :
0.0—l I

I l l I l I
14.5 15.0 15.5 16.0 16.5 17.0

 

Energy (103 cm")

Figure 5.1. Time-resolved fluorescence spectra from ZnCytc in aqueous solution at 22 °C
at six delays. The dashed line depicts the initial red shift of the Q00 vibronic band of the
fluorescence spectrum followed by the blue shift at time-delays longer than 250 ps.

106

5.3.2. Dynamic Stokes Shift

In order to examine the shift of the spectra in more detail, we obtained the
dynamic Stokes shifi, a plot of the center of the Q00 vibronic line as a function of time, by
the method described in detail in chapter 2. Figure 5.2 shows the early time components

of the dynamic Stokes shifi response observed in ZnCytc with Soret-band excitation.

 

 

'0
C . .
, '. '33-“?
17.10_1 C” p * O
n' * °
0 . «a.
. ...
. §'$‘O
o “3'."
_. O
17.08 . 'If
__A ~",
5 ,v'
U 0
"2;, 17.06- '
=2 :
5‘ o
5 -: -
g" 0 o
0 17.04 0
L: 1 Q 0
. O
O
. O
O
17.02: "
l7.00—
I I I I I
0.0 0.5 1.0 1.5 2.0
Time Delay (ns)

Figure 5.2. Early time response of the frequency of the center of the Q00 vibronic line of
the fluorescence spectrum from ZnCytc in aqueous solution at 22 °C. The excitation laser

was tuned to 420 nm, the peak of the Soret absorption band.

107

The dynamic Stokes shifi response with Soret-band excitation shown in Figure
5.2, exhibits a biphasic character. The initial segment from 100—250 ps is a very fast red
shifi. The time constant of 125 ps is significantly shorter than of the fast component
observed with Q-band excitation discussed in chapter 4. At the 200-ps delay point,
however, the response turns around and starts a biexponential shift to the blue with time
constants of 330 ps and 3.1ns that continues throughout the observation window
permitted by the fluorescence lifetime. Figure 5.3 compares the dynamic Stokes shift
responses obtained with Soret-band excitation (top curve) and with Q-band excitation
(bottom curve). The response with Q-band excitation is well modeled with a sum of two
exponentials (as explained in chapter 4), but the response resulting from Soret-band
excitation requires at least three exponentials to be adequately modeled. Note that the
final observed position of the fluorescence spectrum with Soret-band excitation is

200 cm'1 to the blue of the equilibrium position observed with Q—band excitation.

Figure 5.4 compares the dynamic Stokes shift response observed in ZnCytc
following Soret-band excitation in water, 50% glycerol and in 40% methanol. The
compositions of the glycerol and methanol solutions were chosen to obtain the same,
decreased polarity compared to water, but significantly different viscosities. The three
responses allow a separate determination of the effects of the external solvent polarity
and viscosity on the nature of the response. Note first that the initial red shift is shifted to
a shorter time scale in glycerol/water and in methanol/water compared to that in water so
that the amplitude of the initial component is very small (glycerol) or even absent

(methanol) in the >100-ps observation window available in these results.

108

 

 

 

 

 

 

17.20 —
0 . . ”.02" 53...:
° . “‘2. 0:”... ‘o." 03:. ?;.::.::Eu I.
0.. ”1"; .. wm-—-- . . —
-.. “xi-iv 1.. n". . '39:? .fi; ‘o...:~..,."'... true?" 1:.“
.2 are... O... a; .0. '0' :.'. o'o'" . . ".. .':.‘::
17.10 -l . ' Soret band ' '
‘E’ 200 cm"1
to
.9.
V 17.00% '.
>3 ..
o
a
o - .
3' .‘ . Qband . ,
“*3 ”W i' ... .. :.-.~ ..'..:'.'.'
L“ " V....‘-..' °~~- ‘°°‘°'.‘-r'~".°.‘-v
I ' " ' 3' II ”13.3.3 : .:.°~:~oib oo’.’.n‘.". " '3'???
I690 - . ° 7° ' ,.
I I I I I I I
0 2 4 6 8 IO 12

Time Delay (ns)

Figure 5.3. Time evolution of the frequency of the Q00 intensity maximum of the ZnCytc
fluorescence spectrum in aqueous solution at 22 °C. The top curve was obtained with the
excitation laser tuned to 420 nm, the peak of the Soret absorption band. The early time
portion of this trace is shown on Figure 2. The bottom curve is that obtained with Q—band
excitation (chapter 4). Each data trace is superimposed with an exponential fit of the form

y = C + Effie—W" .

109

l7.15‘

17.10"

17.05—

    
 

 

”(5.34. .33"

17.20- ... . .-
17.15_ 0% glycerol

17.10—

  

llns

17.05—

Frequency (103 cm'l)

17.15-

17.10—

 

17.05—

 

 

O
N—
A
O\
00
S
5

Time Delay (ns)

Figure 5.4. Comparison of the time evolution of the frequency of the Q00 intensity
maximum of the fluorescence spectrum from ZnCytc in three solvent systems at 22 °C.
The excitation laser was tuned to 420 nm, the peak of the Soret absorption band. The data

is superimposed with exponential fit of the form y = C + Zi Ate—WI .

llO

Figure 5.5 compares the time evolution of the linewidth of the Q00 vibronic line in
the fluorescence spectrum from ZnCytc in water, 50% glycerol and in 40% methanol. For
the experiments discussed in this chapter we fit the time resolved spectra with the model
described with equation 2.2, allowing the linewidth parameter to float. In the Soret-band
excitation experiments, the width of the fluorescence Q00 line exhibits time dependence
that is not observed in Q-band excitation experiments discussed in chapter 4. The
linewidth data in Figure 5.5 are modeled as exponentials. Initial narrowing of the line
shape is observed in water and to a smaller extent in a glycerol/water mixture. It is
followed with slow line broadening in both solvents. The linewidth time dependence
observed in methanol/water mixture exhibits only a slow line broadening component in

the observation window available in these experiments.

5.4. Discussion

5.4.1. Nature of Dynamical Stokes Shift in ZnCytc with Soret-Band
excitation

This chapter presents the Soret-band excitation of ZnCytc that results in optical
triggering of protein unfolding reactions. As discussed in chapter 2, the Zn(II)-porphyrin
chromophore exhibits two strong absorption bands, the Q band and the Soret (or B) band,
which arise from transitions to the SI and to the S,, (n 2 2) manifold of states (Figure
5.6).28'29 The 81 state has a good fluorescence quantum yield and a long emission lifetime

(2'F > 2 ns).3°’3' With excitation of the Q band, the S. state is prepared directly, and a

photon is emitted subsequent to vibrational equilibration at the ambient temperature.

111

380-

 

    

 

 

IISpS : . J . :‘- c”0‘ . 'V 0 .0 ns..
‘- ‘ ' o u 0'. o I . ’0'. 9..
3401.. I. . . . "X' .a. , 30¢.
‘ 0.5 ns
water
300-
'g 380-
8
:5
g3 340-
d)
.E
._l
300-
o e . . ~ ....00.. . 0".
380- ' ." ° ~ ° "' '\ o
340. , '
40% methanol
10 ns
300-4
I l I I I I
0 1 2 3 4 5
Time Delay (ns)

Figure 5.5. Comparison of the time evolution of the linewidth of the Q00 vibronic line of
the fluorescence spectrum from ZnCytc in three solvent systems at 22 °C. The excitation
laser was tuned to 420 nm, the peak of the Soret absorption band. The data is

superimposed with exponential fit of the form y = C + 2:1.Aie—t/Ti .

112

 

 

 

 

 

 

80

Figure 5.6. Energy level diagram for Zn(II) porphyrins, showing an absorption (A)
transition in the Soret band to the Sn manifold of states followed by internal conversion
(IC) to the S; state and recovery of the ground state So by emission of a fluorescence
photon (F).

Intramolecular vibrational redistribution (IVR) occurs on the sub-ps time scale
and vibrational relaxation occurs on the < lO-ps time scale.32’33 With excitation of the
Soret band at 420 nm, however, an entirely different response is observed. Here, the
initially prepared Sn state rapidly decays nonradiatively on the < S-ps time scale to
produce the S. state (see Figure 5.7). The hydrophobic core of ZnCytc accordingly
obtains a significant vibrational excitation (~6850 cm") when the Soret band is excited.
The time-resolved fluorescence spectra report that this excitation results in an unfolding
event. In the following section, we propose an explanation for the sequence of events that
occur following Soret-band excitation and a discussion of the associated Stokes shift

response.

113

We suggest that the initial response of the fluorescence spectrum to the red
following Soret-band excitation, accompanies unfolding of the protein and entry of water
molecules into the hydrophobic core. This conclusion is warranted by the subsequent
shift of the spectrum to the blue that persists over the fluorescence time window. The
dumping of energy into the hydrophobic core of the protein that accompanies internal
conversion and IVR certainly would be expected to induce a fast red shift of the
fluorescence spectrum owing to an accelerated reorganization of the protein. A faster
reorganizational response would be similarly expected if the temperature of the
surroundings were raised. The following shift of the fluorescence spectrum to the blue,
however, is not anticipated if all that happens is an accelerated sampling of the energy
landscape. The blue shift reports that the average polarity of the protein/solvent medium
that surrounds the Zn(II) porphyrin is decreasing with time, and the simplest explanation
for this response is a reorganizational response of the protein structure that results in
expulsion of water molecules from the hydrophobic core. So, the fast initial red shift
includes a component that corresponds to the rapid entry of water molecules into the

hydrophobic core that accompanies unfolding of the protein.

The results shown in Figures 5.2 and 5.3, then, report a sequence of events that is
consistent with an essentially vertical transition of the protein on the energy landscape
owing to the internal vibrational excitation of the protein that results from internal
conversion and IVR of the energy difference between the excitation laser’s energy and
the energy of the 0—0 vibronic transition of the S; state. Having absorbed the vibrational

excitation, the protein begins to move away from the native structure for a short time,

114

perhaps 200 ps; owing to vibrational energy transfer with the cold bulk solvent water that
surrounds the protein, the vibrational excitation does not last indefinitely. The protein is
then dropped on the energy landscape at a displaced position and begins to evolve at the
ambient temperature. The initial displacement involves unfolding and entry of water
molecules into the hydrophobic core. The trajectory subsequently assumed by ZnCytc
apparently involves motion towards the native configuration; the refolding trajectory
involves a collapse of the unfolded structure and ejection of the water that entered. At the
end of the observation window, the protein has not had enough time to recover to the
initial state; the fluorescence spectrum remains well shifted to the blue with respect to the

equilibrium position indicated by the Q-band response.

Figure 5.4 compares the dynamic Stokes shift response observed in ZnCytc
following Soret-band excitation in water with that in 50% glycerol/water and in 40%
methanol/water. The three responses allow a separate determination of the effects of the
external solvent polarity and viscosity on the nature of the response. Note first that the
initial red shift is shifted to a shorter time scale in glycerol/water and in methanol/water
compared to that in water so that the amplitude of the initial component is very small
(glycerol) or even absent (methanol) in the >100-ps observation window available in
these results. These observations suggest that the initial unfolding reaction occurs more
rapidly in the presence of glycerol or methanol than in water. We suggest that this result
is consistent with a lowering of the activation energy for the unfolding reaction owing to
a reduction in the polarity difference between the external bulk solvent and the native

(anhydrous) hydrophobic core. Next, observe that the shift back to the blue is

115

considerably slower in glycerol/water and considerably accelerated in methanol/water;
the observed time constants are 11 ns and 600 ps, respectively. These observations show
that the initial refolding response that we assign to collapse of the protein is rate limited
by the viscosity of the external solvent; the viscosity of the glycerol/water mixture is
significantly larger compared to that of water, but the viscosity of the methanol/water
mixture is significantly lower compared to water. The blue shift is apparently nearly
completed in water by the l2-ns point; in comparison, the blue shift is completed in the
methanol/water mixture at the 2-ns point, and we can see the beginning of the shift back
to the red that must accompany the approach to the native folding state of the protein on a V

much longer time scale in water.

The dynamic Stokes shift response is accompanied with a change in the width of
the Q00 vibronic line (see Figure 5.5). The initial line narrowing observed in water
corresponds well to the time scale of the initial red shift component of the dynamic
Stokes shift response described above. It could be that this line narrowing arises from a
motional line narrowing caused by averaging between two or more structures that occur
as a consequence of expansion of the protein core due to the unfolding process. The
subsequent broadening of the Q00 vibronic line relates to the collapse of the protein

structure around the porphyrin and concomitant slowing of the dynamics.

5.4.2. The Optical G Jump in ZnCytc: Dynamics of Unfolding and
Refolding following Soret-Band excitation

The results discussed in this chapter show that Soret band excitation provides an

optical trigger for unfolding and misfolding reactions in ZnCytc. We suggest that the
116

dynamic Stokes-shift response exhibited by an electronic probe in the hydrophobic core

of a target protein can be explained by the sequence of events shown in Figure 5.7.

In the first step, the excess vibrational energy of the chromophore is deposited in
the protein by vibrational energy transfer on the < 20-ps time scale. The chromophore is
left in a vibrationally relaxed state from which it can recover the ground electronic state
by emitting a photon; the fluorescence spectrum emitted by an ensemble of proteins can
be used as a zero-background probe of the following protein dynamics. Next, the protein
obtains a nearly instantaneous change in enthalpy without suffering a change in structure
or solvent structure, which would contribute to an associated entropy change, so the
change in enthalpy is equivalent to a change in Gibbs energy (step 2 in Figure 5.7). On
the ~100-ps time scale, however, the vibrationally excited protein is “cooled” by the
surrounding solvent, so the protein is dropped (vibrationally equilibrates) at a displaced
position on the energy landscape after having evolved at the excited energy level for the
intervening period of time (step 3). The next event involves initiation of a refolding (step
4a) or a misfolding trajectory (step 4b). If the energy difference between the absorbed
and emitted photons is made large enough to overcome the unfolding activation-energy
barrier, the optical G jump can launch an unfolding trajectory. It is important to note,
however, that this process occurs without causing a significant change in the temperature
of the surrounding solvent; in optically dilute solutions, the heat capacity of the solvent is

effectively infinite.

ll7

3

 

   

l transfer of enthalpy
. from protein to
2 > fl surrounding solvent
(~100 p8)
structural evolution
A in the excited state 4b\

misfolding
trajectory

l

vibrational relaxation and

vibrational energy transfer refolding

 

 

to the protein (< 20 PS) trajectory
5 H20
hydrophobic exit
collapse
refolding I

hv

Figure 5.7 . Sequence of events following the excitation of the electronic probe in the
hydrophobic core of a target protein in the optical G jump experiment: 1. the initial
vibrational relaxation of the probe and associated vibrational energy transfer to the
protein, 2. structural evolution of the protein at an excited vibrational level, 3. transfer of
enthalpy from the protein to the surrounding solvent, and the associated vibrational
equilibration of the protein at a displaced position on the energy landscape, 4. initiation of
a refolding (4a) or misfolding (4b) trajectory, 5. hydrophobic collapse of the unfolded
protein and ejection of water, 6. slow refolding into the final, native state.

118

These initial four steps accompany the fast initial red shift of the fluorescence
spectrum from the probe. The nature of the structural evolution of the protein at an
excited vibrational energy level described in step 2, depends on whether the amount of
vibrational energy delivered by the probe is large enough to allow the protein to
overcome the activation barrier that retained it in the initial folding state under the
ambient conditions. If the excitation wavelength is too far to the red, the G jump that will
be delivered will be insufficiently large to get the protein over the ban'ier; the red shift of
the fluorescence spectrum will just be a report of the vibrational relaxation of the probe
and the sampling of the energy landscape by the surrounding protein in a region near to
the initial folding state. If the excitation wavelength is far enough to the blue, however,
an unfolding trajectory will be launched, and the red shift will include contributions from
the increased polarity of the region around the probe that accompanies unfolding of the
protein and solvation of the polypeptide. In that case, once vibrational energy transfer
from the protein to the surrounding solvent has occurred and the protein has vibrationally
equilibrated, the dynamic Stokes-shift response that follows is determined by the events

that occur during the refolding or misfolding trajectory:

The observed tum-around and blue shift of the fluorescence spectrum corresponds
to a hydrophobic collapse of the unfolded protein and an ejection of a significant number
of water molecules (step 5 in Figure 5.7). It is likely, however, that some water molecules
stay behind and hydrate the polypeptide in its core region.34 The ejection of these residual
water molecules is likely to correspond to the rate limitation for the folding into the

native fold. This leads to an apparent metastability of the state that follows the blue shift,

119

at least on the fluorescence time scale. This event is followed by the much slower
ultimate red shift of the system into the final, native state (step 6). This phase of the

dynamics is evidently strongly limited by solvent fiiction.

The possible role of solvent friction in the final refolding of the protein is
suggested by the effect of the solvent viscosity on the time constants for the blue and
final red shifts observed in the preliminary work on ZnCytc. Given such an observation,
one is prompted to suggest that the fold that a protein assumes could be kinetically

controlled by fi‘iction with a surrounding surface, such as from a chaperone protein.35

The key to the success of the optical G-jump is the dynamical bottleneck that
evidently exists in the vibrational energy transfer from the protein to the surrounding
solvent. The intramolecular vibrational redistribution (IVR) processes that inject the
excess vibrational energy of the chromophore into the protein occur on a much shorter
time scale than the formally vibrational energy-transfer processes that dissipate the
excitation into the surrounding solvent. This bottleneck provides a short time for the
protein to evolve at an excited energy level before equilibrating at the ambient solution
temperature. At first glance, it might be considered remarkable that perhaps 100 ps of
evolution is sufficient to initiate an unfolding transition that results in entry of a
significant number of water molecules into the hydrophobic core. The rate at which the
protein diffuses away from the initial structure, however, is much faster at the elevated
free energy level provided by the G-jump than it is at the ambient temperature. This
argument suggests that it will be of interest to vary the excitation energy provided to the

protein above the unfolding barrier by tuning the excitation wavelength significantly

120

above the threshold wavelength at which unfolding occurs in order to vary the
displacement from the native structure; larger excitations above the barrier will result in
longer periods of accelerated diffusion prior to the clamping of the dynamics by the

equilibration with the surrounding solvent.

121

5.5.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

References

Sunde, M.; Blake, C. The structure of amyloid fibrils by electron microscopy and
X-ray diffraction. In Adv. Prot. Chem, 1997; Vol. 50; pp 123—159.

Prusiner, S. B. Proc. Natl. Acad. Sci. USA 1998, 95, 13363—13383.

Dobson, C. M. Trends Biochem.Sci. 1999, 24, 329—332.

Bucciantini, M.; Giannoni, E.; Chiti, F.; Baroni, F.; Formigli, L.; Zurdo, J. S.;
Taddei, N.; Ramponi, G.; Dobson, C. M.; Stefani, M. Nature 2002, 416, 507—51 1.

Caughey, B.; Lansbury, P. T. Annu. Rev. Neurosci. 2003, 26, 267—298.

Selkoe, D. J. Nature 2003, 426, 900—904.

Shastry, M. C. R.; Luck, S. D.; Roder, H. Biophys. J. 1998, 74, 2714—2721.

Gianni, S.; Travaglini-Allocateli, C.; Cutruzzolla, F.; Brunori, M.; Shastry, M. C.
R.; Roder, H. J. Mol. Biol. 2003, 330, 1145—1152.

Kumar, R.; Bhuyan, A. K. Biochemistry 2005, 44, 3024—3033.

Callender, R.; Dyer, R. B. Curr. Opin. Struc. Biol. 2002, 12, 628.

Williams, S.; Causgrove, T. P.; Gilmanshin, R.; Fang, K. S.; Callender, R. H.;
Woodruff, W. H.; Dyer, R. B. Biochemistry 1996, 35, 691—697.

Werner, J. H.; Dyer, R. B.; Fesinmeyer, R. M.; Anderson, N. H. J. Phys. Chem. B.
2002, 106, 487—494.

Pandit, A.; Ma, H. R.; van Stokkum, I. H. M.; Gruebele, M.; van Grondelle, R.
Biochemistry 2002, 41, 15115—15120.

122

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

Vu, D. M.; Peterson, E. S.; Dyer, R. B. J. Am. Chem. Soc. 2004, 126, 6546.

Snow, C. D.; Nguyen, N.; Pande, V. S.; Gruebele, M. Nature 2002, 420, 102—106.

Ma, H. R.; Gruebele, M. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 2283—2287.

Bryngelson, J. D.; Wolynes, P. G. Proc. Natl. Acad. Sci. U. S. A. 1987, 84, 7524—
7528.

Chan, H. S.; Dill, K. A. Proteins: Struct. F unct. Genet. 1998, 30, 2—33.

Brooks III, C. L.; Onuchic, J. N.; Wales, D. J. Science 2001, 293, 612.

Dill, K. A.; Chan, H. S. Nature Struct. Biol. 1997, 4, 10—19.

Dill, K. A. Protein Sci. 1999, 8, 1 166—1180.

Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. Science 1991, 254, 1598-1603.

Leopold, P. E.; Montal, M.; Onuchic, J. N. Proc. Natl. Acad. Sci. U. S. A. 1992,
89.

Onuchic, J. N.; Luthey-Schulten, Z.; Wolynes, P. G. Annu. Rev. Phys. Chem.
1997, 48, 545—600.

Shea, J.-E.; Brooks III, C. L. Annu. Rev. Phys. Chem. 2001, 52, 499—535.

Baker, D.; Agard, D. A. Biochemistry 1994, 33, 7505—7509.

Kauzmann, W. Adv. Protein. Chem. 1959, 14, 1—63.

Gouterman, M. Optical spectra and electronic structure of porphyrins and related
rings. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978;
Vol. 3; pp 1—159.

123

(29)

(30)

(31)

(32)

(33)

(34)

(35)

Adar, F. Electronic absorption spectra of hemes and hemoproteins. In The
Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978; Vol. 3; pp 167-
209.

Leeson, D. T.; Berg, 0.; Wiersma, D. A. J. Phys. Chem. 1994, 98, 3913—3916.

Vanderkooi, J. M.; Adar, F.; Erecinska, M. Eur. J. Biochem. 1976, 64, 381.

Miller, R. J. D. Annu. Rev. Phys. Chem. 1991, 42, 581—614.

Henry, E. R.; Eaton, W. A.; Hochstrasser, R. M. Proc. Natl. Acad. Sci. USA 1986,
83, 8296—8696.

Cheung, M. S.; Garcia, A. E.; Onuchic, J. N. Proc. Natl. Acad. Sci. USA 2002, 99,
685—690.

Shtilerman, M.; Lorimer, G. H.; Englander, S. W. Science 1999, 284, 822—825.

124

6. Future Directions

The work in this dissertation opens an exciting possibility for the application of
the optical G-jump method in studying protein unfolding and misfolding dynamics. The
method can be used under controlled solvent and temperature conditions that relate far
better to the conditions relevant to disease than are possible using existing
temperature-jump and denaturant-mixing experiments. We anticipate that it will be
generally possible to use the optical G-jump method with any protein of interest provided
that a suitable electronic probe can be engineered into the structure using site-directed

mutagenesis or by solid-state peptide synthesis.

The optical G jump experiment can be triggered with a large number of intrinsic
or extrinsic probes. The choice of the probe is limited only by the excitation-energy range
that is accessible to the optical G-jump which is restricted to the spectral range to the red
of the ~320-nm onset of the absorption of tryptophan.1 Most organic chromophores emit
fluorescence solely from the lowest excited singlet electronic state, which is prepared by
internal conversion from the higher singlet states S...2 Furthermore, most fluorophores
have non-negligible oscillator strengths for absorption transitions to these higher singlet
states,3 so the optical G-jump excitation scheme will work with any fluorescent probe that

has a red-shifted absorption compared to the 320 nm cut off.

The work presented in this dissertation suggests that the characterization of the

apparent barrier heights for unfolding reactions in the target systems mentioned above

125

could be carried out through a comparison of dynamic Stokes-shift responses obtained as
a function of laser excitation wavelength. Below the barrier for unfolding, the response
will resemble the one observed for ZnCytc with Q-band excitation described in chapter 4.
As the laser wavelength is tuned to the blue, at some point the barrier should be
overcome, and an unfolding transition should occur. The barrier crossing could be
reported by responses like the ones observed for ZnCytc with its Soret band. The
characteristics of this response is fast initial red shift of the fluorescence spectrum, but
the key diagnostic observation is the breaking back to the blue as the refolding begins. At
excitation wavelengths to the blue of the spectral origin but below the unfolding barrier,
the initial red shift could exhibit fast components that arise fi'om vibrational relaxation
that are added to the ones that arise from normal sampling of the energy landscape near

the native state, but no break to the blue should be observed.

Scanning the excitation wavelength farther to the blue than the wavelength that
corresponds to the unfolding threshold may permit longer displacements (in time and in
distance along structural coordinates) from the native minimum. A comparison of the
dynamic Stokes-shift dynamics at and well above the threshold might reveal the onset of
trajectories to alternate minima on the energy landscape that lead to formally misfolded

structures.

The barrier-height determination could be repeated for at least two additional
experimental variable axes: the sample temperature and the solvent mixture. As has been
demonstrated with ZnCytc, addition of alcohols will vary the polarity and viscosity of the

surrounding solvents. The variation of polarity provides a direct means to manipulate the

126

barrier height for the unfolding reaction; this provides a test of the essential ideas of the

hydrophobic effect.‘1'8

The determination of the unfolding barrier height at different temperatures takes
advantage of a unique feature of the optical G-jump technique, that the temperature at
which most of the dynamics take place is that of the ambient, surrounding solvent
temperature. This is an important feature because the trajectories that would be accessible
to the unfolding/refolding system depend on the sample temperature. At low
temperatures, the trajectories would be expected to find the global free-energy minimum
on the energy landscape. As the temperature is raised, however, intermediate barriers that
steer the protein toward the global minimum may be overcome, allowing the system to
find local minima.9 These local minima may be the kinetic traps that are manifested by

the misfolded systems in disease.

127

6.1.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

References

Cantor, C. R.; Schimmel, P. R. Biophysical Chemistry. Part II: Techniques for the
Study of Biological Structure and Function; W. H. Freeman and Company: San
Francisco, 1980.

Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cummings Publishing
Company: Menlo Park, Califomia, 1978.

Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Second ed.; Kluwer
Academic/Plenum Publishers: New York, 1999.

Lesk, A. M. Biophys. Chem. 2003, 105, 179—182.

Frank, H. S.; Evans, M. W. J. Chem. Phys. 1945, 13, 507-532.

Honig, B.; Yang, A. S. Adv. Prot. Chem. 1995, 46, 27—41.

Privalov, P. L.; Gill, S. J. Adv. Prot. Chem. 1988, 39, 193—234.

Kauzmann, W. Adv. Protein. Chem. 1959, 14, 1—63.

Brooks III, C. L.; Onuchic, J. N.; Wales, D. J. Science 2001, 293, 612.

128

   

I"Eli‘lllllfiilllllI