The Clamp loader of Escherichia coli DNA Polymerase III

MISSING IMAGE

Material Information

Title:
The Clamp loader of Escherichia coli DNA Polymerase III kinetics of the ATP-dependent steps in the sliding-clamp loading reaction
Physical Description:
xv, 288 leaves : ill. (in ILLUS in FF) ; 29 cm.
Language:
English
Creator:
Williams, Christopher R., 1974-
Publication Date:

Subjects

Subjects / Keywords:
Research   ( mesh )
DNA Polymerase III -- genetics   ( mesh )
Escherichia coli -- genetics   ( mesh )
DNA Binding Proteins -- genetics   ( mesh )
DNA Replication -- genetics   ( mesh )
Nuclease Protection Assays -- methods   ( mesh )
Kinetics -- methods   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D)--University of Florida, 2003.
Bibliography:
Bibliography: leaves 274-287.
Statement of Responsibility:
by Christopher R. Williams.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 80287708
ocm80287708
System ID:
AA00012899:00001


This item is only available as the following downloads:


Full Text










THE CLAMP LOADER OF Escherichia coli DNA POLYMERASE III:
KINETICS OF THE ATP-DEPENDENT STEPS IN THE SLIDING-CLAMP
LOADING REACTION














By

CHRISTOPHER R. WILLIAMS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA



































Copyright 2003

by

Christopher R. Williams































This document is dedicated to my parents, family, and Heather Runyan.














ACKNOWLEDGMENTS

Foremost I would like to thank my advisor, Linda Bloom, Ph.D., for outstanding

guidance and providing an exceptional working environment in a state-of-the-art research

laboratory. I thank my committee at UF, Drs. Daniel L. Purich, Arthur S. Edison, J. Bert

Flanagen, and Alfred S. Lewin, for their assistance not only in guiding my project, but in

guiding me to become more wise and perceptive as a scientist. I must also acknowledge

my committee at Arizona State University, Drs. Neal Woodbury, Yuri L. Lyubchenko,

and Kenneth J. Hoober, for their assistance in the early stages of this dissertation project.

I acknowledge Manju Hingorani, Ph.D., for helpful discussions at the Keystone

Symposia meetings and protein preparation, Mike O'Donnell, Ph.D. for the generous

gifts of clamp loaders, Martin Webb, Ph.D., Myron F. Goodman, Ph.D., and Jeffery

Bertram for the gift of the phosphate binding protein plasmid, and help in

characterization and use of MDCC-PBP, and Petr Kuzmic, Ph.D., for contribution of the

custom version of DynaFit and assistance with kinetic modeling. I would like to thank

my following colleagues for their friendship and invaluable discussions regarding this

project and science in general: Brandon Ason, Ryan Shaw, John C. Lopez, Gabriel

Montafio, Ph.D., Gregory Uyeda, Jos6 Clement6, Ph.D., Joyce Feller, Ph.D. Finally, I

would like to thank my parents for their unwavering support of my endevours, and for

purchasing my first microscope and chemistry sets years ago.














TABLE OF CONTENTS
Page

ACKNOW LEDGM ENTS...................... ........................................................................ iv

LIST O F TA B LE S ...................................................................... ............ .........................

LIST O F FIG U R E S ...................... .......................................................... ....................... xi

A B ST R A C T .................................................... : ................................................................xiv

CHAPTER

1 INTRODUCTION STATEMENT OF PROBLEM.............................................1...

The Processive Escherichia coli DNA Polymerase In Holoenzyme........................... 1
DNA Polym erase III Holoenzym e..................................................................1...
T he P Sliding C lam p .......................................................................................... 2
The DnaX Clamp Loader Stoichiometry and Organization of Subunits............3
Clam p Loader Subunit Functions.............................................. ...................... 4
The M echanism of P Clamp Loading.......................................... ....................... 5
Importance of the E. coli Model Replication System ........................................7...
The General Problem Under Study and Research Questions Addressed.................. 8
Conformational Dynamics of the Clamp Loading Machine............................ 10
Enhancement of the Clamp Loading Machine by its Clamp............................ 13
The x and W Subunits are Required for Optimum Activity of the Clamp Loaderl4
Application of the Analyses of the Clamp Loading Machine to Other Complex
M olecular M machines ..................................................................................... 17
Novel Hybrid Devices Based on the Clamp Loader and Sliding Clamp............ 19
"Off-the-Wall" Example of the Clamp and Clamp Loader in a Novel Device...20
Design of Research Project .............................. ..........................21


2 LITERATURE REVIEW..............................................................................24

DNA Replication in Escherichia coli.............................................. ..................... 24
DNA Polym erase III Holoenzym e ........................................................................... 25
The T Subunit is the Coordinator of Pol III Holoenzyme Function and
P rocessivity ............................................................... ........................ 29
Structure of the 1 Sliding Clamp Processivity Protein.....................................31
The DnaX Clamp Loading M machine ............................................... ...................... 35
The AAA+ Superfamily of Motor Proteins....................................................40








X-ray Crystal Structure of the Clamp Loading Machine.......................45
Structure of the Nucleotide Binding Site and the Proposed Conformational
C change of the y Subunit................................................................................ 48
X-Ray Crystal Structure of the 8 Subunit Bound to a p Monomer and the '
Mechanism for Opening the P Sliding Clamp.......................... ..........52
DNA structural requirements for P clamp loading by y complex ....................56
Mechanism of the 0 Clamp Loading Reaction Cycle by y Complex......................60
Mutations of the 1 Clamp, and y Complex 8' and y Subunits: Effects on the
Clamp Loading Mechanism..............................................................65
The Clamp Loading Machine Within Polymerase III Holoenzyme........ ..... 68
Clamps and Clamp Loaders of Bacteriophage, Eukaryotic, and
A rchaeal O rganism s.............................................. .......... ................................ 71
Bacteriophage T4 Clamp and Clamp loader ....................................................73
Eukaryotic PCNA Clamp and Replication Factor-C Clamp Loader................75
Archaeal PCNA Clamp and Replication Factor-C Clamp Loader...................80


3 M ATERIALS AND M ETHODS ............................................................................ 84

Proteins, Reagents, and Oligonucleotide Substrates............................................ 84
DNA Polymerase III Proteins ... .................... ......... ............. .......... 84
Reagents .................... ........................................85
O ligonucleotide Substrates.................................................. ................... 85
Purification of Escherichia coli Phosphate Binding Protein............................ 86
Labeling of Phosphate Binding Protein with MDCC.......................................89
Characterization of MDCC-Labeled Phosphate Binding Protein ...........................90
Removal of Free-Inorganic Phosphate (Pi) Contamination with the "Pi-mop" ..91
Concentration and Efficiency of Labeling of MDCC-PBP............... ................ 91
Characterization of the fluorescence-molar response of MDCC-PBP to Pi........92
Active Site Titration of M DCC-PBP ......................................... .................... 93
Fluorescence Anisotropy Binding Assays...................................... ...................... 94
Calculation of A nisotropy ................................................................................. 94
Steady-state Measurement of Clamp Loader RhX-DNA Binding Kinetics.....95
Pre-Steady-State Measurement of Clamp Loader RhX-DNA Binding
K in etics ................................................................................. ...................... 9 9
Fluorescence-based MDCC-PBP ATP Hydrolysis (ATPase) Assay.................... 101
Steady-State Kinetics of ATP hydrolysis............................... ..................... 101
Pre-Steady-State Kinetics of ATP Hydrolysis ............................................... 104
Computer Modeling of ATP Hydrolysis Kinetic Data................................... 109
Correlated Pre-Steady-State MDCC-PBP ATPase Assays and Fluorescence
A nisotropy Binding A ssays................................................. ..................... 110
Stopped-Flow Dead Time Determinations............................................................. 112
Determination of the Dead Time for the Biologic SFM-4 Stopped-Flow........ 112
Determination of the Applied Photophysics SX. I 18MV Stopped-Flow
Reaction Analyzer Sequential- and Single-mix Dead Times ......................114








4 ATP-DEPENDENT CONFORMATIONAL CHANGE IN THE CLAMP LOADER 18

Intro d uctio n .............................................................................................................. 1 18
Steady-State Characterization of y Complex ATP Hydrolysis, DNA Binding and
C lam p Loading A ctivities................................................................................... 120
Enhancement of Steady-State ATP Hydrolysis Kinetics of y Complex by 0
clam p................................................................................................... 120
Steady-State DNA Binding and Clamp Loading Activities of 7 Complex....... 122
Pre-Steady-State Kinetics of DNA-Dependent ATP Hydrolysis by y Complex...... 127
Pre-Steady-State MDCC-PBP ATPasp Assays for y Complex in the Absence
and Presence of 1 Clam p ............................................................................ 127
Pre-Steady-State MDCC-PBP ATPase Assays at Different Concentrations of y
Com plex in the Presence of P1..................................................................... 129
ATPyS-Chase of Pre-Steady-State ATP Hydrolysis Activity by y Complex... 131
Pre-Steady-State Kinetics of 1 Clamp Loading by y Complex Initiated at
Different Steps of the Reaction Cycle .............................................................. 135
Kinetics of Clamp Loading when y Complex is Equilibrated with ATP.......... 135
Kinetics of Clamp Loading when y Complex is Equilibrated with ATP and 3 137
Kinetics of Clamp Loading when y Complex is Added Directly to a Solution
of A TP, 0 and D N A ................................................................................... 141
Kinetics of ATP Hydrolysis During Clamp Loading when y Complex is
E quilibrated w ith A TP ........................................................................................ 142
Kinetics of ATP Hydrolysis when y Complex is not Equilibrated with ATP.......... 144
Kinetics of Formation of the Two Populations of y Complex............................... 146
Computer Modeling of ATP Hydrolysis Reaction Kinetics.................................. 149
D iscussion.............................................................................. .................. .... 153


5 CHARACTERIZATION OF THE MINIMAL CLAMP LOADER COMPLEX
AND COMPARISON TO GAMMA COMPLEX....................... 163

Introduction ................................................................. ...................................... 163
y388' is the Minimal Clamp Loader Complex Which Can Bind DNA................. 165
Analysis of DNA Binding Activity of the Individual Subunits or
Sub-complexes of the Clamp Loader ......................................................... 165
Analysis of the DNA Binding Activity of 7388' Minimal Complex in the
A absence and Presence of 13 ......................................................................... 167
Comparison of P Clamp Binding Affinity of the Minimal Complex and y Complex169
Equilibrium p13'" Binding Activity of the Minimal Complex and y Complexl69
Apparent Dissociation Constant for ATP Binding the Minimal Complex or y
C om plex ..................................................................................................... 17 1
Kinetics of ATP Hydrolysis by the Minimal Complex Measured Using the MDCC-
PB P A TPase A ssay ................................................. .......................................... 173
Steady-State ATP Hydrolysis Kinetics ofy388' Minimal Complex in the
Absence and Presence of P Clam p............................................................... 173
Pre-Steady-State Kinetics of ATP Hydrolysis by y388' Minimal Complex in
the Absence and Presence of 1 Clamp......................... ......................... 176








Kinetics of ATP Hydrolysis when the Minimal Complex is not equilibrated
w ith A T P ...................................................................................................... 180
ATPyS-Chase of Pre-Steady-State ATP Hydrolysis Activity by the Y738'
M inim al C om plex ....................................................................................:... 183
ATPyS-Chase of Steady-State ATP Hydrolysis Activity by the y388' Minimal
Com plex or y Com plex............................................................................... 186
Clamp Loading Activity of the Minimal Complex is More Sensitive to ADP
than y C om plex ......................................................................................... 188
Pre-Steady-State Kinetics of Clamp Loading by the Minimal Complex Initiated at
Different Steps of the Reaction Cycle.............................................................. 190
Kinetics of Clamp Loading when the Minimal Complex is Equilibrated with
A T P an d ........................................................................ ............................. 19 1
Kinetics of Clamp Loading when the Minimal Complex is Equilibrated with
A T P .......................................................................... ............................... 194
Kinetics of Clamp Loading When the Minimal Complex is Mixed Directly
with a Solution of ATP, 1, and DNA.......................................................... 195
Direct Real Time Correlation of the Minimal Complex DNA Binding and ATP
Hydrolysis Kinetics in the Presence and Absence of 0 Clamp......................... 196
D discussion .................................................................... .......................................200
Understanding y Complex Kinetics by Characterization of and Comparison
with y388' M inim al Complex............................................. ...................... 200
7388' is the Minimal Complex with DNA Binding Ability, and Binds ATP and
p with Affinity Similar to y Complex..........................................................202
Pre-Steady-State ATP Hydrolysis and DNA Binding Kinetics: Analyses of the
Active and Inactive Clamp Loader States ........... .................................. 202
The x and W Subunits, Missing from the Minimal Complex, May Facilitate the
Conformational Dynamics of y Complex....................................................210
13 Clamp Enhances the Switch from Inactive to Active Clamp Loader
P populations ................................................................................................. 2 11
Experiments when the Minimal Complex was not Equilibrated with ATP
Reveal Slower Conformational Change Kinetics than y Complex................213
The Nature of Nucleotide Binding to the Minimal Complex and y Complex ..214


6 CONCLUSIONS AND RECOMMENDATIONS..............................................221

Introduction ......................... ............................................. .................................... 221
Steady-State Kinetics of the Clamp Loader in the Absence or Presence of 1..........224
Kinetics of ATP-Dependent Conformational Changes within the Clamp Loader...226
A Possible Mechanism for 0 Clamp Enhancement of Clamp Loader Activity........233
The Missing X and W Subunits are Responsible for the Kinetic Differences
Between the Minimal Complex and y Complex..........................................236
The X and y Subunits are E. coli Clamp Loader AAA+ Adaptor Proteins .......237
Program atic Recomm endations.................................................. ...................... 239
Pre-Steady-State Kinetics of 13 Clamp Binding...............................................239
Fluorescence Lifetime Measurement of MDCC-PBP Titrated with Inorganic
Phosphate ................................................... ............................................. 24 1








Further Investigation of ATP Binding and Nucleotide Exchange by the Clamp
L o ader................................................................................... ...................... 242
Analysis of Clamp Loader Conformational Dynamics by Circular Dichroism
Spectroscopy ................... .......................................................................... 244
Identification of the Putative Clamp Loader DNA Binding Surface ..............245
The X and V AAA+ adaptor hypothesis............................... 246


APPENDIX

COMPUTER MODELING OF EXPERIMENTAL KINETIC DATA ........................248

DynaFit Script For Fitting Shown in Figure 4-9 ........ ........................248
DynaFit Script For Fitting in Shown in Figure 4-10C ...........................................249
DynaFit Script For Fitting in Shown in Figure 5-6 ................................................251
Simulation Mechanisms for Equilibration Steps: KinTekSim Program ................252
D ynaFit O utput Indices ......................................................... .............................. 255
Fitting for y Complex Data in Figure 4-9..................... ............................255
Fitting for y Complex Data in Figure 4-10C................................................258
Fitting for Minimal Complex Data in Figure 5-6..............................................260
Fitting for Minimal Complex Data For Estimation of Conformational Rate
Constants From a Single D ata set................................................................. 262
Alternate Dynafit Model with a "Branch" Step at After Hydrolysis of Second ATP265
Alternate Dynafit Model Applied to y Complex in Figure 4-9 ...................... 265
Alternate Dynafit Model Applied to y Complex in Figure 4-10C ..................267
Alternate Dynafit Model Applied to the Minimal Complex in Figure 5-7 .......269
Alternate Dynafit Model Applied to y Complex in a DNA Binding Assay in the
A absence of P C lam p.................................................. ............................ 271


LIST OF REFERENCES........................ ............................................................274

BIOGRAPHICAL SKETCH............................... ...................................................... 287














LIST OF TABLES


Table p~ag

2-1. Clamps and clamp loaders through evolution.....................................................72

3-1. A ssay and protein buffers............................................................... ...................... 85

3-2. TG-plus media contents........................ .......................................................... 87

4-1. Steady-state ATP hydrolysis kinetics of y complex in the absence and
presence of p ........................................................ .... .......... ....................... 12 1

5-1. Steady-state ATP hydrolysis kinetic parameters for the minimal complex and y
complex in the absence and presence of 1..................................................... 174

5-2. Steady-state ATPyS-chase assay results: comparison of the minimal complex and y
complex in the presence or absence of ............................... ..........................187














LIST OF FIGURES


Figure page

1-1. Schematic of the p clamp loading reactiori for processive DNA synthesis..............6...

2-1. The crystal structure of 3 sliding clamp......................................... ..................... 32

2-2 The crystal structure of a AAA+ motor protein............................. 43

2-3. The crystal structure of the 7388' clamp loader.................................................... 46

2-4. The crystal structure of the 68-p complex......................................... .................... 53

2-5. A schematic cartoon of the basic steps in the clamp loading reaction and initiation
com plex form ation..................................................................... ....................... 62

2-6. Architecture of the polymerase III holoenzyme at the replication fork organized by
the DnaX clam p loading machine................................................ ..................... 69

3-1. Pi titration analysis of M DCC-PBP ................................................ ....................... 93

3-2. A plot of observed reaction decay rates as a function of NBS concentration......... 113

3-3. Fluorescent decay amplitudes plotted as a function of experimentally observed
decay rate constants to determine dead time of the SFM-4 stopped-flow........... 114

3-4. Sequential-mix reaction fluorescent decay amplitudes plotted as a function of
experimentally observed decay rate constants to determine dead time of the
SX. 18M V stopped-flow.............. ........................................................ 116

3-5. Single-mix reaction fluorescent decay amplitudes plotted as a function of
experimentally observed decay rate constants to determine dead time of the
SX. 18M V stopped-flow..................... ..................................................... 117

4-1. Steady-state DNA binding and clamp loading activities of y complex................ 123

4-2. Steady-state kinetics of clamp loading as a function of ATP concentration........... 125

4-3. Kinetics of ATP hydrolysis by y complex in the presence and absence of p.......... 128

4-4. Quantification of the number of ATP molecules hydrolyzed by y complex in the
first turnover of p clam p loading............................................ ....................... 130








4-5. Pre-steady-state kinetics of ATP hydrolysis by y complex in the presence and
absence of the 1 clamp in assays with and without an ATPyS chase............... 132

4-6. Kinetics of clamp loading measured in reactions that were initiated at different
stages of the loading cycle.............. .................................................... 138

4-7. Kinetics of ATP hydrolysis during the clamp loading reaction when y complex is
equilibrated w ith A TP............................................................... ..................... 143

4-8. Kinetics of ATP hydrolysis in reactions initiated by the addition of y complex to
A TP and D N A ........................................................ ...................................... 145

4-9. Kinetics of ATP hydrolysis when y complex is incubated with ATP for a defined
period of time prior to addition of DNA.......................... .......................... 147

4-10. Kinetic modeling of ATP hydrolysis reactions............................................. 150

5-1. Change in steady-state anisotropy for RhX-ss DNA in the presence of individual
subunits, sub-complexes, and y complex with or without ATP ........................... 166

5-2. Steady-state binding of the minimal complex or y complex with RhX-pt DNA with
and w without p ...................... ......... ....... ........................... 168

5-3. Steady-state anisotropy binding activity of y complex or the minimal complex with
p "iy in the presence or absence of ATP........................ .......................... 170

5-4. Minimal complex or y complex binding p13' as a function of ATP concentrationl72

5-5. Kinetics of ATP hydrolysis by the minimal complex or y complex in the presence
and absence of p1................................. ..... ............................................ 178

5-6. Kinetics of ATP hydrolysis of the minimal complex or y complex directly mixed
w ith pt D N A and A TP ............................................................ ....................... 181

5-7. Pre-steady-state kinetics of ATP hydrolysis by the minimal complex when chased
with non-hydrolyzable ATPyS ................................... ........... ................. 184

5-8. Effect of increasing ADP concentration on the steady-state 13 clamp loading
reaction for the minimal complex and Y complex ............................................ 189

5-9. Pre-steady-state kinetics of the clamp loading reaction initiated at different steps 192

5-10. Direct correlation of the kinetics of DNA binding and ATP hydrolysis by the
Y386' minimal complex in the absence and presence of 1 clamp ......................198

5-11. Kinetic modeling of ATP hydrolysis reactions.................................................. 203

6-1. Kinetic modeling of ATP hydrolysis reactions............................ ...................... 229









A-1. KinTekSim simulation of y complex, no equilibration with ATP........................253

A-2. KinTekSim simulation of the minimal complex equilibrated with ATP ...............254

A-3. KinTekSim simulation of the minimal complex, no equilibration with ATP ........255

A-4. DynaFit Fitting of the minimal complex (Figure 5-6), no equilibration time with
A T P .... .................................................. .................................................... 262

A-5. DynaFit fitting of the minimal complex (Figure 5-7B, black trace), Single
experimental data set, 1000 ms equilibration with ATP....................................265

A-6. Figure 4-9 data for y complex fit to alternate model with a "branch" step after
hydrolysis of tw o A TPs .......................................... ...................................... 267

A-7. Figure 4-10C data for y complex fit to alternate model with a "branch" step after
hydrolysis of two ATPs .............................................. 269

A-8. Figure 5-7, black trace data for the minimal complex fit to alternate model with a
"branch" step after hydrolysis of two ATPs.............................. ...........271

A-9. Example of the alternate model applied to DNA binding (anisotropy) data for a
reaction of y complex and Rhx-pt DNA..................................... .....273














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE CLAMP LOADER OF Escherichia coli DNA POLYMERASE III:
KINETICS OF THE ATP-DEPENDENT STEPS IN THE SLIDING-CLAMP
LOADING REACTION

By

Christopher R. Williams

December 2003

Chair: Linda B. Bloom
Major Department: Biochemistry and Molecular Biology

DNA polymerase III holoenzyme, the principal enzyme responsible for E. coli

chromosomal replication, synthesizes stretches of DNA thousands of nucleotides long at

a rate approaching 750 nucleotides s4' without dissociation. A ring-shaped DNA sliding-

clamp "p" topologically links the polymerase to DNA. A clamp loader, "y complex",

assembles P on DNA in an ATP-dependent reaction. This dissertation project was

undertaken for investigation of the mechanism of p clamp loading by the clamp loader for

processive replication. ATP binding and hydrolysis activities of the clamp loader

promote conformational changes modulating its binding affinity for the clamp and DNA.

Using fluorescence-based steady-state and real-time stopped-flow methods, the kinetics

of these dynamic conformational changes were measured. Pre-steady-state ATP

hydrolysis assays performed in the absence or presence of 1 resulted in biphasic or

monophasic kinetics respectively. Biphasic kinetics suggested that the clamp loader,

equilibrated with ATP, exists in a mixture of two dominant species. Addition of P








converted this mixture into a single activated population, effectively increasing its

concentration. These experiments, in addition to adenosine 5'-O-(3-thiotriphosphate)

(ATPyS)-chase assays showed that activated y complex hydrolyzed one ATP per y-

subunit at a rate faster than ATP dissociation. The hypothesis that y complex exists in

multiple species with ATP, was confirmed by measuring pre-steady-state clamp loading

kinetics in DNA-binding assays. When y complex was equilibrated with ATP, a mixture

of two species formed which when mixed with DNA and P exhibited biphasic DNA

binding kinetics. When equilibrated with ATP and 1, rapid monophasic DNA binding

kinetics resulted. Direct mixing of y complex with DNA 1 and ATP displayed slow

monophasic kinetics, limited by the conformational change rate. The rate of the

conformational changes separating the two dominant species was determined by

investigating their evolution in equilibration time with ATP. Computer modeling of

these experimental data revealed a conformational change rate of -4.5 s". Comparison of

the kinetics of a "minimal" clamp loader complex missing x and y subunits, revealed that

these subunits facilitate the conformational changes in y complex required for modulation

of 1 and DNA binding affinities during the clamp loading reaction.













CHAPTER 1
INTRODUCTION STATEMENT OF PROBLEM

The Processive Escherichia coli DNA Polymerase HI Holoenzyme

DNA polymerase III holoenzyme (pol III holoenzyme) is the principal enzyme

involved in replication of the Escherichia coli chromosome (Komberg and Baker, 1992).

The pol III holoenzyme copies the parent chromosome with surprising speed and

processivity, moving along the DNA at a maximum velocity reaching approximately 750

nucleotides per second, without release from the template at distances of over thousands

of nucleotides. Polymerase III holoenzyme must possess these functional characteristics

to complete rapid replication of the chromosome for cell division. Rapid and processive

DNA synthesis is required for all forms of DNA metabolism and genome maintenance

including DNA repair and recombination, and is evolutionarily conserved in all branches

of life, underscoring the significance of advancing the detailed understanding of the

proteins providing these functional mechanisms.

DNA Polymerase mI Holoenzyme

Pol III holoenzyme is a multi-protein complex consisting of ten distinct subunits (a,

s, 0, 0, t, 7, 8, 8', x and y). A core of the holoenzyme consisting of the a-5' to 3'

polymerase, F-3' to 5' exonuclease, and 0 subunits is paired-up by the 2 protein-dimer

(McHenry, 1982; McHenry and Crow, 1979). The r2-dimerized pol III core, termed pol

III', was originally purified from cells and isolated from pol III holoenzyme. The

replication activities of pol III core and pol III' were determined to be distributive in

kinetic and product size determination assays, whereas holoenzyme could synthesize








stretches of DNA with high processivity (Fay et al., 1981; LaDuca et al., 1983). Clearly,

these were incomplete forms of pol III holoenzyme, and were missing important factors

necessary for complete and rapid replication of the chromosome.

Another form of pol III, termed pol II*, contained all of the subunits of pol III

holoenzyme except one, the P subunit (Fay et al., 1982). Addition of this P subunit was

all that was required to convert the distributive pol III* enzyme into a highly processive

enzyme with DNA synthesis activity matching pol III holoenzyme. The P subunit is

initially associated with the replication fork DNA at sites called preinitiation complexes

where polymerase III holoenzyme assembles for DNA synthesis. Several early

biochemical studies revealed that a complex consisting of the X,y,8,8', y and V subunits

within pol III holoenzyme comprised a complex responsible for ATP-dependent

placement of the 1 subunit at primed sites on DNA for formation of the preinitiation

complexes, and were required for conferring processivity in replication (Maki et al.,

1988; Maki and Komberg, 1988).

The P Sliding Clamp

Solution of the X-ray crystal structure of the P subunit revealed that it was a ring-

shaped dimer of crescent-shaped protomers (Kong et al., 1992). P was termed a "DNA

sliding-clamp" due to its ability to topologically link pol III holoenzyme to template

DNA in such a way that holoenzyme remains tightly associated with DNA, yet has

significant freedom of movement on the DNA. The P clamp has an inner pore lined with

a-helices that act akin to "skates," traversing the clefts of the major and minor grooves in

the DNA backbone as the clamp slides along. This inner pore has a diameter large

enough to encircle duplex DNA as well as hybrid RNA-DNA duplex structure found at








the sites of preinitiation complex formation. A combination of hydrogen bonding,

hydrophobic and ionic interactions at the dimer interfaces strengthen and hold together

the 3 clamp subunits. P binds directly to the core of pol mI holoenzyme through the a

subunit (Stukenberg et al., 1991). The P clamp has a dissociation constant for

dimerization in the range of- 6.0 x 10-" molar, and is stable enough to remain on

circular DNA for ~ 100 minutes (Yao et al., 1996). The extraordinary stability of the

circular P3 dimer in solution and on the circular chromosome stresses the need for some

mechanism to open 0 for its loading onto and disassembly from DNA.

The DnaX Clamp Loader Stoichiometry and Organization of Subunits

The other processivity proteins within pol III holoenzyme (t,y,5,5',x and y) form a

complex that performs the duty of opening and loading (3 onto DNA for formation of

preinitiation complexes on primed DNA at the leading and lagging strands at the

replication fork (Kelman and O'Donnell, 1995). These proteins form the DnaX complex

"clamp loader," with the subunit stoichiometry [(DnaX)3,81i5'lX'li], in which each

subunit executes a unique function (Pritchard et al., 2000). The dnaX gene forms both r

and y subunits by a translational frameshift that forms a stop codon defining the y subunit

C-terminus, approximately two-thirds the length of the complete mRNA (Flower and

McHenry, 1990; Tsuchihashi and Kornberg, 1990). Therefore, y and T subunits are

identical in the first two-thirds of their sequence and structure. As a consequence, the N-

terminal domain oft acts akin to the 'y subunit, and the unique C-terminal domain of t

has a specialized function in coordinating the holoenzyme at the replication fork

(Dallmann et al., 2000). Due to the requirement of the T subunit for dimerization of the

pol III core, the DnaX complex associated with pol III holoenzyme most likely has a








stoichiometry of t2y1 (Pritchard et al., 2000). Each of the other subunits (8,8',x and Wi) is

present in a single copy in this clamp loader (Onrust et al., 1995). For in vitro

reconstitution of the DnaX clamp loader, the X and y subunits bind the y subunit through

V-y interaction (Glover and McHenry, 2000), and greatly increase the affinity for the 85'

subunits for the complex (Olson et al., 1995). The structure of the E. coli clamp loader

recently also revealed existence of a trimer of y (DnaX) subunits (Jeruzalmi et al.,

2001a).

Clamp Loader Subunit Functions

The x subunit interacts with single-stranded DNA binding protein (SSB) at the

replication fork (Glover and McHenry, 1998). This x-SSB interaction is thought to be

involved in a primase-to-polymerase switch prior to formation of preinitiation complexes

(Yuzhakov et al., 1999). The y subunit has no known function other than forming a

structural bridging contact for X to the clamp loader through W| interaction with the y

subunit (Glover and McHenry, 2000; Xiao et al., 1993b). The function of the X-SSB

interaction is well characterized, however there is no known function for the X and W

subunits directly in loading the clamp in DNA.

It is the 5 subunit that binds to, and alone has the ability to open, the P clamp

(Jeruzalmi et al., 2001b; Leu et al., 2000). The surface of 6 subunit which binds to p is

concealed by the 6' subunit when the clamp loader is in an "inactive" state (Jeruzalmi et

al., 2001a). The T and y subunits transduce the energy from ATP binding and hydrolysis

into mechanical work within this clamp loading machine to load the clamp on DNA

(Bertram et al., 2000; Onrust et al., 1991).








All of the subunits of the clamp loader, except X and y are members of a diverse

superfamily of AAA+ (ATPases Associated with a variety of cellular Activities)

molecular motors (Neuwald et al., 1999). This AAA+ superfamily contains conserved

sequences and structures for nucleotide binding and hydrolysis that drive conformational

changes in these motor proteins. In the DnaX clamp loader ATP binding and hydrolysis

by the r (i.e., through the N-terminal y-region) and y subunits promote conformational

dynamics that modulate the binding affinity for the P clamp as well as DNA for the

loading reaction. The 8 and 8' have been identified as AAA+ superfamily members

based on sequence and structural homology of 8', and structural homology in the case of

8, however, neither has the ability to bind or hydrolyze ATP, although they are believed

to actively participate in the conformational dynamics of the clamp loader for the loading

reaction (Jeruzalmi et al., 2001a; Podobnik et al., 2003).

The Mechanism of P Clamp Loading

Figure 1-1 depicts a simplified schematic of the clamp loading mechanism (under

study in this dissertation) for processive DNA synthesis by polymerase III. Initially, ATP

binds to the clamp loader's T and y subunits. Binding of up to three ATP molecules

causes conformational changes in the T and y subunits, and therefore changes the overall

structure of the clamp loader. Although the exact nature of these ATP-dependent

conformational changes is not yet known, there are two major consequences following

the conformational changes. The P clamp interaction surface of the 8 subunit becomes

exposed, and a DNA binding surface on the clamp loader either forms or becomes

exposed. Interaction of the 8 subunit with p has been biochemically and structurally

characterized, showing that 8 induces a conformational change in the 0 subunit that











Clamp Loader Core
Complex f



S+ ATP
plate DNA confornational + pi
changes


A.

PClamp




Primed-Ter


Figure 1-1. Schematic of the p clamp loading reaction for processive DNA synthesis. A)
p clamp, and the DnaX clamp loader are present with a primed-template DNA
substrate. B) Binding of ATP increases the affinity of the clamp loader for
both P and DNA by conformational changes producing exposure of the 6
subunit-0 interaction surface, and formation of a putative DNA binding
surface. Interaction of 8 with P induces conformational changes in the clamp
that open it at a single interface. C) Primed-template DNA triggers ATP
hydrolysis and rapid dissociation of the clamp loader from the loaded clamp.
D) DNA Polymerase then binds the same surface of p that was previously
occupied by the clamp loader. E) Incorporation of deoxyribonucleotide-
triphosphates (dNTPs) by polymerase then proceeds without polymerase
dissociation from the template.

triggers opening at a single dimer interface (Jeruzalmi et al., 2001b; Turner et al., 1999).

The nature of the putative DNA binding surface on the ATP-bound clamp loader remains

unknown. However, it is clear that ATP binding is required for the clamp loader to bind

DNA, and further that the clamp loader preferentially places P at the 3'-end of the primer

on a replication-proficient template (Ason et al., 2003). Binding primed-template DNA

triggers the clamp loader to hydrolyze ATP. ATP hydrolysis causes release of the clamp


Polymerase I








loader from 13, which then tightly closes on DNA when the 8 subunit-interaction is

removed. The discharged clamp loader is then in some inactive state at this point in the

cycle, and most likely must undergo some as yet undefined conformational changes in

order to continue loading clamps. Immediate removal of the clamp loader from the

loaded clamp is essential since the clamp loader and the polymerase a subunit share the

same binding surface on the clamp. Once bound to its clamp, polymerase can

processively replicate thousands of nucleotides without further dissociation from the

template.

Importance of the E. coli Model Replication System

In E. coli, the clamp loader and 13 clamp are utilized in DNA replication from

initiation at the origin to DNA partitioning of parent and daughter chromosomes upon

termination (Katayama, 2001; Levine and Marians, 1998). Aside from its interaction

with pol III core, 3 clamp has been found to interact with all five E. coli DNA

polymerases, as well as several other partner proteins in DNA metabolism such as the

MutS, UvrB, and DNA ligase (O'Donnell and Lopez de Saro, 2001; Tang et al., 1999).

The 13 clamp and DnaX clamp loading machine subunits are functionally conserved

across evolution to all branches of life suggesting a fundamentally similar mechanism for

processivity in all DNA metabolism (Ellison and Stillman, 2001). Increasing structural

analyses also reveal that the composition of these clamps and clamp loading machines

from such organisms as diverse as a bacteriophage and a human are remarkably similar,

and that all clamp loaders utilize AAA+ motor proteins (Davey et al., 2002). Using E.

coli as a model replication system has been outstandingly valuable for the determination

of how complex multi-protein machines interact and work together in a well-regulated

efficient manner for DNA replication. It is important to determine how the individual








protein subunits of these complex biological machines carry out their own functions and

communicate with each other for completion of their tasks. Additionally, understanding

of these mechanisms will allow for a more complete comprehension of the important

DNA metabolic tasks in which they are involved.

The General Problem Under Study and Research Questions Addressed

The general problem under study in this project is the many remaining unknown

and questionable aspects of the mechanisms by the clamp loader for p clamp loading and

the results and interpretations of other investigations. Conformational changes in the

clamp loader are a requirement for modulation of clamp loading. However, the kinetics

of clamp loader conformational changes are unknown. Different nucleotide-dependent

clamp loader conformational species have been identified in proteolytic digestion

experiments, but how their abundance and dynamics control the loading reaction cycle is

not understood mechanistically. How do the nucleotide-dependent conformational

dynamics within the clamp loader subunits drive this machine for clamp loading?

Chapter 4 of this dissertation details an investigation that addresses this question. The

studies describe the existence of distinct conformational species of the clamp loader and

the kinetics of their activity in the clamp loading mechanism. Chapter 4 further addresses

the questions concerning how P and ATP affect the kinetics of clamp loader

conformational changes. An additional key question asked in chapter 4 is that of the

nature of nucleotide binding to the clamp loader, and whether there is interdependency

between ATP binding and the conformational dynamics. Computer modeling was

applied to the experimental data for analysis of these important questions.

Sliding clamps of all organisms studied to date are known to enhance the activity of

their respective clamp loader. How the 1 clamp affects the y complex clamp loader for








promotion of its own loading onto DNA was tested in this research project. Although it

is understood that 0 increases the ATP hydrolysis energetic of the clamp loader, it

remains unknown what the kinetic mechanism of this enhancement may be. By

comparison of y complex with the y388' minimal clamp loader in chapter 5, along with

experiments presented in chapter 4, this dissertation addresses how p clamp affects the

specificity and stability of nucleotide binding by the clamp loader, and therefore the

kinetics of the conformational dynamics during the clamp loading reaction for its

enhancement.

In comparison to the AAA+ clamp loading machines of other organisms outlined in

chapter 2, the E. coli clamp loader contains two additional subunits (x and y). It is

known that X and y assist in the assembly of other subunits in the clamp loader, and the X

subunit has a role in primase-to-polymerase switching at the replication fork, but do they

serve any direct function in the clamp loading mechanism by y complex? Chapter 5 of

this work outlines a detailed characterization of the y 35' minimal clamp loader. This

minimal complex is missing the X and y subunits, and is used to assess their roles in the

clamp loading mechanism by direct experimental comparison to y complex. Are the X

and y subunits "AAA+ adaptor" proteins of y complex? Do they provide an example of

a novel adaptor function through their effect on the conformational dynamics of the

clamp loader? Although not originally hypothesized, the findings presented here along

with findings presented elsewhere suggest that they are.

Finally, what can the details learned here about the E. coli clamp loading machine

tell us about other molecular machines driving and regulating complex and diverse

cellular tasks in other organisms? This machine is essentially a molecular switch that








catalyzes a fundamental reaction in all forms of life. Can the knowledge gained through

study of this clamp loading machine be used for development of novel biological or

hybrid molecular machines?

Conformational Dynamics of the Clamp Loading Machine

Solution of the X-ray crystal structures of a minimal 7388' clamp loader complex

and a C-terminal truncated y subunit has recently given information revealing some of the

structure-function relationships of the DnaX clamp loading machine (Jeruzalmi et al.,

2001a; Podobnik et al., 2003). The subcomplex of (y3818') forms a minimal complex

with clamp loading activity similar to the complete DnaX clamp loader (Onrust et al.,

1991). Each of the 73, 8, and 8' subunits is a AAA+ superfamily member, and is

composed of three structural domains that form "C-shaped" molecules. The N-terminal

domains-I and II encompass the conserved nucleotide binding site (in y subunits only),

and the third C-terminal domain-Ill forms an oligomerization region for this

heteropentameric complex. Domain-II generally forms a mobile "hinge" between

domains-I and III. ATP binding and hydrolysis by the y-AAA+ motor subunits directly

trigger conformational changes in these subunits. Although the 8 and 8' subunits are

AAA+ proteins, they do not have the ability to bind and hydrolyze ATP, but most likely

do undergo conformational changes caused by y subunit conformational movements. The

structure of the 7358' clamp loader complex showed that the complex was arranged as a

heteropentameric ring through the C-terminal oligomerization domain, and that there was

extensive contact between each of the five subunits extending "down" in asymmetric

orientation from the oligomerization domain. The nucleotide binding sites were found in

the interfaces between the 8'-y~, 71-72, and Y2-73 subunits. At least one and possibly two

of these interfacial nucleotide binding sites is thought to be constitutively open whereas








the third is deeply buried in an interface. This lead to a model describing a sequential

series of individual y-subunit conformational changes whereupon ATP binding to an

open interface caused a conformational change opening the adjacent interface and so on

(Jeruzalmi et al., 2001a), where the status of ATP binding to the nucleotide binding sites

is communicated between,the subunits of the clamp loader modulating its affinity for p

clamp and DNA.

There are several limitations to the mechanistic inferences made in light of these

structures. The clamp loader structure was determined in the absence of nucleotide.

Therefore, the observed subunit interactions most likely do not represent the functional

complex, as it would appear in the cell, and the overall conformation was thought to be

largely due to a crystal-packing artifact. The truncated-y subunit structure was

determined in the presence of non-hydrolyzable adenosine 5'-O-(3-thiotriphosphate)

(ATPyS), and crystallized as a tetramer (not as it would appear in a functional complex).

The structure contained electron density consistent with ATPyS molecules bound to two

protomers, ADP bound to a third, and no nucleotide in the fourth. ATPyS is sufficient for

both P clamp and DNA binding by the clamp loader but does not confer clamp loading

activity in biochemical assays (Bloom et al., 1996). Therefore, the ATPyS-bound y

subunit structure revealed important features of the conformational change in this motor

subunit. However, the y subunit was truncated and a substantial portion is missing from

its C-terminus. Only the conserved AAA+ nucleotide binding domains-I and 1I were

present. This means that the different conformations observed in each of these structures

may be considerably different than the actual structure in solution, or within the cell. The

experiments presented in this dissertation do not address structure specifically, however








these in vitro studies do directly address the kinetics of the conformational dynamics of

the clamp loader in solution conditions during real-time clamp loading reactions.

There are significant nucleotide-dependent conformational changes in the clamp

loader that must be made during the clamp loading reaction for intersubunit

communication modulating p clamp and DNA binding. Currently, only proteolytic

protection assays (Hingorani and O'Donnell, 1998) have provided biochemical proof of

the nucleotide-dependent conformational changes, however there is no kinetic detail

known regarding the nature of these conformational changes during the clamp loading

reaction. It has been previously hypothesized that the rate-limiting step of the clamp

loading mechanism occurs in the clamp loader, away from DNA (Bloom et al., 1996),

and could be ADP-release or additional conformational changes that "reset" the clamp

loader for resuming the clamp loading cycle. The results of this dissertation address the

kinetics of the nucleotide-dependent conformational dynamics of the clamp loader and

add detail to what is known about the clamp loading mechanism, initially asking, how do

ATP-dependent conformational dynamics drive and regulate this molecular machine for

clamp loading?

The kinetics of nucleotide binding and hydrolysis by the clamp loader, DNA

binding, and p clamp loading have been determined using several biochemical analyses

in this project. Together, these explorations have lead to a hypothesis based on

conformational dynamics for the internal workings of the clamp loading machine. Here,

it is proposed that the clamp loader exists, dominantly, in two distinct conformational

states that are either "activated" or "inactive" for clamp and DNA binding when in

equilibrium with ATP. This project has addressed the kinetics of the ATP-dependent








conformational changes separating the two clamp loader species in the clamp loading

mechanism, and additionally shows that there is complexity in the conformational

dynamics between these two major states. This added complexity probably arises due to

the proposed asymmetry in the three nucleotide binding sites, and the possibility of

having differential affinities for ATP within these sites causing a complex mixture of

conformational intermediates within the clamp loader. Elucidation of the nature of the

additional mixture of conformational states will require further study, but overall this

project reveals, kinetically, higher populations of two major states modulating the

function of the clamp loader in the 1 clamp loading mechanism.

Enhancement of the Clamp Loading Machine by its Clamp

It has been known for some time that the P clamp enhances, but does not trigger,

the ATP hydrolysis activity of the clamp loader in the presence of DNA (Onrust et al.,

1991). This feature of a sliding clamp enhancing the DNA-dependent ATP hydrolysis

activity of its clamp loader is a general characteristic in bacteriophage (Pietroni et al.,

2001), eukaryotic (Yoder and Burgers, 1991), and archaeal organisms (Oyama et al.,

2001); thus it is a fundamental attribute in all clamp loading mechanisms. However, the

mechanism of how a clamp enhances the activity of its clamp loader has remained a

mystery. Part of the originally proposed research project was to determine how the E.

coli 1 clamp affected the kinetics of ATP hydrolysis by the clamp loader. This

dissertation addresses the mechanism of 13 clamp enhancement of clamp loader DNA-

dependent ATP hydrolysis, and reveals several important findings interrelated with

nucleotide binding and the conformational dynamics of the clamp loader. It comes as no

surprise that the mechanism of 0 clamp enhancement of ATP hydrolysis activity is

related to a conversion of the equilibrium between the dynamic conformational states of








the clamp loader. It is hypothesized that the p clamp converts the conformational

equilibrium of the clamp loader with ATP into a completely "active" population (i.e., a

complex of ATP-bound clamp loader and 0). An extension of this hypothesis is that the P

clamp affects the affinity and specificity of the clamp loader for ATP, effectively

"trapping" the nucleotide within the clamp loader promoting formation of the active 3-

bound complex poised for loading onto DNA. Selective binding and trapping the active

clamp loader conformation may also stabilize a putative DNA binding surface on the

clamp loader as well. An increase in the apparent concentration of this active complex

poised for clamp loading would provide a direct mechanism for the increase in ATP

hydrolysis turnover rate observed in kinetic assays. Due to the considerable evolutionary

conservation of clamp loader function across all branches of life (Davey et al., 2002), the

hypothesized mechanism of clamp enhancement of clamp loader activity presented in this

dissertation could also apply to clamp loading for processive DNA synthesis in all other

organisms as well.

The x and W Subunits are Required for Optimum Activity of the Clamp Loader

The X and Wi proteins do not bind DNA, do not bind or hydrolyze ATP, and are

dispensable for clamp loading activity. At the replication fork in vivo, the X subunit is

involved in an important interaction with SSB, and a primase-to-polymerase hand-off

through association with the clamp loader, but not directly involved in preinitiation

complex formation by the clamp loader. A subcomplex of the clamp loader (y388'), of

which the crystal structure was determined, does not contain the x and Wi subunits, yet

maintains nearly the same ATP hydrolysis and clamp loading activity as the clamp loader

with X and W (Onrust et al., 1991). Although the structure of the X-%p dipeptide is known,

it is still a mystery where it binds to the clamp loader. For these reasons the 7388' clamp








loader has been called the minimal clamp loader complex, and in some cases, it has been

called the clamp loader itself.

This dissertation encompasses a detailed biochemical characterization of the (7388')

minimal clamp loader, and comparison of its activities to the "y complex" clamp loader.

This work shows that there are in fact several mechanistic distinctions between the

minimal clamp loader and y complex clamp loader. These differences have generally

revealed that the minimal complex is slightly hindered in its nucleotide binding and

hydrolysis properties and hence clamp loading.activity. Therefore the X and y subunits

are required for optimum clamp loading activity, conceivably by strengthening the

intersubunit communication within the clamp loading machine. Previous studies

revealed that the X and y subunits did in fact stabilize the clamp loader by increasing the

affinity oft and y subunits for 86' (Olson et al., 1995). One of the most obvious

deficiencies in minimal complex function is the loss of the ability to stabilize ATP

binding in an active complex with p clamp. This is hypothesized to be the result of

possibly slower conformational changes in the minimal complex, and perhaps "looser"

conformational dynamics in the minimal complex leading to more conformational

complexity (i.e., more inactive intermediate conformational states) between the major

two states hypothesized for the clamp loader. This research shows that the population of

inactive conformations of the minimal complex is several-fold greater than the population

of active conformation in equilibrium with ATP. P clamp does sustain the ability to

enhance the ATP hydrolysis activity of the minimal complex. However, the results of

kinetic analysis of ATP hydrolysis activity reveal that this enhancement is less than that

of y complex, and that P must do more "work" to convert the large population of inactive








conformational states of the minimal complex into the trapped active conformation. To

date, there is no other steady-state and time-resolved biochemical analysis comparing y

complex to the minimal complex in such a comprehensive manner as that presented in

this dissertation, and the results have allowed a much more clear appreciation for the

clamp loading mechanism catalyzed by the DnaX clamp loader.

The clamp loader is a AAA+ protein containing machine. The x and i subunits are

structurally unrelated and much smaller than the AAA+ subunits of the clamp loader.

Since the X and W subunits seem to be unique to the E. coli clamp loader, and most likely

other gram-negative bacteria (Xiao et al., 1993a; Xiao et al., 1993b), they provide some

unique functions in these prokaryotic clamp loaders compared to those of other

organisms. They present a form of substrate specificity to the clamp loader for the SSB-

coated lagging strand at the replication fork for assistance in preinitiation complex

formation for Okazaki fragment synthesis. These characteristics of the X and W subunits

associate them with a new and growing list of AAA+ adaptor proteins. AAA+ adaptor

proteins provide a simple and effective way for modulation of AAA+ machine function,

giving the machine better control over substrate specificity and rapid redirection of its

specific activity (Dougan et al., 2002). Along with their function at the replication fork,

the x and W subunits are proposed here to provide increased structural stability and

facilitated conformational changes for intersubunit communication within the AAA+

clamp loading machine. It is thus hypothesized that they are in fact AAA+ adaptor

proteins and reveal a novel adaptor function for this newly defined subset of proteins.








Application of the Analyses of the Clamp Loading Machine to Other Complex
Molecular Machines

Like the clamp loader, there are many other enzyme complexes that perform

complicated biological tasks such as assembly and disassembly of new protein-protein

complexes, protein-cofactor complexes, and protein-DNA complexes. Orchestration of

protein-protein and protein-nucleic acid interactions is likely to utilize converging

mechanisms to drive the diverse functions of these multi-protein enzyme complexes. The

intersubunit conformational communication observed structurally and functionally by the

clamp loaders drive several essential jobs within the large replication complexes in all

forms of life. The clamp loader catalyzes assembly of a new protein-nucleic acid

complex without making or breaking any covalent bonds of either component. The

energy for this assembly is derived from within the clamp loader by ATP-dependent

conformational changes that essentially switch the clamp loader between "on" and "off"

states. Understanding of subunit recognition and conformational communication driving

similar switching mechanisms in molecular machines is a complicated task in science,

but explorations such as that presented in this dissertation for the y complex clamp loader

could help accelerate other investigations. Not only for ATP- and GTP-utilizing energase

enzymes, but also for transmembrane transport machinery, cellular cargo-transport

motors, chaperone and protein modulating complexes, proteases, etc.

The investigations outlined in this dissertation are a highly focused and detailed in

vitro examination of the mechanism of a molecular machine's mechanism of switching

between active and inactive states for assembly of a protein-DNA complex at the expense

of energy from ATP binding and hydrolysis. The details learned here about protein

conformational changes and communication of these changes within the machine could








be applied to other solution-based structure-function studies for a broad range of proteins.

A key.result of this and prior work shows that the information for conformational

changes in these proteins is stored in the structure of the proteins themselves, akin to the

storage of protein folding information encoded within secondary and tertiary peptide

structure (Fersht and Shakhnovich, 1998). For example, the 0 clamp is not simply an

unintelligent ring-shaped protein dimer, but contains structural information for release of

a spring-like trigger for opening of one of its sturdy interfaces (Ellison and Stillman,

2001). In the clamp loading machines, nucleotide-dependent intersubunit communication

induces the extraction of intramolecular-stored information for subunit conformational

changes that, in turn, can induce changes in partner proteins.

Molecular motors and switching mechanisms are common in many energase

enzymes other than the AAA+ proteins that make up the clamp loader under study here.

For example, several molecular motors move along the cellular architecture of

microtubule and actin cytoskeletal networks (Mehta et al., 1999; Vale, 2003). These

motors power transport of organelles and other cargo throughout the cell, drive

chromosomal segregation in cellular division during mitosis, give cells the capability of

motility, and power the contraction of muscle fibers that give animals the extraordinary

strength needed for a diverse range of movement, and let their hearts beat for entire

lifetimes (Vale and Milligan, 2000). The machines powering these processes are not

much unlike the clamp loading machine detailed in this dissertation. They each generally

contain several domains including nucleotide binding sites, protein-protein and protein-

nucleic acid binding regions, hinge-like domains, multisubunit oligomerization domains,

and regulatory subunit or cofactor interfaces (Vale and Milligan, 2000). The








conformational movements made by several of these nucleotide dependent energases

drive their functions, give them processive activity, and allow them to generate molecular

power that is matched on a relative scale only by man's largest macromolecular machines

(Baker and Bell, 1998; Ellison and Stillman, 2001). Several molecular switches also act

through conformational dynamics allowing protein-protein and protein-nucleic acid

interactions in signal transduction cascades (Franco et al., 2003; Phillips et al., 2003),

chromatin structural modification (Hakimi et al., 2002), transcriptional and translational

regulation (Mazumder et al., 2003), and ion channels driving action potentials for

neuronal signaling. The ras-GTPase activating protein provides an example of protein-

protein contact directly related to the subunit interfaces within the clamp loader. A

conserved "arginine-finger" residue is positioned between two proteins for catalysis of

nucleotide hydrolysis, inducing conformational changes that switch the activity status of

the protein (Ahmadian et al., 1997). Nearly all cellular processes are driven or regulated

in some way by phosphorylation / dephosphorylation mechanisms, but the activities of

energase molecular motors and switches (Purich, 2001) also appear to have great

importance in maintaining cellular function in all organisms. Therefore the significance

of investigating the inner workings of a molecular machine such as the E. coli clamp

loader extends far beyond the understanding of the mechanisms of DNA metabolism.

Novel Hybrid Devices Based on the Clamp Loader and Sliding Clamp

Understanding of the mechanism of the clamp loader molecular machine;

structurally and functionally how protein subunits come together and communicate with

each other based on nucleotide binding status can be exploited for development of novel

molecular hybrid devices. In even simple terms, the clamp loader is a molecular switch,

activated by ATP binding, and inactivated by DNA-dependent hydrolysis of ATP. The








inactive and active states of this molecular switch are driven by conformational changes,

and further affected by binding a partner protein (the clamp). Essentially, the reaction

mechanism under study in this dissertation is a molecular switch that entails placement or

removal of (protein) rings on nucleicc acid) rods or hoops. This activity produces a

topological connection for another machine (polymerase) to the rod or hoop structure,

essentially increasing its concentration at the preferred catalytic site (template DNA).

The clamp loader switching mechanism could be utilized in other systems that

require regulated activation/deactivation or in systems such as memory or

communications devices that require a simple (binary) on/off signal. By modification of

the protein properties of the clamp loader subunits it is feasible that this energase

machine/switch could be custom-tailored for guided assembly of proteins (or other

materials) other than the clamp, onto specific substrates in order to control the putative-

complex concentration, and activate or deactivate it specifically through NTP binding and

hydrolysis.

"Off-the-Wall" Example of the Clamp and Clamp Loader in a Novel Device

There is a need for understanding and improvement of industrial catalysis systems

(Coontz et al., 2003). Briefly, for homogenous catalysts, where the molecular catalyst

and reactant are dispersed in the same phase, there exists a seemingly simple problem of

separation of the reactants and catalysts from products. The development of

heterogeneous catalysts have provided a solution by direct separation of catalyst and

reactant in different phases, but in most cases do not provide the surface area for cost-

efficient devices. One could envision use of a clamp loading machine to place (and

remove) a ring-shaped clamp, either with the ability to bind a specific catalyst, or

containing the catalyst itself, onto the reactant. This process would allow a substantial








increase the heterogeneous catalytic surface area on an insoluble nano-sized rod- or hoop-

shaped reactant. Perhaps the products would easily be separated in solution or gas

phases. These devices would be "smart" devices where the catalytic components contain

the information needed for their assembly and disassembly for disposal or recycling. For

example, a spring-loaded clamp that can be placed and removed at will by a specialized

machine with easily regulated activity. Improvement of industrial catalysts is only a

single (imaginative) example of how the functional knowledge about sliding clamps and

clamp loading machines could be used for design of a novel hybrid device. With

imagination and lots of money, these and other, perhaps more significant biologically

functional devices could be developed.

Design of Research Project

This research project was developed on two fluorescence-based experimental

methodologies for investigation of the kinetics of DNA dependent ATP hydrolysis and P

clamp loading activities of the y complex clamp loader and the 7385' minimal clamp

loader complex. The two experimental methods utilized were an anisotropy binding

assay, and an E. coli phosphate binding protein-based MDCC-PBP ATP hydrolysis assay.

The anisotropy binding assay was employed for investigation of the dynamics of the

interactions between the clamp loader and fluorescent-labeled DNA or fluorescent-

labeled P clamp. Studies of the DNA dependent ATP hydrolysis kinetics of the clamp

loaders were performed with the MDCC-PBP ATPase assay. Both methodologies

allowed for steady-state and time-resolved measurements of clamp loading kinetics.

For the anisotropy binding assay, depolarization of the emission from a fluorescent

probe reports on the rotational dynamics of the fluorescent-labeled species (Lakowicz,








1999). By monitoring changes in anisotropy, one can measure binding dynamics and

observe intermediates that arise in real time during a given reaction. This enables

elucidation of the kinetics of discrete steps involving interactions with the labeled species

in a reaction mechanism. To study binding kinetics of the clamp loader with P clamp as

well as the binding kinetics of the clamp loader with DNA in the presence and absence of

p clamp on steady-state, and pre-steady-state time scales, the fluorescence depolarization

of X-rhodamine-labeled DNA (RhX-DNA) or pyrene-labeled P clamp (3pyr') was

measured.

ATP hydrolysis kinetics were measured with the MDCC-PBP ATPase assay. This

assay utilizes a site-specific mutant of E. coli phosphate binding protein (phoS gene

product) that allows covalent attachment of an environmentally sensitive fluorescent

probe (N-[2-(1-maleimidyl)ethyl]-7-(diethylamino)coumarin-3-carboxamide, "MDCC")

near the phosphate binding cleft (Brune et al., 1994). Binding of inorganic phosphate (Pi)

product (i.e. from ATP hydrolysis by the clamp loader) to the labeled-phosphate binding

protein is both rapid and tight, and produces a substantial increase in fluorescence

intensity readily measurable in a fluorimeter or stopped-flow real-time detection system.

Computer modeling was performed using two different programs for simulation

and fitting of experimental data for investigation of the clamp loader conformational

change mechanism. The KinTekSim program (Barshop et al., 1983; Dang and Frieden,

1997) was used to test model reaction mechanisms by simulation and visual comparison

to the experimental data. A fitting program (DynaFit) was used to fit the experimental

time course of the reaction with a least-squares regression methodology (Kuzmic, 1996).





23

Computer modeling of the reaction kinetics provided a way to validate, and predict

several of the mechanistic features determined in this project.













CHAPTER 2
LITERATURE REVIEW

DNA Replication in Escherichia coll

A complex assembly of DNA replication proteins forms at each replication fork for

synthesis of a nascent Escherichia coli chromosome. At the leading edge of the

replication fork is a topoisomerase, an enzyme that relieves torsional stress created by

duplex DNA unwinding. Opening up the replication fork and priming the leading and

lagging strand templates is a machine called the primosome. The primosome is made up

of two proteins, DNA helicase, which unwinds parental DNA, and primase, which

synthesizes RNA primers, the sites of initiation of replication of the parent templates.

Stabilizing the single strand template DNA that is exposed upon unwinding by the

helicase is single-stranded DNA binding protein. Associated with the primosome is a

complex of enzymes that synthesize and proofread the nascent DNA. This machine is

DNA polymerase Ill holoenzyme, a dimer of the synthesizing and proofreading units that

coordinate simultaneous high fidelity elongation of the leading strand, and Okazaki

fragments on the lagging strand of nascent DNA. The dimeric DNA polymerase III is

physically linked to the parental template leading and lagging strands by ring-shaped

sliding clamp processivity proteins, such that it cannot easily dissociate from the

template. A clamp loading machine, located within dimerized DNA polymerase III

utilizes the energy of ATP binding and hydrolysis to load the circular sliding clamp on

the leading strand template and on each Okazaki fragment of the lagging strand template.

DNA polymerase I and DNA ligase follow at the tail end of the replication fork, where








DNA polymerase I is poised to excise the RNA primers and replace them with DNA, and

DNA ligase seals the remaining nicks produced between the Okazaki fragments on

lagging strand DNA.

As a whole, this assembly of replication machines has been termed the replisome.

The mass of the replisome approaches 1 million Daltons, and together, two replisomes

move along the parental chromosome in opposite directions simultaneously replicating

the nascent daughter chromosome from a single point of initiation and meeting at the

point of termination. This process of replicating of the entire chromosome lasts about 1

hour for a pair of replisomes; however, under some conditions dichotomous replication of

the chromosome can occur reducing the time of chromosomal replication to nearly 20

minutes. Up to six replisomes, at six separate replication forks, can be simultaneously

synthesizing three daughter chromosomes from the single original parental chromosome

during this process.

DNA Polymerase HI Holoenzyme

At the heart of the replisome is the DNA polymerase III holoenzyme (pol III

holoenzyme). DNA polymerase III was identified as the essential polymerase for E. coli

chromosomal replication, distinct from the two previously discovered activities of DNA

polymerase I, and II by analysis of mutants temperature sensitive for DNA synthesis and

for cell viability (Gefter et al., 1971; Hirota et al., 1972).

Pol III holoenzyme replicates DNA in a semidiscontinuous manner. Pol III

holoenzyme is capable of megabase processivity for continuous synthesis of the leading

strand, and responsible for controlled 1 2 kilobase Okazaki fragment synthesis of the

lagging strand. Pol III holoenzyme performs these activities simultaneously at speeds

approaching 1 kilobase per second, and has an extraordinary fidelity of a single








nucleotide misincorporation in 1 x 109 nucleotides polymerized (Komberg and Baker,

1992).

Through purification and biochemical characterization ofDNA polymerase III and

accompanying proteins found to be involved in synthesis elongation, the individual

subunits of this machine began to be identified. Pol III holoenzyme consists of 10

distinct subunits: a, e, 0, T, y, 8, S', X, y, and P. Five of these subunits are present in two

copies bringing the total number of pol III holoenzyme subunits to 20.

The core ofpol III holoenzyme (pol III core) was later purified and resolved into

individual subunits: a, e, 0 (McHenry and Crow, 1979). The a subunit encoded by the

dnaE gene (Mr = 140,000 Daltons), was found to be the subunit responsible for catalysis

of 5'-3' DNA synthesis activity (Maki et al., 1985; Spanos et al., 1981). The E subunit

encoded by the dnaQ gene (Mr = 25,000 Daltons) is the 3'-5' exonuclease or

proofreading subunit (Scheuermann and Echols, 1984). The 0 subunit encoded by the

holE gene (Studwell-Vaughan and O'Donnell, 1993) (Mr = 10,000 Daltons) is tightly

bound to a and e in pol III core, but a 0-specific function has not been identified. The a

subunit was further characterized after overexpression and purification of the dnaE gene

product (Maki et al., 1985). Gap-filling replication assays and complementation assays

were performed to show that the a subunit was responsible for the 5'-3' polymerase

activity of pol III holoenzyme, and that elevated levels of a in vivo did not increase the

amount of pol III holoenzyme in the cell.

Within pol III core, the a and a subunits are tightly bound and complement each

other's function with the overall effect of increasing the fidelity of chromosomal

replication. Using purified a and a subunits, (Maki and Komberg, 1987) showed that a








complex formed of a-e had increased 5'-3' polymerase activity over a subunit alone, and

highly increased 3'-5' e-exonuclease activity. The affinity for DNA of the a subunit in

pol III core resulted in an increase in apparent affinity of E for the 3'-hydroxyl terminus,

thus stimulating the exonuclease activity. In the cell, this proofreading activity during

synthesis was found to represent a 5-fold stimulation compared to exonuclease activity

uncoupled from synthesis which suggested that the fidelity of DNA replication may be

controlled by the relative abundance of the a and a subunits.

Polymerase III core is not processive in DNA replication whereas polymerase III

holoenzyme is highly processive. Using a kinetic assay, along with product size

determination assays it was originally shown that pol III holoenzyme was processive over

thousands of nucleotides, and pol III core had distributive activity, capable of

synthesizing 10-30 nucleotide stretches only (Fay et al., 1981).

It was found that pol III holoenzyme could be isolated from cells in two distinct

forms other than pol III core, each form catalyzing synthesis of a characteristic length of

product DNA (Fay et al., 1982; LaDuca et al., 1983). Polymerase III' was the first of

these smaller forms ofpol III to be purified. At the time pol I' was purified and

characterized, the r subunit was discovered and predicted to dimerize the pol III core

assemblies forming pol III' (McHenry, 1982). The processivity of pol III' was increased

6-fold over pol III core, and pol III' exhibited greater ability in synthesizing long single-

stranded templates coated with spermidine, a polybasic amine that stabilizes the helical

structure of DNA. The Polymerase III* form was also purified, and contained all of the

same subunits found in pol III holoenzyme except the 3 subunit-dimer. The activity of

pol III* showed an increase in processivity of at least 20-fold over pol III core (LaDuca et








al.; 1983). When pol III* was reconstituted with the p dimer, the characteristic

processivity of the holoenzyme was restored, suggesting 1 as the major factor in

conferring processivity on pol III holoenzyme.

Pol ilI holoenzyme must place the P dimer on primers formed by primase in order

to form preinitiation complexes. The formation of these preinitiation complexes requires

ATP binding and hydrolysis by pol III holoenzyme, and is absolutely required for

initiation of processive synthesis (Burgers and Komberg, 1982). The 1 subunit binds

directly to the a subunit within pol III holoenzyme (Stukenberg et al., 1991), and pol III

holoenzyme is dimerized by the T subunit by direct interaction also with a (Studwell-

Vaughan and O'Donnell, 1991). Through these interactions, pol III holoenzyme

simultaneously and processively synthesizes DNA. Therefore, preinitiation complexes

must be formed at least once on the leading strand for its continuous synthesis, and many

times on the lagging strand for Okazaki fragment synthesis.

Fragmented synthesis of the lagging strand by pol III holoenzyme is in conflict

with the level of processivity gained through binding the p processivity subunit. The

polymerase must detach from each completed Okazaki fragment and rapidly cycle to the

next preinitiation complex on the lagging strand (O'Donnell, 1987). During

chromosomal replication, anywhere from 2,000 to 4,000 Okazaki fragments are

synthesized at a rate of approximately 1 per second (Kornberg and Baker, 1992). Such

polymerase cycling activity requires precise coordination of protein-protein and protein-

DNA interactions at the replication fork. The coordinated efforts of several subunits,

including processivity proteins within pol III holoenzyme are responsible for this rapid

and ordered activity, and consequentially give the holoenzyme structural asymmetry,








while the DNA synthesis activity of the polymerase cores remain symmetric between the

leading and lagging strands.

The T Subunit is the Coordinator of Pol EIl Holoenzyme Function and Processivity

The dnaX gene codes for both the r and y subunits of pol III holoenzyme. The full-

length dnaX gene product is the r subunit (Mr = 71,000 Daltons). The y subunit (Mr =

47,500 Daltons) is created by a -1 translational frameshift adjacent to a hairpin-loop

structure in the mRNA that causes a stop codon (UGA) to appear approximately two-

thirds the way through the dnaX gene (Flower and McHenry, 1990; Tsuchihashi and

Komberg, 1990). Therefore, the y polypeptide is identical in sequence with the first two-

thirds of This translational frameshift occurs with about 50% yield of and y

polypeptides. Although they are extensively identical, the larger T polypeptide has

several important functions specific to its C-terminus, which greatly distinguishes it from

y (Gao and McHenry, 2001).

A dimer of subunits ( 2) acts as the "glue" of pol III holoenzyme, holding it

together and providing a structural scaffold for the asymmetric function of holoenzyme

on the leading and lagging strands. A tight interaction between T 2 and the a subunit of

core occurs through the C-terminus of T bridging two molecules of pol II core (Kim et

al., 1996b; Studwell-Vaughan and O'Donnell, 1991). The C-terminus oft also binds to

the DnaB helicase, physically linking the replication machine of pol III holoenzyme with

the primosome machine composed of DnaB and primase. The physical and

communications link between pol III holoenzyme and the primosome enhances the DNA

unwinding activity of DnaB helicase and connects pol III holoenzyme to the RNA

priming activity involved in Okazaki fragment length determination and polymerase

cycling on the lagging strand (Dallmann et al., 2000; Kim et al., 1996a). A single DnaB








helicase couples with both the leading and lagging strands, and through interaction with

the T-bridged polymerase may help to keep both strands closely associated with the

replication fork.

The T subunit protects the p processivity protein from removal off of the leading

strand indicating a direct role oft in the high processivity of the leading strand (Kim et

al., 1996c). This activity is most likely communicated through the close proximity oft -a,

and P-a binding sites on the C-terminus of a. Within this proximity, t could potentially

contact P providing direct protection from removal, or the t-a interaction may cause a

rearrangement of the P-core complex preventing removal of P. On the lagging strand, T

still protects the 1 processivity protein, but must be able to switch between protective and

non-protective states to allow polymerase cycling. Two distinct triggers for polymerase

cycling on the lagging strand have been identified (Li and Marians, 2000). The dynamic

action of primase binding the replisome through DnaB helicase and synthesizing a primer

is thought to trigger polymerase cycling, and collision of the lagging strand polymerase

with the 5'-end of the previously synthesized Okazaki fragment also is thought to trigger

cycling. With respect to the second case, a processivity switch requiring the T subunit

was identified (Leu et al., 2003; Wu et al., 1992). This "T processivity switch" is "off'

during processive synthesis, conferring protection of the P-core interaction with high

processivity for completed synthesis of the Okazaki fragment. The processivity switch is

turned "on" only upon incorporation of the final dNTP of the Okazaki fragment when the

5'-end of the previously synthesized primer is reached. The resulting nick in DNA

activates the processivity switch, and through actions of the C-terminus of t, core is

released from 1 and allowed to cycle to the next preinitiation complex.








Pol III holoenzyme contains the processivity proteins necessary to form

preinitiation complexes at newly primed sites on template DNA. The T subunit is also

part of this complex of processivity proteins that coordinates the protein-protein and

protein-DNA interactions required for preinitiation complex formation. Unlike the

functions of the r C-terminus, r activity in preinitiation complex formation is located in

the N-terminal region of the protein identical to the y subunit A DnaX-complex forms

within pol III holoenzyme which functions in loading 0 onto primed template DNA. For

preinitiation complex formation, this "clamp loader" is discussed in detail below. The

DnaX complex within holoenzyme contains two r subunits (i.e., same two that dimerize

core) in a complex with y, and the 8, 6', X, & y subunits (stoichiometry: t 27y58'XI)

(Pritchard et al., 2000). The DnaX complex binds and hydrolyzes ATP forming

preinitiation complexes on leading and lagging strand primers in what may be an ordered

activity. In their report, (Glover and McHenry, 2001) showed that formation of the

leading strand preinitiation complex required ATP binding but not hydrolysis by DnaX

complex, and that ATP hydrolysis was required for formation of the lagging strand

preinitiation complex. Further, a non-hydrolyzable analogue of ATP (ATPyS) caused

removal of the polymerase presumably bound to a lagging-strand template, extending the

knowledge of pol III holoenzyme as an intrinsically asymmetric dimer with

distinguishable leading and lagging strand polymerases.

Structure of the p Sliding Clamp Processivity Protein

When the X-ray crystal structure of the 1 processivity protein was solved, it was

immediately clear how this homodimer conferred processivity to polymerase III








holoenzyme (Kong et al., 1992). The P sliding clamp is composed of two identical

crescent-shaped monomers arranged in head-to-tail fashion (Figure 2-1).


Figure 2-1. The crystal structure of P sliding clamp. Two views of the clamp composed
of crescent-shaped protomers (red / yellow) are shown. A) Front view of 0
showing continuous P-sheet structure surrounding an inner core of a-helices.
Asterisks denote the dimer interfaces. Green lines drawn through the left
protomer represent imaginary boundaries between structural subdomains.
From the top of the left (red) protomer, the subdomains are numbered counter
clockwise: 1 (N-terminal), 2 (middle), and 3 (C-terminal). B) Side view of the
p dimer showing the asymmetry of the faces. On the left face is the extended-
loop region to which y complex and pol III a subunit compete for binding.
Images were composed using DeepView / Swiss-Pdb Viewer Ver. 3.7,
http//www.expasy.org/spdbv/, with structural coordinates downloaded from
the Protein Data Bank (Berman, Westbrook et al. 2000),
http//www.rcsb.org/pbd/. (PDB code: 2POL)

The p clamp topologically links the polymerase to template DNA in a non-specific

manner providing a tight interaction with DNA while allowing essentially unlimited two-

dimensional diffusion along the template. Several structural characteristics of P clamp

confer its extraordinary ability to provide processivity to DNA polymerase. Each P








monomer is formed from three structurally similar domains containing 13 sheets on the

outside and a helices on the inside. As a homodimer, the six domains form an inner pore

lined with 12 a helices that are surrounded by an essentially continuous p sheet structure.

The axis of the a helices lining the inner pore are almost exactly perpendicular to the

local direction of the phosphate backbone of DNA facilitating rapid movement along the

duplex. For example, the a helices traverse the major and minor grooves so as not to fall

into them. The height of the P clamp is ~ 80 A. The width of P clamp (~ 34 A) would

cover about one full turn of duplex B-form DNA, or ~ 10 base pairs. It was shown

biochemically that an 11-base pair primer was required by P to form a productive

preinitiation complex synthesized by pol III (Yao et al., 2000). The diameter of the inner

pore is -~ 35 A, enough space to encircle either duplex B-form DNA (-~ 20 A) or a RNA -

DNA hybrid duplex (i.e. a RNA-primed DNA template) (-~ 21 A). The extra space within

the inner pore, between the clamp and DNA, is expected to be filled by one or two layers

of water molecules.

The two faces of the p clamp are asymmetric (figure 2-1B). One face has extended

loop C-terminal structures, and the other face is essentially flat giving the clamp an

overall shallow cone-shaped thickness. The extended loops on the extended C-terminal

face contain the binding sites for both the DnaX clamp loader and the a subunit of pol III

core. The presence of overlapping binding sites for both the clamp loader and

polymerase determine the proper orientation of the p clamp when loaded onto primed-

template DNA and result in competition for the binding sites. DNA mediates this

competition for overlapping binding sites. In the absence of DNA, P3 clamp binds the








clamp loader, but when p is loaded onto DNA, the DNA causes the clamp loader to lose

affinity for the clamp, and polymerase binds (Naktinis et al., 1996).

Calculation of the electrostatic field generated by the P clamp revealed that the

protein has an overall negative electrostatic potential. However, the surface lining the

inner pore has a focused positive electrostatic potential, precisely where the negatively

charged DNA molecule would be. The calculated electrostatic features suggest that the

inner pore of the clamp has inherent affinity for DNA. This electrostatic interaction is

most likely lubricated by water molecules allowing a great degree of freedom for linear

diffusion while still maintaining some affinity for DNA.

The dimer interfaces appear as a continuation of the 1 sheet structure from the outer

surface of each monomer across a molecular boundary, and contribute at least four

hydrogen bonds at each interface. In addition, two distinct sets of interactions between

neighboring a -helical side chains stabilize the p dimer interface. A small hydrophobic

core is formed in the center of the interface by the packing of Phe and Ile side chains with

Ile and Leu side chains between monomers. Six potential intermolecular ion pairs

between Glu, Arg, and Lys side chains surround the small hydrophobic core. The 1

dimer consequentially is very stable with a monomer-dimer dissociation constant of <60

pM, and a half-life on closed-circular DNA of ~ 100 minutes (Yao et al., 1996).

Considering that the cellular concentration of p is ~250-500 nM, no monomeric P would

be present in vivo (Komberg and Baker, 1992). The number of specific and potentially

strong interactions at the dimer interfaces underscores the requirement for a clamp

loading machine to load this clamp on DNA.








The DnaX Clamp Loading Machine

Three major parts make up the pol III holoenzyme within the E. coli replisome.

The replicative polymerase III core, the DnaX complex that contains the T subunit -the

cement of pol III holoenzyme, and the P sliding clamp processivity protein. The DnaX

complex has an additional function in polymerase processivity by loading the P clamp

onto primed-template DNA for preinitiation complex formation. Originally identified as

the y-8 complex or y complex (Maki and Komberg, 1988; O'Donnell, 1987), this DnaX

complex was shown to load P onto DNA in an ATP-dependent manner (Onrust et al.,

1991).

The T and 7 subunits are expressed at roughly equal levels in cells due to a

translational frameshift in the DnaX gene (Flower and McHenry, 1990). Therefore it was

expected that a mixed r / y containing DnaX complex would reside in the pol III

holoenzyme. In fact, several DnaX complexes expressed from an artificial operon were

isolated from cells including two forms of mixed T / y complex, and a y-only complex

(Pritchard et al., 2000). No T-only complex was isolated from cells in that study. One of

the two isolated forms of mixed T / y complex had the stoichiometry of T2 yi of the DnaX

gene products, and this form was predicted to function within the pol III holoenzyme

based on the requirement of the T subunit dimer for holoenzyme function. The separate

y3 only complex, and / or the mixed T I / y2 complex were predicted to function separately

from holoenzyme, for example, in DNA repair activity with DNA polymerase II (Bonner

et al., 1992).

Identification of all five of the subunits of y complex (Maki and Komberg, 1988),

and the genes encoding them (Dong et al., 1993; Xiao et al., 1993a), lead to in vitro

reconstitution and use of this y-only complex as the model 1 sliding clamp loader (Onrust








et al., 1995). Recently, sequence and structural analysis have identified several subunits

of y complex as AAA+ superfamily members and y complex as a AAA+ molecular

machine (Davey et al., 2002). These studies have provided significant insight into the

mechanism of opening the stable ring-shaped dimer structure of P clamp and loading it

onto DNA at the speed required by polymerase III catalyzed Okazaki fragment synthesis.

The y complex contains five subunits with the stoichiometry: y3, 51, 5'i, X1, and WI.

The x and v subunits bind a y subunit through v (Glover and McHenry, 2000), and

greatly increase the affinity for the 88' subunits to the complex (Olson et al., 1995). The

y subunit, alone in solution, was found to exist in a monomer-tetramer equilibrium,

however when formed in a complex with 88' became a trimer of y3 (Pritchard et al.,

2000). The structural asymmetry derived by single copies of the 8, 8, Land y subunits

of the DnaX clamp loader within holoenzyme imparts structural asymmetry in the

holoenzyme, and implicates their function with the lagging strand polymerase.

The 8 and 8' subunits, products of the holA and holB genes, respectively, were

purified and characterized for their interactions and function in y complex (Dong et al.,

1993). These subunits have similar mass, (8: M, = 38,700 Daltons, and 8': M, = 36,900

Daltons), and high affinity for each other and were usually co-purified as a consequence.

The 8 subunit binds P through a "p interaction element" (3IE), while 8' binds tightly to

the y subunit. During the clamp loading cycle, an "inactive" form of y complex, the 8'

subunit is thought to occlude 8 from binding P and a conformational change upon binding

of ATP to y complex causes exposure of the 13IE of the 8 subunit (Naktinis et al., 1995).

Genetic knockouts of the holA or holB genes are not viable revealing that the function of

8 in clamp loading absolutely requires the 8' subunit (Song et al., 2001). Song, Pham et








al. (2001) also showed that 8 and 8' are required not only for ATP-dependent

preinitiation complex formation, but are essential for DNA synthesis elongation as well.

They participate in the DnaX clamp loader complex function in coupling to DnaB

helicase at the replication fork to pol III holoenzyme for leading and lagging strand

synthesis.

The 8 subunit alone was found to have the ability to unload 0 from DNA without

any requirement for ATP binding or hydrolysis (Leu et al., 2000). In this study, the

approximate quantity of each subunit of y complex was determined from whole cell

lysates by protein absorbance measurements and western blot analyses. Surprisingly they

discovered a -4-fold excess of the 8 subunit present separate from any within y complex.

In Okazaki fragment synthesis, about 2000 to 4000 P clamps are used for complete

replication of the lagging strand. However, with only about 300 molecules of p present

in the cell, there is a need to recycle the clamps left behind on the DNA as the lagging

strand polymerase cycles between Okazaki fragments. To keep up with the pace of

Okazaki fragment synthesis P must actively be unloaded at a rate of at least 0.01 s"'. It

was found that the quantity ofy complex within the cell was limited to -140 molecules

by the 8' subunit. Since pol III holoenzyme and any free y complex are present in low

quantities and perform several DNA metabolic functions such as replication and repair, it

was envisioned that the excess 8 subunit aided in removal of P clamps abandoned by pol

HI at the end of each Okazaki fragment. In their investigation (Leu, Hingorani et al.,

2000) also showed that in vitro removal of 1 from circular DNA substrates by purified 8

happened at a rate of 0.011 s-1. This rate was comparable to measured rates of 1

unloading by y complex (0.015 s'1) and pol III holoenzyme (0.007 s') in similar assays.








As all of the 8' subunit is expected to be stoichiometrically associated with y complex,

the 8' subunit would only regulate the activity of 5 in clamp loading, leaving free-8

subunit to recycle leftover P clamps, keeping a sufficient level of free sliding clamp

available for rapid Okazaki fragment synthesis.

Although not necessary for in vitro clamp loading activity, y complex contains the

X and y subunits. The x (M, -16,600 Daltons) and y (Mr -15,000 Daltons) subunits are

the products of the holC and holD genes, respectively, and have important functions

adapting the lagging strand polymerase for Okazaki fragment synthesis (Onrust et al.,

1995; Xiao et al., 1993a). As stated above, the 'y subunit binds to y, bridging the

interaction of the X subunit to the complex. The ,X-y' dipeptide forms a rod-like structure

(Kelman et al., 1998), however it is not known where on the y subunit X-y binds in the

clamp loader complex (Jeruzalmi et al., 2001a). A sub-complex of y-y, or X-' alone

had no independent function in promoting DNA synthesis activity by pol III (Xiao et al.,

1993b).

The X subunit through the i structural bridge to the y subunit in the DnaX complex

strengthens DNA polymerase III holoenzyme interactions with the single-stranded DNA

binding protein (SSB)-coated lagging strand template facilitating preinitiation complex

formation and processive synthesis elongation. Isolated pol III core or pol III', which are

without the X-y subunits, are inhibited in synthesizing DNA on SSB-coated template

DNA (Glover and McHenry, 1998). Removal of x-' from pol III holoenzyme results in

reduced efficiency of clamp loading activity and an increase in salt sensitivity in the

physiological range (40-150 mM) (Kelman et al., 1998). Investigation of a common

point mutation in SSB (SSB-113) revealed that the direct interaction with the X subunit is








located at the C-terminus of SSB, and that this interaction is most likely hydrophobic in

nature. Through this direct x-SSB contact on template DNA, the X subunit was thought to

provide resistance to elevated ionic strength for proper clamp loading activity, as well as

possibly keeping the lagging strand polymerase localized to the SSB-coated template for

greater efficiency in synthesis elongation.

The x subunit has been found to be required in a primase to polymerase handoff for

Okazaki fragment synthesis. A "three-point switch" was proposed where the x subunit,

through direct interaction with SSB, causes primase to dissociate from a newly formed

primer allowing initiation complex formation and synthesis of the Okazaki fragment to

proceed (Yuzhakov et al., 1999). Primase (DnaG) interacts distributively with DnaB

helicase in formation of the primosome for controlled RNA primer synthesis (Tougu and

Marians, 1996). After synthesizing a primer, primase remains attached to the nascent

primer, and is thought to protect it from exogenous exonuclease activity. Primase is also

directly attached to SSB, and therefore blocks initiation complex formation. Through the

X-SSB interaction, but not a direct X-primase interaction, primase is displaced, perhaps

through a conformational change in SSB. The T-DnaB helicase interaction provides a

contact point to hold pol III holoenzyme to the replication fork while lagging strand

polymerase cycles (Kim et al., 1996a). Contact of pol III holoenzyme to the newly

primed template, and bringing the X subunit through the DnaX clamp loader complex into

the vicinity of the SSB-coated lagging strand to displace primase through SSB would

allow preinitiation complex formation for polymerase cycling for rapid lagging strand

Okazaki fragment synthesis.








The Dna X (r or 'y) subunits of the clamp loader bind ATP (Tsuchihashi and

Kornberg, 1989). In the reconstituted y complex clamp loader, interaction with the 58'

subunits causes a solution-monomer-tetramer equilibrium of y subunit to form a trimer

giving the clamp loader a stoichiometry of 73y, 1, 's, it, and xip (Pritchard et al., 2000).

The y subunits in isolation (i.e. in monomer-tetramer equilibrium) do not hydrolyze ATP

at an appreciable rate with respect to activity needed for replication of the lagging strand.

However, when reconstituted into a complex with the 8,8',X and xp subunits, or just the 8,

and 8' subunits, the rate of ATP hydrolysis is much higher, and further enhanced in the

presence of 3 to a rate that would be rapid enough for preinitiation complex formation for

Okazaki fragment synthesis (i.e. -2 s-1) (Onrust et al., 1991). This means that the clamp

loader can bind and hydrolyze a maximum of three molecules of ATP to power the clamp

loading reaction. Since the y / T subunits of the clamp loader are the only subunits which

bind and hydrolyze ATP to power the clamp loading reaction they are known as the

motor subunits of the clamp loader.

The AAA+ Superfamily of Motor Proteins

Through sequence analysis and ultimately structural analysis the y, 8, and 8'

subunits of the clamp loader have been identified as members of the AAA+ superfamily

of ATPases (Davey et al., 2002; Neuwald et al., 1999). As the name of this superfamily

implies: "ATPases Associated with a variety of cellular Activities", proteins of this class

are involved in many cellular processes. Whole genome analyses have indicated that the

AAA+ class is ancient and has undergone substantial functional divergence prior to the

emergence of the major divisions of life, thus this class is large and diverse (Neuwald et

al., 1999). Proteins of the class generally assist in protein transitions, such as remodeling,








assembly, and disassembly. AAA+ proteins assist not only in protein-protein transitions,

but also in protein-nucleic acid transitions.

AAA+ proteins including GTPase proteins, which are mechanochemical proteins

that transduce the energy of nucleotide hydrolysis into useful work, are becoming

recognized as a quickly growing group of "energase" enzymes (Purich, 2001). An

energase can simply be defined as an enzyme that couples the change in energy of a

covalent bonding state directly to changes in non-covalent substrate- and product-like

states. For AAA+ machines such as the y complex clamp loader, the chemical energy of

ATP hydrolysis activity is transduced into mechanical changes in the clamp loader

structure. As a AAA+ machine, y complex clamp loader could be described as a

molecular matchmaker. The y complex modifies a protein (i.e., |3 clamp) such that it

becomes enabled for interaction with DNA in such a way that it would not normally

interact.

The AAA+ superfamily includes chaperone-like ATPases such as regulatory

components of proteases, transcriptional regulators, vesicle synthesis and fusion proteins,

and dynein motor proteins to name a selected few. Aside from AAA+ proteins of y

complex, other important replication proteins are members of the AAA+ class. For

example, bacterial DnaA, DnaC, and RuvB proteins, and the subunits of the eukaryotic

origin recognition complex, Mcm2-7 helicase, and replication factor-C subunits of the

eukaryotic clamp loader (Erzberger et al., 2002; Neuwald et al., 1999).

Increasing investigation of AAA+ proteins reveal that adaptor proteins may

modulate their functions. As a consequence, a growing list of these structurally unrelated

adaptor proteins are being identified for many of the known AAA+ proteins (Dougan et








al., 2002). Generally smaller than their AAA+ partner, AAA+ adaptor proteins provide a

simple and effective way to modulate the function of the AAA+ machine. Adaptor

proteins have been found to give the AAA+ protein better control over substrate

specificity, allowing quick response to changing local conditions and redirection of

AAA+ activity. The y complex has two proteins, X and W, that are not related in sequence

or structure to the other AAA+ class subunits of the clamp loader. Questions concerning

the function of X and W in the clamp loading reaction by y complex are of major

importance in this dissertation, and it is attractive to propose that they are in fact adaptor

proteins in complex with the AAA+ clamp loading machine. As described above, the X

subunit, through V, has important interactions with SSB at the replication fork for

preinitiation complex formation. As will be discussed in the results of this dissertation, X

and Mi are also important for activity beyond interaction with SSB at the replication fork,

and are important for optimal activity of the clamp loading machine itself. This

dissertation explores the structural and conformational stability provided to y complex by

the X and y proteins, and therefore may reveal a novel example of how adaptor proteins

modulate the activity of their partner AAA+ machines. The X-ray crystal structures of a

minimal 7388' clamp loader and a x-V complex have been solved, but it still remains a

mystery as to where the x-r dipeptide binds to the y subunit of the clamp loader. The

limited examples of AAA+-adaptor protein complexes have shown that the adaptor

proteins generally bind to the least conserved N-terminal domain of their AAA+ partner.

This fact may provide direction for further structural analysis of the complete y complex

including the x and W proteins.
































Figure 2-2 The crystal structure of a AAA+ motor protein. The 71-243 crystal structure
(Podobnik, Weitze et al. 2003) is shown here as a representative AAA+ motor
protein. This structure was truncated at the C-terminus, and only the AAA+
domains-I (red) and -II (yellow) are shown. A molecule of ATPyS bound to
the NTP-binding site is shown in space-filling representation. The conserved
structural and functional features of AAA+ motors are indicated: Sensor-2
(blue), P-loop (green), and Sensor-I (pink). Also marked are the conserved
Arg (y-215) of sensor-2 and Thr (y-155) and (SRC) "Arginine (y-169)-finger"
of sensor-1. A Cys-coordinated Zn" atom is also present in Domain-I.
Images were composed using DeepView / Swiss-Pdb Viewer Ver. 3.7,
http//www.expasy.org/spdbv/, with structural coordinates downloaded from
the Protein Data Bank (Berman, Westbrook et al, 2000),
http//www.rcsb.org/pbd/. (PDB code: 1NJF)

AAA+ proteins have several common structural features that define their function

(Figure 2-2). All AAA+ proteins have an approximately 220 amino acid conserved

structural ATPase core, although sequence is not always conserved (Jeruzalmi et al.,

2001b; Neuwald et al., 1999). The AAA+ conserved structure has two domains (I and








II), and a third oligomerization domain (III) is common in multimeric complexes of

AAA+ protein subunits. AAA+ proteins generally contain the phosphate-loop (P-loop)-

type NTP-binding site having the Walker-A and Walker-B motifs that are common in

ATPase and GTPase proteins. The Walker-A motif (GVxxxGKT) is involved in the

phosphate binding of ATP,(or GTP), and the Walker-B motif (DExx) is involved in metal

(Mg) binding and catalysis (Walker et al., 1982).

The P-loop a,P-fold provides a platform upon which other AAA+ structural motifs

are mounted. Above the P-loop, in domain-II is the sensor-2 motif containing a highly

conserved Arg residue (Arg-215 in the y subunit). This Arg residue is predicted to

interact with the 0- and y-phosphate of bound ATP (Podobnik et al., 2003). Domain II

containing the sensor-2 motif is typically a hinge-like domain around which movements

in domains-I and III are made.

In domain-I, flanking the P-loop from the underside is the sensor-1 motif. The

sensor-I motif contains the highly conserved Ser-Arg-Cys (SRC) sequence that is

predicted to interact with the y-phosphate of the bound ATP through hydrogen bonding.

The SRC sequence includes an "arginine-finger". This arginine-finger is thought to

stimulate catalysis of ATP hydrolysis by analogy to the similar, conserved switch H

region of GTPase activating proteins (Ahmadian et al., 1997). In an oligomeric AAA+

complex such as y complex, the SRC-sensor-1 motif of one AAA+ subunit interacts with

the nucleotide-binding site of a neighboring AAA+ subunit, and is implicated in

interfacial stimulation of ATP hydrolysis (Jeruzalmi et al., 2001a; Podobnik et al., 2003).

The X-ray crystal structure of the y complex AAA+ clamp-loading machine is described

below. It has five AAA+ proteins that make up three parts of the clamp loading machine:








three y-ATPase-motor subunits, the 8 subunit, harboring the |I interaction element (DIE),

and the 8' "stator", upon which movement of the other subunits is supported.

X-ray Crystal Structure of the Clamp Loading Machine

The X-ray crystal structure of a reconstituted minimal clamp loader complex

(7358') was solved at a resolution of 3.0 A using multiwavelength anomalous diffraction

methodology (Jeruzalmi et al., 2001a). The subunits crystallized as a heteropentameric

circle having the stoichiometry (6'1-73-81) (Figure 2-3).

While the amino acid composition of the 8 and 8' subunits was complete, all three y

subunits were truncated by 57 amino acids from protease-labile C-terminii. All subunits

showed a three domain "C"- shape previously observed in the structure of the 8' subunit

alone (Guenther et al., 1997), giving the complex an overall form analogous to five

fingers extending down from the palm of a hand. Both of the AAA+ domains I and II

were observed for all five subunits. This was a surprising observation for the 6 subunit,

as it shared only 7-8 % sequence identity with other AAA+ proteins including those

within the clamp loader. All of the ATPase motor y subunits crystallized in nucleotide-

free form, and the 8' and 8 subunits were already known not to bind ATP. The X-ray

crystal structure of a truncated form of the y subunit (_1-243) including the AAA+ domains

I and II only was later solved in the presence of nucleotide (Podobnik et al., 2003), and

will be used for the discussion of the nucleotide binding sites and consequences of

nucleotide binding the y motor subunit.

The C-terminal domain III of all five subunits of the clamp loader forms a tight

circular collar at the "top" of the complex (Figure 2-3B). The subunits are arranged

where the 8' and 5 subunits surround the three y subunits.










Figure 2-3. The crystal structure of the 7385' clamp loader. The clamp loader
forms a circular heteropentameric structure, each subunit having three
domains (I II and III). A) "Front"-view: The point of view is directly
into the "open" interface between the 8' (white) and 8 (blue) subunits.
The three y subunits (different shades of red) are present behind 8' and
8. Structural domains are indicated: Domain-HI (the C-terminal
"collar"), Domain-H ("hinge"), and Domain-I (N-terminal
"functional"). B) The clamp loader has been tilted forward 90 for a
slab-view of the "top" showing the tight pentameric "collar" of the C-
terminal domain-III only. Subunit nomenclature is indicated. C) The
clamp loader has been tilted back 900 from the view in (A), and
domains-II and III have been dimmed for the "bottom" view. The N-
terminal domain-I of each subunit is colored for emphasis on their
asymmetrical disposition. The NTP binding sites of the y subunits are
occupied with sulfate molecules from crystallization (space-filling
representation), and the Zn+ atoms (yellow) bound to 8', yl, y2, and '3
subunits are shown. The P-interaction element (pIE), and the
conserved hydrophobic-wedge residues of the pIE (8-Leu-73 and -Phe-
74) are shown (green). Images were composed using DeepView /
Swiss-Pdb Viewer Ver. 3.7, http//www.expasy.org/spdbv/, with
structural coordinates downloaded from the Protein Data Bank
(Berman, Westbrook et al, 2000), http//www.rcsb.org/pbd/. (PDB
code: 1JR3)










































Hydrophobic interactions within a helical packing arrangement strengthen subunit

interactions in this oligomerization domain. Only the C-terminal domain of the 6 subunit

appeared "loose" within this oligomerization domain, and was proposed to have

consequences for movement of the 8 subunit within the complex.

The N-terminal domains of the five subunits containing the AAA+ structural

domains-I and II are highly asymmetric with respect to each other extending down from

the C-terminal collar (Figure2-3C). Each N-terminal domain is increasingly splayed out








with respect to the circular C-terminal collar forming an "open" o-shape. Splaying-out of

N-terminal domains in this structure resulted in considerable exposure of the 5 subunit

and thus the PIE. Based on biochemical studies, a nucleotide-free structure would be

predicted to be more "closed" where the 8 subunit-PIE would be obstructed by the 8'

subunit. The authors speculate that the observed structure of the clamp loader may be an

artifact of the crystal packing forces, but overall would reflect that of an "open" or

activated complex.

The nucleotide binding sites of the y subunits are buried within the subunit

interfaces as is common with other oligomeric AAA+ ATPases. The AAA+ domains are

arranged such that domains-I and II of one subunit cradle domain-I of the neighboring

subunit providing a mechanism for conformational communication based on nucleotide

binding status. Three interfacial nucleotide binding sites are between the 5'-yl, yi-y2, and

Y2-Y3 subunits. The N-terminal AAA+ domains of the y subunits were highly asymmetric

in the complex, resulting in asymmetry in these interfacial nucleotide binding sites. The

8'-yi interface and Y2-73 interface were open in this nucleotide-free structure, while the y7-

72 interface had the most extensive contact and therefore appeared inaccessible to

nucleotide.

Structure of the Nucleotide Binding Site and the Proposed Conformational Change
of the y Subunit

The X-ray crystal structure of C-terminal truncated y1-243 was solved in the presence

of ATPyS to a resolution of 2.2 A using multiwavelength anomalous dispersion

methodology (Podobnik et al., 2003). The resulting structure had four y subunits

arranged in the asymmetric crystal unit. One of these y subunits had complete electron

density corresponding to ATPyS in the nucleotide binding site, and another was missing








electron density corresponding to the y-phosphate, and therefore was taken as the

structure of ADP-bound y. Two other y subunits in the asymmetric unit had no

nucleotide bound. Slow hydrolysis of ATPyS through the crystallization procedure was

thought to result in producing the ADP-bound y. This structure allowed identification of

several as yet unknown features of nucleotide binding and the conformational changes

induced due to nucleotide binding. The results have important implications for the

mechanism of communication between the AAA+ domains within the clamp loader and

therefore the overall clamp loading mechanism.

The structure showed that the nucleotide was bound in the expected position in the

cleft between domains-I and II formed by the Walker-A P-loop (see figure 2-2). The

adenine ring was located in a non-polar environment, and formed hydrogen bonds from

the N-6 position with the main chain carbonyl groups of Val-19 and Val-49. This feature

provides a mechanism for distinguishing ATP from GTP by the y subunit. Both the 2'-

and 3'-hydroxyl oxygens of the ribose moiety were hydrogen-bonded to the carbonyl

oxygen of Ala-7 of the N-terminal helix extending from domain I. Three residues of the

conserved Walker-A P-loop motif coordinate the p- and y- phosphates of the nucleotide.

The Lys-51 residue forms a salt-bridge, and Thr-52 and Ser-53 form hydrogen bonds to

the phosphate groups from the sensor-1 loop.

Comparison of the nucleotide-bound 71-243 structure with the nucleotide-empty

7388' structure reveals that the y subunit domain-II a-helix containing the sensor-2 motif

blocked the region corresponding to the position of the ribose ring. This observation was

extended to all of the empty nucleotide-binding sites in the y388' structure. Thus it was

predicted that nucleotide binding to a given y subunit alters the position of domain-I with








respect to domain-III by conformational change. The reorientation of domain-I would

open up the nucleotide binding site from a collapsed state, and clearly affect the positions

of the sensor-1 and sensor-2 motifs within the complex. Also, rotation of the sensor-2

region containing the conserved Arg-215 residue occurs upon nucleotide binding and

provides space for the phosphate groups of the nucleotide. The sensor-1 motif does not

move with respect to the P-loop, and therefore follows the general movement with the N-

terminal domain-I upon nucleotide binding.

Conformational changes induced by nucleotide binding to the P-loop region are

directly linked to movement of the catalytic "arginine-finger" of sensor-1, thereby

coupling (i.e. communicating) changes in nucleotide binding status to the neighboring

subunits of the complex. In the y388' structure, the sensor-1 motif containing the

arginine-finger, was either too far away from the neighboring nucleotide binding site that

it is thought to stimulate in the case of the 5'-y, interface, or in steric-clash with the

neighboring nucleotide binding site as in the case of the YI-72 interface. Therefore,

conformational changes in individual subunits must be made to correctly position the

sensor-I arginine-finger with the neighboring nucleotide for catalysis. These

conformational changes that occur upon ATP binding not only bring the subunits into

proper orientation for ATP hydrolysis, but more importantly cause exposure of the 8

subunit from occlusion by the 6' stator so that it can bind 0 clamp.

"Crude" modeling of the p clamp bound to the clamp loader structure was

performed using information from the 5-P structural interaction (described in detail

below). This "simplistic" docking of dimeric P with the clamp loader revealed that the

orientation of the S-p interaction would cause the clamp to make contact with all subunits








in the clamp loader (Jeruzalmi et al., 2001a). One of the key points addressed in this

dissertation is the question of how P affects the clamp loader in the clamp loading

mechanism. The multipoint contact suggested by the modeling of P bound to the clamp

loader provides a structural basis for several conclusions concerning individual subunit

conformational changes and how 1 may "trap" an ATP-bound "activated" clamp loader.

This contact may potentially explain how the clamp loader stabilizes the open P ring

while a single dimer interface remains open (see below) until a primer template junction

is located.

The exact nature of how the clamp loader binds to DNA is not yet fully understood.

A cysteine-coordinated zinc module is present in the 5' and all three y subunits of the

complex (figure 2-3) and may provide a interaction zone for duplex DNA, however, the

function of the zinc modules in the clamp loader complex has not been investigated.

Calculation of the surface electrostatic potential of the clamp loader structure invites

some interesting speculation on the nature of where DNA binds and stimulates the

activity of the clamp loader. Two extensive regions of positive electrostatic potential

were observed on the structure. Both were seen on the N-terminal asymmetric "bottom"

side of the clamp loader that is predicted to interact with the 0 clamp. One of these

regions was on the 5 subunit, adjacent to the P1E presumably where DNA would thread

through the P clamp when bound to 5. The other region was observed as a partially

continuous belt that tracks along the outer surface of the N-terminal domains of the y

subunits. Each of these regions of positive electrostatic potential predict that when DNA

is encircled within the 1 clamp under the clamp loader, single-stranded DNA could be

positioned to interact with the outer regions of the ATP binding domains causing








additional rearrangement of sensor-2 and sensor-1 motifs stimulating hydrolysis of ATP.

The requirements for formation of the DNA binding site of y complex are further

explored in the results of this dissertation.

The X-ray crystal structure of the 8 subunit bound to a p monomer was also solved,

and provides a possible mechanism of clamp opening by the 8 subunit as well as how the

p clamp may interact with y complex through binding the 8 subunit-pIE.

X-Ray Crystal Structure of the 8 Subunit Bound to a p Monomer and the
Mechanism for Opening the P Sliding Clamp

The X-ray crystal structures of two complexes between the 8 subunit and a mutant

monomeric form of 0 (31) were solved by multiwavelength anomalous diffraction

(Jeruzalmi et al., 2001b). Attempts to crystallize the 8 subunit with wild-type P2 were

unsuccessful, and therefore a mutant P that remained monomeric due to mutations within

the dimer interface was used (Stewart et al., 2001). The 8: 1 structures either contained

full-length 8 subunit, composed of all three domains observed in the y complex structure

including the AAA+ region, or truncated 8 subunit (amino acids 1 140) 81-'40 containing

the N-terminal domain I only. Both complexes of 8:13 and 81140:01 crystallized in 1:1

complexes at resolutions of 2.9 A and 2.5 A respectively (Figure 2-4).

Examination of both structures revealed a mechanism by which p clamp is opened

upon the 8 subunit inducing or trapping a conformational change in 0 such that a stable

dimeric interface cannot be formed. Some possibilities for ring opening were that 8

forcibly breaks the dimer interface, but due to the fact that ATP is not required by 8 to

open P, this was not very probable. Another possibility was that the |3 clamp could

oscillate between conformations and the 8 subunit could have high affinity for a

conformation inconsistent with ring closure.































Figure 2-4. The crystal structure of the 8-1p complex. The full structure of the P subunit
(blue) is shown bound to p1 monomer. Domains (I II and III) of 8 are
indicated. The subdomains of 1i are colored: N-terminal subdomain-1
(white), middle subdomain-2 (pink), and C-terminal subdomain-3 (red). The
conserved hydrophobic-wedge, pIE (green) Leu-73 and Phe-74 of 5 subunit
are marked. The a-helix of the P subunit involved in the "spring-loaded"
conformational change is colored yellow. Images were composed using
DeepView / Swiss-Pdb Viewer Ver. 3.7, http//www.expasy.org/spdbv/, with
structural coordinates downloaded from the Protein Data Bank (Berman,
Westbrook et al.), http//www.rcsb.org/pbd/. (PDB code: 1JQJ)

This also was not a very probable mechanism due simply to the high stability of the P

dimer alone on circular DNA. A third possibility was that the 8 subunit could induce and

stabilize a conformation in p that is inconsistent with ring closure. The results from

analysis of these structures are consistent with the third possibility, and it was proposed

that 8 subunit binding induces a conformational change in the P clamp that results in








straightening of a crescent-shaped P monomer such that a single dimer interface is

opened to allow DNA passage into the clamp.

The 3IE is on the N-terminal domain of the 8 subunit, and consistent with earlier

biochemical studies, makes contact with the C-terminal-extended-loop face of the p

subunit (Naktinis et al., 1995). Only one 8 subunit bound p1 in the full length 8: p1

structure, and it was observed that domains II and III of the 8 subunit impeded binding of

a second molecule of 8 to P1. In the 81'140:13 structure showed that the p1E forms a

hydrophobic wedge that binds to a hydrophobic pocket on the p1 subunit between

subdomains 2 and 3 (of the P subunit) (Figure 2-4). Subdomains 2 and 3 are the middle

and C-terminal structural domains of the P protomer (Kong et al., 1992). The Met-71,

Leu-73, and Phe-74 residues form the hydrophobic wedge on the 8 subunit. The

hydrophobic pocket on 13, into which this wedge is inserted, is formed by Leu-177, Pro-

242, Val-247, Val-361, and Met-362 of 1. The PIE hydrophobic wedge of the 5 subunit

is highly conserved among all bacterial 8 subunit sequences examined. Likewise, the

residues forming the hydrophobic pocket on P are also highly conserved across evolution,

and therefore found on sliding clamps of bacteriophage, eukaryotes, and archaea (see

below). Conservation of this interaction between the clamp loader and sliding clamp

suggests a common mechanism for clamp opening in all known organisms.

These structures have revealed that interaction of 8 with P requires conformational

changes in p that weaken the dimer interface, and importantly, 8 does not interact directly

with the dimer interface. The structure of the 0 dimer interface consists of a hydrophobic

core surrounded by ionic interactions (Kong et al., 1992). An a-helix participating in

these interactions at the dimer interface is distorted in the wild-type dimeric 0 clamp such








that additional hydrophobic residues contained on this distorted helix do not extend into

the interface. The hydrophobic pocket binding site where the 8-subunit-PIE binds is near

a 5 -residue loop adjacent to the distorted a-helix in p. The binding of 6 is thought to

impose strain on the distorted helix that causes its relaxation into a more ideal a-helical

conformation. This results in extension of the hydrophobic residues from the previously

distorted portion of the a-helix directly into the dimer interface. Removal of dimer

interface stability was proposed to trigger release of spring-like tension in the crescent-

shaped P monomer such that it becomes straightened. Thus the clamp is not simply an

inert ring, but contains structural information for its own opening. The crescent shape of

the P subunit to which 8 bound was calculated to straighten by 120 at the unopened

interface and by 50 at the opened interface by comparison with the structure of dimeric 3.

Comparison of the 6 subunit in the 8:P1 structure with the structure of the 8 subunit

from the y358' structure revealed that the N-terminal domain of 8 is in a considerably

different position when bound to p. An a helix (a4), containing the PIE, in domain-I

translates by -5.5 A and rotates -450 with respect to the rest of domain-I, and the pIE or

hydrophobic wedge is extended further away from the a4-helix for contact with P. The

"reverse" conformational change in 5, from the p-bound form to the P-free form of results

in exposure of specific hydrophobic residues that are important for interaction with the 5'

stator of the clamp loader.

Although the two structures were solved with mutants of P that do not form dimers,

the above conclusions regarding release of spring-like tension in opening an interface

were consistent with molecular dynamics simulations performed on the dimeric P clamp.

In these simulations, a 1 protomer was simply released from dimeric constraints








determined from the original 02 clamp structure. The P protomer followed a trajectory in

the simulation that resulted in straightening of its crescent-shape such that it

superimposed well with the structure determined from the S-bound P monomer.

DNA structural requirements for P clamp loading by I complex

The 3 clamp loading mechanism and preinitiation complex formation for

processive synthesis by pol III holoenzyme has been studied on several different DNA

substrates including synthetic poly-dT-primed poly-dA-templates, as well as circular

phage genomes in different replicative forms (i.e. single- or multi-primed, nicked-duplex

etc.). It was shown that short synthetic primed-template DNA (pt DNA) substrates (i.e.

an 80-105 nucleotide template with a 30 nucleotide primer) whose sequence was based

on the M13 phage where sufficient for proficient elongation by pol III core in the

presence of P clamp and reconstituted y complex, and therefore for the study of the clamp

loading mechanism (Bloom et al., 1996).

Using these short synthetic pt DNA substrates in experiments along side circular

phage genome substrates containing supercoiled or relaxed-structures, the DNA structural

requirements for 1 clamp loading where investigated (Ason et al., 2000; Ason et al.,

2003; Yao et al., 2000). For preinitiation complex formation at the replication fork it was

known that RNA primers were required. The exact structural nature of these primers

were unknown except that a 3'-hydroxyl group was annealed at the single-stranded DNA

(ss DNA) / double-stranded RNA/DNA (ds RNA/DNA) junction where the preinitiation

complex was formed. In experiments with supercoiled DNA and closed relaxed circular

DNA substrates it was shown that reconstituted y complex did not require a 3'-end to

load p, as it was able to load the clamp onto supercoiled DNA, but not closed relaxed








circular DNA (Yao et al., 2000). It was hypothesized that supercoiled DNA may produce

unwound regions due to superhelical tension, and at these ds RNA/DNA / ss DNA

junctions y complex could load 0 forming the preinitiation complex.

Using circular single-stranded M13 phage DNA primed with synthetic DNA

primers differing in specific structural features such as length and unannealed 15-

nucleotide "flaps" at 5'- or 3'-ends, or both 5'- and 3'-ends, it was shown that y complex

can load P onto a wide range of these substrates (Yao et al., 2000). The minimal primer

length requirement was found to be 10 base pairs, consistent with the size of the inner

pore width of 3, and with the known fact that in vivo, primers formed by primase, are ~

10-12 nucleotides in length (Kitani et al., 1985; Kong et al., 1992). Yao, Leu et al. also

showed that y complex could load 0 onto primers with 5'- and/or 3'- 15-nucleotide-long

unannealed flaps, supporting their conclusion that y complex only requires a ds DNA / ss

DNA junction to load the clamp.

By placing protein- or DNA secondary-structural blocks on or within the primer in

replication assays, the polarity and primer spatial requirement for clamp loading were

determined. Using physical (i.e. protein) blocks tightly bound at the annealed primer 5'-

or 3'-ends it was shown that y complex exhibits polarity in loading the 1 clamp

specifically at the 3'-end of the primer. This result was expected due to the fact that

polymerase synthesis elongation extends from the primer 3'-end in the 5' to 3' direction.

Placement of DNA secondary structural elements at different distances upstream from the

3'-end of the primer, allowed determination of the spatial requirement of the primer for

P-bound y complex and p-bound pol III core. It was shown that with y complex, 14-16

base pairs of primer were needed to load the clamp, and 20-22 base pairs were needed for








P to bind pol 111 core. This result indicated that y complex interacts directly with 4 to 6

base pairs of the primer, and that the clamp is pushed back an additional -6 base pairs

when pol III core binds p at the ds DNA / ss DNA junction (Yao et al., 2000).

Utilizing fluorescence-based steady-state and pre-steady-state kinetic anisotropy

assays with a X-Rhodamine-labeled short synthetic pt DNA in solution, a DNA-triggered

switch in y complex was discovered (Ason et al., 2000). This DNA-triggered switch

caused a change in affinity of y complex for DNA, and was found to be pt DNA- and

ATP hydrolysis-dependent. Initially, in steady-state assays, it was observed that y

complex had higher affinity for ss DNA than pt DNA. Even more surprising, was that

the 5'-single-stranded template overhang of the pt DNA was the same sequence and

length as the ss DNA substrate examined in these experiments. However, pre-steady-

state real-time assays revealed DNA binding kinetics suggesting that y complex bound

transiently with high affinity to the pt DNA substrate. Use of non-hydrolyzable ATPyS

in both steady-state and pre-steady-state assays with pt DNA substrates removed this

dynamic switch in y complex affinity for DNA resulting in a high affinity state only. It

was therefore proposed that ATP hydrolysis by y complex was required for cycling

between high and low affinity states, and that the low affinity state may have been an

ADP-bound y complex. With ss DNA substrates y complex with p maintained high

affinity for the DNA, hydrolyzed ATP, but did not load p (Bloom et al., 1996; Turner et

al., 1999).

The dynamic interaction with pt DNA and the DNA triggered switch in y complex

accomplishes two major goals in the clamp loading reaction. The primer-template, ds

DNA / ss DNA junction, provides the proper site for clamp loading and preinitiation








complex formation, and modulates the DNA binding affinity of the clamp loader, so as

not to allow y complex competition with pol III core bound to p at previously formed

initiation complexes. The nature of the low affinity state of7 complex remains a

mystery, and this dissertation addresses several possibilities [and questions] including the

dynamics of conformational changes defining different states ofy complex with respect

to the clamp loading reaction.

A recent investigation addressed the possibility that the pt DNA-triggered

modulation of y complex provides a dynamic mechanism for recognition of appropriate

sites for P clamp specific for proficient DNA synthesis. This was accomplished by

investigation of the DNA structural features required to trigger y complex into the low

affinity state and release P (Ason et al., 2003). In fluorescence-based steady-state and

pre-steady-state anisotropy assays in solution, 7 complex binding and clamp loading were

studied with elongation-proficient DNA substrates (i.e. synthetic template DNA primed

on its 3'-end or center), or elongation-deficient DNA substrates (i.e. synthetic template

DNA primed such that a blunt duplex was formed at the 5'-end of the template, or

unprimed-ss DNA). In steady-state titrations of complex to the elongation-proficient or

deficient DNA substrates, results similar to the previous investigation defining the DNA-

triggered switch were observed. The y complex showed high affinity for elongation-

deficient DNA substrates, but did not appear to bind well to the elongation-proficient

DNA substrates. Pre-steady-state assays revealed biphasic "up-down"-kinetics (i.e., a

rapid rise to a defined peak followed by a decay phase into steady-state DNA binding

activity), representing transient high affinity ofy complex for the elongation-proficient

DNA substrates as well as characteristic changes in anisotropy for the clamp loading








reaction when 1 was present. For elongation-deficient DNA substrates, pre-steady-state

binding kinetics were monophasic suggesting a simple high affinity bimolecular

equilibrium, whether p was present or not. Additionally, a correlated DNA-binding and

ATP hydrolysis assay with 7 complex and elongation-proficient pt DNA clearly showed

that a binary complex consisting of y complex with pt DNA transiently formed just prior

to DNA-triggered ATP hydrolysis and release of y complex without any further binding.

Overall, the results show that y complex uses a dynamic mechanism driven by ATP

binding and hydrolysis for targeting P3 clamp only to DNA that can serve as a template for

synthesis elongation by pol III core, and prevent further interaction with polymerase-

bound p.

Mechanism of the P Clamp Loading Reaction Cycle by y Complex

Formation of a processive pol III holoenzyme requires the formation of

preinitiation complexes on the leading and lagging strands at the replication fork. The

ring-shaped sliding clamp must be opened and loaded onto the circular E. coli

chromosome in order to form a topological link between pol III and DNA. The clamp

loading machine performs this task through dynamic protein-protein and protein-DNA

interactions. Some 26 years ago, Sue Wickner presented a model for the mechanism of

DNA elongation catalyzed by DNA polymerase III that still generally holds true today

(Wickner, 1976). Wickner's model described a mechanism where primed-template DNA

was activated by the ATP-dependent transfer of "DNA elongation factor I" (the sliding

clamp) by "DNA elongation factor III" (the clamp loader), to which DNA polymerase III

then bound in an ATP-independent manner, followed by DNA elongation. Combined

biochemical analysis and structural details have in recent times given us a detailed, yet








still incomplete, view of the clamp loading mechanism required for processive synthesis

by pol III holoenzyme.

The y complex (y3,8,8',y,V), or a minimum complex of five subunits (y3,8,8') is the

energase complex that transduces the energy from ATP binding and hydrolysis into

mechanical work to load P onto DNA (Figure 2-5).

Within this machine, the three y subunits serve as the motor subunits, 6 has the p-

interaction element (PIE), and 8' is the "stator" upon which movement of the other

subunits is thought to be supported. In general terms, ATP binding and hydrolysis by the

y subunits promote conformational changes that modulate the dynamic protein-protein

and protein-DNA interactions involved in clamp loading. To work at the speed and

efficiency required at the replication fork within the cell, there must be precise

communication between all subunits of the clamp loading machine to power and regulate

its activity.

Initially, ATP binding to the y subunits, but not hydrolysis, promotes changes in the

clamp loader that modulate the binding affinity for the clamp and DNA, and therefore

powers most of the steps in the clamp loading reaction (Bertram et al., 1998; Hingorani

and O'Donnell, 1998). Asymmetry of the interfacial nucleotide binding sites in the clamp

loader structure provided a model for activation upon ATP binding for p clamp loading.

The S'-yi and y2-y3 interfacial nucleotide binding sites were open in the structure and

could provide space for the first one or two molecules of ATP to bind. These

conformational changes could also result in the opening of the y1-y2 ATP binding site,

and further "splaying-out" of the y subunits from the "backbone" of the 8' stator, and

ultimately exposure of the 8 subunit PIE (Figure 2-5B). This model would bring the








maximum number of ATP molecules needed for exposure of the 5 subunit PIE to three in

"open" y complex.


A. B. C.





r-IE
ATP ATp TP T/



D. E. F.


P ol HI 7
core

primed-template
DNA Pi




Figure 2-5. A schematic cartoon of the basic steps in the clamp loading reaction and
initiation complex formation. The cartoon of the clamp loader in "closed" and
"open" states is based on the y388' crystal structure, and the y complex
subunits are colored as in figure 2-3. A) ATP binding to the y subunits. B)
ATP-dependent conformational changes occur within the clamp loader
exposing the 8 subunit PIE (green). C) P clamp binding to the clamp loader
through the PIE induces a conformational change in 3, opening the clamp at a
single interface. D) The p-bound clamp loader complex binds pt DNA and
positions the clamp at the ds DNA / ss DNA junction at the 3'-OH of the
primer. E) pt DNA triggers ATP hydrolysis, the release of inorganic
phosphate (Pi), and dissociation of ADP-bound clamp loader from loaded p
clamp. F) Polymerase III core (a, ,, and 0 subunits) can then bind P and
commence processive elongation of the template.

Originally it was shown that ATP binding caused a conformational change in y complex

that resulted in characteristic proteolytic cleavage of the exposed 8 subunit, and that this

change did not require 0 (Hingorani and O'Donnell, 1998; Naktinis et al., 1995). The








interaction of the exposed 8 subunit PIE with p then induces or stabilizes conformational

changes in P that result in ring opening at a single dimer interface (Figure 2-5C). The

kinetics of ATP binding and the dynamics of the conformational changes within the

clamp loader are a major component of this dissertation and will be discussed further in

chapters 4 and 5.

ATP binding to y complex is also required for interaction with DNA in the clamp

loading reaction (figure 2-5D) (Bloom et al., 1996; Stukenberg and O'Donnell, 1995).

Studies described in this dissertation show that none of the individual subunits of the

clamp loader have significant DNA binding affinity, and it is hypothesized here and that

the interaction of all clamp loader subunits form the DNA binding site. With a clamp

bound and opened, the clamp loader then binds DNA at a ds DNA / ss DNA junction,

specifically on elongation proficient primed-template DNA (Ason et al., 2003; Yao et al.,

2000). The DNA binding kinetics and positioning of 3 at the 3'-hydroxyl of the primer

happen in less than 100 ms with an approximate bimolecular binding constant on the

order of 2.0 4.0 x 108s M1 s1 just prior to the hydrolysis of one molecule of ATP for

each y subunit in the clamp loader (Ason et al., 2000).

Primed-template DNA triggers ATP hydrolysis and dissociation of the clamp

loader (Figure 2-5E). Earlier investigations of ATP hydrolysis by the clamp loader

showed that mutations of the conserved lysine in the y subunit Walker-A motif of

domain-I, to alanine or arginine (K51A or KS1R), resulted in complete abrogation of

ATP hydrolysis activity by the clamp loader (Xiao et al., 1995). Our laboratory has also

shown that these mutations of the conserved lysine in y complex abolished both P and

DNA binding in solution (unpublished), therefore it is likely that the lysine-mutants








cannot bind ATP. Use of non-hydrolysable ATPyS allowed y complex to bind 0 and

DNA, consistent with the notion that nucleotide binding powers most of the steps in the

clamp loading reaction. In these reactions, the complex of P-bound y complex remained

in a dynamic steady-state interaction with DNA for periods of time extending to 90

minutes (Bertram et al., 1998; Bloom et al., 1996). These investigations showed that

although ATP binding can bring p-bound y complex to DNA; ATP hydrolysis was

absolutely necessary for release of 7 complex, and completion of the clamp loading

reaction. The pre-steady-state ATP hydrolysis assays showed that the ATPs were

hydrolyzed at a minimal rate of 20-34 s-' (Bertram et al., 2000). Closure of the p clamp

on DNA was found not to require any further energy (Turner et al., 1999), therefore ATP

hydrolysis by the clamp loader is not needed to close P onto DNA. The y complex crystal

structure predicted a mechanism wherein ATP hydrolysis drives final conformational

changes in which 3 is pushed off the 8 subunit PIE as 8 returns to its occluded interaction

with the rigid 8' subunit. This is consistent with an earlier investigation that showed ATP

hydrolysis was required for release of y complex from P3 on DNA, leaving the clamp

behind (Bloom et al., 1996). Dissociation of y complex from the loaded clamp is

essential, and would provide space for the interaction of p with polymerase III, which is

known to bind the same surface of p that y complex had occupied (Figure 2-5F). Pre-

steady-state analysis of DNA binding activity and ATP hydrolysis have shown that a

single "turnover" of the clamp loading reaction takes approximately 300 ms before

entering a steady-state that continues at a rate of -2-2.5 s-1 (Ason et al., 2000).

Ultimately ATP hydrolysis would then allow the clamp loading machine to reset

itself for continued clamp loading, most likely through nucleotide exchange and








conformational changes. The y complex released from the clamp loaded on DNA is

presumably in some inactive "closed" state that would have to release ADP and become

able to bind ATP again. ADP release could cause conformational changes that may

transiently bring y complex through some state in which there is no nucleotide bound.

ATP binding would then cause the conformational changes that reactivate y complex for

another clamp loading reaction. In previous pre-steady-state ATP hydrolysis assays,

there was a "pause" in hydrolysis activity observed during the last 200-250 ms of the first

turnover of ATP during transition into the steady-state (Ason et al., 2000). Given the

way these assays were initiated (i.e., with preincubation of 7 complex with ATP and 1

before mixing with pt DNA), during this transition, the rate-limiting step of the reaction

cycle was believed to occur. The exact nature of the rate-limiting step of the clamp

loading reaction is not yet known. Possibilities include ADP-release, ATP binding,

conformational dynamics within y complex, and P clamp binding. The steady-state and

pre-steady-state kinetics of DNA binding and ATP hydrolysis by the clamp loader are

further studied in this dissertation, and reveal some novel kinetic features that predict that

conformational changes within y complex are most likely the rate-limiting step in the

reaction and regulate the clamp loading mechanism.

Mutations of the p Clamp, and y Complex 8' and y Subunits: Effects on the Clamp
Loading Mechanism

Mutations in the p clamp as well as in subunits of y complex have been studied to

provide greater detail for the interactions modulating this clamp loading mechanism.

The steady-state and pre-steady-state kinetics of DNA binding and ATP hydrolysis

activities were investigated using a 1 clamp with mutations at two positions within the

dimer interface. The leucine to alanine (L273A, L108A) mutations were thought to








weaken the dimer interfaces, although this mutant P was still able to form dimers in

solution (Bertram et al., 1998). The P interface mutants were found to bind y complex

and DNA similar to wild-type p, but they remained bound to DNA in a ternary complex.

The DNA binding assay results appeared similar to assays performed in the presence of

non-hydrolyzable ATPyS where the P-bound clamp loader appeared "stuck" in a dynamic

steady-state interaction with DNA, without loading P. In correlated pre-steady-state

DNA binding and ATPase assays, the P interface mutants bound to DNA and hydrolyzed

ATP with nearly identical kinetics as wild-type-P. However, the clamp loader steady-

state ATPase activity with the P mutants was significantly decreased, coincident with

extremely slow dissociation of y complex from mutant p on DNA (Ason et al., 2000).

The results confirmed the idea that the clamp loader must cycle off of P that is loaded

onto DNA, and undergo a rate-limiting step that is separated from the DNA-bound state

in order to continue loading clamps as previously suggested by Bloom et al. 1996. The

results of these investigations also showed that although p does not require any energy

from ATP hydrolysis to close on DNA, ATP hydrolysis is tightly coupled to release of

the clamp on DNA. It seemed interesting that mutations within the 1 dimer interface so

greatly affected the steady-state DNA binding and ATP hydrolysis activities of the clamp

loader. Now, with better knowledge of the interaction between the clamp loader-S

subunit and p clamp it is more clear how P interface mutations would cause these effects

in the clamp loader. Perhaps the conformational change of p cannot be properly

promoted by the 8 subunit-p1IE interaction, disallowing loading onto DNA. This

possibility could result in a decrease in the interaction of y complex with DNA, and

reduced DNA stimulated ATP hydrolysis activity.








To gain further insight into the AAA+ machine activity of y complex, mutations in

the sensor-1 (SRC) motif were made, and steady-state ATP binding, hydrolysis, and

clamp loading activities were examined (Johnson and O'Donnell, 2003). In one of these

mutations, arginine-158 of the 6' stator subunit (8'-R158A) SRC motif was changed to

alanine. This mutation removed the "arginine-finger" from the 8'-yi nucleotide binding

interface. The results showed that the reconstituted 8'-R158A-y complex still bound

approximately three molecules of ATP, but had reduced ATP hydrolysis activity and

diminished ability to load P in DNA synthesis assays. These results implied that the

"arginine-finger" contributed from the 8' subunit to the 6'-7y nucleotide binding site was

in fact catalytic in nature and therefore, this "stator" subunit was coupled with the motor

function of the clamp loader, as well as performing as the motor support subunit. Recent

results from our laboratory have shown that this reconstituted 8'-R158A-y complex

mutant has little ability to bind P, as well as significantly reduced DNA binding and

clamp loading abilities (Snyder, A.K., personal communication). This is consistent with

the reduced DNA synthesis activity observed in the original study.

The "arginine-finger" was removed from the 7 subunits to create a second clamp

loader mutant (Johnson and O'Donnell, 2003). The y-R169A mutation removed the

"arginine-finger" from the 71-72 and 72-73 nucleotide binding interfaces, thus only

affected two nucleotide binding sites in the clamp loader. This reconstituted 73-R169A-

complex still maintained the ability to bind three molecules of ATP and 0, but had no

ATP hydrolysis activity, clamp loading activity, and was unable to stimulate DNA

synthesis. Our laboratory has since shown that this 73-R169A-complex can bind p at a

level near that of wild-type y complex in an ATP-dependent manner in solution, but








showed no DNA binding activity in steady-state and pre-steady-state solution-based

assays (Snyder, A.K., personal communication). Johnson and O'Donnell discussed the

8'- and y-subunit mutants in terms of an ordered ATP binding and hydrolysis mechanism

where ATP molecules are hydrolyzed in the reverse order than they were bound,

requiring proper positioning of the "arginine-fingers" for catalysis. Our laboratory's

results indicate that there is more than just a disruption in catalysis of ATP hydrolysis in

these mutants. Taken together, all of the mutation analysis results exemplify the need for

precise subunit conformational communication within the clamp loading machine for P

binding and also for formation of the putative DNA binding surface on the clamp loader.

The Clamp Loading Machine Within Polymerase HI Holoenzyme

Most of the biochemical analyses and all of the structural analyses of the clamp

loader and clamp loading mechanism have been performed with reconstituted y complex.

How do these results apply to the clamp loader within pol HI holoenzyme? The clamp

loader within holoenzyme contains both the t and y subunits encoded by the dnaX gene.

In the T subunit, the region of the DnaX protein structure extending from the C-terminus

of the y subunit region into the C-terminal domain of r contains many proline residues

and is thought to be unstructured and therefore a highly flexible region (O'Donnell et al.,

2001). This flexible region divides the DnaX (t) protein into the y-AAA+ motor region,

and the r-C-terminal replisome interaction region. Within holoenzyme, there are at least

two T subunits, and a single y subunit (Figure 2-6).

The structural arrangement of the DnaX subunits of the clamp loader within

holoenzyme is likely to be [86, T2,7i,S'1,,X,~l]or [6'1,7y, T2,61,X1,~1]. The t and

subunits only bind to the y subunit through xy, and the 6' subunit also has been shown to

bind only to the y subunit with high affinity (Glover and McHenry, 2000).
































Figure 2-6. Architecture of the polymerase III holoenzyme at the replication fork
organized by the DnaX clamp loading machine. The leading- and lagging-
strands are unwound by hexameric DnaB helicase (green), shown bound to the
SSB-coated (yellow) lagging strand. The DnaX clamp loader containing two
T subunits (violet) (8'1,Yi, T2,8i,Xi,Wj) is connected within the holoenzyme
through the flexible linker regions of the T subunit C-terminal domains that
also dimerize pol HI core (a e and 0), and further cement the holoenzyme
through contact with DnaB. The clamp loader x and T subunits are not shown
for clarity. Two p-rings (red) clamp the pol III cores to the templates, and a
third is shown bound to the clamp loader. In this schematic, primase would
bind and synthesize a new RNA-primer on the lagging strand. Then it is
predicted that the clamp loader would "swing" over to the nascent primer,
displace primase through the X subunit, and load a fresh clamp forming a
preinitiation complex. When the lagging-strand polymerase reaches the 5'-
end of the previously synthesized Okazaki fragment, it would then release P
and cycle to the newly formed preinitiation complex and commence
elongation of DNA.

The two r subunits cement the array of proteins at the replication fork by dimerizing the

core polymerases in the holoenzyme, binding to the replicative DnaB helicase, and

protecting the P subunit from removal when it is attached to the polymerase a subunit.








Through the flexible linker region, these functions of the T subunits are connected to the

y-AAA+ motor region of the clamp loader. In this way, it is envisioned that the clamp

loader associated with the holoenzyme in vivo has a swinging range of motion on the

flexible tether loading at least one clamp on the leading strand and providing continuous

clamp loading activity on the lagging strand for Okazaki fragment synthesis (O'Donnell

et al., 2001).

Duplex DNA is anti-parallel, and therefore the core polymerases of the holoenzyme

must work in some asymmetric fashion to simultaneously complete replication of the

chromosome (Figure 2-6). The single y subunit of the DnaX clamp loader, and

consequential asymmetric distribution of the 8,8',Xand Vi subunits gives the holoenzyme

structural asymmetry. The y subunit is the only clamp loader subunit to which i and

therefore X adaptor proteins bind. This provides the holoenzyme clamp loader both

structural and functional asymmetry through T-DnaB helicase contact, and the x-SSB

interactions necessary for lagging strand synthesis.

Recently it was proposed that the asymmetry in the clamp loader provides

asymmetry to the core polymerases through nucleotide binding to the clamp loader

(Glover and McHenry, 2001). Use of non-hydrolyzable ATPyS was sufficient to allow

the DnaX complex to load a clamp for only one polymerase, presumably the leading

strand polymerase. ATP hydrolysis was absolutely required for clamp loading on the

"lagging strand" polymerase in their assays, and subsequent addition of ATPyS was able

to disassociate this "lagging-strand" polymerase. Whether the nucleotide binding status

of the clamp loader, or simply the structural asymmetry of the clamp loader provide

distinct leading and lagging strand functions to the core polymerases will require further








investigation. Overall, the structurally and functionally asymmetric holoenzyme has the

means to stay continuously linked to the leading strand while performing discontinuous

synthesis on the lagging strand at the replication fork by the distinct structural and

functional characteristics of the DnaX clamp loader. Polymerase processivity is required

for DNA replication in all free living cells, and nature has provided a consistent

mechanism to provide the means for polymerase processivity though conserved evolution

of these replication proteins across all branches of life.

Clamps and Clamp Loaders of Bacteriophage, Eukaryotic, and Archaeal Organisms

The combined biochemical and structural investigations of the E. coli clamp and

clamp loader have yielded many great details into the mechanism of processivity clamp

assembly on DNA. These intimate details have allowed complementary studies of

clamps and clamp loaders of other organisms to thrive due to the evolutionary

conservation of these processivity proteins across all branches of life, and suggest a

common mechanism of clamp loading in all life forms. It is well known that sliding

clamps contact polymerases with DNA for processive synthesis of complete

chromosomes or replicons in a fundamentally similar way (Ellison and Stillman, 2001;

Hingorani and O'Donnell, 2000). Now, structural evidence is revealing the extraordinary

functional similarities in clamps and clamp loading machines of viral replicases,

prokaryotes, eukaryotes and archaea for processive DNA elongation (O'Donnell et al.,

2001). All organisms through these evolutionary branches of life appear to utilize AAA+

proteins as the energase of their clamp loading machines (Table 2-1). Of those organisms

studied, each clamp loader appears to, or is predicted to utilize a AAA+ clamp loading

machine, like y complex, that contains several ATPase motor subunits, a clamp

interacting subunit, and a supporting stator subunit (Davey et al., 2002). This








conservation of sequence and, to a greater extent, structural similarities of these AAA+

proteins illustrate the evolutionary importance of these clamp loading machines.

Table 2-1. Clamps and clamp loaders through evolution
Evolutionary Processivity AAA+ Function. Some Characterized
Branch Proteins protein Organisms
Prdkaryotic 3 -clamp

/ + -motor Escherichia colf
+ -stator Bacillus subtilus
8 + -clamp- Bacillus-
interacting staerothermophilus
subuni te Aquifex aeolicusf
-AAA+ adaptor Thermus thermophilusf
-AAA+ adaptor
Bacteriophage gp45 -clamp
(viral) gp44 + -motor (stator?) T7 phage

gp62 +(?) -clamp RB69 phage
interacting
subunit
Eukaryotic PCNA -clamp Homo sapiens

RFC_2d + -motor Saccharomyces cerevisiae
RFC-3 + -motor Schizosaccharomyces-
RFC-4 + -motor pombe
RFC-5 + -stator Caenorhabditis elegans
RFC-1 + -clamp Drosophila melanogaster
interacting Arabidopsis thaliana
subunigt Oryza sativa (rice)
Archaeal PCNA -clamp Pyrococcusfuriosus
Archaeoglobus fulgidus
RFC-S + -motor (stator?) Archaeoglobusfulgidus
RFC-L + -motor / clamp Sulfolobus solfataricus
RFC-L + -motor clamp Methanobacterium-
subunit thermoautotrophicum AH
by analogy to the y complex clamp loader
gp, gene product
has conserved hydrophobic wedge residues of clamp interaction element
d nomenclature used is for yeast replication factor-C (RFC)
clamp loader structure known
'Thermophile bacteria








Bacteriophage T4 Clamp and Clamp loader

When infecting bacteria, the bacteriophage has the necessary components to form

its own replication machinery. The replisome of T4 bacteriophage contains gene product

(gp) gp43 polymerase, gp41 / gp59 helicase and its accessory factor, respectively, gp61

primase, and gp32 single-stranded DNA binding protein. T4 also provides the

processivity clamp gp45 and clamp loader complex gp44/62 to complete the replisome

(Salinas and Benkovic, 2000). The gp45 clamp interacts with gp43 polymerase through a

C-terminal region of the polymerase. This interaction between the clamp and polymerase

is highly conserved across evolution (Ellison and Stillman, 2001), and the structure of a

T4 clamp bound to polymerase on DNA has been modeled based on the nearly identical

RB69 bacteriophage sliding clamp bound to polymerase (Alley et al., 2001; Shamoo and

Steitz, 1999). To date, these are the only structural views of a polymerase bound to its

clamp on DNA.

The gp45 processivity protein is a ring-shaped clamp formed of six similar

structural domains like 0 clamp, but unlike dimeric 3, is a trimeric clamp. Although there

is less than 10 % sequence identity between gp45 and p, (and PCNA, see below), the

clamp structure is amazingly similar (Jeruzalmi et al., 2002). The crystal structure of

gp45 shows that its properties are similar to p clamp and PCNA (Moarefi et al., 2000).

The gp45 clamp has an overall negative charged P-sheet outer surface surrounding an a-

helical inner pore of positive electrostatic potential. The dimensions of the gp45 clamp

show an inner pore diameter of -35 A, and a width of~25 A, however, the clamp has an

overall triangular topology compared to the nearly circular structures P clamp and PCNA.

Consistent with data that show gp45 clamp is the least stable of the known sliding clamps

on DNA (Yao et al., 1996), the gp45 clamp is predicted to be slightly open in solution








through a pucker which gives it a "lock-washer"-type structure (Alley et al., 1999). It is

still possible that the clamp loader must open the clamp further for the loading reaction

(Alley et al., 2000).

The T4 clamp loader is a pentameric complex of the gp44 and gp62 subunits with a

subunit stiochiometry of four gp44 subunits to one gp62 subunit (Jarvis et al., 1989). The

T4 gp44/62 complex binds and loads gp45 onto DNA in an ATP-dependent manner and

requires hydrolysis of bound ATP. In earlier work, it was shown that four molecules of

ATP were hydrolyzed by gp44/62 during formation of the T4 holoenzyme, and that some

step following ATP hydrolysis was rate limiting in the reaction (Sexton et al., 1998). A

highly detailed investigation using fluorescent resonance energy transfer techniques with

fluorescent-labeled gp45 clamp, combined the pre-steady-state kinetic analysis of clamp

opening and closing by the gp44/62 clamp loader with analysis of binding and hydrolysis

of ATP (Alley et al., 2000). This investigation showed that hydrolysis of ATP was

required to open the gp45 clamp, and then additional hydrolysis was required to load the

clamp on DNA. A more recent study on the ATP hydrolysis activity of gp44/62 in clamp

loading conflicted with previous work, and showed that ATP binding alone is sufficient

for gp44/62 to bind gp45, and at least one ATP is required to complete the loading

reaction (Pietroni et al., 2001). This study also revealed that ATP hydrolysis was not

required to open the gp45 clamp, and that the rate-limiting step in the loading reaction

was either ADP release from gp44/62, or a conformational change before clamp loading.

This investigation is more consistent with the detailed study of the y complex in 0 clamp

loading.








The gp44 subunits of the complex have homology with AAA+ proteins and contain

the Walker-A P-loop and sensor-1 SRC motifs needed for ATP binding and hydrolysis

(Davey et al., 2002; Neuwald et al., 1999). By analogy to y complex, the gp44 subunits

are expected to be the motor subunits of the T4 clamp loader that undergo ATP-

dependent conformational changes that drive gp45 binding and interaction with DNA in

the loading reaction. The gp62 subunit shares no sequence homology to other AAA+

subunits, and does not bind or hydrolyze ATP. Until its structure is known it can only be

predicted that gp62 is a AAA+ homologue, and the subunit that interacts with the gp45

clamp during the clamp loading reaction (i.e. like the 8 subunit of y complex). There is

no support-like stator subunit in the gp44/62 complex as all of the gp44 "motor" subunits

are identical. However, even in y complex, the 8' stator subunit has been shown to be

directly involved in the catalytic motor function of the clamp loader through precise

positioning of its SRC motif by conformational changes (Johnson and O)Donnell, 2003).

One of the gp44 subunits of the T4 clamp loader may work in a similar manner, and it

may turn out that this putative, distinct, gp44 subunit may also be incapable of ATP

hydrolysis like 8' in y complex. These predictions await further biochemical and

structural analyses of the T4 clamp loader.

Eukaryotic PCNA Clamp and Replication Factor-C Clamp Loader

The essential replication machinery of eukaryotic organisms was originally

identified by investigations using the SV40 DNA virus as a model system. SV40 uses

host cell machinery in replication of its DNA, and requires only its own "T-antigen" for

replication initiation and DNA helicase activity (Waga and Stillman, 1994). Processivity

proteins, proliferating cell nuclear antigen (PCNA) and replication factor-C (RFC) were








identified as the sliding clamp and clamp loader, respectively, for the replicative

polymerases 5 /I (Waga and Stillman, 1998). Like the y complex clamp loader, RFC has

DNA-dependent ATP hydrolysis activity that is further stimulated by the presence of the

PCNA clamp (Lee et al., 1991). RFC is also directly involved in a primase to polymerase

switch, similar to y complex, and influences the length of primer synthesis by polymerase

a primase (Mossi et al., 2000; Tsurimoto and Stillman, 1991).

The PCNA sliding clamp structure was solved from budding yeast S. cerevisiae

(Krishna et al., 1994). Once again, even though sequence homology between the

bacteriophage T4 clamp, prokaryotic 1 clamp, and PCNA was low, the structure of

PCNA is nearly identical to the other clamps (Moarefi et al., 2000). Unlike the 1 clamp,

but like T4 gp45 clamp, PCNA found to be trimeric. All three clamps have six

structurally similar subdomains, dimensions (i.e. the PCNA inner pore is ~ 35 A), and

similar electrostatic characteristics on their structurally conserved outer p-sheets, and

inner a-helices. The overall topology of the PCNA trimer is more hexameric or circular

than gp45, remains closed in solution, and is more stable on DNA than gp45, but not as

stable as dimeric 1 clamp (Yao et al., 1996).

RFC has been extensively studied from yeast and human origin, and has been

identified as a heteropentameric complex consisting of one large subunit and four small

subunits of different mass. The subunit nomenclature is: RFC- (yeast/human): l/p140,

2/p37, 3/p36, 4/p40, and 5/p38. For simplicity the yeast RFC nomenclature will be used

here. Deletion analysis of RFC subunits showed that the C-terminus of each subunit was

required for formation of the pentameric complex, and that this complex must form to

allow DNA stimulated ATP hydrolysis activity (Uhlmann et al., 1997). It was also








shown that none of the individual subunits alone had DNA binding activity (Uhlmann et

al., 1996). Using surface plasmon resonance (SPR) and filter magnetic-bead binding

assays, it was shown that ATP mediated all interactions of RFC with PCNA and DNA

(Gomes and Burgers, 2001). This study also showed that ATP binding but not hydrolysis

was required for the interaction with DNA, and that use of non-hydrolyzable ATPyS

resulted in a "stuck" complex of PCNA-bound RFC on DNA. These results are

consistent with complementary investigations performed with y complex. The DNA

structural requirements for PCNA loading by RFC were also tested and showed that RFC

could load PCNA on ds DNA / ss DNA junctions not requiring an annealed 3'-OH

similar to the DNA structural requirement for y complex catalyzed clamp loading

activity.

Recent studies with yeast RFC (Schmidt et al., 2001a), as well as previous

investigations with human RFC (Cai et al., 1998) identified the subunits involved in ATP

hydrolysis activity and those essential for DNA recognition by mutational analysis of the

conserved Walker-A lysine residue in each subunit. The results of these mutational

analyses revealed that RFC subunits 2,3, and 4 were required for ATP binding and

hydrolysis in efficient PCNA loading. Mutations in RFC-1, the large subunit, had no

effect on PCNA binding and loading abilities. RFC-5 has modified Walker-A and B

sequences suggesting that it does not bind or hydrolyze ATP, similar to the 8' subunit of

y complex (Cai et al., 1998; Schmidt et al., 2001a). Consequentially, these Walker-A

mutations of RFC-5 did not effect PCNA binding and loading activities.

As in the T4 clamp loading reaction cycle, the use of ATP by RFC in PCNA

loading is also not fully understood. This may be due to limitations of the SPR and filter








binding analyses that were being performed along side steady-state ATP hydrolysis

analysis of these reactions. Despite these limitations, a pathway of multiple stepwise

ATP binding events is thought to be required for RFC loading PCNA onto DNA, and that

this proposed reaction resembles the well studied E. coli clamp loading reaction pathway

(Gomes et al., 2001). In the yeast PCNA loading reaction, 2 ATP molecules initially

bind RFC, PCNA then binds RFC, and RFC gains affinity for a third ATP.

Conformational changes within RFC are most likely taking place as each ATP binds an

RFC subunit. Finally, RFC-PCNA binds pt DNA, and apparently one more molecule of

ATP binds RFC, perhaps to the RFC-1 subunit. Upon interaction with pt DNA, PCNA is

released, coincident with hydrolysis of up to all four bound ATP molecules. It remains

unknown how many ATP molecules are used for this reaction.

The four small subunits of RFC (2-5) form a functional core, and are used in other

cellular processes. The large RFC-1 subunit swaps with another protein, specific for

functions such as replication termination (Kouprina et al., 1994), and cell cycle

checkpoint control (Green et al., 2000). Differing the nature of ATP binding affinities

and putative conformational changes in the RFC-2-5 core may provide differential

reaction mechanisms for these diverse functions. Perhaps, the first two ATP molecules

bind and cause conformational changes that allow a presentation of general binding site

for PCNA and DNA only when RFC-1 is present, or form and expose different binding

sites when another protein takes the place of RFC-1. Then, other ATPs could bind and

produce conformational changes specific for each function. Another possibility could be

that when a specific protein such as PCNA binds, it alters the affinity for ATP at

additional sites in RFC specific for, in this case, clamp loading.








All five RFC subunits are AAA+ homologues, and are related in sequence to the y

and 8' subunits of y complex (Neuwald et al., 1999; O'Donnell et al., 1993). The

heteropentameric stiochiometry of the RFC subunits predicts that this eukaryotic clamp

loader will have structural similarity to y complex. The RFC-1 subunit is the least

conserved of the RFC subunits, and its specificity in clamp loading with the RFC-2-5

core, and dispensability from the RFC-2-5 core in functions other than clamp loading

suggest that it may be the PCNA clamp interacting subunit of the RFC complex by

analogy to the 8 subunit of y complex, (see Table 2-1). Natural modifications of the

Walker-A and B nucleotide-binding domain of the RFC-5 subunit suggest that this

subunit may not bind or hydrolyze ATP, but it still has a sensor-1 SRC motif that could

interact with the other RFC subunits (Cullmann et al., 1995). Like the 8' subunit of y

complex, RFC-5 is predicted to function as the stator in this eukaryotic clamp loading

machine (Davey et al., 2002). The remaining RFC-2, 3, and 4 subunits, when mutated in

the conserved Walker-A sequence for ATP binding and hydrolysis show the most severe

in vitro defects and in vivo phenotypes (Schmidt et al., 2001a; Schmidt et al., 2001b).

Therefore RFC-2, 3, and 4 were proposed to be the motor subunits of this machine. In

fact, as heterotrimeric complexes, both yeast and human RFC-2-3-4 complexes had

DNA-dependent ATPase activity (Ellison and Stillman, 1998).

This functional analogy of the different RFC AAA+ subunits to the y complex

subunits for prediction of RFC clamp loader structure and mechanism of clamp loading

awaits solution of an atomic structure of the RFC clamp loader. Although, as described

below, the atomic structure of part of the archaeal RFC clamp loader which has








significant homology to eukaryotic RFC, has been solved, and closely resembles the

structure of y complex motor subunits.

Human RFC alone and in a complex with PCNA has been viewed by transmission

electron microscopy (TEM) and atomic force microscopy (AFM) (Shiomi et al., 2000).

The TEM results showed that RFC forms a "closed-U-form" structure in the absence of

ATP. When viewed in the presence of ATP, the RFC molecules appeared as a more

"open-C-form" structure demonstrating that ATP caused RFC to open. Partial

proteolysis results with ATP, non-hydrolyzable ATPyS or ADP confirmed that there was

induction of a distinct structural change in RFC only in the presence of ATP. Additional

TEM imagining also showed that ATPyS or ADP had virtually no effect on the structure

of RFC (i.e. RFC remained in "U-form" similar to that without nucleotide). Although the

images were low resolution and did not allow specific distinction of RFC or PCNA

molecules in the imaged RFC-PCNA complexes, the authors concluded that the "open"

ATP-bound RFC "C-form" was most likely bound to PCNA in the TEM images.

Archaeal PCNA Clamp and Replication Factor-C Clamp Loader

Biochemical analyses and recent structural analysis of the clamps and clamp

loaders of several archaeal organisms are providing important information that is helping

bridge the gap between the prokaryotic structural and functional clamp loading

characteristics and eukaryotic structural and functional clamp loading characteristics.

Archaeal PCNA clamps and RFC clamp loaders share -30-40 % sequence

homology to their eukaryotic counterparts (Seybert et al., 2002). Archaeal organisms

have a PCNA clamp and a RFC clamp loader complex (Table 2-1). Archaeal RFC and

PCNA have been biochemically characterized for Sulfolobus solfataricus (De Felice et

al., 1999; Pisani et al., 2000), Methanobacterium thermoautotrophicum AH (Kelman and








Hurwitz, 2000), Archaeoglobusfulgidus (Seybert et al., 2002), and Pyrococcusfuriosus

(Oyama et al., 2001).

The RFC clamp loading machine from each of these archaeons is composed of two

subunits, RFC-L (large), and RFC-S (small), and each has AAA+ homology (Davey et

al., 2002). Biochemical analyses of RFC and PCNA of these organisms have identified

several general characteristics, and these characteristics are consistent with the clamp

loading mechanisms of both eukaryotic and prokaryotic organisms. Each archaeal RFC

must form a complex to acquire DNA-dependent ATP hydrolysis activity. This ATPase

activity is associated with the RFC-S subunits, and has shown preference for ds DNA / ss

DNA primer-template junctions (Kelman and Hurwitz, 2000; Pisani et al., 2000). For the

Archaeoglobusfulgidus aJRFC clamp loader, it was further shown that the ATP

hydrolysis activity was enhanced 2 to 3-fold by a/PCNA, and that ATP binding, not

hydrolysis was required to stimulate the interaction of a/PCNA with DNA (Seybert et al.,

2002). Each archaeal RFC clamp loader was also shown to support processive DNA

synthesis with its respective PCNA and pol B-family archaeal DNA polymerase. The

major difference in the characterization of these clamp loaders was their subunit

stiochiometry. All have been shown to form pentameric complexes similar to the

eukaryotic and prokaryotic clamp loaders, except the Methanobacterium

thermoautotrophicum AH RFC, which has been shown to form a hexameric complex of

two RFC-L subunits with 4 RFC-S subunits. The general pentameric subunit

stoichiometry resembles the T4 bacteriophage clamp loader and the eukaryotic RFC,

having four RFC-S subunits with 1 RFC-L subunit.








About the same time the y complex clamp loader crystal structure was solved, the

crystal structure of a RFC-S subunit complex of Pyrococcusfuriosus was solved at 2.8 A

resolution (Oyama et al., 2001). The RFC-S subunit formed a "dimer-of-trimers", not

necessarily the functional assembly of the clamp loader. Each RFC-S subunit showed

significant AAA+ structural homology. Superimposition of a RFC-S subunit with the y

complex 8' subunit structure had a root-mean-square displacement of only 3 A for the

two conserved AAA+ domains (I and II), and showed marked overall similarity. The

RFC-S subunits had the Walker-A P-loop and Walker-B motifs flanked by sensor-2 and

sensor-1 motifs forming the AAA+-structurally conserved nucleotide binding site, and

also had the SRC "arginine-finger" motif. Each RFC-S subunit had an a-helical C-

terminal Domain-Ill corresponding to the circular "collar" oligomerization domain of the

y complex clamp loader. The subunit interfaces within each trimer of RFC-S subunits

was arranged such that each nucleotide binding site was in contact with the neighboring

subunit, and the interfaces differed between each subunit, suggesting asymmetry among

the subunits in the crystal, a feature distinguishing the y complex structural-functional

characterization (Jeruzalmi et al., 2001 a).

The RFC-S subunits are AAA+ proteins, and based on the extraordinary structural

similarity with the y AAA+ subunits of the prokaryotic clamp loader, suggest that the

mechanism of ATP-dependent conformational changes that drive the y complex is most

likely the same mechanism driving the archaeal clamp loader. The RFC-S subunit has

the highest sequence similarity to the eukaryotic RFC-3 AAA+ subunit (Cann and Ishino,

1999), and therefore two predictions can be made. First, the RFC-S subunit of archaeal

organisms is the "motor" subunit of the clamp loader, and second, that the mechanism of








the eukaryotic clamp loader may be closer to the mechanism of the prokaryotic clamp

loader than originally expected. To add to this, the RFC-L subunit generally lacks an

SRC motif, and shows highest sequence similarity to the eukaryotic RFC-1 subunit (Cann

and Ishino, 1999). Thus it has been proposed that like the eukaryotic RFC-1 and

prokaryotic 8 subunit, RFC-L is the clamp interacting subunit of the archaeal clamp

loading machine (Davey et al., 2002).

The functional conservation of the sliding-clamps and clamp loaders in all branches

of life is readily understandable considering the importance of their functions in DNA

replication and therefore in cell division for propagation of any given organism. These

are ancient proteins that have not changed considerably for billions of years. The use of

the E. coli model system for studying the mechanisms of clamp loading for processive

replication has thus given us the fundamental basis for study of the clamp loading

mechanisms in all forms of life. In this dissertation project, the conformational dynamics

of the E. coli clamp loader were under investigation, and it is hoped that the results can

be applied for understanding of other AAA+-based clamp loaders and provide additional

functional insights for other AAA+ motors.













CHAPTER 3
MATERIALS AND METHODS

NOTE: The anisotropy binding assays and MDCC-PBP ATPase assays were

performed with both y complex and the minimal complex clamp loader, usually in "back-

to-back" (i.e. hours apart) assays. Therefore, the term "clamp loader" will be used

throughout this chapter to describe both, except where different concentrations of each

were used or where only the activity of y complex was studied.

Proteins, Reagents, and Oligonucleotide Substrates

DNA Polymerase m Proteins

DNA polymerase III proteins were a generous gift from Mike O'Donnell's

laboratory at The Rockefeller University, New York, NY. Proteins were stored in 20

mM Tris-HCI pH7.5, 2 mM DTT, 0.5 mM EDTA, and 10% glycerol. Assay buffer for

all experiments contained 20 mM Tris-HCI pH 7.5, 50 mM NaCI, 40 igg/mL Bovine

serum albumin (Invitrogen Corp., Carlsbad, CA), 5 mM DTT, and 8 mM MgCI2 (Table

3-1).

Concentrations of complex were determined from absorbances at 280 nM in 6 M

guanidine hydrochloride and the calculated extinction coefficient. This concentration

was verified by amino acid analysis following acid hydrolysis of the protein (performed

by the Protein Chemistry Core Facility, Biotechnology Program, University of Florida).

A sample for amino acid analysis was prepared by dialyzing y complex against 20 mM

sodium phosphate buffer pH 7.5 at 4 C to remove glycerol and other buffer components

that interfere with the analysis. The concentration of this dialyzed protein was also








determined by three other methods for comparison; 1) its absorbance at 280 nM in 6 M

guanidine hydrochloride pH 6.5, 2) its concentration in a Bradford assay using prepared

reagents from BioRad and a BSA standard, and 3) its concentration in a Bradford assay

using an IgG standard. The amino acid analysis and absorbance measured under

denaturing conditions yielded concentrations of, 1.6 tM and 1.7 tM, respectively, within

experimental error. Both Bradford-type assays overestimated the concentration of

protein by a factor of 1.2 using the BSA standard and a factor of 3.0 using the IgG

standard. The concentration of p was determined from its absorbance at 280 nm and the

extinction coefficient for native protein of (17 900 M'cm 1) (Johanson et al., 1986).

Reagents

ATP was from Sigma-Aldrich Co. (St. Louis, MO) and 5'-O-(3-thiotriphosphate)

(ATPyS) was from Roche Molecular Biochemicals. Bacterial purine nucleoside

phosphorylase (PNPase) and 7-methylguanosine (MEG) were from Sigma-Aldrich and

stored as solutions at -80 C.

Table 3-1. Assay and protein buffers
Standard assay buffer Protein storage Protein dilution
20 mM Tris-HCI pH 7.5 20 mM Tris-HCI pH 7.5 20 mM Tris-HCI pH 7.5
50 mM NaCl 0.5 mM EDTA 40 ig/mL BSA
40 tg/mL BSA 2 mM DTT 5 mM DTT
5 mM DTT 10 % glycerol 0.5 mM EDTA
8 mM MgCl2 10% glycerol 10 % glycerol

Oligonucleotide Substrates

Synthetic oligonucleotides were made on an ABI 392 DNA/RNA synthesizer using

standard 3-cyanoethylphosphoramidite chemistry and reagents from Glen Research Corp.

(Sterling, VA). Denaturing polyacrylamide gel electrophoresis was used to purify new

DNA. The 30-mer primer has the sequence: 5'-GAG CGT CAA AAT GTA GGT ATT




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ET3R7BFT8_X88VBU INGEST_TIME 2012-12-07T18:37:45Z PACKAGE AA00012899_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES