The University of Toledo

The University of Toledo Digital Repository Theses and Dissertations

2005

Crystallographic studies of the E. coli DNA replication restart primosome Aude Izaac The University of Toledo

Follow this and additional works at: http://utdr.utoledo.edu/theses-dissertations Recommended Citation Izaac, Aude, "Crystallographic studies of the E. coli DNA replication restart primosome" (2005). Theses and Dissertations. Paper 1435.

This Thesis is brought to you for free and open access by The University of Toledo Digital Repository. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of The University of Toledo Digital Repository. For more information, please see the repository's About page.

A Thesis entitled

Crystallographic studies of the E. coli DNA replication restart primosome

by Aude E. Izaac

Submitted as partial fulfillment of the requirements for the Master of Science in Chemistry

Adviser: Dr. Timothy C. Mueser

Graduate School

The University of Toledo May 2005

Copyright © 2005 This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author.

An Abstract of

Crystallographic studies of the E. coli DNA replication restart primosome

Aude Izaac

Submitted as partial fulfillment of the requirements for the Master of Science in Chemistry

The University of Toledo May 2005

Understanding the mechanisms of DNA replication has been one of the main research topics in biochemistry for decades, due to the tremendous potential applications involved, such as medical treatments and better knowledge of the biological evolution of organisms. In this perspective, macromolecular X-ray crystallography provides a very powerful tool. By solving the 3-dimensional structure of the proteins involved in replication mechanisms, one can get a good insight at complicated biological systems at the atomic level. It is also important to study the protein-protein and protein-substrate interactions when they form biological complexes. iii

This work focused on studying several proteins which are associated with the assembly of the replication restart primosome in Escherichia coli, during the repair of damaged DNA. The crystallization of PriA, the most important of the restart primosome proteins, was investigated. Two truncations of PriA (the PriA N and PriA NI domains) were expressed, purified and characterized to be prepared for crystallization. Three additional replication restart proteins: PriB, PriC and DnaT, were expressed and purified. The solubility of each protein was optimized by identifying the best solution conditions. The proteins were characterized, alone and in complexes, using several biophysical techniques. DnaT was successfully crystallized and screened for X-ray diffraction. Finally, using ten standard test proteins, the correlation between the optimization of protein solubility (and solution conditions) and the success of crystal screening was studied.

iv

ACKNOWLEDGEMENTS

First, I would like to thank my adviser Dr. Mueser for welcoming me in his lab and for all his expertise, knowledge and guidance in this project. Thank you so much, Sir, for patiently answering the million questions I had and teaching me so much! I would also like to thank the members of my committee, Dr. Viola and Dr. Funk for their helpful suggestions and discussions. I would like to thank Dr. Hiroshi Nakai for his collaboration and for providing protein constructs. Many thanks to the UT Department of Chemistry and their staff, especially Leif Hanson for all his help with the high-throughput crystallization facility and the X-ray diffractometer. I am also incredibly grateful to my family and friends back in France for their love and support during all the time I was away and also for their wonderful encouragements. I cannot forget to thank my friends and the “French Connection” here in Toledo for the great times, the parties, the food and the humor. Finally, I would like to thank all my labmates, past and present, with who I spent so much time in the last three years, for making my life and my work so enjoyable every day: Steve, Vinu, Juliette, Laurence, Jennifer, Brandon, Anne, Deepa, Pooja, Wilawan and Kelly. I cannot express how glad I am to have met you all. I will never forget the good laughs (veg’ and non-veg’), the lunches, the Cedar Point trips, the movings, there’s too much to say! I had such an amazing time and I hope you now have a little bit of France in your hearts. I would never have made it through without you guys, thank you so much! v

TABLE OF CONTENTS

ABSTRACT ......................................................................................................... iii ACKNOWLEDGEMENTS .................................................................................... v TABLE OF CONTENTS ...................................................................................... vi LIST OF TABLES ................................................................................................ x LIST OF FIGURES............................................................................................. xii LIST OF ABBREVIATIONS ............................................................................... xv PREFACE ....................................................................................................... xviii CHAPTER 1 Introduction ................................................................................... 1 1.1.

Research goals ...................................................................................... 1

1.2.

Overview of the research process .......................................................... 2

1.3.

DNA replication restart mechanisms in E. coli........................................ 2

1.3.1.

Why do cells need replication restart mechanisms?.............................................. 2

1.3.2.

The two independent pathways in E. coli .............................................................. 4

1.3.3.

The replication restart primosome ......................................................................... 6

CHAPTER 2 Methodology ................................................................................ 11 2.1.

Extraction studies of expressed proteins.............................................. 11

2.2.

Protein expression ............................................................................... 12

2.2.1.

Large scale preparation of bacteria cells ............................................................. 12

2.2.2.

SDS-PAGE gel electrophoresis ........................................................................... 13

2.3.

Cell lysis ............................................................................................... 13

2.3.1.

Low salt cell lysis.................................................................................................. 14

2.3.2.

Cell lysis with ammonium sulfate precipitation .................................................... 14

2.4.

Protein purification ............................................................................... 17

2.4.1.

Principle of High Performance Liquid Chromatography (HPLC) ......................... 17

2.4.2.

Chemistry of the different chromatography media............................................... 17

2.4.3.

Design of purification schemes ............................................................................ 19

2.4.4.

Experimental procedures ..................................................................................... 20

2.5.

Preparation of protein solutions............................................................ 22

2.5.1.

Dialysis................................................................................................................. 22

vi

2.5.2.

2.6.

Concentration....................................................................................................... 22

Solubility screen ................................................................................... 23

2.6.1.

Standard solubility screen.................................................................................... 23

2.6.2.

Optimized buffer and maximum solubility ............................................................ 24

2.7.

Crystal screens .................................................................................... 25

2.8.

Optimization of crystal growth conditions ............................................. 27

2.9.

Crystal manipulation for X-ray diffraction.............................................. 29

2.10. Dynamic Light Scattering (DLS) ........................................................... 30 2.11. Differential Scanning Calorimetry (DSC) .............................................. 32 2.12. Retardation gels ................................................................................... 32 2.13. Test for endogenous nuclease activity ................................................. 33 2.14. DNA purification and annealing............................................................ 34 CHAPTER 3 E. coli PriA .................................................................................. 36 3.1.

Introduction .......................................................................................... 36

3.1.1.

Function of PriA in the primosome....................................................................... 36

3.1.2.

DNA binding ......................................................................................................... 37

3.1.3.

Characteristics of PriA ......................................................................................... 39

3.1.4.

Limited proteolysis ............................................................................................... 41

3.2.

Full-length PriA..................................................................................... 43

3.2.1.

Dialysis and concentration ................................................................................... 43

3.2.2.

Dynamic Light Scattering ..................................................................................... 43

3.2.3.

Crystallization....................................................................................................... 44

3.2.4.

Purification from endogenous nucleases............................................................. 46

3.2.5.

Co-crystallization with DNA.................................................................................. 49

3.3.

PriA N-terminal domain ........................................................................ 52

3.3.1.

Protein expression ............................................................................................... 52

3.3.2.

Cell lysis ............................................................................................................... 52

3.3.3.

Purification ........................................................................................................... 53

3.3.4.

Solubility screen ................................................................................................... 57

3.3.5.

Concentration and dialysis................................................................................... 58

3.3.6.

Dynamic Light Scattering ..................................................................................... 58

3.3.7.

Size-exclusion chromatography........................................................................... 59

3.3.8.

Crystal screens .................................................................................................... 60

3.3.9.

Test for endogenous nuclease activity ................................................................ 63

vii

3.3.10.

3.4.

Fluorescence experiment..................................................................................... 64

Strategies for studying the NI domain of PriA....................................... 65

3.4.1.

Rationale .............................................................................................................. 65

3.4.2.

Protein extraction studies..................................................................................... 67

3.5.

PriA NI domain with a C-terminal HisTag ............................................. 69

3.5.1.

Protein expression ............................................................................................... 69

3.5.2.

Cell lysis with ammonium sulfate precipitation .................................................... 69

3.5.3.

Purification ........................................................................................................... 70

3.5.4.

Dynamic Light Scattering ..................................................................................... 72

3.6.

PriA NI domain ..................................................................................... 72

3.6.1.

Protein expression ............................................................................................... 72

3.6.2.

Cell lysis with ammonium sulfate precipitation .................................................... 73

3.6.3.

HPLC purification ................................................................................................. 74

3.6.4.

Test for endogenous nuclease activity ................................................................ 79

3.7.

Discussion............................................................................................ 80

CHAPTER 4 E. coli PriB and PriC.................................................................... 85 4.1.

Introduction .......................................................................................... 85

4.1.1.

Role of PriB and PriC in the primosome .............................................................. 85

4.1.2.

Comparison of PriB and PriC............................................................................... 86

4.2.

PriB ...................................................................................................... 90

4.2.1.

Protein expression ............................................................................................... 90

4.2.2.

Cell lysis and ammonium sulfate precipitation..................................................... 91

4.2.3.

HPLC purification ................................................................................................. 92

4.2.4.

Solubility screen ................................................................................................... 97

4.2.5.

Dialysis................................................................................................................. 98

4.2.6.

Concentration....................................................................................................... 99

4.2.7.

Dynamic Light Scattering ................................................................................... 100

4.2.8.

Study of the protein-protein interactions involving PriB ..................................... 101

4.3.

PriC .................................................................................................... 106

4.3.1.

Protein expression ............................................................................................. 106

4.3.2.

Cell lysis and ammonium sulfate precipitation................................................... 107

4.3.3.

HPLC purification ............................................................................................... 108

4.3.4.

Solubility screen ................................................................................................. 111

4.3.5.

Protein preparation ............................................................................................ 111

4.3.6.

Secondary structure prediction .......................................................................... 112

viii

4.3.7.

4.4.

Protein-protein complexes ................................................................................. 113

Discussion.......................................................................................... 115

CHAPTER 5 E. coli DnaT................................................................................ 119 5.1.

Introduction ........................................................................................ 119

5.2.

Role of DnaT in the primosome.......................................................... 119

5.3.

Protein expression ............................................................................. 121

5.4.

Cell lysis and ammonium sulfate precipitation.................................... 121

5.5.

HPLC purification ............................................................................... 122

5.6.

Solubility screen ................................................................................. 124

5.7.

Protein preparation............................................................................. 125

5.8.

Dynamic Light Scattering ................................................................... 125

5.9.

Differential Scanning Calorimetry ....................................................... 128

5.10. Crystallization..................................................................................... 132 5.10.1.

Crystal screens .................................................................................................. 132

5.10.2.

Optimization of crystal growth conditions .......................................................... 134

5.10.3.

Crystal manipulation and X-ray diffraction screening ........................................ 137

5.11. Discussion.......................................................................................... 138 CHAPTER 6 A preliminary screen used to optimize protein solubility and improve crystallization results ...................................................................... 140 6.1.

Introduction ........................................................................................ 140

6.2.

Experimental methods........................................................................ 142

6.2.1.

Preparation of the protein solutions ................................................................... 143

6.2.2.

Solubility screen ................................................................................................. 144

6.2.3.

Optimized buffer and maximum solubility .......................................................... 145

6.2.4.

Design of the Additive/Precipitating Agent crystallization screen ...................... 145

6.2.5.

Crystal screens .................................................................................................. 150

6.3.

Results ............................................................................................... 153

6.4.

Discussion.......................................................................................... 168

BIBLIOGRAPHY .............................................................................................. 171 APPENDICES .................................................................................................. 175

ix

LIST OF TABLES

Table 2x1: Antibiotics used..............................................................................................12 Table 2x2: Buffers for the cell lysis followed by ammonium sulfate precipitation ............16 Table 2x3: HPLC programs .............................................................................................22 Table 2x4: Standard solubility screen..............................................................................23 Table 2x5: Commonly used crystal screens ....................................................................26 Table 2x6: DLS results and analysis ...............................................................................31 Table 3x1: PriA structural and functional motifs ..............................................................40 Table 3x2: Characteristics of PriA truncations.................................................................43 Table 3x3: DLS results for PriA .......................................................................................44 Table 3x4: Crystal screens of PriA ..................................................................................46 Table 3x5: HPLC purification of PriA N - Buffers .............................................................54 Table 3x6: DLS results for PriA N....................................................................................58 Table 3x7: Crystal screens of PriA N...............................................................................61 Table 3x8: DLS results for PriA NI QC ............................................................................72 Table 3x9: HPLC purification of PriA N - Buffers .............................................................75 Table 4x1: Characteristics of PriB and PriC ....................................................................87 Table 4x2: HPLC purification of PriB - Buffers.................................................................93 Table 4x3: DLS results for PriB .....................................................................................100 Table 4x4: HPLC purification of PriC - Buffers ..............................................................108 Table 4x5: Molar extinction coefficients.........................................................................115 Table 5x1: DLS results for DnaT ...................................................................................127 Table 5x2: Crystal screens of DnaT ..............................................................................132 Table 6x1: Test proteins ................................................................................................143

x

Table 6x2: The 17-condition solubility screen................................................................145 Table 6x3: Additive/Precipitating Agent screen - List of conditions ...............................148 Table 6x4: Automated preparation of crystal trays ........................................................150 Table 6x5: Optimized buffers and maximum solubility ..................................................159 Table 6x6: Crystallization results ...................................................................................161

xi

LIST OF FIGURES

Figure 1x1: Overview of the research process ..................................................................3 Figure 1x2: DnaA- and PriA-directed replication pathways ...............................................6 Figure 1x3: Replication restart at D-loops .........................................................................9 Figure 2x1: Flow-chart of the two cell lysis protocols ......................................................15 Figure 2x2: HPLC purification schemes ..........................................................................21 Figure 3x1: DNA substrates for PriA................................................................................39 Figure 3x2: PriA domains and truncations.......................................................................42 Figure 3x3: PriA crystal screen hits .................................................................................45 Figure 3x4: HPLC purification of PriA - POROS PE ........................................................47 Figure 3x5: Nuclease activity test for PriA .......................................................................48 Figure 3x6: Flap-DNA substrate used for co-crystallization with PriA .............................49 Figure 3x7: HPLC purification of the Flap-DNA substrate ...............................................50 Figure 3x8: SDS-PAGE gel of the expression and cell lysis of PriA N ............................53 Figure 3x9: HPLC purification of PriA N - Chromatograms and SDS-PAGE gels ...........55 Figure 3x10: Solubility screen of PriA N at room temperature.........................................57 Figure 3x11: PriA N - Size exclusion ...............................................................................60 Figure 3x12: PriA N crystal screen hits............................................................................62 Figure 3x13: Nuclease activity test for PriA N .................................................................63 Figure 3x14: DNA retardation gel ....................................................................................64 Figure 3x15: Strategies used for the PriA NI domain ......................................................66 Figure 3x16: SDS-PAGE gel of the solubility studies of PriA NI......................................68 Figure 3x17: SDS-PAGE gels of the expression and cell lysis of PriA NI QC .................70 Figure 3x18: HPLC purification of PriA NI QC - Chromatogram and SDS-PAGE gel .....71

xii

Figure 3x19: SDS-PAGE gels of the expression and cell lysis of PriA NI .......................74 Figure 3x20: Purification scheme for PriA NI...................................................................75 Figure 3x21: HPLC purification of PriA NI - Chromatograms and SDS-PAGE gels ........76 Figure 3x22: Nuclease activity test for PriA NI ................................................................80 Figure 4x1: Pathways in PriA-directed replication restart mechanisms...........................87 Figure 4x2: 3-dimensional structure of the E. coli PriB (PDB code 1WOC) ....................89 Figure 4x3: SDS-PAGE gels of the expression and cell lysis of PriB ..............................91 Figure 4x4: HPLC purification of PriB - Chromatograms and SDS-PAGE gels...............94 Figure 4x5: Solubility screen of PriB at room temperature ..............................................97 Figure 4x6: Solubility screen of PriB at 4 °C...................................................................98 Figure 4x7: Diagram of a native gel...............................................................................102 Figure 4x8: Native gel electrophoresis ..........................................................................103 Figure 4x9: SDS-PAGE gel of the expression and cell lysis of PriC..............................107 Figure 4x10: HPLC purification of PriC - Chromatograms and SDS-PAGE gels...........109 Figure 4x11: Solubility screen of PriC at room temperature ..........................................111 Figure 5x1: Formation of the preprimosome on φX174-type DNA.................................120 Figure 5x2: SDS-PAGE gel of the expression and cell lysis of DnaT............................122 Figure 5x3: HPLC purification of DnaT - Chromatogram and SDS-PAGE gel ..............123 Figure 5x4: Solubility screen of DnaT at room temperature ..........................................124 Figure 5x5: DSC curves of DnaT...................................................................................129 Figure 5x6: DnaT crystal screen hits .............................................................................133 Figure 5x7: DnaT crystals..............................................................................................135 Figure 6x1: Additive/Precipitating Agent screen - 96 condition block ............................147 Figure 6x2: The Instrumentation Center of the OMCC ..................................................151 Figure 6x3: Solubility screen results ..............................................................................154 Figure 6x4: Maximum solubility .....................................................................................160

xiii

Figure 6x5: Crystal pictures in standard vs. optimized conditions .................................162 Figure 6x6: Catalase......................................................................................................167 Figure 6x7: Xylanase at high concentration...................................................................168

xiv

LIST OF ABBREVIATIONS

% (v/v) ………….. % volume/volume % (w/v) ………..... % weight/volume % poly ………...... % Polydispersity 0 (or 3) hr …….... 0-(or 3-)hour sample 1-(2- or 3-)D ....... 1-(2- or 3-)Dimensional aa ……………….. Amino Acid AEBSF …………. 4-(2-Aminoethyl)-benzenesulfonylfluoride Amp …………….. Ampicillin ATP …………….. Adenosine Triphosphate Ax ……………….. Absorption, at x nm (for a 1 mg/mL solution) Bis-tris ………….. 2,2-Bis(hydroxymethyl)-2,2’,2’’-nitrilotriethanol BME …………….. β-Mercaptoethanol bp ……………….. base-pairs BSA …………….. Bovine Serum Albumin CAM …………….. Chloramphenicol CAPS ………....... 3-Cyclohexamino-1-propanesulfonic acid CCD …………….. Charge-Coupled Device Cond. ………….... Conductivity DLS ……………... Dynamic Light Scattering DSC …………….. Differential Scanning Calorimetry DTT ……………... Dithiothreitol xv

E. coli …………… Escherichia coli EDTA ………….... Ethylenediaminetetraacetic acid FT ……………….. Flow-Through fraction GLS …………….. Gel Loading Solution HEPES ……….... 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HPLC ………….... High-Performance Liquid Chromatography IPTG …………..... Isopropyl-β-D-thiogalactopyranoside Kan ……………… Kanamycin L ……………….... Load fraction LB ……………….. Luria Broth MES …………….. 4-morpholineethanesulfonic acid monohydrate MPD …………….. 2-Methyl-2,4-Pentanediol MW ……………… Molecular Weight MWCO …………. Molecular Weight Cut-Off nt ......................... nucleotide OB-fold …………. oligosaccharide/oligonucleotide-binding fold ODx ……………... Optical Density, at x nm OMCC ………….. Ohio Macromolecular Crystallography Consortium PAS …………….. Primosome Assembly Site PDB …………….. Protein Data Bank PEG …………….. Poly(ethylene glycol) PEI ……………… Poly(ethylenimine) pI ……………….. Isoelectric point

xvi

PIPES ………….. Piperazinebis(ethanesulfonic) acid Rh ………………. Hydrodynamic Radius RT ………………. Room Temperature SDS-PAGE …….. Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis SOS …………….. SOS error value SSB …………….. Single-Stranded DNA-Binding protein ss(or ds)DNA ….. single-stranded (or double-stranded) deoxyribonucleic acid TAPS …………… N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic acid TBE …………….. Tris, Boric acid, EDTA TCEP …………… Tris(2–carboxyethyl)phosphine TE ………………. Tris, EDTA TEDG …………... Tris, EDTA, DTT, Glycerol Tm …………….... Melting temperature Tris ……………… Tris(hydroxymethyl)aminomethane UV-Vis ………….. Ultraviolet-Visible V ………………… Void fraction ∆H …………….... Heat change per mole ∆Hv ……………... Heat change per unfolding unit εx ………………… Molar extinction coefficient, at x nm

xvii

PREFACE

Through my master’s thesis I wanted to gain knowledge in the fields of biochemistry and protein crystallography. This gave me the opportunity to become familiar with the way academic research is conducted in general and to learn how to design and conduct projects on my own. I now have a good experience of biochemistry and macromolecular crystallography laboratory techniques and a broad knowledge of protein chemistry. I was also very fortunate to become one of the advanced users of the high-throughput crystallization facility of the Ohio Macromolecular Crystallography Consortium. I then provided training of these instruments to other people. The way the projects are designed in our laboratory enables everyone to get involved at more or less any stage of the research process and to learn a very broad array of techniques. This includes various areas of expertise from molecular biology all the way to structure solution.

xviii

CHAPTER 1 Introduction

1.1. Research goals Dr. Mueser’s macromolecular crystallography group is interested in studying DNA replication and repair enzymes from different types of organisms including several thermophilic Archaea and two Prokaryotic organisms: E. coli and the Bacteriophage T4. The work presented here focuses on the E. coli DNA replication restart proteins, for which little is known from the literature in a structural point of view3. The limited amount of studies of these mechanisms is partly due to the fact that E. coli is a more distant model system to humans than the other organisms but also because the findings rely on indirect measurements. The E. coli replication restart primosome functions as a rescue pathway on stalled DNA replication forks and at recombination intermediates4-6. This project consisted in studying four of the seven E. coli primosomal proteins: PriA, PriB, PriC and DnaT. To date, there is no known equivalent to this type of replication repair and restart mechanisms in Eukaryotes. So if they exist, the functional equivalents to these proteins are yet to be discovered4,7. As a result, the approach to this research was more fundamental. It aimed at gathering

1

2 general information on the system and hopefully at solving the structure of some of these proteins.

1.2. Overview of the research process Most of the molecular cloning had been completed on the E. coli primosomal proteins and this work started at the protein expression level. After that, it included all other aspects of the research, up to the X-ray diffraction experiments (Figure 1x1, steps in blue on the chart). Although each part of the project was concluded at a different level, this research provides enough preliminary information to serve as a starting point for further structural analysis of the primosome.

1.3. DNA replication restart mechanisms in E. coli 1.3.1.

Why do cells need replication restart mechanisms?

DNA is the universal carrier of genetic information in all living organisms and is an extremely stable macromolecule that is transmitted from one generation to the next with very high fidelity. However within the cell, the DNA is subject to damage from various sources. DNA damage is usually classified into two categories: spontaneous and environmental8. The first refers to any type of alteration of the DNA as a result of a chemical reaction such as nucleotide degradation, rearrangement, oxidation, mismatch or radical-induced lesions.

3 Figure 1x1: Overview of the research process

PCR

Molecular cloning

Transformation into an expression host

Small scale expression

Sequencing

Extraction studies

Large scale expression

HPLC purification

Solubility screen and buffer optimization

Dialysis and concentration

Nuclease activity test

Crystal screens

Bio-physical studies

Optimization of crystal growth conditions

X-ray data collection

Structure solution

4 The latter is due to environmental stress caused by physical damaging agents like ionizing and UV radiations. In order to survive, it is essential that cells maintain genetic stability during DNA replication mechanisms. To do so, living organisms have evolved a number of cell response mechanisms, in order to repair DNA breaks and damaged replication forks. Even with these complex machineries in place, DNA replication is not perfect. In prokaryotes, the replication fork proceeds at the incredible rate of at least 1000 nt/sec (almost 20 times faster than in eukaryotes). Consequently, the probability for the whole genome to be replicated without errors is very low and it has been observed in vitro that most replication forks collapse when encountering DNA damage before they can complete the synthesis of the entire chromosome4,9. The kind of DNA damage triggers different SOS responses in the cell, dictating which protein system will be involved in replication restart. Full recovery of an arrested fork after UV irradiation was measured to take between 30 and 45 minutes.

1.3.2.

The two independent pathways in E. coli

The initiation of DNA synthesis in E. coli takes place in two different and complementary ways. Normal DNA replication is initiated at the origin of replication (OriC), which is a 245 bp region containing four identical DNA sequences recognized specifically by a protein called DnaA. Once bound to the DNA, DnaA promotes the assembly of the primosome by recruiting the DnaB/DnaC protein complex. DnaC is released; simultaneously the DnaB

5 helicase is loaded onto the DNA and unwinds the duplex through ATP hydrolysis. The resulting single-strands are coated with SSB proteins and the DnaG primase is loaded on the DNA through a transient interaction with DnaB to synthesize short RNA primers complementary to the parent DNA strands. The primers are then extended in the 5’  3’ direction by the DNA polymerase III holoenzyme (pol III HE)8,10. The other situation in which the initiation of DNA synthesis is essential is when the replication fork stalls and/or falls apart, after running into a site of DNA damage. In that case, a pre-existing replication complex needs to be reestablished on the DNA. This process, called DNA replication restart, was first introduced by K. Marians and co-workers5 and involves a complex of seven proteins referred to as the φX174-type primosomal proteins11 or replication restart primosome (see also § 1.3.3). The primosome is composed of PriA, PriB, PriC, DnaT, DnaC, DnaB and DnaG. The role of this large protein complex is the same as that of DnaA in the OriC mechanism: load the DnaB helicase onto the laggingstrand DNA for subsequent synthesis of RNA primers by the DnaG primase. This second pathway is directed by PriA, which recognizes a specific sequence of DNA called the Primosome Assembly Site (PAS) and then recruits the other components of the primosome, just like DnaA does at the OriC site5. The replication restart primosome was discovered about 30 years ago12 but only recently did the importance of such mechanisms become apparent. It was found in the last decade that PriA-directed replication restart occurs not only at damaged replication forks, but also at recombination intermediates and during

6 the replication by transposition of the Bacteriophage Mu6,13,14. Figure 1x2 shows a simplified comparison of the two pathways (the PAS site is represented as a duplex for simplicity but the actual PAS is a hairpin).

Figure 1x2: DnaA- and PriA-directed replication pathways

OriC

DnaB

DnaA SSB

PriB, PriC, DnaT DnaB

PAS

PriA SSB

1.3.3.

The replication restart primosome

The E. coli primosome was discovered at the same time as the DNA polymerase III holoenzyme during the in vitro reconstitution of bacteriophage φX174 DNA replication and was first reported in 197412. The bacteriophage φX174 is an icosahedral phage that infects certain enteric bacteria like E. coli10. It has been involved historically in several landmark experiments: its 5386 bp genome was the first DNA molecule to be purified to homogeneity15 and also the first genome to be completely sequenced16. Biochemical studies in E. coli have

7 shown that the primosome, a large complex of enzymes that travels along the lagging-strand template, can both unwind the duplex DNA and synthesize RNA primers for the initiation of Okazaki-fragment synthesis, was required for stable DNA replication10,17. More recent terminology refers to the proteins of this complex as the φX174-type primosomal proteins11. There are seven of them: PriA, PriB, PriC, DnaT, DnaC, DnaB and DnaG. The detailed composition of the primosome will be discussed below. It was also found that strains carrying PriA null mutations had severely reduced viability and chronic induction of SOS response, as well as high sensitivity to UV irradiation in particular18-20. This leads to the belief that PriA was the protein that directed the assembly of the entire replication restart primosome. As mentioned above, the PriA-directed mechanism complements the normal initiation of DNA replication by DnaA at the OriC site. The PriA-directed primosome was found to operate at different sites where DNA replication had been interrupted, especially at stalled replication forks that had encountered DNA damage and at unresolved recombination intermediates after the establishment of a D-loop DNA structure by the RecA family of recombination proteins21,22. The primosome was found to assemble on many different DNA substrates, but amazingly the binding was always very specific23. It was later found that the specificity resided in the PriA protein only, which is able to recognize specific sequences on various DNA substrates (see § 3.1.2), among which the φX174 PAS site, a stable hairpin structure17,24. PriA then recruits the other primosomal proteins in an ordered manner. The assembly of a complete primosome in a 5-step sequential mechanism was proposed3. First, PriA

8 recognizes and binds the DNA substrate. Then, PriB or PriC bind on the PriAyDNA complex and stabilize it significantly, probably by triggering a conformational change on PriA. Next, DnaT binds to the complex, most likely by a protein-protein interaction with PriA. The fourth step is the addition of the DnaB helicase onto the DNA complex through the ATP-dependent dissociation of the DnaByDnaC complex and release of the free DnaC. This step is the same as in the OriC mechanism. The DNAyPriAyPriB/PriCyDnaTyDnaB complex formed so far is called the preprimosome. Finally, the DnaG primase is loaded onto the lagging-strand of the DNA by a transient interaction with DnaB, which completes the formation of the entire primosome. The stoichiometry of the primosome has been measured25. It is composed of one or two monomers of PriA (this has not been clearly defined yet), two PriB dimers, one PriC monomer, most likely one monomer of DnaT, one DnaB hexamer and one DnaG monomer. In this complex, PriA provides DNA-binding specificity as well as ATPase and 3’5’ helicase activity, whereas DnaB provides 5’3’ helicase activity. This unique two helicase feature allows the primosome to translocate bidirectionally on the DNA26. The mechanisms involved at D-loops are now better understood and it was proposed that PriA binds to the 3-stranded junction27, directs the assembly of the primosome and then starts unwinding the duplex DNA to provide sufficient ss-DNA space for DnaB. DnaB then unwinds the lagging-strand template in the opposite direction, allowing a stable replisome to be re-established. This mechanism is shown on Figure 1x3. The primosome was found to be very stable, since it remained unaltered after priming and also after DNA replication25. Thus,

9 the same primosomal complex can be “recycled” to be used at several sites of DNA damage.

Figure 1x3: Replication restart at D-loops

Preprimosome at the 3-stranded junction

PriA PriB or PriC DnaT DnaB DnaG SSB RNA primer

Since the discovery of the replication restart primosome, many genetic and biological studies have been conducted to better understand the role of the different proteins involved, especially PriA. However, there is an evident lack of structural information about the replication restart complex. In this relatively new field of study, many crucial points remain unclear (such as the exact function of PriC in the primosome or the nature of the interactions between proteins). In many cases, suppositions were made that are contradictory and need to be

10 proved by further biochemical studies. In order to resolve some of the arguments, crystallization of the primosomal proteins on their own or in binary complexes was attempted in this study. X-ray crystallography is indeed the method of choice if we want to answer some of the main questions raised over the years: how do these proteins interact with each other, are there conformational changes involved, how does PriA specifically recognize its DNA substrate, etc… In order to study the primosomal proteins structurally, much work had to be done. This thesis presents the preparation of PriA and two of its truncations (PriA N and PriA NI), PriB, PriC and DnaT in large amounts in a pure form. This lead to the development of improved purification protocols, as well as the optimization of the solubility of the proteins. Different biophysical analyses were also performed to better characterize the proteins in solution. Subsequent crystallization studies were conducted on PriA, PriA N and DnaT. Finally, some preliminary findings are reported on several binary protein complexes that are part of the primosome. The last chapter is a study of the correlation between the optimization of protein solubility and the success of crystallization experiments.

CHAPTER 2 Methodology

2.1. Extraction studies of expressed proteins This procedure was used before making large amounts of any expressed protein, to determine if it was produced in a soluble form. After the protein was expressed on a small scale in different hosts, the cells were centrifuged and the pellets frozen at -20 °C. For extraction studies, the cell pellets were thawed out in ice and weighed. A volume of 10 mL of lysis buffer (100 mM Tris-HCl pH 8.0, 150 mM NH4Cl, 10 mM EDTA, 5 mM DTT, 10% sucrose, 0.3% PEI and 1 mM AEBSF) was then added per gram of cells. A small amount of lysozyme was added to the suspension, which was stirred at room temperature for 20 min. It was then placed back on ice and sonicated for 2 min. Next, 50 µL of the solution was pipetted into each of three centrifuge tubes. The first tube served as a control. To the second and third tubes, 50 µL of 1M NaCl and 50 µL of Bug Buster were added respectively. The samples were mixed and incubated in ice for 10 min. They were then centrifuged at 10000xg for 2 min and the lysate was separated from the pellet. For SDS-PAGE analysis, 2X SDS was added to each sample and the gel was run as described in § 2.2.2.

11

12

2.2. Protein expression 2.2.1. Large scale preparation of bacteria cells For crystallization experiments, large amounts of pure protein are required. Therefore, the proteins were overexpressed in bacteria cells in large scale experiments as follows. Using a sterile wooden stick, a small amount of frozen cells was taken from the glycerol stock of the protein (stored at -80 °C) and transferred into each of three small Erlenmeyer flasks containing 100 mL of 25 g/L LB medium and 1 mM of the required antibiotics (Table 2x1). The cultures were grown overnight at 37 °C in a New Brunswick Scientific Innova 4000 Incubator Shaker. In the morning, 50 mL of solution was transferred into each of six 2 L Erlenmeyer flasks, containing 1 L of 25 g/L LB and 1 mM of the required antibiotics. Once again, the cells were grown at 37 °C in the shaker. They were induced with 1 mM IPTG when the OD600 of the solution had reached 0.4 to 0.6 and then left to grow for another 3 hr. The cells were harvested by centrifugation at 5000xg using a Beckman Coulter™ TJ-25 centrifuge and the supernatant was discarded. The cell pellet was recovered, weighed and stored at -20 °C until used.

Table 2x1: Antibiotics used Plasmid

Antibiotic

Cell line

pET 21a

pET 28

XL10

Rosetta

RIL

BL21(DE3) Gold

Amp

Kan

None

CAM

CAM

None

13

2.2.2. SDS-PAGE gel electrophoresis The success of protein expression (and many other experiments) was checked by SDS-PAGE gel electrophoresis. For each large scale preparation, seven 1mL samples were taken out of the solution for analysis. One sample was taken before induction with IPTG (called the “0 hr sample”) and six samples (one from each 2 L flask) were taken after 3 hr of growth (called the “3 hr samples”). Each sample was centrifuged at 20000xg for 5 min using a bench-top centrifuge 5417C (Eppendorf). The supernatant was discarded and the pellet was resuspended in 25 µL of BugBuster™ Protein Extraction Reagent (Novagen). Finally, 25 µL of 2X NuPage™ LDS Sample Buffer (Invitrogen) was added to each sample. The samples were boiled and loaded on a NuPage™ 4-12% BisTris gel. The gel was run in 1X NuPage™ MES SDS Running buffer, at 200 V for 35 min using an XCell SureLock™ system. It was then stained overnight in Commassie® G250 stain (Bio-Rad), rinsed with water, destained in a solution of 30% methanol, 10% glacial acetic acid and transferred in a gel-dry solution (30% methanol, 5% glycerol) before drying.

2.3. Cell lysis Once it had been established that protein overexpression occurred, the frozen bacterial cells were thawed out in ice and lysed open. There are several protocols for doing this. For the primosomal proteins, two were used: the low salt cell lysis protocol and the cell lysis protocol with subsequent ammonium sulfate

14 precipitation. In each case, a 25 µL sample was taken out at each branch-point in the procedure ( in Figure 2x1) for SDS-PAGE analysis, to ensure that the protein had been correctly extracted in the corresponding fraction.

2.3.1.

Low salt cell lysis

The lysis buffer used in this procedure was composed of 100 mM Tris-HCl pH 8.0, 150 mM NH4Cl, 10 mM EDTA, 5 mM DTT, 10% sucrose, 0.3% PEI and 1 mM AEBSF. A volume of 10 mL of lysis buffer was used per gram of cells to be lysed. The cells were kept in ice and manually mixed in the lysis buffer. A small amount of hen egg white lysozyme (Sigma) was added to the solution and the suspension was stirred at room temperature for 20 min. The cells were put back on ice and lysed open by sonication for a duration of 2 min using a Branson™ sonifier 250. The soluble material (lysate) was separated from the cell debris (pellet) by centrifugation at 4 °C at 20000xg for 30 min. A sample of the pellet and the lysate was then run on SDS-PAGE gel to verify that the protein of interest was soluble, that is, found in the lysate. Glycerol was added to the lysate to a final concentration of 15%. This crude material was flash-frozen on dry-ice and stored at -80 °C until HPLC purification.

2.3.2.

Cell lysis with ammonium sulfate precipitation

This procedure was based on the work of K. Marians28. The composition of the two lysis buffers used in this procedure is given in Table 2x2.

15 Figure 2x1: Flow-chart of the two cell lysis protocols Figure 2x1a: Low salt cell lysis

Thaw cells in ice

Add lysis buffer

Centrifuge

Discard pellet (Cell debris)

Add 0.3% PEI

Stir at RT with lysozyme

Sonicate in ice

Collect and freeze lysate

Figure 2x1b: Cell lysis followed by ammonium sulfate precipitation

Thaw cells in ice

Add lysis buffer

Centrifuge

Discard pellet (Cell debris)

Add 0.3% PEI

Stir at RT with lysozyme

Sonicate in ice

Add saturated (NH4)2SO4 dropwise

Collect lysate

Centrifuge

Resuspend in TEDG buffer

Collect pellet

Discard supernatant

16 For each gram of cells, 10 mL of lysis buffer A was used, following the above low salt lysis protocol until the sonication step. At that point, 0.3% PEI was added to the solution to precipitate nucleic acids and the suspension was stirred for 10 min. The cell pellet was separated by centrifugation at 4 °C and 20000xg for 30 min. The supernatant was carefully decanted and kept on ice. The protein was then precipitated on ice by the dropwise addition of a saturated solution of (NH4)2SO41 to a final concentration of 45%. The solution was stirred for 10 min and the precipitate collected by centrifugation at 20000xg for 10 min. Finally, the protein was gently resuspended in TEDG buffer until a clear solution was obtained. For this, a volume of TEDG buffer between 5% and 15% of the total volume of lysate was required. The solution was then frozen in dry ice and stored at -80 ˚C.

Table 2x2: Buffers for the cell lysis followed by ammonium sulfate precipitation

1

Lysis buffer A

Lysis buffer B (TEDG)

50 mM Tris-HCl pH 7.5

50 mM Tris-HCl pH 7.5

150 mM KCl

-

20 mM EDTA

1 mM EDTA

10 mM DTT

5 mM DTT

10% sucrose

15% glycerol

~ 3.5 M, pH adjusted to 7.5 using 6 M NH4OH

17

2.4. Protein purification 2.4.1. Principle of High Performance Liquid Chromatography (HPLC) HPLC refers to a purification technique where a chromatography medium, the stationary phase, is packed into a column of specific height and diameter and through which a liquid, the mobile phase, is pumped at high pressure. The sample to be purified is injected on the column and eluted with running buffers. Proteins are separated from other species depending on their velocity and/or affinity for the stationary phase. The different types of chromatography medium used are presented in section 2.4.2. All the proteins studied in this project were purified inside a 4 °C cabinet using a BioLogic DuoFlow™ HPLC system (Bio-Rad) controled by the BioLogic DuoFlow™ software version 5.0. For each type of HPLC column used, a specific program was written to separate the protein of choice from the contaminants (see § 2.4.4). The program was run and the fractions were collected into test tubes using a fraction collector. The conductivity, the OD260 and the OD280 of the solution were monitored in real-time throughout the run to visualize the protein elution peak.

2.4.2. Chemistry of the different chromatography media Four types of chromatography techniques were used to purify the different

primosomal

proteins:

ion-exchange,

size-exclusion,

interaction and immobilized metal ion affinity chromatography.

hydrophobic

18 There are two types of ion-exchange chromatography media: cation-exchange and anion-exchange. A cation-exchange stationary phase is coated with negatively charged groups; whereas an anion-exchange stationary phase is coated with positively charged groups. The charged amino acids on the surface of the protein interact with the groups on the resin and the protein binds to the column. The cation-exchange media used here were the low-resolution SP Sepharose (Amersham Biosciences 17-5073-01) and the high-resolution POROS 20 HS (Applied Biosystems). SP Sepharose has a cross-linked agarosedextran matrix coated with sulfopropyl groups (-OCH2CH2CH2SO3-). POROS HS is a perfusion medium that can tolerate much higher pressure limits, and has a poly(styrene-divinylbenzene)

polymer

matrix

(PS/DVB)

also

coated

with

sulfopropyl groups. The anion-exchange media were not used, since the pI of the proteins used in this study was basic. However, hydroxyapatite chromatography (HA) was used to separate the protein from nucleic acids. The CHT™ ceramic HA (Bio-Rad 157-0040) column is a resin composed of calcium phosphate groups (Ca5(PO4)3OH) that will compete very efficiently with the phosphate backbone of DNA and RNA29. The Superdex® 75 size-exclusion column (Amersham Biosciences 171044-10) was used to separate a protein from impurities that have a very different molecular weight. The 75 stands for 75 kDa and refers to the highest size of molecule that can be accurately separated by the column, based on the pore size of the beads of the stationary phase. Any larger macromolecule elutes in the flow-through fraction. The matrix is a macroporous gel of agarose and

19 dextran29. During the run, the proteins travel at different speeds through the pores, depending on their hydrodynamic radius. Therefore a standardization plot can be made from standard proteins that will be used to determine or verify the size of the protein studied. Hydrophobic interaction chromatography (HIC) was used to purify the DNA-binding proteins from contaminating endogenous nucleases. HIC is based on the separation of proteins by hydrophobic interactions between the hydrophobic, non-polar surface of the stationary phase and the mobile phase that consists of a high salt buffer (also to facilitate hydrophobic interactions). In this method, the nucleases will stick to the column, whereas the protein of interest will be found in the flow-through fraction. The medium used was POROS PE, which is composed of PS/DVB beads coated with phenyl ether hydrophobic groups29. Finally, Immobilized Metal Ion Affinity Chromatography (IMAC) was used to purify the proteins containing a His-tag tail. This was done on a Nickel-NTA column, where the resin is composed of nitrilotriacetic acid (NTA) bound to a Sepharose matrix and charged with Ni2+ ions29,

30

. NTA is a tetra-

dentate chelating group that will bind Ni2+ ions to specifically interact with the histidine residues on the protein. As a result, the purity achieved with this type of affinity chromatography is very high.

2.4.3. Design of purification schemes The purification scheme for each protein was designed theoretically and then refined upon experimentation, because the protein did not always

20 behave exactly as expected, or because a step in the process became useless. Generally, the protein would be run on a low-resolution ion-exchange column first. As a general rule, if the pI of the protein is below 7, the Q Sepharose anion-exchange column should be used and if the pI is above 7, the SP Sepharose cation-exchange column should be used. Next, DNA-binding proteins would be run on the HA column to remove contaminating nucleic acids. For DNAbinding proteins, if nucleases needed to be removed from the sample, the POROS PE column was run as well. The following step was the high-resolution ion-exchange column (either POROS HQ or POROS HS, again depending on the pI of the protein), to separate any impurity left at this point. If necessary, the protein could then be run on the Superdex 75 column to clean the solution even further. After each run, the purity of the fractions was checked by SDS-PAGE gel electrophoresis. This was also helpful in determining exactly which fractions to collect.

2.4.4. Experimental procedures All the programs for ion-exchange chromatography consisted of a linear salt gradient from 100% Buffer A (low salt) to 100% Buffer B (high salt). In the case of Nickel affinity chromatography, the program was a step gradient from the wash buffer to the elution buffer. The programs are summarized in Table 2x3. For each run, the column was first equilibrated with buffer A for about 5 column volumes (CV) and the baseline was set. The crude samples were loaded into the super-loop with a filter syringe and the somewhat clean samples

21 were directly injected on the column through the HPLC pump. The OD260 and the OD280 of the solution were monitored and the flow-through was collected. When the whole sample had been loaded, some buffer A was injected on the column in the same way to clean the lines and the corresponding program was run. At the end of the run, the column was rinsed with storage buffer and the lines were cleaned with deionized water.

Figure 2x2: HPLC purification schemes

SP Sepharose (or Q Sepharose)

POROS PE

Hydroxyapatite

POROS HS (or POROS HQ)

DNA-binding proteins Other proteins Optional step

Superdex

22 Table 2x3: HPLC programs Type of gradient Linear (A  B) then steady (B)

Buffers

Volume

Flow-rate

Fractions

A= low salt B= high salt

5 CV

3 mL/min

50 x 3 mL

HA

Linear (A  B) then steady (B)

A= low salt B= high salt

5 CV

3 mL /min

50 x 1.5 mL

POROS HS

Linear (A  B)

A= low salt B= high salt

5 CV

8 mL /min

50 x 4 mL

Superdex 75

Constant (A)

A= low salt

2 CV

0.5 mL/min

60 x 1 mL

Ni-NTA

Step (A  B)

A= low salt B= Imidazole

10 CV

1 mL/min

25 x 1 mL

SP Sepharose

2.5. Preparation of protein solutions 2.5.1. Dialysis Slide-A-Lyzer® dialysis cassettes were purchased from Pierce and utilized when the sample volume was small, ranging from 0.1 mL to 12 mL. All the proteins were dialyzed in 10000 MWCO dialysis cassettes except PriB, for which the 3500 MWCO model was preferred. When the volume was larger, SnakeSkin® Pleated Dialysis Tubing with the same pore size was used. All the dialysis experiments were done with medium-speed stirring.

2.5.2. Concentration Protein solutions of small volume were concentrated in 1.5 mL Microcon™ concentrators (Millipore, models YM-10 green and YM-3 yellow) and the final sample was recovered by inverting the concentrator over a centrifuge

23 tube and spinning. Larger samples were concentrated in 4 mL or 15 mL Amicon™ Ultra concentrators (Millipore) and recovered directly by pipetting the solution out of the well. The concentration of each sample was calculated by measuring the absorbance of the solution at 280 nm (protein samples) or at 260 nm (DNA samples) with an Agilent 8453 UV-Visible spectrophotometer (Agilent Technologies). Finally, the concentrated protein solution was run through a single-use Ultrafree®-MC centrifugal filter (Millipore, red 0.45 µm pore size).

2.6. Solubility screen 2.6.1. Standard solubility screen This method, developed in our lab31, is used to find the best solution conditions for each protein, which often leads to more success in the crystallization. The solubility of the protein is measured in sixteen solution conditions (Table 2x4) including six chloride salts to test different cations, six sodium salts to test different anions and four buffers to get a pH profile.

Table 2x4: Standard solubility screen Cationic salts (0.1 M)

Anionic salts (0.1 M)

Buffers (0.1 M)

1 - NH4Cl

7 - Na Formate

13 - Na MES pH 5.6

2 - NaCl

8 - Na Acetate

14 - Na PIPES pH 6.5

3 - KCl

9 - Na Cacodylate

15 - Na HEPES pH 7.5

4 - LiCl

10 - Na Sulfate

16 - Na TAPS pH 8.5

5 - MgCl2

11 - Na Phosphate

6 - CaCl2

12 - Na Citrate

17 - Supernatant (control)

24 First, a few mg of protein was dialyzed against deionized water. The precipitated protein solution was then aliquoted into 16 centrifuge tubes labeled for each condition and they were all centrifuged at 20000xg for 4 min. The supernatants of all the tubes were collected together to serve as a control experiment. Next, 25 µL of the corresponding salt or buffer (100 mM stock solution) was added to each tube and the pellet was resuspended by mixing thoroughly with a pipette. The tubes were incubated at room temperature for 20 min and centrifuged again at 20000xg for 4 minutes. The amount of protein that redissolved in each condition was measured using a BioRad protein assay. For each condition, 5 µL of protein sample was mixed with 995 µL of 1X BioRad Protein Assay reagent in a disposable cuvette. The samples were incubated for 5 min and the absorbance of each one was recorded at 595 nm by the UVVisible spectrophotometer. The results were plotted and compared on a relative scale.

2.6.2. Optimized buffer and maximum solubility The buffer and salt that gave the highest solubility values were deduced directly from the screen. These conditions constitute the optimized buffer of the protein. In case several conditions gave similar results, the combinations could be tested accordingly and the final solubility might be even higher. After the optimized buffer was determined, the protein was dialyzed in it and concentrated. The maximum solubility was estimated by concentrating a test-sample until the protein started precipitating or until the concentration

25 leveled-off. We believe that, for most proteins, the optimal concentration range for crystal screens is between 30% and 50% of the maximum solubility.

2.7. Crystal screens Preliminary crystal screening was done manually at first. Single well Corning™ trays were used when only one protein solution had to be screened and 3 well round-bottom Greiner™ trays were used when two or three protein solutions were screened simultaneously (Hampton Research). Several 96condition commercial screens were used to setup the trays, as well as some screens prepared in-house (Table 2x5). To setup a tray, 100 µL of mother liquor solution was dispensed from a deep-block into each reservoir. This was done row by row using an 8-channels LTS pipette (Rainin). Next, 1 µL of mother liquor was transferred from the reservoir to the well in the same way. Using a 10 µL programmable pipette (Eppendorf Research Pro), 1 µL of protein solution was added to each well without mixing, to prevent air bubbles. In the end, the tray was sealed with clear tape using a manual roller and centrifuged for a few seconds for proper mixing of the drops and ensure that they fell at the bottom of each well. The tray was stored in a closed cabinet at room temperature or in a cold room at 4 °C. The results were manually recorded with a Nikon SMZ1500 microscope after 1, 3 and 5 days and then once a week for 4 to 6 weeks. Pictures of crystal hits were taken with a Nikon CoolPix™ 990 digital camera.

26 Table 2x5: Commonly used crystal screens Name of Screen

Type of Screen

Distributor

Crystal Screen I™

Sparse matrix

Hampton Research

Crystal Screen II™

Sparse Matrix

Hampton Research

Natrix™

Sparse Matrix

Hampton Research

Wizard I™

Random sparse matrix

Emerald BioStructures

Wizard II™

Random sparse matrix

Emerald BioStructures

Cryo I™

Sparse matrix

Emerald BioStructures

Cryo II™

Sparse Matrix

Emerald BioStructures

Ion Screen

Ion/pH matrix

Prepared in-house

Additive screen

Precipitating agent/Additive screen

Prepared in-house

Index™

Combination of Grid screen, Sparse-matrix and Incomplete factorial

Hampton Research

Membfac™

Sparse-matrix of detergents

Hampton Research

Crystal Screen Cryo™

Sparse-matrix

Hampton Research

When the instruments of the Ohio Macromolecular Crystallography Consortium became available, the process became completely automated. The 100 µL of mother-liquor was dispensed from the deep-block to the reservoir by the Cartesian robot (Genomic Solutions). Crystallization drops, this time containing 0.5 µL of well solution and 0.5 µL of protein solution, were setup using the Honeybee robot (Genomic Solutions). The trays were sealed using an automatic plate sealer (Brandel) and checked for crystal hits as described above.

27 The pictures were taken with the Rhombix Imager (Kendro) in bright and polarized light (see also § 6.2.5). Every time a hit was found, it was identified as a protein crystal (as opposed to a mineral salt crystal) by adding 0.3 µL of Izit Crystal Dye™ (Hampton Research) in the drop. The dye was left to react for about 30 min and the drop was visualized under the microscope. A protein crystal would turn dark blue when picking-up the dye and the background solution would be clear or light blue. On the other hand, a salt crystal would remain colorless.

2.8. Optimization of crystal growth conditions There are three steps in the preparation of diffraction quality crystals from the conditions found in the screens. The first one is to reproduce and then improve the crystals in home-made solutions, using the hanging-drop vapor diffusion technique in what is referred to as “expansion trays”. This is a very tedious process that is critical to the success of the entire crystallization procedure. The second step, which happens only when optimal growth conditions have been determined, is to make as many “production crystals” as possible for X-ray diffraction experiments. The last step is to find the proper cryoprotectant in order to freeze the crystals in liquid nitrogen, since most of the X-ray diffraction is done in cryo conditions to limit damage to the crystal and extend its lifetime. In our laboratory, all the expansion trays were prepared by the A/B gradient technique32, which allows the fast preparation of any kind of wide or

28 shallow gradient. Only solution A (the lower end of the gradient) and solution B (the higher end) need to be prepared and easy pipetting maps can be used, reducing drastically the amount of time needed and the waste generated. The format used for each expansion tray depends on the type of gradient made. A 4x6 expansion (by rows) allows testing four different and rather wide gradients in the same tray and requires the preparation of four sets of A and B solutions (numbered A1/B1, A2/B2, A3/B3, A4/B4). A 2x12 expansion (by 2 rows) allows testing two narrower gradients per tray and requires the preparation of two sets of A and B solutions (numbered A1/B1, A2/B2). Finally, a 1x24 expansion enables to prepare a fine-tuned gradient and requires making only one set of A and B solutions. With this method, any variation of the growth conditions can be refined individually and rapidly. For expansions, two types of trays were used: the 24well Costar™ tray (Hampton Research) with 18 mm cover-slides and the 24-well Nextal™ tray (Nextal). The trays were made by pipetting both A and B solutions in the wells throughout the gradient (total volume = 1 mL), using an EDP 10 mL programmable pipette (Rainin). The trays were then placed on a slow, continuous stirring device for a few minutes for mixing. When using Costar™ trays, the edge of each well was greased. The drops were then prepared on the cover-slides by pipetting 1 µL of protein solution and 1 µL of well solution. Using forceps, the cover-slides were turned upside-down and sealed on top of the wells. The trays were stored at room temperature and scored under the microscope after a few days. The same procedure was followed in the cold room using the Nextal trays.

29

2.9. Crystal manipulation for X-ray diffraction Once good quality crystals were grown, that were large enough to be used on the X-ray diffractometer, they had to be transferred in a cryoprotectant solution before freezing. However this step was not necessary when the cryoprotectant was part of the crystallization conditions. A 2X solution containing all the components of the protein dialysis buffer and of the crystallization drop was made and a few microliters were placed on a microscope slide. It was mixed with the same amount of 2X cryoprotectant solution. The crystal of choice was then looped out of the drop very carefully using a nylon crystal mounting loop of the same size (Hampton Research) and quickly transferred into the cryoprotectant solution. The drop was checked with the microscope to make sure that the crystal had not cracked or partially dissolved. The crystal was then picked-up inside the loop, flash-frozen in liquid nitrogen and stored in a Dewar until it was screened for X-ray diffraction. Then, the loop was mounted directly onto the goniometer head of the diffractometer in the cryostream and centered in the beam. A different crystal mounting method was used for room temperature diffraction experiments that consisted in preparing the crystal inside a glass capillary. The bottom of the capillary was attached to a pin support with melted wax and filled with a solution containing all the components of the protein dialysis buffer and of the crystallization drop from the bottom to somewhere in the middle. An air gap of a few mm was left and more solution was added until the capillary was full. The crystal was looped out of the drop as described previously and

30 dipped into the top part of the solution in the capillary. When it had sunk to the bottom, the liquid was removed very carefully, leaving only the smallest amount possible around the crystal. The solution was then added back into the capillary a few millimeters above the crystal until the capillary was full and the top end was sealed with wax. For X-ray diffraction, the mounting pin was inserted into the goniometer head and the crystal was centered in the beam. X-ray diffraction screening was done at 100 K at the APS synchrotron (Argonne National Laboratories) on Bio-CARS beamline 14-BM-C and on the FR-E Rigaku diffractometer of the OMCC at the University of Toledo.

2.10. Dynamic Light Scattering (DLS) For DLS experiments, protein samples of low concentration were made (typically 1 to 5 mg/mL). Each one was carefully filtered through a single-use Ultrafree®-MC centrifugal filter (Millipore, orange type, 0.1 µm pore size). A fresh sample of sterile ultra-pure water was also made every time using the Protein Solutions microfilter system with a Whatman Anodisc 13 filter (0.02 µm pore size). The measurements were made on a DynaPro-801 DLS instrument (Protein Solutions) and the data was acquired using the Dynamics™ software version 5.25.44. The temperature was set on the control panel and left to stabilize for 5 min. The cuvette used was a 16.45F-Q-3 quartz spectrophotometer cell, submicro Fluorimeter (45 µL, 3 mm, Starna Cells). It was cleaned several times with sterile water, refilled, capped and placed into the sample cell. A blank run was done to verify that the counts were below 20 (background level). If not, the

31 cuvette was taken out and cleaned again. When this was done, the water was removed from the cuvette, the protein sample was injected instead and the cuvette was placed into the sample cell. The run was done and was stopped after 30 to 33 measurements were taken on the “cumulants datalog” window. The results were analyzed in % mass from the “regularization histogram” and the “results summary” windows33, using the Regu M peak model and suppressing the first peak when unrealistic (radius Rh< 1 nm). The parameters obtained from a DLS experiment are detailed in Table 2x6.

Table 2x6: DLS results and analysis

Rh (nm)

MW (kDa)

33

% poly

Distribution

Baseline

SOS

Meaning

Average hydrodynamic radius of the molecule

Average molecular weight

Percent of polydispersity in solution (estimates the homogeneity)

Indicates the presence of aggregates or not

Measure of the distribution

Sum Of Squares (signal to noise ratio)

Wanted value for a protein

2-10 nm

Variable

15-40% (15-20% ideal for crystallization)

Monomodal

0.997-1.005

1.0-20.0

Analysis

< 1nm = unrealistic (suppress 1st peak)

Usually underestimated by a few kDa

< 20% = monodisperse solution

Monomodal = no aggegates

0.997-1.003= monomodal

1.0-5.0 = low noise

Bimodal / Multimodal = presence of aggregates

1.003-1.005= bimodal

5.0-20.0 = some noise

20-30% = low polydispersity > 30% = polydisperse solution

> 1.005 = multimodal

> 20.0 = high noise

32

2.11. Differential Scanning Calorimetry (DSC) DSC measures the heat released when a system is heated or cooled at a constant rate. The melting temperature (Tm) at which a protein unfolds can be calculated and it can be determined whether there is one or several unfolding units in the molecule. The instrument used was a Microcal VP-DSC™ (temperature range -10 °C to 130 °C). There are two 500 µL cells in the instrument, a reference cell, which contains only the buffer in which the protein is prepared and a sample cell, which contains the protein sample. The two cells are heated or cooled at a constant rate and the temperature difference between them is recorded. First, the reference cell was filled with the degassed buffer solution. Several buffer scans were done, heating from 10 °C to 90 °C and cooling down from 90 °C to 10 °C. Next, the sample cell was filled with the degassed protein solution. The protein scan was then completed by heating from 10 °C to 90 °C and the data was analyzed using the software package Origin™.

2.12. Retardation gels This experiment was used to compare the DNA binding ability of the PriA truncations to that of the full-length protein. A fork DNA substrate that contained a fluorescent label was used to visualize the band shifts on a DNA retardation gel. This type of sample is light sensitive and required working in the

33 dark. The reaction buffer used was composed of 25 mM Tris acetate, 65 mM NaOAc, 2 mM EDTA, 20 mM DTT and 200 µg/mL of BSA. The samples were prepared on ice and the different components were always added in the same order: DNA first, then reaction buffer, then protein. They were incubated on ice for 5 min and 2 µL of 50% glycerol was added to each one. The samples were loaded on a 6% DNA retardation gel and run in 1X TBE buffer at 4 °C, 100 V for 75 min. The gel was then photographed under fluorescent light34.

2.13. Test for endogenous nuclease activity The presence of endogenous nucleases in a protein sample is an important problem because experiments cannot be performed with DNA, since it would be degraded by the nucleases. The presence of nucleases in the solution was assessed by running different DNA samples on an agarose gel with and without the protein. The samples were prepared on ice by mixing 5 µL of DNA with 0.5 µL of protein (when required) and were incubated at 37 °C for an hour. After this, 4 µL of NN stop dye was added to each one and they were heated at 95 °C for 3 min. Then 2 µL of glycerol was added to the samples. They were loaded on a 1% agarose gel and run in 1X TAE running buffer at 4 °C, 90 V for 45 min using a Power Pack Basic system (Bio-Rad). The gel was then stained for 30 min with SYBR Gold™ nucleic acid gel stain (Molecular Probes).

34

2.14. DNA purification and annealing The DNA substrates used in crystallization experiments can be very expensive when ordered pre-purified. Therefore, we order them in standard desalted conditions and purify them in the laboratory. The single-stranded oligos were purchased from Integrated DNA Technologies (IDT) in a lyophilized form (1 µmol per synthesis). They were dissolved in 2 mL of sterile ultra-pure water upon arrival and their OD260 was measured to determine the amount received. The purification was done according to the protocol previously developed in our laboratory35 using a BioCAD/SPRINT Perfusion chromatography system and an anion-exchange POROS HQ column. After equilibration, the DNA sample was injected on the column at 8 mL/min and eluted by running a linear salt gradient from 100% Buffer A (10 mM NH4OH, 0.2 M NH4OAc) to 100% Buffer B (10 mM NH4OH, 3 M NH4OAc). The real-time absorbance at 296 nm was monitored during the run and the fractions of interest were collected and aliquoted in centrifuge tubes for concentration. For this, 1 µL of 100X TE buffer was added to each tube and they were transferred to a Speed-Vac concentrator system with a refrigerated condensation trap. The samples were spun overnight in medium or high heat. In the morning, an oily DNA sample remained in each tube that was dissolved in 50 µL of autoclaved ultra-pure water. The different fractions were pooled together and the OD260 of the final sample was measured to calculate its concentration. The purity of each oligo was checked by running a 20% TBE polyacrylamide gel (Invitrogen) at 4 °C, 180 V for 75 min. The bands were viewed under UV light.

35 The pure single-stranded DNA samples were either annealed fresh or frozen at -20 °C until used. For DNA annealing, 1 µL of 100X TE and 100 mM NaCl were added to each pure single-stranded oligo. The required oligos (3 for a flap DNA substrate) were mixed together in a centrifuge tube in the proper molar ratio. The tube was placed to float in a beaker of water just below boiling (95 °C) and the system was transferred into a Styrofoam box. It was left to cool to room temperature overnight and placed at 4 °C until cold. Successful annealing was checked by running a 20% TBE gel as described above.

CHAPTER 3 E. coli PriA

3.1. Introduction 3.1.1.

Function of PriA in the primosome

PriA is the most important protein of the replication restart primosome in prokaryotes. The PriA phenotype has a poor viability, low assimilation by homologous recombination and increased sensitivity to many DNA damaging agents, especially UV radiations18-20. This underlines the necessity for proper replication restart mechanisms in living cells. PriA is the only protein in the primosomal complex, besides DnaB and DnaG, to have the ability to bind DNA3. Thus it provides specific recognition of replication restart sites and triggers the assembly of the primosome. PriA provides many functions to the primosome: ATPase and 3’5’ helicase activities and DNA recognition26,36,37. As a result, the primosome can both unwind duplex DNA and translocate along ssDNA. In the event that PriA is unavailable, the helicase function can be carried out by other helicases involved in replication mechanisms, but the process is much slower and the replication restart is much less efficient14. The DNA recognition of PriA is broad but specific and the different substrates share common features. 36

37

3.1.2.

DNA binding

Many studies have been carried out over the years on the DNA-binding properties of PriA because the replication restart primosome was found to form at various places on different DNA substrates. PriA has the ability to bind many types of DNA structures: the φX174 PAS hairpin24, ssDNA, dsDNA with a 3’ overhang, D-loops, forks, flaps, Y-junctions and even the “chicken-foot”23,38-40. A trend emerged from these analyses: for any of these substrates, there is an absolute requirement for a small region of ssDNA or a small gap. If not, PriA is unable to bind the DNA. Further proof came when it was determined that even though PriA is involved in recombination mechanisms, it cannot not bind the Mu fork or Holliday junctions38, two structures that exclusively contain continuous duplex DNA. The DNA-binding properties of PriA are summarized in Figure 3x1. A recent study was the first to examine in detail the kinetics of PriA binding to ssDNA41. PriA strongly binds to oligos that are 12-nt long at least, the binding affinity increasing with the length of the substrate. One monomer of PriA occludes 20 ± 3 nt but only 8 ± 1 nt are actually involved in direct interactions with the protein. High salt disfavors the binding, so they are most likely electrostatic interactions. The proposed model for the binding suggests two possible recognition patterns: either the DNA-binding domain of PriA physically protrudes from the rest of the structure, recognizing a short specific DNA sequence, or PriA binds to the DNA and bends it outward on both sides. Other types of PriA binding to DNA have been extensively studied23. PriA binds to duplex DNA substrates that contain a 3’ overhang of at least 16 nt and shows

38 directionality recognition, since it does not bind to any substrate with a 5’ overhang. PriA also binds to almost any nicked or gapped duplex structure, as well as to forks and flaps, provided that the single-stranded region is at least 8 nt long (Figure 3x1). Interestingly, PriA was found to have the strongest affinity for DNA structures containing a sharp bend. This was already observed when the replication restart primosome was first isolated on the φX174 PAS sequence, which is a 70 nt-long fragment with considerable secondary structure that can be folded into several hairpin shapes, undoubtedly responsible for its specific recognition10. The binding of PriA on the PAS has a kD of 11.4 nM (which is further reduced to 6.2 nM when PriB binds). Similarly, PriA has high affinity for D-loops with a 3’ invading strand, which form during recombination42. In all cases, the rescue of stalled replication forks by the PriA-directed assembly of the replication restart primosome occurs through the binding of PriA to a 3-stranded junction containing substantially bent DNA. PriA binds to all three strands5 and unwinds the lagging-strand duplex DNA, creating an SSB-free region of ssDNA large enough for DnaB to bind. Similarly, it assists in the RecG-dependant resolution of blocked replication forks during homologous recombination by relaxing short pieces of supercoiled / overwound duplex DNA of a chicken-foot structure back into a Y-junction39. PriA also plays this role of opening duplex DNA to provide binding space for DnaB at the Bacteriophage Mu fork with the aid of a protein complex called MRFα214. The versatility of PriA DNA-binding properties probably relates directly to some of the unique features of the protein and accounts for its very distinct function.

39 Figure 3x1: DNA substrates for PriA

ssDNA

3’ overhang dsDNA

Flap

Fork

Y-junction

φX174-PAS

Chicken-foot

D-loop

Mu fork

Holiday junction

3.1.3.

Characteristics of PriA

The E. coli PriA is a large protein composed of 732 amino acids with a molecular weight of 81.8 kDa. Its calculated pI of 9.18 makes it basic overall. PriA is found as a monomer in solution, which is also the biologically active unit in the replication restart primosome. For the UV measurements, the extinction

40 coefficient used was ε280 = 105150 M-1 cm-1 and the calculated A280 was 1.285 (for a 1 mg/mL solution). This protein was classified as unstable (Swiss-Prot information file #P17888 PRIA_ECOLI). Based on primary sequence analysis, PriA belongs to the superfamily 2 of DEXH-type DNA helicases43 and defines its own subfamily, due to its two unique zinc-finger motifs interrupting the hypothetical helicase domains. Six motifs are conserved among the prokaryotic PriA proteins, at least two of them being well-known functional and structural motifs common to many helicases (Table 3x1, Figure 3x2)44-46.

Table 3x1: PriA structural and functional motifs Motif

Position

Sequence

Supposed function

Walker A

aa 224-231

GVTGSGKT(E)

Helicase activity

Walker B

aa 320-323

DEEH

Helicase activity / ATP binding

“SAT”

aa 356-358

(G)SAT(P)

?

Zinc-finger #1

aa 336-348

CHDCX5CPRC

?

Zinc-finger #2

aa 363-379

CHHCX9CPSC

?

“Motif VI”

aa 580-587

QVAGRAGR

? (not DNA-binding)

The functions of these domains are still unclear for the most part. Even though a DNA-binding domain distinct from the helicase domain of PriA has been reported, it remains questionable (see Figure 3x2). Site-directed mutagenesis was performed on many residues to try to isolate those responsible for activity. But

41 with so many features present on the protein, the interactions turned out to be much more complicated than anticipated and researchers have proposed many different interpretations to their results, sometimes contradicting one another. The most likely explanation seems that several parts of the protein cooperate with the rest of the molecule and there may not be any clearly separated domains that carry out only one function.

3.1.4.

Limited proteolysis

Previous

attempts to

crystallize the native

E.coli

PriA

were

unsuccessful. As an alternative, three PriA domains were identified by limited proteolysis using trypsin digestion. Each truncation produced was analyzed, cloned and transformed in various expression hosts, in collaboration with Dr. Hiroshi Nakai at Georgeteown University (Washington, D.C.). A diagram showing the position of the different domains and the cleavage sites is represented on Figure 3x2. The domains were called the “N domain” (for N-terminal) or PriA N,

from amino acid 1 to 201; the “I domain” (for Internal) or PriA I, from amino acid 200 to 516 and the “C domain” (for C-terminal) or PriA C, from amino acid 516 to 732. Two other truncations were also made: the “NI domain” or PriA NI (from amino acid 1 to 516), for which the C domain of the protein has been removed and the “IC domain” or PriA IC (from amino acid 200 to 732), for which the N domain of the protein has been removed. General information for the truncations is given in Table 3x2.

42 Gel shift experiments suggested that the DNA-binding domain of PriA was the N domain and that the helicase domain was located somewhere in the IC domain (Dr. H. Nakai, personal communication). Indeed, PriA I contains the two Walker motifs found in many ATP-dependant helicases. The only identified feature of the C domain is the “Motif VI”. Its function is unknown but it does not participate in interactions with nucleic acids44. After protein expression, our group found both PriA I and IC domains to be highly insoluble. Furthermore, the molecular cloning of PriA C was more challenging. As a consequence, this work focused on the other two truncations available: PriA N and PriA NI. The primary goal was to express and purify both domains for crystallization trials.

Figure 3x2: PriA domains and truncations DNA binding domain?

Helicase domain?

Walker B Walker A

1

N

200

Zinc fingers

SAT

I

“Motif VI”

516

IC

NI

C

732

43 Table 3x2: Characteristics of PriA truncations

Domain

MW (kDa)

Length

Cysteines

pI

Calc. Α280

State in solution

Stability

PriA N

22.6

201 aa

2

7.8

1.798

?

unstable

PriA NI

57.4

516 aa

11

8.7

1.269

?

unstable

PriA I

35.0

317 aa

9

8.7

0.919

?

unstable

PriA IC

59.3

533 aa

9

9.2

1.088

?

unstable

PriA C

24.3

217 aa

0

9.7

1.327

?

unstable

3.2. Full-length PriA 3.2.1.

Dialysis and concentration

Stocks of purified PriA had been made previously, stored at -80 °C and made available for this research. The solution was thawed out on ice and dialyzed at 4 °C overnight in 10 mM Na PIPES pH 6.5, 50 mM NH4Cl and 25 mM MgSO4 (the dialysis buffer deduced form a solubility screen experiment). The protein was then concentrated at 4 °C until about 10 mg/mL.

3.2.2.

Dynamic Light Scattering

The quality of the protein solution was assessed by DLS to see whether it was suitable for crystallization experiments. For this experiment, the

44 PriA sample was prepared at a concentration of 5 mg/mL and was filtered in a single-use Ultrafree®-MC filter (yellow type, pore size = 0.1 µm, Millipore). The experiment was performed at both 20 °C and 4 °C. The results are given in Table 3x3. Interestingly, at room temperature, the polydispersity is much lower

than at 4 °C and the solution is very homogeneous.

Table 3x3: DLS results for PriA

Temperature

Rh (nm)

MW (kDa)

% poly

Baseline

SOS

Distribution

4 °C

3.481

59.60

25.1

1.009

24.6

Bimodal

20 °C

3.789

73.21

15.1

-

-

-

3.2.3.

Crystallization

According to the DLS, PriA seemed more likely to crystallize at room temperature than in the cold. Nonetheless, crystal screening was usually tried in both conditions to enhance the chances of getting protein hits. PriA crystal screens were setup at room temperature and at 4°C as summarized in Table 3x4. The first series of trays were done manually (AI 001 to AI 006) and no protein crystal hits were found in any of the conditions. Later, more trays were setup automatically (AI 068 to AI 074) with the new commercial kits. This time a few hits were found and stained with Izit Crystal Dye™. The pictures of some of the protein crystals are shown in Figure 3x3.

45 Expansion of several crystallization conditions from the hits at room temperature were done in Costar trays with a 4x6 A/B gradient format. No crystals grew and most of the drops precipitated. The expansions at 4 °C were not done but they should be tried in the future, as well as the other room temperature conditions.

Figure 3x3: PriA crystal screen hits

Additive Screen #80, RT

Additive Screen #28, 4 °C

30% (v/v) MPD,

2 M K, Na phosphate pH 8.9

10% (w/v) Xylitol

Additive Screen #4, RT

Index #24, 4 °C

40% (w/v) PEG-1000

2.8 M Na Acetate pH 8.0

46 Table 3x4: Crystal screens of PriA Tray ID

Temperature

Type of tray

Kits

AI 001

RT

3-well Greiner

Wizard I and II

No

AI 002

RT

3-well Greiner

Crystal Screen I and II

No

AI 003

RT

3-well Greiner

Cryo I and II

No

AI 004

RT

3-well Greiner

Ion screen and Natrix

No

AI 005

4 °C

3-well Greiner

Wizard I and II

No

AI 006

4 °C

3-well Greiner

Crystal Screen I and II

No

AI 068

RT

Corning

Additive screen

AI 069

RT

Corning

Index

AI 070

RT

Corning

AI 074

RT

Corning

Crystal Screen Cryo and MembFac Salt RX

1?

AI 071

4 °C

3-well Greiner

Additive screen

5

AI 072

4 °C

3-well Greiner

Index

2

AI 073

4 °C

3-well Greiner

Crystal Screen Cryo and MembFac

1

3.2.4.

Hits?

6 No 1

Purification from endogenous nucleases

Even though the PriA samples had been well purified, they still contained contaminating nuclease activity. This prevented co-crystallization experiments with DNA substrates because the nucleases would rapidly degrade the DNA. To solve this problem, an additional purification step was performed that would remove the contaminating nucleases, which consisted of running the

47 protein solution on a POROS PE column. As explained in Section 2.4.2, the nucleases should stick to the media very tightly whereas the protein should elute in the void fraction34. For this experiment, the POROS PE column was equilibrated at 15mL/min for 5 CV with a high salt elution buffer containing 25 mM Tris pH 7.5, 50 mM NaCl, 1 mM EDTA, 2 mM DTT and 600 mM (NH4)2SO4. Then the baseline was set. A frozen sample of PriA (10-20 mg) was thawed out in ice and saturated ammonium sulfate was added to it until the conductivity matched that of buffer A (~ 90 mS/cm). The protein was loaded on the column at 12 to 15 mL/min and the void fraction was collected when a peak started appearing on the UV detector and until the OD280 went back to the baseline level. The chromatogram of the flow-through trace is shown on Figure 3x4. The presence (or the absence) of nucleases could then be assayed by agarose gel electrophoresis: different DNA substrates were run with and without the protein and the bands compared. The gel obtained is presented in Figure 3x5.

Figure 3x4: HPLC purification of PriA - POROS PE

48 Figure 3x5: Nuclease activity test for PriA

1

2

3

4

5

6

Lanes: 1

pET 28

2

pET 28 + PriA before POROS PE

3

pET 28 + PriA after POROS PE

4

λ-DNA

5

λ-DNA + PriA before POROS PE

6

λ-DNA + PriA after POROS PE

The results with the pET 28 vector were inconclusive because it was not concentrated enough. However, the lanes with λ-DNA clearly showed that the sample that had not been treated on POROS PE had a lot of nucleases in it because the DNA band disappeared. But after PriA was treated on POROS PE, the DNA was intact, which means that the solution was nuclease free. At this point, the protein was very dilute and in high salt conditions. So it had to be dialyzed and concentrated as described in § 3.4.1 before it could be utilized.

49 3.2.5.

Co-crystallization with DNA

After the protein solution had been purified away from contaminating nucleases, crystallization experiments with DNA were possible. It has been known for many years that PriA binds quite well to flap-DNA structures containing a leading strand duplex with a 3’ end at the junction of the fork (this was used as a control on the gel shift assay in § 3.3.10). So this type of flap-DNA substrate was employed in the crystal screening of the PriAyDNA complex. The three single-stranded oligos were designed (Figure 3x6) and purchased from IDT in a standard desalted form. They were purified and annealed according to the method given in section 2.14. The chromatograms are given in Figure 3x7.

Figure 3x6: Flap-DNA substrate used for co-crystallization with PriA

5’ Parental Lagging strand

3’ 3’ 5’

Leading strand duplex

Parental Leading strand

5’ 3’

50 Figure 3x7: HPLC purification of the Flap-DNA substrate Figure 3x7a: Parental Leading strand oligo

***

Figure 3x7b: Parental Lagging strand oligo

**

51 Figure 3x7c: Leading strand oligo

**

Fractions 13, 14 and 15 were collected for the parental leading strand. For the parental lagging strand, fractions 10 and 11 were collected and for the leading strand, fractions 8 and 9 were collected. The purity of these fractions was verified on a 20% TBE gel and indeed, they all had a single-band meaning that they were pure. Next, the three oligos were annealed together in a 1:1:1 molar ratio and the final solution was concentrated to 0.5 mM. All the stock solutions were then frozen at -20 °C until used. For crystal screens, the PriA solution was concentrated to 15.6 mg/mL and mixed with the DNA substrate in a centrifuge tube in a 1:1 ratio, so that the final concentration in the screens was about 10 mg/mL. Three crystal screens were setup at room temperature in 96-well Corning trays, using the automated

52 crystallization robots, with the Wizard™ I and II, the Crystal Screen™ I and II and the Index™ kits. Crystal hits may have not been found yet and this work needs to be followed-up for more results.

3.3. PriA N-terminal domain 3.3.1.

Protein expression

The N-terminal domain of E. coli PriA (PriA N) was cloned in pET 21a vectors and transformed into BL21 (DE3) Gold cells for expression. The primary sequence was confirmed by DNA sequencing. A glycerol stock of the cells was made and kept at -20 °C. After that, large amounts of the protein were obtained by the large scale preparation method (§ 2.2.1). The cells were grown at 37 °C in 25 g/L LB and 1 mM Amp until the OD600 reached about 0.6. The culture was then induced with 1 mM IPTG, grown for another 3 hr, harvested and kept at -20 °C until cell lysis. The expression of PriA N is shown in Figure 3x8 (lanes 1 to 6), it was always very good. A typical large scale preparation yielded about 15 g of cell pellet for 6 L of LB medium.

3.3.2.

Cell lysis

The cells were lysed open using the low salt cell lysis protocol described in § 2.3.1. The pellet and lysate obtained were tested on SDS-PAGE gel electrophoresis. PriA N was mostly found in the lysate, as seen on the gel

53 (Figure 3x8, lane 8). This material was flash-frozen in dry-ice, after addition of 15% glycerol and was ready for HPLC purification.

Figure 3x8: SDS-PAGE gel of the expression and cell lysis of PriA N

Lanes:

kDa: 31.0 21.5 14.4

3.3.3.

1

0 hr sample

2

3 hr sample (1)

3

3 hr sample (2)

4

3 hr sample (3)

5

3 hr sample (4)

6

3 hr sample (5)

7

Cell lysis pellet

8

Cell lysis lysate

Purification

The purification of PriA N consisted of two cation-exchange chromatography steps: first SP Sepharose (Low-resolution cation exchange) and then POROS HS (High-resolution cation exchange). Before each column, the conductivity of the protein sample was measured and matched to that of buffer A by addition of the necessary amount of 50 mM Tris-HCl pH 7.5. A linear salt gradient was used for both runs (see § 2.4.4). After each run, the fractions of

54 interest were pooled together and their purity was estimated by SDS-PAGE gel electrophoresis. The chromatograms and the gels are shown in Figure 3x92.

Table 3x5: HPLC purification of PriA N - Buffers Anion exchange

Buffer A

Buffer B

Composition

50 mM Tris-HCl pH 7.5

50 mM Tris-HCl pH 7.5

-

1.5 M NaCl

100 mM NH4Cl

100 mM NH4Cl

10 mM MgCl2

10 mM MgCl2

1 mM DTT (or BME)

1 mM DTT (or BME)

9.5 mS/cm

75 mS/cm

Conductivity

2

On all HPLC chromatograms, the fractions tested on the corresponding SDS-PAGE gel were

labeled with a ¸.

55 Figure 3x9: HPLC purification of PriA N - Chromatograms and SDS-PAGE gels Figure 3x9a: Step 1 SP Sepharose 1

2

3

4

100.0% Buffer B 2.50

2.00

5

6

7

8

9

Fractions 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

*

* * * *

*

2.50

2.00

75.0

OD260 1.50

1.50

OD280

50.0

1.00

1.00

Cond. 25.0

0.50

0.50

0.00

0.00

0.0

-0.50 AU

-0.50 0.00

10.00

SP Sepharose

20.00

Min.Tenth

30.00

40.00

50.00

POROS HS

3x9a: After SP Sepharose, the protein was found in fractions 10 to 18 in a very large peak. The other peak represents DNA (OD260 in purple > OD280 in green). L= Load FT= Flow-Through

21.5 kDa

AU

56 Figure 3x9b: Step 2 POROS HS 1

2

3

4

100.0% Buffer B

5

6

7

8

9

Fractions 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

* * * *

*

2.00

2.00

75.0

1.50

1.50

50.0 1.00

1.00

Cond. 25.0 0.50

0.00

AU

0.50

OD260

OD280 0.00

0.0

0.00

10.00

Min.Tenth

20.00

AU

3x9b: The protein peak on POROS HS was also huge. Fractions 12 to 15 were pooled together. However, a ladder was visible and the addition of a potent reducing agent was needed to remove disulfide bond cross-linking. This was done by treating the solution with 2 mM TCEP-HCl. The lane on the right shows that the main bands were successfully removed.

The protein was at least 95% pure at that point. The yields for PriA N were exceptional: 200 to 230 mg of pure protein can be obtained in one preparation from 15 g of bacterial cells.

57

3.3.4.

Solubility screen

A solubility screen was performed on PriA N at room temperature and the result is shown on Figure 3x10. The solubility of PriA N was rather high in most of the conditions tried but particularly in Na citrate and at low pH. As a result, it was decided that the optimized buffer would be composed of 50 mM Na Citrate pH 5.5 and 2 to 5 mM TCEP. The maximum solubility was measured to be at least 40 mg/mL in those conditions. As a result, the optimal concentration for crystallization should be around 20 mg/mL.

Figure 3x10: Solubility screen of PriA N at room temperature

0.11

Supernatant

0.38

H2O 0.32

TAPS pH 8.5

0.55

HEPES pH 7.5 0.48

PIPES pH 6.5

0.80

MES pH 5.6

0.84

Salt / buffer

Na Citrate 0.40

Na Phosphate

0.39

Na Sulfate

0.48

Na Cacodylate 0.20

Na Acetate

0.28

Na Formate

0.37

CaCl2

0.50

MgCl2

0.52

LiCl 0.42

KCl 0.25

NaCl

0.32

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

Absorbance at 595nm

0.60

0.70

0.80

0.90

58 3.3.5.

Concentration and dialysis

The purified PriA N sample was dialyzed for 3 hours at 4 °C against its optimized buffer, concentrated until a value between 20 mg/mL and 40 mg/mL was reached and then filtered. Part of the sample was kept at 4 °C to be used fresh and the remaining solution was aliquoted in 1mL fractions and flash-frozen in dry ice after addition of 15% glycerol. This solution was stored at -80 °C until later use.

3.3.6.

Dynamic Light Scattering

DLS measurements were done on PriA N at 20 °C and at 4 °C, to determine the homogeneity of the protein in solution and if the quality of the solution was good enough for crystal growth. For this analysis, a 5 mg/mL PriA N sample was made in 50 mM Na citrate pH 5.5, 10 mM TCEP. It was filtered in a single-use Ultrafree®-MC filter (yellow type, pore size = 0.1 µm, Millipore). The results are given in Table 3x6.

Table 3x6: DLS results for PriA N

Temperature

Rh (nm)

MW (kDa)

% poly

Baseline

SOS

Distribution

4 °C

3.495

60.16

17.0

1.002

7.95

Monomodal

20 °C

3.272

51.30

19.5

1.012

10.1

Bimodal

59 Although the results at 20 °C were acceptable, it was clear from these values that PriA N behaved better at 4 °C than at room temperature. The quality of the solution at 4 °C was very good with a low polydispersity and a monomodal distribution. This made it likely to crystallize. One thing that was noticed though was the large predicted molecular weight. The protein seemed to form a trimer in solution, even in the presence of reducing agent because a monomer of PriA N would be 22.5 kDa and the main species was around 60 kDa. Since DLS cannot estimate the molecular weight of a macromolecule very accurately, sizeexclusion chromatography was used to verify this result.

3.3.7.

Size-exclusion chromatography

A sample of PriA N in its optimized buffer was run on a Superdex 75 column at 4 °C to determine the exact molecular weight of the protein in those conditions. The chromatogram obtained is shown in Figure 3x11. The top of the protein peak was found at the limit between fractions 12 and 13. Using the standardization plot for Superdex 75, it was deduced that this fractions corresponds to a macromolecule of size between 60 kDa and 68 kDa. This is indeed three times the size of a PriA N monomer and proves that even in the presence of a strong reducing agent, PriA N forms a trimer in solution. So there must to be other types of protein-protein interactions than covalent cysteine linkage occurring with the N domain of PriA.

60 Figure 3x11: PriA N - Size exclusion 100 90 80 70 y = 537632x -3.5536 R2 = 0.939

60

MW (kDa)

Í

50 40 30 20 10 0 0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Fraction number

Molecular weight (kDa)

Protein

12 Limit of 12 and 13 14 16 20

96 ~ 65 36 23 14.4

Soybean LOX (Std #1) PriA N T4 RNase H (Std #2) T4 gp 59 (Std #3) HEW lysozyme (Std #4)

3.3.8.

Crystal screens

Many crystal screens were setup for PriA N. For the most part, this was done in the same trays as the full-length PriA using 3-well Greiner plates at room temperature and at 4 °C, as recorded in Table 3x7. In the first series of trays (AI 001 to AI 006), the optimized buffer for PriA N had not been determined yet and the protein was screened in PIPES buffer at 10 mg/mL. There were no crystal hits. So the same trays were setup again later with the protein in its

61 optimized buffer at a concentration of 19 mg/mL. This time, several crystal hits were found and turned dark blue when stained with Izit Crystal Dye™, clearly demonstrating the importance of optimizing the solution conditions prior to crystal screening. The pictures are shown in Figure 3x12, the best hits were found in the cold.

Table 3x7: Crystal screens of PriA N Tray ID

Temperature

Type of tray

Kits

Hits?

AI 001

RT

3-well Greiner

Wizard I and II

No

AI 002

RT

3-well Greiner

Crystal Screen I and II

No

AI 003

RT

3-well Greiner

Cryo I and II

No

AI 004

RT

3-well Greiner

Ion screen and Natrix

No

AI 005

4 °C

3-well Greiner

Wizard I and II

No

AI 006

4 °C

3-well Greiner

Crystal Screen I and II

No

AI 034

RT

Corning

Crystal Screen I and II

2

AI 035

RT

Corning

Cryo I and II

AI 036

RT

Corning

Wizard I and II

1

AI 037

RT

Corning

Ion screen and Natrix

2

AI 045

4 °C

Corning

Crystal Screen I and II

No

AI 046

4 °C

Corning

Cryo I and II

1?

AI 047

4 °C

Corning

Wizard I and II

1?

AI 048

4 °C

Corning

Ion screen and Natrix

No

AI 057

RT

3-well Greiner

Index

No

AI 058

RT

3-well Greiner

Additive screen

AI 059

RT

3-well Greiner

AI 071

4 °C

3-well Greiner

Crystal Screen Cryo and Membfac Additive screen

AI 072

4 °C

3-well Greiner

Index

2

AI 073

4 °C

3-well Greiner

Crystal Screen Cryo and Membfac

1

No

1 No 1

62 Figure 3x12: PriA N crystal screen hits

Crystal Screen I #29, RT

Ion Screen #31, RT

0.8 M K, Na Tartrate

20% (w/v) PEG-4000

0.1 M Na HEPES pH 7.5

0.1 M Na HEPES pH 7.5 0.2 M Na Formate

Index #28, 4 °C

Membfac #35, 4 °C

35% (v/v) Tacsimate pH 7.0

12% (v/v) MPD 0.1 M Na HEPES pH 7.5 0.1 M Na citrate

Index #24, 4 °C

Additive Screen #77, 4 °C

2.8 M Na Acetate pH 8.0

30% (v/v) MPD 10% (w/v) D-glucose

63 Expansion of the crystallization conditions from the hits were done at 4 °C in Nextal trays using 4x6 A/B gradients. No crystals grew over a few weeks but there was no precipitation observed in the majority of the drops, which is a rather good sign and the crystals may just take months to grow. Expansions of the room temperature hits still need to be done.

3.3.9.

Test for endogenous nuclease activity

Since the PriA solutions had contained contaminating nucleases, this also had to be checked for the PriA N samples. The experiment would also tell whether the protein needed to be run on the POROS PE column for further purification, if co-crystallization with DNA was to be tried. The results are shown below in Figure 3x13. It can be seen that the DNA bands were identical with and without the protein, which means that the solution did not contain any nucleases.

Figure 3x13: Nuclease activity test for PriA N

Lanes: 1

pET 28+ PriA N (no POROS PE)

2

pET 28

3

λ-DNA+ PriA N (no POROS PE)

4

λ-DNA

64 3.3.10. Fluorescence experiment This gel shift experiment was performed to test the DNA-binding ability of the N-terminal domain of PriA and compare it to the native protein. The substrate used was a flap-DNA with a fluorescent label at the end of the leading strand. It was similar to the substrate used in the co-crystallization of PriA with DNA and had the same number of nucleotides on each strand. The retardation gel was run as described in section 2.12 and is shown in Figure 3x14.

Figure 3x14: DNA retardation gel

Í Lanes: 1

Free DNA

2

DNA : PriA

1:1

3

DNA : PriA

1:3

4

DNA : PriA N

1:1

5

DNA : PriA N

1:3

As seen on the gel, PriA does bind the flap-DNA at a 1:1 ratio but it binds even tighter when it is in excess (lanes 1 and 2). On the other hand, lanes 3 and 4 show that the N domain of PriA binds the DNA very weakly, since the fluorescent band is found mostly at the bottom of the gel, corresponding to free

65 DNA. It was concluded from this experiment that the co-crystallization of PriA N with DNA would probably not be successful. At this point, the other option was to study the larger PriA NI domain, where only the C-terminal domain of the protein had been truncated.

3.4. Strategies for studying the NI domain of PriA 3.4.1.

Rationale

The NI domain of PriA had been previously cloned, expressed and purified. Yet there was a major drawback: the solution was always full of nucleases and at that time, the purification procedure involving a POROS PE column was not routinely utilized. So a different approach was taken that consisted of adding a hexahistidine tag (HisTag) on either end of the protein. The purification would then be done using nickel-affinity chromatography with very high efficiency. To do so, the gene coding for PriA NI domain was cloned in two different vectors: in pET 28 to insert an N-terminal HisTag and also in pET 21a to insert a C-terminal HisTag. Both plasmids were transformed into various expression hosts and solubility studies were done on each one, in order to identify the cell line that would be the most efficient in producing soluble protein. Later, a protocol was adapted from similar work on the T4 gp 32 protein, to remove nucleases using POROS PE purification. This was attempted in parallel with the HisTag strategies, using the “native” PriA NI domain that had been made previously. These experiments are summarized in Figure 3x15.

66

Figure 3x15: Strategies used for the PriA NI domain

PriA NI domain (PCR)

PriA NI (Quikchange™) Cloned in pET 21a

PriA NI Cloned in pET 28

PriA NI Cloned in pET 21a

C-terminal HisTag

N-terminal HisTag

No HisTag

BL21 RIL

Soluble

Rosetta

BL21 RIL

BL21 (DE3)Gold

Expression and purification (3 steps)

X

X

X

X

Insoluble

Insoluble

Insoluble

Nucleases

Expression and purification with POROS PE (4 steps)

Test for nuclease activity D None

67

3.4.2.

Protein extraction studies

This experiment was done on the two HisTag forms of PriA NI, to determine in which cell line the protein would be the most soluble and also to find out what kind of cell lysis procedure should be used. Four samples were used: 1. PriA NI C-terminal HisTag (pET 21a) expressed in BL21 RIL cells 2. PriA NI C-terminal HisTag (pET 21a) expressed in Rosetta cells 3. PriA NI N-terminal HisTag (pET 28) expressed in BL21 RIL cells 4. PriA NI N-terminal HisTag (pET 28) expressed in BL21(DE3)Gold cells

Each sample was produced on a small scale and examined for expression (§ 2.1.1). Three cell lines worked fine but the Rosetta cells of the Cterminal HisTag sample never grew despite several attempts. The pellets and lysates obtained after cell lysis were all run on the same SDS-PAGE gel for comparison (Figure 3x16). All the cell lysis experiments on the N-terminal HisTag samples failed to produce soluble protein, since the band corresponding to PriA NI is only found in the pellet (lanes 8 and 12, ~ 58 kDa). On the other hand, for the C-terminal HisTag sample expressed in RIL cells, the protein was found in the lysate, but only after using the high salt lysis. Although some of the protein was still insoluble, this was by far the best result. Therefore, it was decided to pursue only the large scale preparation of the PriA NI C-terminal HisTag that was expressed in BL21 RIL cells and that the cells should be lysed in high salt conditions.

68 Figure 3x16: SDS-PAGE gel of the solubility studies of PriA NI

P L P L P L P L

L L P L L

L

55.4 kDa

Lanes: 1

Molecular weight

2

PriA NI C-terminal HisTag RIL

Lysis buffer

Pellet

3

PriA NI C-terminal HisTag RIL

Lysis buffer

Lysate

4

PriA NI C-terminal HisTag RIL

1 M NaCl

Pellet

5

PriA NI C-terminal HisTag RIL

1 M NaCl

Lysate

6

PriA NI C-terminal HisTag RIL

Bug Buster

Pellet

7

PriA NI C-terminal HisTag RIL

Bug Buster

Lysate

8

PriA NI N-terminal HisTag RIL

Lysis buffer

Pellet

9

PriA NI N-terminal HisTag RIL

Lysis buffer

Lysate

10

PriA NI N-terminal HisTag RIL

1 M NaCl

Lysate

11

PriA NI N-terminal HisTag RIL

Bug Buster

Lysate

12

PriA NI N-terminal HisTag Gold

Lysis buffer

Pellet

13

PriA NI N-terminal HisTag Gold

Lysis buffer

Lysate

14

PriA NI N-terminal HisTag Gold

1 M NaCl

Lysate

15

PriA NI N-terminal HisTag Gold

Bug Buster

Lysate

69

3.5. PriA NI domain with a C-terminal HisTag 3.5.1.

Protein expression

Site-directed

mutagenesis

had

been

performed

using

the

Quikchange™ kit (Stratagene) on the NI domain of E. coli PriA. It was cloned in pET 21a vectors to change the stop codon to a glutamine allowing read through to the HisTag on the C-terminus. This truncation is referred to as “PriA NI QC”. The protein was produced from a glycerol stock using the large scale preparation method (§ 2.2.1). The cells were grown at 37 °C in 25 g/L LB, 1 mM Amp and 1 mM CAM until the OD600 of the solution was between 0.4 and 0.6. Induction was done with 1 mM IPTG and the cells were grown for another 3 hr, harvested and kept at -20 °C until cell lysis. The over-expression of PriA NI QC was successful (Figure 3x17, lanes 2 to 8), with a large band appearing after 3 hr of growth around 58 kDa. The yield was generally around 12 g of cell pellet for 6 L of LB medium.

3.5.2.

Cell lysis with ammonium sulfate precipitation

First, a high-salt cell lysis was performed, in which the lysis buffer contained 1 M NaCl (see § 2.1), as deduced from the extraction studies. This was unsuccessful and the only way to obtain a soluble fraction of PriA NI QC was to use the cell lysis protocol with subsequent ammonium sulfate precipitation (§ 2.3.2). The final result is shown on Figure 3x17 (lanes 10 and 11). The lysate was frozen in dry-ice in 15% glycerol and stored at -80 °C.

70 Figure 3x17: SDS-PAGE gels of the expression and cell lysis of PriA NI QC

Lanes: 1

MW

2

0 hr sample

3

3 hr sample (1)

4

3 hr sample (2)

5

3 hr sample (3)

6

3 hr sample (4)

7

3 hr sample (5)

8

3 hr sample (6)

9

Lysis pellet with 1M NaCl

10

Pellet after (NH4)2SO4 precipitation

11

Lysate after (NH4)2SO4 precipitation

The amount of protein obtained in the lysate was small but a lot of material had been lost in the failed high-salt cell lysis (lane 9). In order to obtain more protein, this experiment needs to be repeated on a larger sample with the proper cell lysis protocol and should work fine.

3.5.3.

Purification

The crude lysate of PriA NI QC was purified in one step by IMAC chromatography using the Ni-NTA resin. The column was first regenerated with the proper “regeneration buffer” and then equilibrated at 2 mL/min with “lysis buffer” for 5 CV. The sample was injected on the column at the same speed and the flow-through was collected for about 10 CV, until the OD280 went back to the

71 baseline level. The program was then run (a step gradient between the “wash buffer” and the “elution buffer”) and the fractions collected. The chromatogram and the SDS-PAGE gel are given in Figure 3x18.

Figure 3x18: HPLC purification of PriA NI QC - Chromatogram and SDS-PAGE gel 8

9

10

11

12

13

Fractions 14

15

16

17

18

19

20

21

22

23

24

25

100.0% Buffer B

0.350

0.300

*

* * * * * * *

0.350

90.0

80.0

0.300

70.0 0.250

0.250

OD280

60.0

0.200

0.200 50.0

0.150

40.0

0.150

30.0 0.100 20.0

OD260

Cond.

0.050

0.100

0.050 10.0

-0.000

AU

0.0

-0.000

50.00

55.4 kDa 36.5 kDa

60.00

Min.Tenth

70.00

80.00

AU

3x18: As seen on the gel (left), the fraction that was loaded on the column was small. Yet the protein peak was significant (F.14 to 18). The protein was almost 100% pure. On the graph, the ODs remained high even after the protein eluted because of the intrinsic absorption of the imidazole used to elute the protein.

72

3.5.4.

Dynamic Light Scattering

DLS experiments were done after the protein was purified without any further optimization. The PriA NI QC sample was concentrated to 6 mg/mL and carefully filtered as describe before. The measurements were done at room temperature and at 4 °C. The results are shown below.

Table 3x8: DLS results for PriA NI QC Temperature

Rh (nm)

MW (kDa)

% poly

Baseline

SOS

Distribution

4 °C

5.507

181.4

34.5

1.003

10.7

Monomodal

20 °C

5.884

212.9

26.0

1.014

2.73

Bimodal

From these results it was hard to determine which condition was best. The 4 °C data were of better quality, with no aggregate, but the polydispersity of the solution was high, making it unlikely to crystallize. At 20 °C on the other hand, the polydispersity was a lot lower but the formation of a large aggregate was observed. The molecular weight suggested that PriA NI QC was a trimer in solution but this was not investigated further.

3.6. PriA NI domain 3.6.1.

Protein expression

The “native” PriA NI domain was cloned in pET 21a vectors and transformed into BL21 (DE3) Gold cells for expression. DNA sequencing

73 confirmed that the primary sequence was correct. The large scale preparation of the protein was done using the method in § 2.2.1 but at 24 °C instead of 37 °C, otherwise the protein was insoluble. The cells were grown at 24 °C in 25 g/L LB and 1 mM Kan until the OD600 reached about 0.6. The culture was then induced with 1 mM IPTG, grown for another 3 hr, harvested and kept at -20 °C until cell lysis. The expression of PriA NI was very good and the band was found around 58 kDa (Figure 3x19, lanes 1 to 4). The large scale preparation yielded only about 8 to 10 g of cells for 6 L of LB medium but the amount of overexpressed protein was still satisfactory.

3.6.2.

Cell lysis with ammonium sulfate precipitation

The lysis of PriA NI BL21(DE3)Gold cells was done using the cell lysis with subsequent ammonium sulfate precipitation method (§ 2.3.2). The results are shown on Figure 3x19. As seen in lanes 5 and 6, only half of the material was extracted in a soluble form during cell lysis but then it was precipitated out quantitatively and the protein was found in the pellet (lane 7). At the end, the pellet was collected and gently resuspended in SP Sepharose buffer A until a clear solution was obtained. This material was frozen in dry-ice with 15% glycerol and stored at -80 °C until it was purified.

74 Figure 3x19: SDS-PAGE gels of the expression and cell lysis of PriA NI

Lanes:

PriA NI

3.6.3.

1

0 hr sample

2

3 hr sample (1)

3

3 hr sample (2)

4

3 hr sample (3)

5

Lysis Pellet

6

Lysis Lysate

7

Pellet after (NH4)2SO4 precipitation

8

Lysate after (NH4)2SO4 precipitation

9

MW

HPLC purification

The purification scheme for PriA NI had been devised previously. It consisted in three ion-exchange steps (Figure 3x20). My task was to include an extra step in order to remove endogenous nucleases by running the solution on a POROS PE column between the HA and POROS HS. It was inserted in that order, so that the high salt sample from HA could be loaded directly on POROS PE and then could be concentrated in only a few fractions using POROS HS, which produces sharp, narrow protein peaks. As always, the conductivity of the sample was adjusted before every ion-exchange step. The POROS PE run was done in the same way as for PriA (see § 3.2.4) and the void fraction was collected manually. The results of each run are given in Figure 3x21.

75 Figure 3x20: Purification scheme for PriA NI

SP Sepharose (Low-resolution cation exchange)

HA (Anion exchange with nucleic acids)

POROS PE (Hydrophobic interactions)

POROS HS (High-resolution cation exchange)

Table 3x9: HPLC purification of PriA N - Buffers Anion exchange

Buffer A

Buffer B

Composition

50 mM Tris-HCl pH 7.5

50 mM Tris-HCl pH 7.5

-

1.5 M NaCl

100 mM NH4Cl

100 mM NH4Cl

10 mM MgCl2

10 mM MgCl2

1 mM DTT (or BME)

1 mM DTT (or BME)

Conductivity

9.5 mS/cm

75 mS/cm

HA

Buffer A

Buffer B

Composition

25 mM Tris-HCl pH 7.5

25 mM Tris-HCl pH 7.5

-

1 M (NH4)2SO4

50 mM NaCl

50 mM NaCl

1% Glycerol

1% Glycerol

11 mS/cm

63 mS/cm

Conductivity

76 Figure 3x21: HPLC purification of PriA NI - Chromatograms and SDS-PAGE gels Figure 3x21a: Step 1 SP Sepharose 1

2

3

4

100.0% Buffer B 2.00

5

6

7

8

9

Fractions 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

* ** ****

*

*

*

*

*

*

2.00

90.0 1.75

1.75 80.0

OD260 1.50

1.50

70.0

60.0

1.25

1.25

OD280

50.0

1.00

1.00

40.0 0.75

0.75 30.0

Cond.

0.50

0.50

20.0

0.25

0.00

0.0

0.00

-10.0

-0.25 AU

0.25

10.0

0.00

-0.25 10.00

20.00

Min.Tenth

30.00

40.00

3x21a: The protein peak was very large. The fractions of interest (from 10 to 25) were pooled together to be run on HA.

55.4 kDa

AU

77 Figure 3x21b: Step 2 HA 1

2

3

4

5

6

7

8

9

10 11 12

13 14

15 16 17

18 19 20

21

Fractions 22 23 24 25 26

* * *

0.70 35.0% Buffer B

27

28 29 30

31 32 33

*

34 35

36 37 38 39

40

41

42 43 44

45 46 47

*

0.70

0.60

0.60 30.0

0.50

0.50

OD280

25.0

0.40

0.40 20.0

0.30

0.20

0.30

15.0

OD260 0.20

10.0

0.10

5.0

-0.00

0.0

AU

Cond.

0.10

-0.00

5.00

10.00

Min.Tenth

15.00

20.00

kDa: 55.4 36.5

3x21b: After the HA, the protein sample was already much cleaner. A sharp peak was observed in fractions 20 to 28. The main impurity should be efficiently removed by the high-resolution cation-exchange medium.

AU

78 Figure 3x21c: Step 3 POROS PE

3x21c: The flow-through and the void fractions were collected manually until the end of the peak and this solution was then loaded on the POROS HS (see SDS-PAGE gel on the following page).

Figure 3x21d: Step 4 POROS HS 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

Fractions 18 19

20

21

22

23

24

25

26

27

28

29

30

31

* * *

0.70 35.0% Buffer B

32

33

34

35 36

0.70

0.60

0.60 30.0

0.50

0.50 25.0

OD280

0.40

0.40 20.0

0.30

0.20

0.30

15.0

0.20

10.0

Cond.

OD260

0.10

5.0

0.10

-0.00

0.0

-0.00

-0.10

-5.0

AU

-0.10 5.00

Min.Tenth

10.00

15.00

AU

79

3x21d: After the POROS PE and HS runs, the purity of the protein is acceptable. Fractions 14 to 18 were collected, concentrated to 8mg/ml and filtered. Some of the solution was stored at 4 °C and used fresh, the rest of it was frozen in dry-ice and stored at -80 °C.

PriA NI

During this purification, the amount of protein was very large in the SP Sepharose run and in the HA run. But somehow, after the POROS PE the protein band was much smaller. While this fraction was very dilute, making the bands fainter, some protein must have been lost in the flow-through because the small band from POROS HS in the last step was also unusual.

3.6.4.

Test for endogenous nuclease activity

The presence of contaminating nucleases has been the major problem encountered with the PriA NI samples. It was critical to check whether the introduction of an extra purification step was actually efficient in removing the nucleases from the protein solution. The experiment was performed using the protocol in § 2.13 and the results are shown Figure 3x22.

80 It can be seen that the DNA bands were almost identical with and without the protein. A trace degradation band was observed in the lanes containing the protein but it was faint. This shows that the solution had reduced activity but was not entirely nuclease-free.

Figure 3x22: Nuclease activity test for PriA NI

1

2

3

4

Lanes: 1

pET 28+ PriA NI after POROS PE

2

pET 28

3

λ-DNA+ PriA NI after POROS PE

4

λ-DNA

3.7. Discussion The native E. coli PriA protein was investigated, as well as two of its truncations: the N-terminal domain and the NI domain. Since there is no structural information available on PriA, the primary focus of this work was crystallization, but a lot more knowledge was also gained along the way on its behavior in solution. First of all, the solubility of the protein seemed to be strongly dependent on the presence of MgSO4 in the dialysis buffer: it was observed that

81 if the concentration of MgSO4 was dropped below 25 mM, the protein was no longer soluble. However, the precipitation was reversible upon addition of higher concentrations of MgSO4. This might be an indication that the zinc-finger motif present in the core of the protein requires a divalent metal for stability and proper folding. The DLS data suggested that the homogeneity of the solution was better at room temperature than at 4 °C, something that had not been anticipated. But this effect may not be critical because there was no real difference in crystallization results. Indeed, the crystal screens produced only a few protein hits and unfortunately these could not be reproduced. The optimization of the crystallization conditions at 4 °C certainly needs to be further investigated and it would be interesting to try some slightly different A/B gradients before any definitive conclusion can be made on the “crystallizability” of PriA. In some respects, the co-crystallization of PriA in complex with DNA seems much more promising. First, the purification of the DNA oligos is now well established and the cost of DNA is not an obstacle to these experiments anymore. Furthermore, the problem of contamination by endogenous nucleases can now be overcome for virtually any sample through the use of a POROS PE column in the purification. So in the case of PriA, the co-crystallization with DNA should be successful sooner or later. So far, only a few crystallization screens have been tried and this can surely be studied more extensively. At the same time, since the DNA-binding properties of PriA are well documented in the literature, other DNA substrates could provide an alternative if the flap-DNA is unsuccessful.

82 The other approach was to prepare different truncations of PriA in the hope that at least one of them would crystallize. For this, three domains were considered (N, I and C) and the different truncations were made. The only two that were soluble were the N domain and the NI domain. Of course, there was no information available on these domains, since this was the first time they were made. So the preparation protocols had to be designed and tested and some preliminary studies had to be done before crystallization trials were possible. PriA N turned out to be a great protein to work with and this made the process a lot less tedious. The expression levels were high and a simple low salt cell lysis could be used with high recovery. The efficiency of the HPLC purification was remarkable in terms of yield and this was truly the best I was ever able to achieve. It was found that cross-linking occurred between the two cysteine residues, which was prevented by the addition of reducing agent. The solubility of the PriA N was good and optimal buffer conditions were easily found. It is interesting to note that the composition of the buffer was decisive for subsequent studies. Having PriA N in its optimized buffer not only increased the solubility limit, but it also allowed recording DLS results and made the crystal screening successful (see chapter 6 for a study of these observations on other proteins). Indeed, DLS studies were impossible in the HPLC buffer and worked perfectly after the optimized buffer was determined. It showed that the protein solution was of good quality, especially at 4 °C and that PriA N was present as a trimer of 68 kDa in solution. This was confirmed by size-exclusion chromatography. Similarly, the crystal screens prepared in standard buffer failed to produce any

83 protein crystal. However, when the same thing was done in the optimized buffer, many hits were found. The crystal growth conditions still have to be refined because crystals did not grow in the expansion trays. Most drops were clear though, even after a long time, which suggests that the protein may not be concentrated enough. The protein solution was tested for endogenous nuclease activity and it was found free of nucleases. As a result, co-crystallization with DNA was possible. This line of work was rapidly abandoned though, since the DNA retardation gel showed that PriA N could not bind the flap-DNA substrate on its own (even in a 3-fold excess). This contradicted suggestions from the literature that the N domain of PriA was the DNA binding domain. If cocrystallization was to be pursued, a DNA substrate would have to be found, for which PriA N has high affinity. As far as PriA NI domain is concerned, many experiments were done to find a host in which the protein would be expressed in a soluble form when a HisTag was added on either side. From this, it was concluded that the PriA NI with an N-terminal HisTag was insoluble and that the PriA NI QC with a C-terminal HisTag was soluble in BL21 RIL cells. This system was chosen to overproduce the protein. The protein expression was successful but problems were encountered during cell lysis. In the future, ammonium sulfate extraction and precipitation on ice should be used to obtain as much soluble protein as possible.

The

protein

can

then

be

easily

purified

by

nickel

affinity

chromatography in quantitative yields. The DLS results for this protein solution were not very good. One explanation could be that the HisTag tail may be

84 introducing some instability and may need to be cleaved-off before further experiments are done. But more importantly, the solubility profile of PriA NI QC has not been determined yet and the protein solution conditions have not been optimized. This may have a decisive effect on the behavior of the protein, just like it did for PriA N. So a lot of work remains to be done on this domain in the future but it would be worth the effort to take advantage of the presence of the HisTag. For example, the protein-protein interactions of the primosome could be studied by making binary complexes using PriA NI QC instead of the native PriA. These complexes could be purified by nickel affinity chromatography and then run on a size-exclusion column. This would be a very nice way to determine the stoichiometry of the binding, as an alternative to the native gel shift assays (see § 4.2.8) which are more complicated experiments. Of course, the crystallization of those complexes could also be tried. Finally, an important problem was solved regarding the “native” PriA NI domain, when contaminating nucleases were successfully removed from the solution. But like for PriA NI QC, the solubility of the protein has not been tested and optimized, so the solution is not yet ready for crystallization trials.

CHAPTER 4 E. coli PriB and PriC

4.1. Introduction 4.1.1.

Role of PriB and PriC in the primosome

PriB and PriC are two similar proteins with almost identical behavior and function. Their role is to serve as intermediates in the formation of the large primosomal complexes, first identified on the specific PAS site of the φX174 DNA, during PriA-induced replication restart mechanisms in E. coli47, 48. Results from the literature show that the binding of PriB and/or PriC constitutes the second of five steps in the sequential formation of the primosome3. Early studies of the complexes showed that neither of these proteins have intrinsic DNAbinding ability but the presence of either protein will stabilize the DNAyPriA complex and facilitate the subsequent binding of DnaT by triggering a conformational change of PriA3,

48

. The exact nature of these interactions

remains unclear, mostly because there are few experimental studies available on both proteins. This work aims at gathering more information on these mechanisms by studying the solution properties of PriB and PriC, as well as

85

86 reproducing in vitro binary protein complexes involving PriB and PriC. Very recently, the first crystal structure of PriB (a Histag fusion form of the protein) was solved to 2.1 Å by Liu and co-workers in Taiwan49. Two other PriB structures rapidly followed50, 51, all of which should provide further information to rationalize the results found here and give directions for future work.

4.1.2.

Comparison of PriB and PriC

It has been determined that two parallel pathways involving PriB and PriC are occurring during the second step of the formation of the primosome4, 5. Either PriB or PriC would bind the DNAyPriA complex and a subsequent conformational change of PriA then allows DnaT to bind to the complex. These pathways are redundant and it was argued that the two proteins were interchangeable (Figure 4x1). There is no evidence showing that both proteins would be or need to be present simultaneously in the same complex. Yet there is an important difference between PriB and PriC: if PriC is absent from the site of replication restart, the PriB pathway is still active, whereas if PriB is absent, the primosome cannot form on the DNA. In fact, PriC was found to be present during the assembly of the primosome on the φX174 DNA but was not required for the mechanism to occur. PriC only seems to have an overall stabilizing effect on the primosomal complex. As a consequence, the PriB pathway is referred to as the “major pathway”, since it is always taking place during DNA replication restart processes, whereas the PriC pathway is called the “minor pathway”. Moreover, it was proposed that PriC was involved in a PriA-independent pathway at

87 recombination intermediates, where the Rep protein substitutes for PriA and the rest of the mechanism is unchanged5.

Figure 4x1: Pathways in PriA-directed replication restart mechanisms φX174 -type DNA DnaC/DnaB

+ PriB

+ DnaT

+ PriA

+ DnaG

+ DnaB

+ PriC - DnaC Replication restart (DNA polymerase III holoenzyme)

„ Major pathway „ Minor pathway

The general features of PriB and PriC are recapitulated in Table 4x1 (from Swiss-Prot information files #P07013 PRIB_ECOLI and #P23862 PRIC_ECOLI). Figure 4x2 shows the 3-dimensional structure of the PriB dimer solved in 2004. The crystal structure of PriC is unknown.

Table 4x1: Characteristics of PriB and PriC MW (kDa)

Length

Number of cysteines

pI

Α280

State in solution

Stability

PriB

11.3

103 aa

4/monomer

8.5

0.524

dimer

unstable

PriC

20.2

174 aa

3

10.0

1.193

monomer

unstable

88 As seen on Figure 4.2a, PriB is a homodimer with twisted β-sheet structure (it was also reported as a β-barrel). The overall structure is much entangled. Each monomer has an OB-fold and is covalently linked to the other by two symmetrical disulfide bonds in the middle of the β-sheet. The interface between the two monomers is tightly packed by hydrophobic interactions in the β-sheet. Interestingly, the structure of the PriB monomer is almost identical to the N-terminal DNA-binding domain of the E. coli SSB, although they only share 13% sequence identity and PriB is unable to form homotetramers51. With this recent structural evidence, the binding of PriB to ssDNA was reexamined. Although the typical ssDNA binding-pocket found in numerous SSB structures was not present in PriB, it still appears to bind ss-DNA in a sequence-specific manner, as well as ss-RNA49. In addition, it was established that PriB is involved in protein-protein interactions with SSB on SSB-coated φX174 DNA. Figure 4x2b shows the charge distribution on the dimer. One side of the β-sheet is positively charged, containing ten of the twelve arginines and the eight lysine residues. The opposite side is negatively charged with the four aspartates and ten of the twelve glutamate residue. Preliminary results show that these surfaces are good candidates for protein-protein interactions with the basic PriA and acidic DnaT.

89 Figure 4x2: 3-dimensional structure of the E. coli PriB (PDB code 1WOC) Figure 4x2a: Ribbon diagram of the PriB dimer 1, 2

Figure generated with the programs MOLSCRIPT and RENDER using Raster 3D

NA

CA

NB

.

CB

Figure 4x2b: PriB surface potentials Adapted from Lopper, M., J. M. Holton, et al. Structure (Camb) 2004, 12, (11), 1967-75.

“TOP” view

„ Positive charge (10 R + 8 K)

in Figure 4x2a (A)

„ Negitive charge (10 E + 4 D)

When

beginning

this

work,

no

structural information was available. The cloning of PriB and PriC had already

90 been done and the project started at the level of protein expression. Protocols for the expression and purification of the primosomal proteins had been published28 but they involved a large number of steps and were not very efficient in terms of yield. To overome these limitations, my task was to design new, simplified and more efficient experimental methods for the expression and purification of these two proteins. Once this was achieved, further studies (like solubility studies, crystallization trials and biophysical characterization) could be carried out to understand their function in greater detail.

4.2. PriB 4.2.1.

Protein expression

The E. coli PriB gene (EG 10764, 315 bp) was cloned in pET 21a vectors and sent to us by our collaborator Dr. Nakai. The plasmids were transformed into Rosetta cells. Large amounts of the protein were obtained from a glycerol stock of the cells using the large scale preparation technique (§ 2.2.1). The cells were grown at 37 °C in 25 g/L LB, 1 mM Amp and 1 mM CAM until the OD600 reached about 0.6. The culture was then induced with 1 mM IPTG, grown for another 3 hr, harvested and kept at -20 °C until cell lysis. PriB showed relatively limited overexpression at 11 kDa but this was enough to produce a good amount of pure protein (Figure 4x3, lanes 2, 3 and 4). A less intense band could also be seen in the 0 hr sample presumably due to

91 leaky expression control. A typical large scale preparation of PriB yielded 15 g to 18 g of cells for 6 L of LB medium.

4.2.2.

Cell lysis and ammonium sulfate precipitation

The PriB Rosetta cells were lysed using the cell lysis with ammonium sulfate precipitation protocol (section 2.3.2). The precipitated material (pellet, Figure 4x3, lane 5) was collected and gently resuspended in TEDG buffer until a

clear solution was obtained. This material was frozen at -80 °C until purification.

Figure 4x3: SDS-PAGE gels of the expression and cell lysis of PriB

Lanes:

11 kDa

1

MW standard

2

0 hr sample

3

3 hr sample (1)

4

3 hr sample (2)

5

Pellet after (NH4)2SO4 precipitation

6

Lysate after (NH4)2SO4 precipitation

As seen on Figure 4x3, the ammonium sulfate precipitation step was quantitative (lanes 5 and 6). Using this protocol, PriB could be separated from more soluble proteins, prior to HPLC purification.

92

4.2.3.

HPLC purification

The HPLC purification of a crude extract of PriB obtained after ammonium

sulfate

precipitation

consisted

in

two

cation-exchange

chromatography steps followed by a size-exclusion chromatography step:

SP Sepharose (Low-resolution cation exchange)

POROS HS (High-resolution cation exchange)

Superdex 75 (Size exclusion)

For cation-exchange chromatography, the purification of the protein was achieved using a linear salt gradient (section 2.4.4). Buffer A was a standard low salt buffer and Buffer B was a high salt buffer. For the size exclusion chromatography, a single running buffer was used (Table 4x2). Before each cation-exchange column was run, the conductivity of the protein sample was measured and lowered if necessary to that of Buffer A by dilution with 50 mM Tris-HCl pH 7.5. The chromatograms obtained after each purification step are given in Figure 4x4. The fractions of interest were collected and their purity was checked by SDS-PAGE gel electrophoresis.

93 By using this purification method, it was possible to get a large amount of almost pure protein rather quickly. The yield was typically 30 mg of protein from 25 g of bacteria cells. This was a major improvement compared to previous methods.

Table 4x2: HPLC purification of PriB - Buffers Anion exchange

Buffer A

Buffer B

Composition

50 mM Tris-HCl pH 7.5

50 mM Tris-HCl pH 7.5

-

1.5 M NaCl

100 mM NH4Cl

100 mM NH4Cl

10 mM MgCl2

10 mM MgCl2

1 mM DTT (or BME)

1 mM DTT (or BME)

Conductivity

9.5 mS/cm

76 mS/cm

Size exclusion

Buffer A

Composition

50 mM Tris-HCl pH 7.5 50 mM NH4Cl 10 mM MgCl2 1 mM DTT(or BME)

Conductivity

5.5 mS/cm

94 Figure 4x4: HPLC purification of PriB - Chromatograms and SDS-PAGE gels Figure 4x4a: Step 1 SP Sepharose

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Fractions 25 26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

25.0 % Buffer B

* * * *

0.450

* *

0.450

0.400

0.400

Cond.

20.0 0.350

0.350

OD280

0.300 15.0

0.300

0.250

0.200

0.250

OD260

10.0

0.150

0.100

0.150

0.100

5.0

0.050

-0.000

0.200

0.050

0.0

-0.000

-0.050

-0.050

-5.0 AU 0.00

10.00

20.00

30.00

Min.Tenth

40.00

4x4a: The protein was found in fractions 9 to 21 (peak on the left). Most of the impurities were found in the flow-through fraction. The peak at the end of the run represented DNA (OD260 in purple > OD280 in green).

PriB

AU

95 Figure 4x4b: Step 2 POROS HS

1

2

3

4

5

6

7

8

9

10

11

12

13

14

0.70

15

16

17

18

19

20

21

*

35.0 % Buffer B

22

23

24

Fractions 25 26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

0.70

* *

0.60

0.60

OD280

30.0

50

0.50

0.50 25.0

0.40

0.40 20.0

OD260

0.30

0.30

15.0

0.20

10.0

Cond.

0.20

0.10

5.0

0.10

-0.00

0.0

-0.00

-5.0 AU

0.00

10.00

Min.Tenth

20.00

4x4b: The gel indicates that PriB eluted in fractions 21 to 28, although some impurities were still present. These fractions were collected and run on Superdex 75, including fraction 21 because the high molecular weight impurity should be easily removed by size exclusion chromatography.

PriB

AU

96 Figure 4x4c: Step 3 Superdex 75

1

2

3

4

5

6

7

0.70

8

9

10

*

35.0 % Buffer B

11

12

13

14

15

16

*

0.60

17

18

19

20

21

22

Fractions 23 24

25

26

27

28

29

30 31

35

37

39

41

43

45

47

49

51

53

55

57

59

0.70

******

0.60

OD280

30.0

33

0.50

0.50 25.0

OD260

0.40 20.0

0.30

0.20

0.40

0.30

15.0

0.20

10.0

Cond.

0.10

5.0

0.10

-0.00

0.0

-0.00

-0.10

-5.0

AU

-0.10 30.00

Min.Tenth

60.00

AU

4x4c: The maximum of the peak was found at the limit between fractions 20 and 21, which corresponds exactly to a mass of PriB (11-12 kDa). The main impurity was removed almost entirely (fraction 10). Interestingly, several peaks can be seen between fractions 16 and 24, but PriB was the main component. These fractions were all collected. In the end, the protein was about 95 % pure.

97

4.2.4.

Solubility screen

A solubility screen was performed on the purified PriB to obtain the solubility profile of the protein at room temperature. It can be seen on Figure 4x5, that the solubility of PriB was low in any of the conditions screened. So the experiment was repeated at 4 °C to determine if the temperature would affect the results. But it was clear from Figure 4x6 that decreasing the temperature did not improve the solubility of PriB. The results at room temperature suggest using NH4Cl and MgCl2 in a buffer like Na PIPES at pH 6.5 to optimize the protein solution.

Figure 4x5: Solubility screen of PriB at room temperature

0.02

Supernatant H2O

0.00

Taps pH 8.5

0.00

Hepes pH 7.5

0.00 0.10

Salt / Buffer

Pipes pH 6.5 Mes pH 5.6

0.09

Na Citrate

0.09

Na Phosphate

0.09 0.16

Na Sulfate 0.06

Na Cacodylate Na Acetate

0.00 0.05

Na Formate

0.09

CaCl2

0.23

MgCl2 LiCl KCl NaCl

0.01 0.07 0.09 0.13

NH4Cl 0.00

0.10

0.20

0.30 Absorbance at 595 nm

0.40

0.50

0.60

98

Figure 4x6: Solubility screen of PriB at 4 °C

0.05

Supernatant H2O

0.00

Taps pH 8.5

0.00

Hepes pH 7.5

0.00 0.10

Pipes pH 6.5 0.06

Mes pH 5.6

0.09

Salt / Buffer

Na Citrate

0.09

Na Phosphate 0.03

Na Sulfate

0.04

Na Cacodylate 0.00

Na Acetate

0.08

Na Formate

0.06

CaCl2

0.05

MgCl2 0.00

LiCl

0.02

KCl

0.04

NaCl 0.00

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

0.60

Absorbance at 595 nm

4.2.5.

Dialysis

As expected, difficulties arose in finding good buffer conditions to keep PriB soluble. About 500 µL of the HPLC purified sample was dialyzed for 3 to 4 hr in the optimized conditions deduced from the solubility screen. This dialysis buffer was composed of 50 mM Bis-Tris pH 6.5, 50 mM NH4Cl and 10 mM MgCl2. The protein precipitated in the dialysis cassette and the concentration of ammonium chloride was increased stepwise to 100 mM and then 150 mM in the dialysis buffer. This also failed, even after an overnight dialysis. Later, three

99 experiments were setup in parallel to try to solve this problem: a small amount of precipitated PriB solution was aliquoted into three micro-centrifuge tubes. The first tube was kept at room temperature and NaCl was slowly added to the solution by increments of 100 mM until a concentration of 1 M was attained. Each time, the solution was mixed very gently to redissolve the precipitate. The second tube was also kept at room temperature and the third tube was kept on ice. NH4Cl was added to both tubes as described above for NaCl. By slowly mixing the sample over a long period of time, the protein went gradually back into solution in all three experiments. It was concluded from these experiments that the best way to stabilize PriB in solution in its native form was to use high salt conditions. Thereafter, the buffer used was 50 mM Bis-Tris pH 6.5, 1 M NH4Cl and 10 mM MgCl2.

4.2.6.

Concentration

It was difficult to concentrate PriB on its own, even in the high salt buffer, since it could easily precipitate. The maximum concentration that could be reached was about 11 mg/mL but most of the time is was around 5 mg/mL. The solution was concentrated in a 3500 MWCO Amicon® Ultra concentrator (15 mL, Millipore) at 5000xg. The exact concentration of the solution was measured by UV-Visible spectroscopy. This “high salt” protein sample was rather stable and could be stored at 4 °C for a few weeks. Freezing the protein in dry-ice with 15% glycerol was not attempted.

100 4.2.7.

Dynamic Light Scattering

Since PriB was so unstable in solution, it seemed essential that further studies were done. DLS would be of great help in a case like this, because it provides information on the homogeneity of the solution and on possible aggregation states. Two DLS measurements were done on PriB: one at 20 °C and one at 4 °C. A sample of PriB was made in 50 mM Bis-Tris pH 6.5, 150 mM NH4Cl and 10 mM MgCl2 at a concentration of 1 mg/mL. It was filtered in a single-use Ultrafree®-MC filter (yellow type, pore size = 0.1 µm, Millipore) to remove any large particle that would interfere with the measurement. The results are presented in Table 4x3. In both cases, the molecular weight of the predominant species was about twice that of PriB (the monomer is ~11 kDa), meaning that PriB was indeed present as a dimer in solution. At 20 °C, the polydispersity of the solution was really good, in the ideal range of 15-20%,

but it increased significantly at 4 °C. This means that the solution was very homogeneous at room temperature and the protein solution was likely to crystallize (see discussion).

Table 4x3: DLS results for PriB

Temperature

Rh (nm)

MW (kDa)

%poly

Baseline

SOS

Distribution

4 °C

2.552

28.07

27.4

1.005

10.2

Bimodal

20 °C

2.318

22.22

19.5

1.005

37.0

Bimodal

101 4.2.8. Study of the protein-protein interactions involving PriB After attempting many times to produce a stable sample of the native PriB with limited success and since its crystal structure had just been solved, it seemed more important at this point to study the protein-protein interactions involved in the formation of the primosome. PriB is believed to be found in two binary protein complexes, namely PriAyPriB (in a 1:2 molar ratio) and PriByDnaT (in a 2:3 molar ratio). However, this has only been demonstrated in the presence of DNA25, 48. Attempts were made to show if the same interactions are happening in the absence of DNA. Native gel mobility shift assays were designed to determine the proper stoichiometry of the two complexes. The technique consists in running an Agarose gel under non-denaturing conditions to allow each protein (or complex of proteins) to migrate in the electric field according to its isoelectric point (pI). Proteins that have an acidic pI would run toward the cathode (+) and proteins that have a basic pI would run toward the anode (-). So there is a difference in the distances traveled if a protein is free in solution or if it is bound in a complex, especially when it carries a distinct overall electric charge.

The procedure was as follows: a 0.6% (w/v) Agarose gel was prepared by dissolving 0.24 g of Analytical SeaKem GTG® Agarose into 40 mL of 1X TAE pH 8 running buffer (gel #1) or into 40 mL of 1X Tris-Acetate pH 6.5 running buffer (gels #2 and 3). It was boiled and poured in the gel compartment, with the wells placed in the middle and left to solidify for half an hour. Each protein

102 sample to be run was prepared in ice. It was composed of 5 µL of proteins + buffer (total volume), 1 µL of GLS and 2 µL of 80% glycerol and was loaded entirely on the gel. The gel was run at 4 °C and 50 V for 2 hr (gel #1) or for 5 hr (gels #2 and #3). All the gels were fixed in 0.05% SDS for 30 min and stained overnight in a fresh solution of SYPRO® Orange dye (Molecular Probes), as described in the manufacturer’s manual. The gels were photographed the next day under fluorescent light. The experiments on PriB binary complexes were done using a high salt sample of PriB (5.5 mg/mL stock solution) and a PriA and DnaT sample in their optimized buffers (11.5 mg/mL and 22.5 mg/mL respectively). The results are shown on Figure 4x8.

Figure 4x7: Diagram of a native gel

(+) Cathode

Low pI proteins

High pI proteins

Anode (-)

103 Figure 4x8: Native gel electrophoresis Figure 4x8a: Native gel #1 - PriAyPriB and PriByDnaT

1 2 3 4 5 6

7 8 9 10 11 12 13 14 15 16

+

-

Lanes: 1

PriA

0.1 mM

9

PriB:DnaT

1:1

2

PriB

0.1 mM

10

PriB:DnaT

1:2

3

PriA:PriB

1:1

11

PriB:DnaT

1:3

4

PriA:PriB

1:2

12

PriB:DnaT

1:6

5

PriA:PriB

1:3

13

PriB:DnaT

2:1

6

PriA:PriB

1:4

14

PriB:DnaT

3:1

7

PriA:PriB

1:6

15

PriB:DnaT

6:1

8

PriA:PriB

1:10

16

DnaT

0.1 mM

On the first half of the gel, it can be seen that neither PriA nor PriB migrated effectively, although PriA seemed to move toward the anode (lane 1). PriB could not be detected by itself (or stayed in the well) but a band shifting toward the cathode was visible on lanes 7 and 8. More interestingly, the right half of the gel showed that DnaT ran well (lane 16) and a definite band shift was observed when PriB is in a 3:1 or a 6:1 excess (lanes 14 and 15).

104 Figure 4x8b: Native gel #2 - PriByDnaT

1

2

3

4

5

6

7

+

-

Lanes: 1

PriA

0.1 mM

2

PriB

0.1 mM

3

DnaT

0.1 mM

4

PriB:DnaT

1:1

5

PriB:DnaT

1:2

6

PriB:DnaT

1:3

7

PriB:DnaT

2:1

8

PriB:DnaT

2:3

For this gel, modified running conditions were used (Tris Acetate pH 6.5, longer running time), which were more appropriate for the study of PriAyPriB and PriByDnaT complexes. Indeed, at this lower pH, it can be seen that PriA ran towards the anode, that DnaT ran towards the cathode and that PriB was now detectable and ran slightly towards the cathode. However, the lanes testing the PriByDnaT complex did not show any conclusive result.

105 Figure 4x8c: Native gel #3 - PriByDnaT

1 2 3 4 5 6

7 8 9 10 11 12 13 14 15 16

+

Lanes: 1

PriB

0.1 mM

9

PriB:DnaT

2:1

2

DnaT

0.1 mM

10

PriB:DnaT

3:1

3

PriB:DnaT

1:1

11

PriB:DnaT

6:1

4

PriB:DnaT

1:2

12

PriB:DnaT

9:1

5

PriB:DnaT

1:3

13

PriB:DnaT

12:1

6

PriB:DnaT

1:6

14

PriB:DnaT

2:3

7

PriB:DnaT

1:9

15

PriB:DnaT

2:6

8

PriB:DnaT

1:12

16

PriB:DnaT

2:9

This gel showed the same migration for the free PriB and DnaT as the previous one. A new band was visible that may represent the complex at about half the distance that DnaT would normally travel (lanes 3, 4 and 6 to 10) but the migration pattern was rather complicated. First of all, PriB could only be detected at higher concentrations. Secondly, an excess of PriB was needed to observe a band shift. Partial binding was seen at a ratio of at least 3:1 and full binding was

106 seen at a ratio of 6:1. How does this relate to the fact, confirmed by my own results, that PriB is a dimer and DnaT is a trimer in solution? Indeed there was no band shift at the 2:3 ratio (lane 14). A PriB dimer may have bound only one DnaT monomer and masked its signal partially (see also gel #1, lane 14). Complete binding occurred when PriB was present in a 6-fold excess. This would mean that there are two or three dimers of PriB binding to a single monomer of DnaT. In fact, this can be compared to the stoichiometries reported for the primosome in the presence of DNA. In the case of DnaT, a 1:0.33 ratio was measured25. Since DnaT forms a trimer in solution, two scenarios were proposed, but the actual ratio has not been proved yet. The above value raised the question whether one monomer of dnaT binds to each primosome or if one trimer of DnaT binds every three primosomes. This native gel experiment would strongly support the first possibility and the fact that even though DnaT is a trimer by itself, the biologically relevant form is the monomer. As far as PriB is concerned, two dimers were reported to be present in each primosome. Since the 4:1 ratio was not tested, a definitive conclusion is not possible but it would concur with the observation that full binding happens between the 3:1 and 6:1 ratios of PriB to DnaT.

4.3. PriC 4.3.1.

Protein expression

The E. coli PriC gene (EG 10765, 528 bp) was cloned in pET 21a vectors and was also sent to us by Dr. Nakai. The plasmids were transformed

107 into Rosetta competent cells. Large amounts of the protein were obtained from a glycerol stock of the cells using the method described in Section 2.2.1. The PriC Rosetta cells were grown at 37 °C in 25 g/L LB, 1 mM Amp and 1 mM CAM until the OD600 reaches about 0.6. The cells were induced with 1 mM IPTG and grown for 3 hr, harvested and kept at -20 °C until cell lysis was performed. The expression of PriC was acceptable (Figure 4x9, lanes 2 and 3). As for PriB, a leaky expression band was seen before induction with IPTG. The amount of cells collected was around 8 g to 10 g for 6 L of LB medium.

4.3.2.

Cell lysis and ammonium sulfate precipitation

The lysis of PriC Rosetta cells was the same as for PriB (§ 4.2.2). Most of the protein was found in the pellet after the ammonium sulfate precipitation (Figure 4x9, lane 4) and many soluble impurities could be removed (lane 5).

Figure 4x9: SDS-PAGE gel of the expression and cell lysis of PriC

Lanes:

20 kDa

1

MW standard

2

0 hr sample

3

3 hr sample

4

Pellet after (NH4)2SO4 precipitation

5

Lysate after (NH4)2SO4 precipitation

108 4.3.3.

HPLC purification

The HPLC purification of a crude extract of PriC consists in two cationexchange steps: a low-resolution SP Sepharose column followed by a highresolution POROS HS column. Before each run, the conductivity of the protein sample was checked and lowered to that of buffer A by addition of the necessary amount of 50 mM Tris-HCl pH 7.5. The same salt gradient was used for both runs (Buffer A, low salt to Buffer B, high salt). At the end, the fractions of interest were pooled together and their purity was estimated by SDS-PAGE gel electrophoresis.

Table 4x4: HPLC purification of PriC - Buffers Anion exchange

Buffer A

Buffer B

Composition

50 mM Tris-HCl pH 7.5

50 mM Tris-HCl pH 7.5

-

1.5 M NaCl

100 mM NH4Cl

100 mM NH4Cl

1 mM DTT (or BME)

1 mM DTT (or BME)

9.5 mS/cm

75 mS/cm

Conductivity Composition

50 mM Tris-HCl pH 7.5

(alternative buffer B) 3

750 mM NH4Cl 1 mM DTT (or BME)

Conductivity

3

85 mS/cm

Second buffer B used, as a result of the solubility screen. Both buffers worked similarly.

109 Figure 4x10: HPLC purification of PriC - Chromatograms and SDS-PAGE gels Figure 4x10a: Step 1 SP Sepharose

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Fractions 25 26 27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49 50

2.00

2.00 100.0% Buffer B

* * * * * * *

1.75

1.50

1.75

1.50

OD260

1.25

1.25

OD280

1.00

1.00

50.0

0.75

0.75

Cond. 0.50

0.50

0.25

0.25

0.00

0.0

0.00

-0.25

AU

-0.25

0.00

10.00

20.00

Min.Tenth

30.00

40.00

50.00

AU

4x10a: A very good purification was achieved after SP Sepharose already. The protein was collected in fractions 19 to 29. Most of the contaminating proteins seen in the load fraction (L) ended up in the flowthrough.

PriC

110 Figure 4x10b: Step 2 POROS HS 1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

Fractions 25 26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

40.0 % Buffer B

*

0.350 35.0

*

0.300

0.350

0.300 30.0

0.250

0.250 25.0

OD260

0.200 20.0

0.150

0.100

OD280 0.200

0.150

15.0

10.0

Cond.

0.100

0.050

5.0

0.050

-0.000

0.0

-0.000

-0.050

-5.0

-0.050

AU

10.00

Min.Tenth

20.00

4x10b: After the high efficiency POROS HS, a 99% pure sample was obtained in fractions 32 to 50. The quantity of protein purified with this method can be up to 50 mg starting with 10 g of cells, depending on the level of protein expression. This is a much more efficient protocol than those from the literature.

PriC

AU

111

4.3.4.

Solubility screen

A solubility screen was performed on the purified PriC at room temperature. From the solubility profile (Figure 4x11), it was deduced that the optimized buffer to prepare the protein solution should be made of 25 mM TAPS pH 8.5 and 50 mM NH4Cl.

Figure 4x11: Solubility screen of PriC at room temperature

0.14

Supernatant

0.26

H2O

0.85

Taps pH 8.5 0.62

Hepes pH 7.5 0.14

Pipes pH 6.5

0.44

Mes pH 5.6

0.46

Salt / Buffer

Na Citrate 0.14

Na Phosphate Na Sulfate

0.00 0.53

Na Cacodylate 0.45

Na Acetate 0.40

Na Formate CaCl2

0.02

MgCl2

0.00

LiCl

0.00

KCl

0.00

NaCl

0.00 0.63

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

Absorbance at 595 nm

4.3.5.

Protein preparation

A small sample of purified PriC was successfully dialyzed in a dialysis cassette for 3 hours in the above dialysis buffer. However, when the same experiment was repeated on a larger sample (also using a snake skin dialysis

112 tubing instead of a dialysis cassette due to the volume of the sample), the protein precipitated and despite many attempts, it was never recovered. As a result, the remaining sample from the HPLC purification was not dialyzed into these low salt conditions and was stored directly in the elution buffer at 4 °C. The concentration of the solution was measured to be 0.2 mg/mL. This very dilute sample remained stable for at least two weeks and it was used to make the PriAyPriC and PriCyDnaT protein complexes (§ 4.3.7 below). In fact we believed that it was by forming these complexes, rather than by keeping it on its own, that we would be able to stabilize PriC in solution. This type of approach had been successful in the past and since PriC (and PriB) are not enzymes, the only reason for them to exist is as part of the primosome. Therefore it might be very difficult to keep them separated in vitro at such high concentrations.

4.3.6. Secondary structure prediction In E. coli, PriB and PriC are two redundant proteins. As seen in this study, their similarities are not limited to their function: they also behave the same for the most part. Sometimes during structural evolution, proteins with the same function appeared from completely different shapes and vice versa. Given the 3-dimensional structure of PriB (a dimer of OB-folds with extensive β-sheet), one would expect to find a similar tertiary structure in PriC although the two proteins have essentially no primary sequence identity. The results should be taken with caution though, since the accuracy of the secondary structure

113 prediction models used for the calculations is still limited. The data would only be useful in case some striking pattern was found. The secondary structure prediction of PriC was calculated from its primary

sequence

on

the

PredictProtein

Server

website

(www.embl-

heidelberg.de/predictprotein/predictprotein.html). The results obtained were in fact rather interesting. According to that model, PriC would be a non-globular protein composed of 86.2% of α-helices, 0% of β-strands and 13.8% of other secondary features (e.g. loops). The reliability index for most of the prediction was really high: 63% of the entire sequence and 80% of the supposed α-helical regions were scored with an 8 or a 9 on a scale from 0 to 9. This was quite unexpected and would have to be confirmed by circular dichroism or by X-ray diffraction. If it is the case, PriB and PriC would represent a very nice example of convergent evolution, where the two proteins evolved from entirely different structures to perform similar functions.

4.3.7. Protein-protein complexes The work described in this section focuses on the two binary complexes of interest: PriAyPriC and PriCyDnaT. These complexes are part of a much larger entity (the primosome) when PriA is bound to the DNA. In that case the molar ratios are 1:1 for PriAyPriC and 1:3 for PriCyDnaT25. But there is no information showing if the complexes could still form in the absence of DNA, or whether PriC and DnaT would even interact in the absence of PriA. For instance, their affinity for each other is still unknown. Thus, the goal was to reproduce the

114 two

complexes

in

vitro

for

subsequent

crystallization

and

biophysical

characterization. The PriAyPriC complex was made by directly mixing the two proteins together in a 1:1 molar ratio at room temperature. That is, the correct volume of PriC (high salt, 0.2 mg/mL = 0.010 mM) was measured and put in a centrifuge tube. The corresponding amount of PriA (low salt, 10 mg/mL = 0.122 mM) was added to the tube and slowly mixed with PriC. The solution was left at room temperature for about 30 min. The sample was then dialyzed against PriA dialysis buffer (10 mM Na PIPES pH 6.5, 100 mM NH4Cl, 25 mM MgSO4) for 4 hr. finally, the solution was concentrated to about 8 mg/mL, filtered and stored at 4 °C or flash-frozen in dry-ice with 15% glycerol to be stored at -80 °C. The PriCyDnaT complex was made in a similar way with a molar ratio PriC:DnaT of 1:3. The correct volume of PriC (high salt, 0.2 mg/mL = 0.010 mM) was measured and put in a centrifuge tube. The corresponding amount of DnaT (low salt, 16.7 mg/mL = 0.858 mM) was added to the tube and slowly mixed with PriC. The solution was left at room temperature for half an hour and dialyzed against DnaT dialysis buffer (25 mM Na HEPES pH 7.5, 100 mM NH4Cl) for 4 hr. The solution was concentrated to 13 mg/mL, filtered and stored as described above. In both cases, the concentration of the complex was determined by measuring the OD280 of the solution. The molar extinction coefficients, given in 52

Table 4x5, were calculated by the following method :

ε280 = aεTyr,280 + bεTrp,280 + cεCys,280

115 with

a = number of tyrosine residues in the macromolecule, εTyr,280 = 1280 M-1cm-1, molar extinction coefficient of tyrosine at 280 nm, b = number of tryptophan residues in the macromolecule, εTrp,280=5690 M-1cm-1, molar extinction coefficient of tryptophan at 280 nm, c = number of cysteine residues in the macromolecule, εCys,280 = 120 M-1cm-1, molar extinction coefficient of cysteine at 280 nm.

Table 4x5: Molar extinction coefficients Tyrosine residues

Tryptophan residues

Cysteine residues

ε280(M-1cm-1)

Calc. A280

PriA

15

15

11

105150

1.285

PriC

1

4

3

24160

1.193

DnaT

4

4

0

27880

1.432

PriAyPriC (1:1)

16

19

14

129430

1.268

PriCyDnaT (1:3)

13

16

3

107800

1.379

4.4. Discussion The main goals of my research project on PriB and PriC were achieved and I learned a lot about protein chemistry with these two proteins. The expression, cell lysis and purification protocols available in the literature provided a base for my work. Both proteins could be produced and purified satisfactorily in amounts large enough for crystallographic studies. However, major solubility

116 problems were encountered all along. The first strategy employed was to stabilize the proteins in high salt solution conditions. Unfortunately this would prevent any subsequent crystallization experiments, since the salt would mask other crystallization parameters. In the articles where the structure of PriB was solved, the solution was also stored in high salt conditions. But in all three cases, the protein had a HisTag or was a seleno-methionine derivative, which apparently allowed dialyzing the protein in relatively low salt conditions prior to crystallization. The reported crystal structure of PriB confirmed most of my results and helped make a lot more sense of all the findings. In particular, the DLS results agreed very well with a dimeric structure in solution, a protein hydrodynamic radius of ~23 Å and the presence of some sort of equilibrium in solution between the dimer and other forms of the protein (low polydispersity, bimodal)51. But this work also showed that PriB is subject to aggregation at low temperature (higher polydispersity, larger particle size and higher molecular weight), which is also consistent with the results of the solubility screens. Perhaps it can be explained by the net charges carried on each side of the dimer. Low salt conditions might favor electrostatic interactions between molecules and promote the aggregation phenomenon; in contrast, high salt concentrations would disrupt these interactions thus stabilizing the protein. This charge distribution also explains the behavior of PriB in the native gels. With such a basic pI (8.5), PriB was expected to shift very clearly toward the anode. In fact, it was quite the opposite. The PriB samples stayed in the wells with a smear shifting slightly toward the cathode.

117 This proves that the protein behaves more like a neutral species and that the two surface potentials probably balance each other out. The study of PriAyPriB using native gel electrophoresis was not completed because PriB could no be concentrated enough at the time. But the work that was done on PriByDnaT shows that the complex is definitely not forming in a 2:3 molar ratio as it might have been expected. It seems rather that two or three dimers of PriB are bound to only one monomer of DnaT. The experimental approach used for PriC was the same as PriB, with similar results. It can be mentioned though, that a PriC monomer is about the same size as a PriB dimer. They have the same function and they interact in the same manner in the primosomal complex. Does that imply comparable 3-dimentional shapes and interfaces for protein-protein interactions? This question remains to be answered. On the other hand, it was an amazement to see how smoothly the PriAyPriC and PriCyDnaT complexes were made starting from the high salt PriC solution. Only minimal optimization was required for the dialysis conditions. The concentrations obtained were suitable for crystallization trials. These experiments were performed on a test scale and can be repeated on larger amounts in the future, in order to setup crystal screens. Overall, the preliminary results on the binary complexes are very promising and should be the main focus of future investigations. Higher concentrations of PriB can be used in native gel shift assays to determine accurately the stoichiometry of the PriAyPriB complex. The stoichiometry of PriAyPriC and PriCyDnaT complexes should also be confirmed easily (the pI of

118 PriC is 10) by the same technique. Fluorescence polarization or Isothermal Titration Calorimetry experiments could be carried out to measure the different association constants. Last but not least, the two binary complexes involving PriC practically ready for crystallizations trials and those involving PriB could probably be made the same way.

CHAPTER 5 E. coli DnaT

5.1. Introduction The E. coli DnaT protein is composed of 179 amino acids and has a mass of 19.5 kDa. Since it does not have any cysteines, it cannot be subject to cross-linking or be involved in protein-protein interactions through disulfide bridges. It is a rather acidic protein with a calculated pI of 5.14. The calculated extinction coefficient used was ε280 = 27880 M-1cm-1 and the calculated absorbance is A280 = 1.432 for a 1 mg/mL solution (Swiss-Prot information file #P07904 DNAT_ECOLI). The structure of DnaT is unknown and the biologically active form of the protein is still subject to debate, but it was found that DnaT forms a stable trimer in solution. CD analysis showed that DnaT was a protein with α/β secondary structure53.

5.2. Role of DnaT in the primosome Similarly to PriB and PriC, DnaT does not have any DNA recognition on its own3 but it is believed to be a protein link between the DNAyPriAyPriB complex (or the DNAyPriAyPriC complex) and the DnaB helicase protein, 119

120 allowing a complete preprimosome to form (Figure 5x1). The assembly of the complex occurs in a sequential manner: a ternary complex of PriB or PriC bound the DNAyPriA complex undergoes a conformational change (supposedly on PriA) that favors the binding of DnaT, which occurs next. DnaT can provide a binding site for DnaB only after it is bound the large DNA-protein complex. It was also observed that if DnaT is in very large excess, it has the ability to bypass the role of PriB or PriC and bind directly to the DNAyPriA complex48. However, it is still unclear whether only one DnaT monomer binds to each primosome or if the whole trimer binds only one out of three primosomes25.

Figure 5x1: Formation of the preprimosome on φX174-type DNA

pas-DNA

121

5.3. Protein expression DnaT constructs were received from Dr. Nakai. The E. coli DnaT gene (EG 10244, 540 bp) had been cloned in pET 21a vectors. The plasmids were transformed into Rosetta competent cells. Large amounts of the protein were obtained from a glycerol stock of the cells using the large scale preparation method (§ 2.2.1). The cells were grown at 37 °C in 25 g/L LB, 1 mM Amp and 1 mM CAM until the OD600 reached about 0.6. The cells were induced with 1 mM IPTG and grown for 3 more hr, harvested and kept at -20 °C until cell lysis was performed. DnaT always showed high levels of overexpression (Figure 5x2, lanes 1 and 2). Typically, 10 g to 13 g of bacterial cells were obtained from 6 L of LB.

5.4. Cell lysis and ammonium sulfate precipitation The DnaT Rosetta cells were lysed using the cell lysis with ammonium sulfate precipitation method (§ 2.3.2) in the same way as PriB and PriC. At the end of the procedure, the protein was recovered quantitatively and it had already been separated from many soluble impurities (Figure 5x2, lane 3). The final pellet fraction was resuspended in TEDG buffer, frozen in dry-ice and stored at -80 °C until purification.

122 Figure 5x2: SDS-PAGE gel of the expression and cell lysis of DnaT

Lanes:

DnaT

1

0 hr sample

2

3 hr sample

3

Pellet after (NH4)2SO4 precipitation

4

Lysate after (NH4)2SO4 precipitation

5

MW standard

21.5 kDa

5.5. HPLC purification Given the very good results of the ammonium sulfate precipitation, the resolubilized pellet was already partially clean. So the purification of DnaT was done in one step using the high-resolution POROS HS cation-exchange column. A linear salt gradient was used from Buffer A (50 mM Tris-HCl pH 7.5, 100 mM NH4Cl, 10 mM MgCl2, 1 mM DTT; conductivity = 10.5 mS/cm) to Buffer B (50 mM Tris-HCl pH 7.5, 1.5 M NaCl, 100 mM NH4Cl, 10 mM MgCl2, 1 mM DTT; conductivity = 75 mS/cm). The chromatogram and SDS-PAGE gel obtained are shown on Figure 5x3.

123 Figure 5x3: HPLC purification of DnaT - Chromatogram and SDS-PAGE gel Fractions 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 2.00

2.00 100.0% Buffer B

1.75

*******

1.75

OD280

1.50

1.50

1.25

1.25

OD260

1.00

1.00

50.0

0.75

0.75

0.50

0.50

Cond. 0.25

0.25

0.00

0.0

0.00

-0.25

-0.25

AU 0.00

kDa: 31.0 21.5

10.00

Min.Tenth

20.00

5x3: The protein was found in fractions 17 to 22, which were pooled together. It was almost 100% pure at this point. The total yield was usually high, with about 120 mg of pure protein from 12 g of cells.

AU

124

5.6. Solubility screen The standard 16-condition solubility screen was performed on DnaT at room temperature to optimize the protein solution prior to crystal screening. The results are presented in Figure 5x4. DnaT actually has a very nice solubility profile because in some of the conditions, it is not soluble at all, whereas in other ones it is highly soluble. For instance, the pH profile shows that it is far more soluble at high pH than at low pH. The trend with cations is also interesting: DnaT is not soluble in divalent cationic salts; whereas it has a rather high solubility in any of the monovalent cationic salts tried. From this, it was deduced that the optimized buffer should contain Na HEPES pH 7.5 and either KCl or NH4Cl. However, if possible, KCl should be avoided because it is insoluble in the presence of SDS.

Figure 5x4: Solubility screen of DnaT at room temperature 0.25

Supernatant

0.41

Taps pH 8.5

0.43

Hepes pH 7.5 Pipes pH 6.5 Mes pH 5.6

0.00 0.01 0.04

NaCitrate

Salt / Buffer

Na Phosphate

0.01 0.25

Na Sulfate 0.09

Na Cacodylate

0.28

Na Acetate 0.13

Na Formate CaCl2

0.00

MgCl2

0.00 0.59

LiCl

0.71

KCl 0.43

NaCl

0.48

NH4Cl 0.00

0.10

0.20

0.30

0.40

Absorbance at 595nm

0.50

0.60

0.70

0.80

125

5.7. Protein preparation The purified DnaT solution was dialyzed for 3 to 4 hours at room temperature against the optimized buffer composed of 25 mM Na HEPES pH 7.5, 50 mM NH4Cl and 2 mM EDTA. In these solution conditions, the maximum solubility of DnaT was about 45 mg/mL. This indicated that the optimal concentration for crystal screen experiments would be around 20 to 22 mg/mL. So the final sample was concentrated at room temperature until a concentration of about 20 mg/mL was reached and then filtered. Some of the solution was frozen in dry-ice with 15% glycerol and stored at -80 °C until later use. The rest was stored at 4 °C. The DnaT solution was very stable, even when kept at room temperature for a few weeks.

5.8. Dynamic Light Scattering DLS studies on DnaT were used to determine if the solution was homogeneous and likely to crystallize and also to investigate possible aggregation phenomenon at low temperature. For the DLS, DnaT samples were made in 25 mM HEPES pH 7.5, 50 mM NH4Cl and 2 mM EDTA at the concentrations indicated in Table 5x1 and filtered in a single-use Ultrafree®-MC filter (yellow type, pore size = 0.1 µm, Millipore). Several initial measurements were done at 20 °C and at 4 °C. The results are summarized in Table 5x1 (see also section 2.10). As they were not very satisfying, many salts and additives were added to the protein

126 as indicated, to try to improve the homogeneity of the solution. It can be seen that none of the experiments worked really well (very high % poly, baseline and SOS values) and the presence of big aggregates was observed (large molecular weight > 20 times the size of the protein). In some cases, the measurements could not even be taken, due to excessive counts that were attributed to very large aggregates. A monomodal solution with low polydispersity could not be obtained. Due to the overall poor quality of the measurements, it could not be confirmed that DnaT was a trimer in solution. However, in the two experiments done using the DnaT sample that had not been frozen, some particles of very large weight were saturating the signal but represented less than 3% of the mass. The other 97% were found to have a size of about 50 to 60 kDa and could be attributed to the DnaT trimer (the polydispersity was also lower for those peaks). Finally, it can be noted that the addition of 20% glycerol and 25% ethylene glycol helped reduce the polydispersity a lot, bringing it down to an acceptable value around 22-26% rather than 45-60% in the other measurements. This suggested that these solution conditions are much more likely to produce crystals than any others.

127

Table 5x1: DLS results for DnaT T (°C)

Conc. (mg/mL)

Rh (nm)

MW (kDa)

% Poly

Distribution

Baseline

SOS

BSA standard

20

2.0

3.791

73.29

15.1

Monomodal

1.002

2.4

DnaT (not frozen)

4

5.0

9.868

746.6

59.7

Bimodal

1.027

11.4

DnaT (not frozen)

20

5.0

8.429

509.3

58.5

Bimodal

1.005

22.0

DnaT (frozen)

20

5.0

10.08

786.7

52.1

Bimodal

1.010

10.2

DnaT + 20% Glycerol

20

1.6

8.388

503.3

26.1

Bimodal

1.022

87.5

DnaT + 25% Ethylene Glycol

20

1.5

10.02

775.7

22.1

Bimodal

1.032

78.7

DnaT + 10% 1,6-Hexanediol

20

1.8

N/A

N/A

N/A

Multimodal

N/A

N/A

DnaT + 50 mM (NH4)2SO4

20

2.0

7.278

356.7

49.2

Bimodal

1.012

63.6

DnaT + 100 mM (NH4)2SO4

20

2.0

7.167

343.7

50.0

Bimodal

1.015

54.2

DnaT + 150 mM (NH4)2SO4

20

2.0

N/A

N/A

N/A

Multimodal

N/A

N/A

DnaT + 50 mM LiCl

20

2.0

7.068

332.2

49.2

Bimodal

1.010

70.8

DnaT + 100 mM LiCl

20

2.0

6.963

320.4

50.8

Bimodal

1.015

63.8

DnaT + 150 mM LiCl

20

2.0

7.117

337.9

49.3

Bimodal

1.019

61.7

DnaT + 100 mM NH4Cl

20

2.0

8.283

488.2

52.7

Bimodal

1.010

27.8

DnaT + 50 mM NH4OAc

20

2.0

7.173

344.4

49.2

Bimodal

1.010

62.2

DnaT + 100 mM NH4OAc

20

2.0

7.992

447.6

45.0

Bimodal

1.014

69.8

DnaT + 150 mM NH4OAc

20

2.0

N/A

N/A

N/A

Multimodal

N/A

N/A

DnaT + 50 mM KCl

20

2.0

7.220

349.8

46.7

Bimodal

1.013

66.8

DnaT + 100 mM KCl

20

2.0

7.038

328.8

48.3

Bimodal

1.016

65.6

DnaT + 150 mM KCl

20

2.0

7.059

331.3

48.1

Bimodal

1.027

65.6

128

5.9. Differential Scanning Calorimetry DSC was used to study the denaturation of DnaT as a function of the temperature. Since it was clear that the protein was unstable at low temperature, this would also provide information on the upper temperature limit. It should show if the trimer dissociates into monomers and unfolds all at once, around the same temperature, or if there are two distinct phenomenon and/or transition states. The experimental procedure is detailed in Section 2.10. The protein sample was made in DnaT optimized buffer (25 mM HEPES pH 7.5, 50 mM NH4Cl and 2 mM EDTA). The concentration in monomer was 60 µM and consequently the concentration in trimer was 20 µM. First, a blank run was done with the buffer in the analysis cell (without protein) and then the same run was done with the protein solution in the cell. In each case, the temperature was ramped up from 10 °C to 90 °C and back down and the heat released was measured. The calorimetric curves obtained are shown in Figure 5x5, along with the calculated thermodynamic data. Figure 5x5a shows that the transition happens between 50 °C and

55 °C. When the baseline is subtracted (Figure 5x5b), a sharp denaturation peak is observed at a melting temperature of 51.4 °C (Tm). The thermodynamic data was analyzed using two different models: the “2-state, 1 peak” model (Figure 5x5c) and the “non-2-state, 1 peak” model (Figure 5x5d). They give the value of

∆H, the heat change per mole of protein and of ∆Hv, the heat change per unfolding unit.

129 Figure 5x5: DSC curves of DnaT Figure 5x5a: Raw data 0.0000

50.5 °C

-0.0005

DnaT raw data Reference

o

Cp(cal/ C)

-0.0010

-0.0015

55.8 °C

-0.0020

-0.0025

-0.0030 0

20

40

60

80

100

o

Temperature ( C) Figure 5x5b: Denaturation curve

30

51.4 °C 25

o

Cp (kcal/mole/ C)

20

DnaT (monomer) 60µM 15

10

5

0

-5 10

20

30

40

50

60

70 o

Temperature ( C)

80

90

100

130 Figure 5x5c: Thermodynamic analysis - Model: 2-state 1 peak

30 DnaT m onom er Model: M2State Chi^2 = 969235 Tm 51.48 ±0.0365 Delta H 1.53E5 ±1.1E3

25

o

Cp (kcal/mole/ C)

20

DnaT monomer 60 µ M Fit 1 peak

15

10

5

0

-5 10

20

30

40

50

60

70

80

90

100

o

Tem perature ( C)

Figure 5x5d: Thermodynamic analysis monomer/trimer - Model: non-2-state 1 peak

30 Data: DnaT monomer Model: MN2State Chi^2 = 829163 Tm 51.50 ±0.0320 DeltaH 1.43E5 ±2.01E3 DeltaH v 1.66E5 ±2.91E3

25

o

Cp (kcal/mole/ C)

20

DnaT monomer 60µM Fit 1 peak

15

10

5

0

-5 10

20

30

40

50

60

70 o

Temperature ( C)

80

90

100

131

90 80 70

Data: DnaT trimer Model: MN2State Chi^2 = 7.51952E6 51.54 ±0.0321 Tm Delta H 4.24E5 ±6.03E3 Delta Hv 1.67E5±2.96E3

DnaT trimer 20µM Fit 1 peak

o

Cp (kcal/mole/ C)

60 50 40 30 20 10 0 10

20

30

40

50

60

70

80

90

100

o

Temperature ( C)

Both models fit the data very well for the monomer concentration. In that case, the thermodynamic values were: ∆H = 1.53E5 ± 1.1E3 cal/mol (2-state), ∆H = 1.43E5 ± 2.0E3 cal/mol and ∆Hv = 1.66E5 ± 2.9E3 cal/mol (non-2-state). But only the non-2-state model gave a satisfying analysis when using the trimer concentration: ∆H = 4.24E5 ± 6.0E3 cal/mol and ∆Hv = 1.67E5 ± 3.0E3 cal/mol. These results are consistent with the precedents (∆Hv is unchanged and ∆H is roughly three times bigger). Since ∆H/ ∆Hv ~ 1 for the monomer and ~ 3 for the trimer, it can be concluded that the dissociation of the trimer and the unfolding of the protein are happening simultaneously (with no transition state), at Tm = 51.4 °C.

132

5.10. Crystallization 5.10.1. Crystal screens Crystal screens of DnaT were setup according to Table 5x2. The first trays were setup manually with using a 12 mg/mL protein solution and the later ones were prepared using the crystallization screening robots, with an 18 mg/mL protein solution. The trays were checked regularly and the crystal hits recorded. Overall, about 50 protein hits were found. Figure 5x6 shows the pictures of some of the best crystals obtained in these preliminary screens with the corresponding crystallization conditions.

Table 5x2: Crystal screens of DnaT Tray ID

Temperature

Type of tray

Kits

AI 007

Hits?

RT

Corning

Wizard I and II

4

AI 008

RT

Corning

Cryo I and II

3

AI 009

RT

Corning

Crystal Screen I and II

9

AI 010

RT

Corning

Ion screen and Natrix

7

AI 011

4 °C

Corning

Cryo I and II

8

AI 012

4 °C

Corning

Crystal Screen I and II

4

AI 013

4 °C

Corning

Wizard I and II

AI 014

4 °C

Corning

Ion screen and Natrix

2

AI 057

RT

3-well Greiner

Index

5

AI 058

RT

3-well Greiner

Additive screen

AI 059

RT

3-well Greiner

Crystal Screen Cryo and MembFac

No

No 7

133 Figure 5x6: DnaT crystal screen hits

Wizard II #13 15% (v/v) Ethanol,

Crystal Screen II #3 25% (v/v) Ethylene glycol

0.1 M Citrate pH 5.5, 0.2 M Li2SO4

Cryo II #46

MembFac #39

30% (v/v) PEG-200,

0.1 M K, Na Tartrate

0.1 M CAPS pH 10.5,

0.1 M HEPES pH 7.5

0.2 M (NH4)2SO4

0.1 M Li2SO4

Natrix #20

Index #14

10% (v/v) 1,6-Hexanediol,

0.5 M Mg Formate

0.05 M Cacodylate pH 6.5,

0.1 M Bis-Tris pH 6.5

0.1 M Ammonium Acetate, 0.015 M Mg Acetate

134

5.10.2. Optimization of crystal growth conditions Each condition had to be refined precisely from the results of the crystal screens, to try to produce the best crystals possible for X-ray diffraction. The expansions of DnaT were done in 24-well Costar trays with 18 mm coverslides, using the hanging-drop vapor diffusion technique. The protein concentration was between 18 and 22 mg/mL. At first, the trays were prepared with 4x6 A/B gradients and then when good crystals were grown, the 2x12 and 1x24 gradient formats were used to find the exact growth conditions required. The two best conditions from the screens were investigated first (previous page, top two pictures). The ethylene glycol condition did not turn out as well as expected, although it indicated that ethylene glycol would be the cryoprotectant of choice, once good crystals were obtained. The “ethanol” condition (15% (v/v) ethanol, 0.1 M Na Citrate pH 5.5, 0.2 M Li2SO4) on the other hand gave promising results. As seen in Figure 5x7a, medium plate-like crystals were easily grown in the refined conditions: 1% to 5% Ethanol, 100 mM to 200 mM Na citrate pH 5.5, 200 mM to 400 mM Li2SO4 (with or without 20% ethylene glycol as the cryoprotectant). However, single 3-D crystals could not be obtained, despite all the attempts made (modifying the pH of the solution, the size of the drops, the type of tray used, using microseeding, etc…). Still, several of these crystals were frozen in 20% ethylene glycol and tested for X-ray diffraction (see § 5.10.3). Expansions of all the other hits found in the crystal screens were also performed, without producing any better result. So, being able to set up new crystal screens using the robotic systems a few months later was the occasion to start over with

135 new crystallization conditions. Indeed, it was very exciting to see that several drops produced rather big 3-dimentional crystals directly in the screens (see p.133, colored pictures). After expansion of these conditions, the best crystals were made from drop #39 of the Membfac™ screen. These were nice single crystals, large enough to be tested on the diffractometer (Figure 5x7b, 0.2x0.3 mm in size). The optimized growth conditions were: 100 mM K, Na Tartrate, 100 mM Li2SO4 (up to 200 mM) and 100 mM Na HEPES with a pH gradient between 7.5 and 8.2. These crystals were frozen in 25% glucose for X-ray diffraction experiments.

Figure 5x7: DnaT crystals Figure 5x7a: Plate crystals obtained when expanding the Wizard II condition #13

The crystals were checked using the blue Izit crystal dye™.

Plate-like crystals used for X-ray diffraction (0.1 mm to 0.3 mm)

136 Figure 5x7b: 3-D crystals obtained when expanding the Membfac condition #39 (pH gradient)

pH 8.19

pH 8.08

pH 7.80

pH 8.10

pH 8.13

pH 8.16

137 5.10.3. Crystal manipulation and X-ray diffraction screening Protein crystals are by nature very fragile and handling them properly requires a lot of practice, patience and concentration. Each crystal to be looped has to be measured approximately to choose the correct size of nylon loop that will be used (from 0.1 mm to 0.7 mm). The DnaT crystals suitable for X-ray diffraction were looped out of the crystallization drop and quickly dipped into a small drop containing the well solution and the cryoprotectant, as described in section 2.9. They were then flash-frozen in liquid nitrogen and stored until tested on the diffractometer. The first crystals to be screened were some of the plates seen on Figure 5x7a. This was done at 100 K at the APS synchrotron (Argonne National

Laboratories) on Bio-CARS beamline 14-BM-C. No diffraction was observed. The 3-D crystals obtained later (Figure 5x7b) were screened on the FR-E diffractometer of the OMCC at the University of Toledo. A set of four crystals was tested for diffraction at 100 K using the CCD detector. Each time 3 images were taken at 0, 45 and 90° angles (5 to 10 seconds exposure). The same was done with another set of four crystals using the image plate detector to allow for longer exposure (several minutes). There was no diffraction observed for any of the crystals. Finally, one crystal was mounted in a 1 mm capillary and put on the Image plate side of the FR-E diffractometer to test for room temperature diffraction (Figure 5x8). In that case, very faint diffraction was visible, but the spots were scattered and of very low resolution, showing that the crystal was disordered.

138 Figure 5x8: DnaT crystal mounted in a capillary

Crystal size: ~ 0.05 x 0.1 mm (Plate)

5.11. Discussion The production of large amount of pure DnaT worked very well. High levels of expression were observed and the cell lysis was quantitative. But the protein was found to be unstable at low temperature, something that was also noticed during the dropwise addition of saturated ammonium sulfate: by doing it in ice, the amount of precipitate was so large that the increasing cloudiness of the solution could be easily seen by the naked eye. Much less precipitate was obtained when the same was done at room temperature. The protein was then purified very efficiently in one step on a POROS HS cation exchange column. It should be mentioned that even though the pI of DnaT is 5.14, it did not stick to the anion exchange POROS HQ but had good affinity for POROS HS. The dialysis buffer was easily deduced from the solubility screen and the preparation of the protein solution was straightforward. The DSC experiment showed that the dissociation of the trimer and the unfolding of the protein occurred simultaneously in one step at 51.5 °C. A study

139 of the denaturation of DnaT was published in 199653, also showing that there was only one unfolding event with no transition state but at 57 °C. They found that the unfolding path depended on many parameters. For example, heat denaturation was compared to Gu-HCl denaturation and they had different outcomes. The DLS showed that the protein solution was not homogeneous and that DnaT encountered huge aggregation at low temperature, making the crystallization experiments at 4 °C less likely to succeed than at room temperature. Any attempt to solve this problem was unsuccessful and a monodisperse solution could not be made. The focus for crystallization was on room temperature screening. Many crystal hits were found. The best ones were reproduced, in an effort to determine the best range of conditions that would produce large single diffraction quality crystals. This step was extremely time-consuming because the results were never entirely satisfying. A total of four conditions were investigated in detail and many parameters were modified. In each case, while the aesthetic morphology of the crystals improved, none of them diffracted, either in cryogenic conditions or at room temperature. This could be due to loose or disordered crystal packing or very high solvent content and would also explain why most of the 3-D crystals redissolved spontaneously after few days. Even though the crystals were already present in the drop, the solution was still in a dynamic equilibrium. Another explanation could be that the crystals were packed in such a manner that “patches” come randomly next to one another because they are complementary in shape, but the overall structure is completely random. The diffraction would then be incoherent.

CHAPTER 6 A preliminary screen used to optimize protein solubility and improve crystallization results

6.1. Introduction Crystallization of proteins and protein complexes is a multi-parametric problem that involves the investigation of a vast number of physical and chemical parameters. Among other factors, the purity of the protein sample plays a major role in the quality of the crystals produced and can strongly influence the nucleation processes54. If better purity can be achieved, the protein solution will be more homogeneous, therefore enhancing the chances of getting good quality crystals. The traditional approach to crystallization screening experiments consists in setting up several sitting-drop crystal trays, in order to test as many conditions as possible and to use the smallest amount of protein. The trays are set aside for a couple of days and checked for hits regularly over the following weeks. If the conditions are suitable, crystals will grow by vapor diffusion over a long period of time. Once crystal hits have been identified, they can be replicated using the hanging-drop vapor diffusion technique which allows growing larger, fewer crystals of X-ray diffraction quality. But even with the latest technological 140

141 advances, crystal screening remains the bottleneck of every crystallization process. However, our task has been made a lot easier in the past years by the use of commercially available crystallization kits. These kits contain hundreds of unique crystallization conditions that are known to have worked for other proteins in the past, designed according to sparse-matrix distributions or incomplete factorials55,

56

. Recently, a breakthrough was made thanks to high-tech

crystallization robotic systems57. This is now saving crystallographers hours of tedious hand work. Even so, there is still no rationalized method that ensures the success of the screening experiments for a given protein. And some proteins will just not crystallize in any of those conditions. On the other hand, the buffers, salts and additives used to prepare the protein solution are inherent to the sample and will be present in every crystallization condition. In many cases, we have observed a correlation between maximizing the solubility of a protein and the success of subsequent crystal screening. Using a quick and simple preliminary solubility screen31, the solubility profile of a protein can be determined in various salt and buffer conditions. The protein is dialyzed into the optimal buffer and salt solution deduced from the solubility screen and then concentrated. First of all, this method makes possible to greatly improve the solubility of most proteins in their native state prior to crystallization experiments. Moreover, since optimal solution conditions can be identified, it is possible to uncouple the presence of the buffer and salt (needed primarily to stabilize the protein in solution) from some very different parameters that will directly affect its crystallization. Therefore, screening the effect of

142 precipitating agents and/or additives in crystal screens can be done, without involving irrelevant parameters from the protein solution itself. To test this hypothesis, a new 96-condition crystal screen was designed, using two precipitating agents (PEG-4000 and MPD) and various types of additives. Ten commercially available proteins were used. The comparison of crystallization drops in optimized solution conditions vs. standard solution conditions (Tris-HCl, NaCl) showed a clear improvement in the quality of crystals in the initial screens. For the proteins that effectively crystallized, in many cases, the optimized protein solution yielded large diffraction quality crystals directly in the screens, whereas in the same condition, the standard solution produced only clear drops, precipitates or micro-crystals at best. This technique was also successfully applied to several other proteins involved in the group’s research projects on DNA replication. It became a routine technique to improve the quality of the protein solutions and the success of preliminary crystal screens.

6.2. Experimental methods To perform this study, ten test proteins were purchased (see Table 6x1). The stocks were either in a powder form or in solution. These proteins

were chosen because they can be used as standards for crystallization experiments, since they are known to crystallize easily and the conditions have been well characterized.

143 Table 6x1: Test proteins Protein

Organism

Form

Source

Ovalbumin

Chicken

Powder

Sigma

Myoglobin

Horse

Powder

Sigma

Catalase

Bovine

Powder

Sigma

α-Lactalbumin

Bovine milk

Powder

Sigma

Pepsin

Porcine stomach

Powder

Sigma

Trypsin

Bovine pancreas

Powder

Sigma

Thaumatin

Thaumatococcus daniellii

Powder

Sigma

Subtilisin Carlsberg

Bacillus licheniformis

Powder

Sigma

Xylanase

Tricoderma sp.

Solution

Hampton research

Glucose Isomerase

Streptomyces rubiginosus

Solution

Hampton research

6.2.1. Preparation of the protein solutions All the experiments were carried out at room temperature. The Xylanase and Glucose Isomerase commercial solutions were dialyzed against distilled water overnight and filtered. The other protein solutions were prepared by dissolving the powder in distilled water until saturation and then filtered. To prevent its degradation, protease inhibitor (AEBSF) was added to Trypsin before filtration. In the case of Myoglobin however, the solution turned dark brown because it was getting oxidized with the air58. To avoid this, a fresh sample was prepared in water and a few crystals of sodium azide were added. The protein solution was then filtered over a Sephadex gel column. No subsequent oxidation was visible and the solution stayed bright red for days after treatment.

144 6.2.2. Solubility screen This method is based on a “reverse solubilization” technique. Starting from a precipitated protein sample, it is possible to measure how much protein goes back into solution upon addition of a particular buffer or salt. As a result, a solubility profile can be determined directly from absorbance measurements and the best conditions can be identified for every protein. First, 150 µL of protein solution was forced to precipitate by adding 190 µL of 40% PEG 8000. An 18 µL aliquot of this material was added to each of 17 Eppendorf tubes. Then, 2 µL of the corresponding 1 M salt/buffer solution was added to each tube and mixed well to resuspend the precipitated protein (see Table 6x2). At this point, the concentrations in each 20 µL sample are 20% PEG 8000 and 100 mM salt/buffer. The solutions were incubated at room temperature for 20 minutes and centrifuged at 20000xg for 4 minutes. The supernatant of each tube was tested for soluble protein using a BioRad Protein Assay: 5 µL of protein sample (or 10 µL when the signal was too small to be measured accurately by the spectrophotometer) was mixed with 995 µL of 1X BioRad Protein Assay reagent in a disposable cuvette. The samples were incubated for 5 minutes and absorbances were measured at 595 nm using an Agilent 8453 UV-Vis spectrophotometer (Agilent Technologies).

145 Table 6x2: The 17-condition solubility screen Cationic salts

Anionic salts

Buffers

1 - NH4Cl

7 - Na Formate

13 - Na MES pH 5.6

2 - NaCl

8 - Na Acetate

14 - Na PIPES pH 6.5

3 - KCl

9 - Na Cacodylate

15 - Na HEPES pH 7.5

4 - LiCl

10 - Na Sulfate

16 - Na TAPS pH 8.5

5 - MgCl2

11 - Na Phosphate

17 - H2O

6 - CaCl2

12 - Na Citrate

6.2.3. Optimized buffer and maximum solubility The salt and buffer giving the highest solubility values for each protein comprise the “optimized buffer” for that protein. After this was determined, two solutions were prepared for each protein (as described above): one in a standard chromatography buffer composed of 50 mM Tris-HCl pH 7.5 and 100 mM NaCl; the other in the optimized buffer (100 mM salt, 50 mM buffer). The maximum solubility value was then measured for both solutions by concentrating a small sample in a 10000 MWCO concentrator at 10000xg until it either precipitated or the concentration leveled off. Crystal screens will be set up with a protein concentration around ½ the maximum solubility.

6.2.4. Design of the Additive/Precipitating Agent crystallization screen In order to investigate the importance of precipitating agents and additives in the crystallization conditions, a new type of crystal screen was designed. It is composed of 96 different conditions (standard 12x8 format) to be

146 easily dispensed from a deep-well block into a tray. The block is divided into three major sections, each containing 32 conditions. The first section screens the effect of four types of precipitating agents (PEGs, phosphates at different pH, high concentration salts and alcohols) in the absence of additives. The second section screens various additives in the presence of 20% PEG 4000 as the precipitating agent. The third section screens the same additives in the presence of 30% MPD as the precipitating agent. There are 10 types of additives tested in the screen: polyamines, polyacids, sugars, zwitterions, metallic ions, reducing agents, organic compounds, detergents, cofactors and cryoprotectants. Figure 6x1 shows a diagram of the 96-condition screen and Table 6x3 lists the

composition of every crystallization condition.

147

Figure 6x1: Additive/Precipitating Agent screen - 96 condition block

Precipitating Agents

PEGs

(6*2)

20% PEG 4000 + Additives

Na, K Phosphate (4)

Poly amines

30% MPD + Additives

Poly acids

Zwitterions

Metallic ions

(1+4)

(5)

(4)

(7) High salts

Alcohols

Poly amines

Poly acids

Zwitterions

Metallic ions

(1+4)

(5)

(4)

(7) Reducing Agent

(4*2)

(4*2) Poly acid

Reducing Agent

Sugars

Detergent

(2)

Sugars

Detergent

(2)

(4)

Cofactor

Organics

(4)

Cofactor

Organics

Cryoprotectant

(2)

Cryoprotectant

(2)

Poly acids

148 Table 6x3: Additive/Precipitating Agent screen - List of conditions Condition

Precipitating agent

Additive

1

20% w/v PEG 400

-

2

40% w/v PEG 400

-

3

20% w/v PEG 1000

-

4

40% w/v PEG 1000

-

5

1M Ammonium Sulfate

-

6

2M Ammonium Sulfate

-

7

10% Ethanol

-

8

20% Ethanol

-

9

10% w/v PEG 4000

-

10

20% w/v PEG 4000

-

11

10% w/v PEG 6000

-

12

20% w/v PEG 6000

-

13

1M Ammonium Phosphate Dibasic

-

14

2M Ammonium Phosphate Dibasic

-

15

10% Isopropanol

-

16

20% Isopropanol

-

17

10% w/v PEG 8000

-

18

20% w/v PEG 8000

-

19

10% w/v PEG 20000

-

20

20% w/v PEG 20000

-

21

1M Na, K Tartrate

-

22

2M Na, K Tartrate

-

23

15% Ethylene Glycol

-

24

30% Ethylene Glycol

-

25

2M Na, K Phosphate pH 5.8

-

26

2M Na, K Phosphate pH 6.5

-

27

2M Na, K Phosphate pH 7.9

-

28

2M Na, K Phosphate pH 8.9

-

29

1M Lithium sulfate

-

30

2M Lithium sulfate

-

31

15% v/v 2-Methyl-2,4-pentanediol

-

32

30% v/v 2-Methyl-2,4-pentanediol

-

33

20% w/v PEG 4000

25mM Spermine

34

20% w/v PEG 4000

25mM Spermidine

35

20% w/v PEG 4000

25mM Imidazole

36

20% w/v PEG 4000

25mM Urea

37

20% w/v PEG 4000

25mM Guanidine-HCl

38

20% w/v PEG 4000

25mM L-Arginine

39

20% w/v PEG 4000

10% v/v Triethylamine

40

20% w/v PEG 4000

25mM Sodium Thiosulfate

41

20% w/v PEG 4000

25mM Sodium Malonate

42

20% w/v PEG 4000

25mM Sodium Oxalate

43

20% w/v PEG 4000

25mM Sodium Citrate

44

20% w/v PEG 4000

25mM Taurine

45

20% w/v PEG 4000

10% w/v D-Glucose

46

20% w/v PEG 4000

10% w/v D-Sucrose

47 48

20% w/v PEG 4000 20% w/v PEG 4000

10% w/v Dextrose 10% w/v Xylitol

149

Condition

Precipitating agent

Additive

49

20% w/v PEG 4000

25mM Glycine

50

20% w/v PEG 4000

25mM D-Alanine

51

20% w/v PEG 4000

25mM γ-aminobutyric acid

52

20% w/v PEG 4000

25mM ε-amino caproic acid

53

20% w/v PEG 4000

25mM Sodium Thiocyanate

54

20% w/v PEG 4000

2mM Octyl-β-glucoside

55

20% w/v PEG 4000

2mM Adenosine Triphosphate

56

20% w/v PEG 4000

10% v/v Glycerol

57

20% w/v PEG 4000

10mM Calcium Chloride

58

20% w/v PEG 4000

10mM Zinc Chloride

59

20% w/v PEG 4000

10mM Potassium Fluoride

60

20% w/v PEG 4000

10mM Sodium Iodide

61

20% w/v PEG 4000

2mM DTT

62

20% w/v PEG 4000

2mM TCEP-HCl

63

20% w/v PEG 4000

10% v/v Dioxane

64

20% w/v PEG 4000

10% v/v Dimethylsulfoxide

65

30% v/v 2-Methyl-2,4-pentanediol

25mM Spermine

66

30% v/v 2-Methyl-2,4-pentanediol

25mM Spermidine

67

30% v/v 2-Methyl-2,4-pentanediol

25mM Imidazole

68

30% v/v 2-Methyl-2,4-pentanediol

25mM Urea

69

30% v/v 2-Methyl-2,4-pentanediol

25mM Guanidine-HCl

70

30% v/v 2-Methyl-2,4-pentanediol

25mM L-Arginine

71

30% v/v 2-Methyl-2,4-pentanediol

10% v/v Triethylamine

72

30% v/v 2-Methyl-2,4-pentanediol

25mM Sodium Thiosulfate

73

30% v/v 2-Methyl-2,4-pentanediol

25mM Sodium Malonate

74

30% v/v 2-Methyl-2,4-pentanediol

25mM Sodium Oxalate

75

30% v/v 2-Methyl-2,4-pentanediol

25mM Sodium Citrate

76

30% v/v 2-Methyl-2,4-pentanediol

25mM Taurine

77

30% v/v 2-Methyl-2,4-pentanediol

10% w/v D-Glucose

78

30% v/v 2-Methyl-2,4-pentanediol

10% w/v D-Sucrose

79

30% v/v 2-Methyl-2,4-pentanediol

10% w/v Dextrose

80

30% v/v 2-Methyl-2,4-pentanediol

10% w/v Xylitol

81

30% v/v 2-Methyl-2,4-pentanediol

25mM Glycine

82

30% v/v 2-Methyl-2,4-pentanediol

25mM D-Alanine

83

30% v/v 2-Methyl-2,4-pentanediol

25mM γ-aminobutyric acid

84

30% v/v 2-Methyl-2,4-pentanediol

25mM ε-amino caproic acid

85

30% v/v 2-Methyl-2,4-pentanediol

25mM Sodium Thiocyanate

86

30% v/v 2-Methyl-2,4-pentanediol

2mM Octyl-β-glucoside

87

30% v/v 2-Methyl-2,4-pentanediol

2mM Adenosine Triphosphate

88

30% v/v 2-Methyl-2,4-pentanediol

10% v/v Glycerol

89

30% v/v 2-Methyl-2,4-pentanediol

10mM Calcium Chloride

90

30% v/v 2-Methyl-2,4-pentanediol

10mM Zinc Chloride

91

30% v/v 2-Methyl-2,4-pentanediol

10mM Potassium Fluoride

92

30% v/v 2-Methyl-2,4-pentanediol

10mM Sodium Iodide

93

30% v/v 2-Methyl-2,4-pentanediol

2mM DTT

94

30% v/v 2-Methyl-2,4-pentanediol

2mM TCEP-HCl

95

30% v/v 2-Methyl-2,4-pentanediol

10% v/v Dioxane

96

30% v/v 2-Methyl-2,4-pentanediol

10% v/v Dimethylsulfoxide

150 6.2.5. Crystal screens All the crystallization experiments were performed using the state-ofthe-art robotic systems from the Ohio Macromolecular Crystallography Consortium (OMCC) at the University of Toledo, Ohio. This facility, open since May 2004, includes a Cartesian dispensing system (Model 96C-550-4S, Genomic Solutions, Irvine,CA), a Honeybee sitting drop crystallization robot (Genomic Solutions, Irvine,CA), an automatic RS-3000 Plate sealer (Brandel, Gaithersburg,

MD)

and

a

Rhombix

digital

imager

(Kendro,

Asheville, NC) with hotel capacity to take pictures and store the trays (Table 6x4, Figure 6x2). X-ray diffraction experiments are carried out on the FR-E X-ray

diffractometer (Rigaku/MSC, Inc.) equipped with an R-Axis IV image plate detector with Max Flux multilayer mirrors and a Saturn 92 CCD detector with Osmic multilayer mirrors.

Table 6x4: Automated preparation of crystal trays

Instrument

Task

Time needed

1 - Cartesian dispensing system

Pour well solution into trays

< 1 min

2 - Honeybee sitting drop crystallization robot

Dispense protein and mother liquor drops

3 - Plate sealer

Seal trays with crystal clear tape

Corning ~ 20 min (1 well) Greiner ~ 25 min (3wells) Seconds

4 - Rhombix Digital Imager

Take high resolution pictures and store trays

A few minutes/tray (schedulable)

151 Figure 6x2: The Instrumentation Center of the OMCC

6x2a: Cartesian dispensing system (Genomic Solutions)

6x2b: Honeybee sitting drop crystallization robot (Genomic Solutions)

6x2c: Rhombix Digital Imager (Kendro)

6x2d: RS-3000 Plate sealer (Brandel)

6x2e: High brilliance FR-E X-ray diffractometer (Rigaku/MSC)

152 The crystal screens were set up at room temperature, in 96 well three-drop Greiner trays, using protein solutions at a concentration of about ½ their maximum solubility (refer to Table 6x5). For each protein, the standard buffer solution and optimized buffer solution were tested in the same tray (respectively in the left drop and right drop). Two crystallization screens were tested: the commercially available sparse-matrix Index™ Screen from Hampton Research and the Additive/Precipitating Agent screen designed for this study (see Section 7.2.4). The Cartesian robot was used to pour 100 µL well solution from a deepblock to the reservoir of the tray. The Honeybee robot was then used to dispense the drops in the tray (1 µL drops = 0.5 µL of well solution + 0.5 µL of protein solution). Once they were ready, the trays were stored at room temperature. The results were recorded after a week of growth at room temperature. Pictures were taken in bright and polarized light using the Rhombix imaging system. The drops were scored as follows: •

Clear drop = C

•

Precipitate (Light, heavy or granulate) = ppt

•

Phase separation = ps

•

Micro-crystals = µ

•

Crystals that need optimization (1-D, 2-D and small 3-D) = Op

•

Diffraction quality crystals (Medium and Large 3-D) = DQ

Finally, the number of drops from each category above was counted for all ten proteins and then compared. The pictures provide a direct comparison

153 of the size and quality of the crystals in standard vs. optimized solution conditions.

6.3. Results The results of the solubility screen for each protein are given in Figure 6x3. The relative absorbance values at 595 nm can be directly correlated to the

protein concentrations in solution (the same dilution factor was used for all conditions). The optimized conditions and respective maximum solubility for each protein are presented in Table 6x5 and a comparison with standard chromatography buffer conditions is shown in Figure 6x4. Throughout this study and during other work on DNA replication proteins, it became clear that for many proteins, sodium citrate gives really good results by increasing the solubility significantly. The results of the crystallization screening experiments were mixed. However, this is rather a good thing because it allows describing what happens for a wide range of proteins that have very different behavior. Out of ten proteins, eight crystallized in the INDEX™ screen and two gave only micro-crystals (Pepsin and Subtilisin) and in the Additive/Precipitating Agent screen, seven proteins crystallized and three produced micro-crystals (Ovalbumin, Pepsin and Trypsin). The overall crystallization results are presented in Table 6x6.

154 Figure 6x3: Solubility screen results OVALBUMIN 0.40

H2O

0.90

Taps pH 8.5

0.91

Hepes pH 7.5 0.32

Pipes pH 6.5 0.25

Mes pH 5.6

0.36

Salt / Buffer

Na Citrate

0.53

Na Phosphate 0.09

Na Sulfate

0.32

Na Cacodylate 0.15 0.15

Na Acetate Na Formate 0.06

CaCl2

0.08

MgCl2

0.13

LiCl KCl

0.02

NaCl

0.01

NH4Cl

0.03

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Absorbance at 595nm

MYOGLOBIN 0.04

H2O

0.64

Taps pH 8.5 0.05 0.08

Hepes pH 7.5 Pipes pH 6.5

0.09

Mes pH 5.6

0.07

Salt / Buffer

Na Citrate 0.04

Na Phosphate Na Sulfate

0.00 0.07

Na Cacodylate Na Acetate Na Formate CaCl2 MgCl2 LiCl

0.03 0.01 0.01 0.02 0.03

KCl

0.00

NaCl

0.00 0.00

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

Absorbance at 595nm

0.60

0.70

0.80

155

CATALASE 0.04

H2O

0.29

Taps pH 8.5 Hepes pH 7.5 Pipes pH 6.5 Mes pH 5.6

0.08 0.00 0.01 0.35

Salt / Buffer

Na Citrate Na Phosphate Na Sulfate

0.14 0.00 0.02

Na Cacodylate Na Acetate

0.00

Na Formate

0.00 0.04

CaCl2 MgCl2 LiCl KCl NaCl NH4Cl 0.00

0.02 0.01 0.00 0.02 0.00 0.10

0.20

0.30

0.40

0.50

0.60

Absorbance at 595nm

α-LACTALBUMIN H2O

0.06 0.38

Taps pH 8.5 0.28

Hepes pH 7.5 0.16

Pipes pH 6.5 0.12

Mes pH 5.6

0.51

Salt / Buffer

Na Citrate 0.27

Na Phosphate 0.11

Na Sulfate

0.26

Na Cacodylate Na Acetate

0.16

c

0.09

Na Formate

0.29 0.30

CaCl2 MgCl2 LiCl

0.08 0.21

KCl 0.18

NaCl

0.21

NH4Cl 0.00

0.10

0.20

0.30

0.40

Absorbance at 595nm

0.50

0.60

156

PEPSIN 0.02

H2O

0.63

Taps pH 8.5 Hepes pH 7.5

0.00

Pipes pH 6.5

0.00 0.00

Salt / Buffer

Mes pH 5.6 Na Citrate

0.00

Na Phosphate

0.00

Na Sulfate

0.00

Na Cacodylate

0.00

Na Acetate Na Formate

0.00 0.00

CaCl2

0.00

MgCl2

0.00 0.00

LiCl

0.00 0.00

KCl NaCl

0.02

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

Absorbance at 595nm

TRYPSIN H2O

0.18 0.11

Taps pH 8.5 Hepes pH 7.5

0.09

Pipes pH 6.5

0.09 0.09

Mes pH 5.6

0.15

Salt / Buffer

Na Citrate Na Phosphate Na Sulfate

0.05 0.03 0.13

Na Cacodylate 0.10

Na Acetate

0.11

Na Formate

0.10

CaCl2

0.11

MgCl2

0.10 0.12

LiCl KCl

0.12

NaCl

0.11

NH4Cl 0.00

0.10

0.20

0.30

0.40

Absorbance at 595nm

0.50

0.60

157

SUBTILISIN 0.25

H2O

0.32

Taps pH 8.5 0.27

Hepes pH 7.5

0.27

Pipes pH 6.5

0.26

Mes pH 5.6 0.15

Salt / Buffer

Na Citrate

0.23

Na Phosphate

0.23 0.24

Na Sulfate Na Cacodylate c

Na Acetate

0.27 0.25

Na Formate

0.40

CaCl2 0.22

MgCl2

0.26

LiCl 0.22

KCl

0.22

NaCl

0.20

NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

0.60

Absorbance at 595nm

THAUMATIN 0.48

H2O 0.28

Taps pH 8.5

0.36

Hepes pH 7.5

0.47

Pipes pH 6.5 0.37

Mes pH 5.6 0.20

Salt / Buffer

Na Citrate

0.25

Na Phosphate 0.19

Na Sulfate 0.13

Na Cacodylate

0.16

Na Acetate

0.16

Na Formate CaCl2

0.16 0.17

MgCl2

0.18

LiCl

0.16 0.15

KCl NaCl

0.17

NH4Cl 0.00

0.10

0.20

0.30

0.40

Absorbance at 595nm

0.50

0.60

158

XYLANASE 0.32 0.31

H2O Taps pH 8.5 0.23 0.23

Hepes pH 7.5 Pipes pH 6.5 Mes pH 5.6

0.29 0.23

Salt / Buffer

Na Citrate 0.17

Na Phosphate

0.25

Na Sulfate 0.17

Na Cacodylate 0.12

Na Acetate

0.25

Na Formate 0.22

CaCl2 0.12

MgCl2

0.19

LiCl 0.13

KCl

0.17 0.16

NaCl NH4Cl 0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.5

0.6

Absorbance at 595nm

GLUCOSE ISOMERASE 0.01

H2O

0.40

Taps pH 8.5 Hepes pH 7.5

0.21

Pipes pH 6.5

0.21 0.18

Mes pH 5.6

0.26

Salt / Buffer

Na Citrate

0.30

Na Phosphate 0.02

Na Sulfate

0.39

Na Cacodylate 0.31

Na Acetate 0.12

Na Formate 0.08

CaCl2 0.00

MgCl2

0.02 0.02

LiCl KCl

0.03 0.04

NaCl NH4Cl 0

0.1

0.2

0.3

0.4

Absorbance at 595nm

159

Table 6x5: Optimized buffers and maximum solubility Maximum Solubility in Optimized buffer (mg/mL) 120

Concentration used for crystal screens (mg/mL) 75

50mM Na TAPS pH 8.5 -

73

N/A

51

50mM Na TAPS pH 8.5 -

86

40

11

5

50mM Na TAPS pH 8.5 100mM Na Citrate

40

20 (10*)

α-Lactalbumin

22

11

50mM Na TAPS pH 8.5 100mM Na Citrate

41

20

Pepsin

66

30

50mM Na TAPS pH 8.5 -

58

30

Trypsin

46

23

50mM Na TAPS pH 8.5 100mM Na Citrate

50

25

Thaumatin

17

8

50mM Na PIPES pH 6.5 -

50

23

Subtilisin Carlsberg

17

8

50mM Na TAPS pH 8.5 10mM CaCl2

52

23

Xylanase

28

14

50mM Na TAPS pH 8.5 100mM Na Formate

76

38 (20*)

Glucose Isomerase

144

75

50mM Na TAPS pH 8.5 100mM Na Cacodylate

178

90

Ovalbumin

Maximum Solubility in Tris/NaCl (mg/mL) 148

Concentration used for crystal screens (mg/mL) 75

Myoglobin

13

N/A

Myoglobin (after azide treatment)

122

Catalase

Protein

* Concentration used after adjustment (see results)

Composition of the Optimized Buffer

50mM Na HEPES pH 7.5 -

G

M

aL

lb u m in

in

lb um

um

at in

Is om

se

na se

er a

Xy la

Su bt ilis in

Th a

Tr yp si n

Pe ps in

ac ta

as e

at ed )

C at al

tre

yo gl ob in

va

(a zi de

lu co se

M yo gl ob in

O

160

Figure 6x4: Maximum solubility Standard chromatography buffer Optimized buffer

200

180

160

140

120

100

80

60

40

20

0

161

Table 6x6: Crystallization results

OVALBUMIN

Additive screen

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

Tris/ NaCl

59

7

29

1

0

0

65

3

16

12

0

0

56

1

9

0

21

10

86

2

4

2

1

1

83

2

6

5

0

0

Opt.

53

4

36

3

0

0

70

1

17

8

1

0

32

32

10

22

1

0

83

0

7

0

5

1

64

4

4

24

0

0

Tris/ NaCl

72

13

11

0

0

0

45

6

29

15

1

0

60

1

17

1

13

4

87

2

5

2

0

0

86

2

7

1

0

0

Opt.

33

33

28

2

0

0

34

2

36

23

1

0

49

4

13

30

0

0

84

2

6

2

1

1

81

3

6

6

0

0

4

THAUMATIN

SUBTILISIN

GLUCOSE ISOMERASE

XYLANASE

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

C

ppt

ps

µ

Op

DQ

Tris/ NaCl

58

11

22

5

0

0

92

0

3

0

0

1

85

3

8

0

0

0

29

0

1

14

52

0

56

13

4

2

5

16

Opt.

66

11

9

10

1

0

75

4

2

5

5

6

69

9

18

1

0

0

63

4

9

2

3

16

47

10

4

1

9

26

Tris/ NaCl

94

1

0

1

0

0

96

0

0

0

0

0

90

0

5

1

0

0

63

1

11

14

7

0

13

20

41

13

8

2

Opt.

93

1

0

2

0

0

92

0

0

3

0

1

77

0

8

9

3

0

74

3

4

6

5

4

77

4

4

3

3

6

Type of result 4

Additive screen

PEPSIN

ppt

TRYPSIIN

INDEX screen

α-LACTALBUMIN

CATALASE

C

Type of result4

INDEX screen

MYOGLOBIN

See scoring p.151

162 The crystallization results obtained in these two preliminary screens are very interesting. In most cases, there is a substantial improvement both in the number and in the quality of the crystals obtained, as a direct result of using the solubility screen optimization method. With the exception of Catalase, which behaved opposite to what was expected and Pepsin and Trypsin, which only gave micro-crystals, all other proteins showed good results. Some of the crystals certainly need further optimization to be useful for diffraction but many of them grew directly to an impressive size (sometimes up to 0.6x0.5 mm), considering that every drop is only 1 µL. In several cases, there was a dramatic effect on the size and the shape of the crystals and new crystal forms were found, that had never been observed before. Several conditions even produced different crystal morphologies within the same drop. A sample gallery of these pictures is presented in Figure 6x5. In each case, a set of two pictures is shown for direct comparison: the picture on the left is the drop in standard chromatography buffer conditions; the picture on the right is the drop in the optimized buffer conditions.

Figure 6x5: Crystal pictures in standard vs. optimized conditions Figure 6x5a: Ovalbumin

INDEX (F1, #6)

163 Figure 6x5b: α-Lactalbumin

Additive (F1, #6)

Additive (D4, #28)

Additive (F3, #22)

Figure 6x5c: Myoglobin

Additive (H8, #64)

164 Figure 6x5d: Thaumatin

INDEX (G4, #31)

Additive (E3, #21)

INDEX (A2, #9)

INDEX (F11, #86)

165 Figure 6x5e: Xylanase

Additive (C4, #27)

INDEX (C10, #75)

INDEX (A4, #25)

Additive (A12, #89)

Additive (A9, #65)

166 Figure 6x5f: Glucose Isomerase

INDEX (G11, #87)

Additive (D4, #28)

INDEX (D10, #76)

INDEX (H7, #56)

Additive (A9, #65)

167 As mentioned earlier, Catalase did not behave as predicted. In fact, optimizing the solubility lead to opposite results, as far as crystallization is concerned. The standard solution drops produced small needle-like and platelike crystals in many conditions, whereas the optimized conditions produced heavy precipitates in about 80% of the drops. For this reason, the crystal screens were redone after cutting the concentration of the protein in half. But the results were the same. The pictures of Catalase crystal hits are shown on Figure 6x6.

Figure 6x6: Catalase

Additive (G5, #39)

INDEX (E6, #45)

INDEX (H10, #80)

168 Something similar happened with Xylanase and the trays were also redone using half the concentration of protein. The results of that second try were really good and large crystals grew, like those shown in Figure 6x5. Following are a few pictures of the drops when the concentration was too high.

Figure 6x7: Xylanase at high concentration

INDEX (B10, #74)

INDEX (C8, #59)

6.4. Discussion The solubility screen represents a simple and reliable way to maximize protein solubility, to identify the buffer components essential to their stability in solution and to learn their pH profile. Moreover, it can be performed in less than two hours and saves the hassle of tedious DLS experiments, although coupling the two techniques would give a good idea of how the protein behaves in

169 solution. It should also be mentioned that the starting material used in the solubility screen is a precipitated protein sample. Thus, it provides an opportunity to use precipitated material that might be otherwise lost but is often valuable, especially when only small amounts of protein are available. The 17-condition solubility screen is representative of the main monoand divalent salts and buffers used in protein chemistry, but it can easily be extended or modified to include some interesting compounds such as malonate. For some proteins, more than one salt or buffer will give high solubility. In that case, different solution combinations can be tested and their absorbance measured, in order to decide which one is the best. This study also revealed that citrate seems to increase the solubility of many proteins and should therefore be strongly considered when preparing a protein solution for crystallization experiments, especially since it would not produce mineral crystals. When looking at the maximum solubility values (Figure 6x4), it is interesting to note that the proteins that did not really crystallize or gave a very limited number of (small) crystals are those for which the optimized conditions did not help (or even decreased) the solubility. That is the case for Ovalbumin, Myoglobin, Trypsin and Pepsin. Furthermore, these proteins have been shown to have unique suitable conditions for crystallization and are dependent on the presence of a particular compound: horse Myoglobin and Pepsin require malonate; bovine Trypsin needs magnesium sulfate and Ovalbumin needs cacodylate54,

58-60

. On the other hand, the solubility screen can help determine

unsuspected conditions for many proteins. The example of Xylanase is striking. It

170 was thought to only crystallize in magnesium sulfate but the optimized solution used here (TAPS pH 8.5, Na formate) produced a lot of crystals of very large size and many different crystal morphologies. This work was a good opportunity to develop some knowledge on the relation between protein behavior in solution and the success of crystallization experiments. It was presented last July at the 2004 ACA meeting in Chicago during a poster session5 and got good feedback. Hopefully, this new and original approach to rationalize preliminary crystal screening experiments will become useful to many protein chemists and crystallographers.

5

Ref. poster WO442: A preliminary screen to optimize protein solubility and

improve crystal screen results; Aude Izaac, Stephen J. Tomanicek and Timothy C.

Mueser. Department of Chemistry, The University of Toledo, Toledo, OH 43606.

171

BIBLIOGRAPHY

1. 2. 3.

4. 5. 6. 7. 8. 9.

10. 11. 12.

13.

14.

15. 16.

Kraulis, P. J., J Appl Cryst 1991, 24, 946-50. Merritt, E. A.; Bacon, D. J., in Methods Enzymol. Academic Press: New York, 1997; Vol. 277, p 505-525. Ng, J. Y.; Marians, K. J., The ordered assembly of the phiX174-type primosome. I. Isolation and identification of intermediate protein-DNA complexes. J Biol Chem 1996, 271, (26), 15642-8. Marians, K. J., PriA-directed replication fork restart in Escherichia coli. Trends Biochem Sci 2000, 25, (4), 185-9. Sandler, S. J.; Marians, K. J., Role of PriA in replication fork reactivation in Escherichia coli. J Bacteriol 2000, 182, (1), 9-13. Marians, K. J., PriA: at the crossroads of DNA replication and recombination. Prog Nucleic Acid Res Mol Biol 1999, 63, 39-67. Haber, J. E., DNA recombination: the replication connection. Trends Biochem Sci 1999, 24, (7), 271-5. Friedberg, E. C.; Walker, G. C.; Siede, W., DNA repair and mutagenesis. ASM Press: Washington, D.C., 1995. Cox, M. M.; Goodman, M. F.; Kreuzer, K. N.; Sherratt, D. J.; Sandler, S. J.; Marians, K. J., The importance of repairing stalled replication forks. Nature 2000, 404, (6773), 37-41. Kornberg, A.; Baker, T. A., DNA replication. 2nd ed.; W.H. Freeman and Co.: New York, 1992. Marians, K. J., Prokaryotic DNA replication. Annu Rev Biochem 1992, 61, 673-719. Wickner, S.; Hurwitz, J., Conversion of phiX174 viral DNA to doublestranded form by purified Escherichia coli proteins. Proc Natl Acad Sci USA 1974, 71, (10), 4120-4. Nakai, H.; Doseeva, V.; Jones, J. M., Handoff from recombinase to replisome: insights from transposition. Proc Natl Acad Sci USA 2001, 98, (15), 8247-54. Jones, J. M.; Nakai, H., Duplex opening by primosome protein PriA for replisome assembly on a recombination intermediate. J Mol Biol 1999, 289, (3), 503-16. Sinsheimer, R. L., A single-stranded DNA from bacteriophage phi X174. Brookhaven Symp Biol 1959, No 12, 27-34. Sanger, F.; Coulson, A. R.; Friedmann, T.; Air, G. M.; Barrell, B. G.; Brown, N. L.; Fiddes, J. C.; Hutchison, C. A., 3rd; Slocombe, P. M.; Smith, M., The nucleotide sequence of bacteriophage phiX174. J Mol Biol 1978, 125, (2), 225-46.

172 17. Arai, K.; Kornberg, A., Unique primed start of phage phi X174 DNA replication and mobility of the primosome in a direction opposite chain synthesis. Proc Natl Acad Sci USA 1981, 78, (1), 69-73. 18. Nurse, P.; Zavitz, K. H.; Marians, K. J., Inactivation of the Escherichia coli priA DNA replication protein induces the SOS response. J Bacteriol 1991, 173, (21), 6686-93. 19. Sandler, S. J.; Samra, H. S.; Clark, A. J., Differential suppression of priA2:kan phenotypes in Escherichia coli K-12 by mutations in priA, lexA, and dnaC. Genetics 1996, 143, (1), 5-13. 20. Masai, H.; Asai, T.; Kubota, Y.; Arai, K.; Kogoma, T., Escherichia coli PriA protein is essential for inducible and constitutive stable DNA replication. Embo J 1994, 13, (22), 5338-45. 21. Cox, M. M., A broadening view of recombinational DNA repair in bacteria. Genes Cells 1998, 3, (2), 65-78. 22. Kogoma, T.; Cadwell, G. W.; Barnard, K. G.; Asai, T., The DNA replication priming protein, PriA, is required for homologous recombination and doublestrand break repair. J Bacteriol 1996, 178, (5), 1258-64. 23. Nurse, P.; Liu, J.; Marians, K. J., Two modes of PriA binding to DNA. J Biol Chem 1999, 274, (35), 25026-32. 24. Shlomai, J.; Kornberg, A., An Escherichia coli replication protein that recognizes a unique sequence within a hairpin region in phi X174 DNA. Proc Natl Acad Sci USA 1980, 77, (2), 799-803. 25. Ng, J. Y.; Marians, K. J., The ordered assembly of the phiX174-type primosome. II. Preservation of primosome composition from assembly through replication. J Biol Chem 1996, 271, (26), 15649-55. 26. Lee, M. S.; Marians, K. J., The Escherichia coli primosome can translocate actively in either direction along a DNA strand. J Biol Chem 1989, 264, (24), 14531-42. 27. Jones, J. M.; Nakai, H., The phiX174-type primosome promotes replisome assembly at the site of recombination in bacteriophage Mu transposition. Embo J 1997, 16, (22), 6886-95. 28. Marians, K. J., Phi X174-type primosomal proteins: purification and assay. Methods Enzymol 1995, 262, 507-21. 29. Janson, J.-C.; Ryden, L., Protein Purification. 2nd ed.; Wiley-Liss A John Wiley & Sons: New York, 1998. 30. Qiagen, The QIAexpressionist (Protocol13). 5th ed.; 2001; p 83-84. 31. Collins, B. K.; Tomanicek, S. J.; Lyamicheva, N.; Kaiser, M. W.; Mueser, T. C., A preliminary solubility screen used to improve crystallization trials: crystallization and preliminary X-ray structure determination of Aeropyrum pernix flap endonuclease-1. Acta Crystallogr D Biol Crystallogr 2004, 60, (Pt 9), 1674-8. 32. Senger, A. B.; Mueser, T. C., Rapid preparation of custom grid screens for crystal growth optimization. J Appl Cryst (in press) 2004. 33. Moradian-Oldak, J.; Leung, W.; Fincham, A. G., Temperature and pHdependent supramolecular self-assembly of Amelogenin molecules: a dynamic Light-scattering analysis. J Struct Biol 1998, 122, 320-327.

173 34. Jones, C. E.; Mueser, T. C.; Nossal, N. G., Bacteriophage T4 32 protein is required for helicase-dependent leading strand synthesis when the helicase is loaded by the T4 59 helicase-loading protein. J Biol Chem 2004, 279, (13), 12067-75. 35. Senger, A. B. A Study of DNA Replication and Repair Proteins from Bacteriophage T4 and a Related Phage. Master Thesis, Univertity of Toledo, 2004. 36. Lasken, R. S.; Kornberg, A., The primosomal protein n' of Escherichia coli is a DNA helicase. J Biol Chem 1988, 263, (12), 5512-8. 37. Lee, M. S.; Marians, K. J., Escherichia coli replication factor Y, a component of the primosome, can act as a DNA helicase. Proc Natl Acad Sci USA 1987, 84, (23), 8345-9. 38. Jones, J. M.; Nakai, H., PriA and phage T4 gp59: factors that promote DNA replication on forked DNA substrates microreview. Mol Microbiol 2000, 36, (3), 519-27. 39. Jaktaji, R. P.; Lloyd, R. G., PriA supports two distinct pathways for replication restart in UV-irradiated Escherichia coli cells. Mol Microbiol 2003, 47, (4), 1091-100. 40. McGlynn, P.; Al-Deib, A. A.; Liu, J.; Marians, K. J.; Lloyd, R. G., The DNA replication protein PriA and the recombination protein RecG bind D-loops. J Mol Biol 1997, 270, (2), 212-21. 41. Galletto, R.; Jezewska, M. J.; Bujalowski, W., Multistep sequential mechanism of Escherichia coli helicase PriA protein-ssDNA interactions. Kinetics and energetics of the active ssDNA-searching site of the enzyme. Biochemistry 2004, 43, (34), 11002-16. 42. Liu, J.; Marians, K. J., PriA-directed assembly of a primosome on D loop DNA. J Biol Chem 1999, 274, (35), 25033-41. 43. Gorbalenya, A. E.; Koonin, E. V.; Donchenko, A. P.; Blinov, V. M., Two related superfamilies of putative helicases involved in replication, recombination, repair and expression of DNA and RNA genomes. Nucleic Acids Res 1989, 17, (12), 4713-30. 44. Tanaka, T.; Taniyama, C.; Arai, K.; Masai, H., ATPase/helicase motif mutants of Escherichia coli PriA protein essential for recombinationdependent DNA replication. Genes Cells 2003, 8, (3), 251-61. 45. Tanaka, T.; Mizukoshi, T.; Taniyama, C.; Kohda, D.; Arai, K.; Masai, H., DNA binding of PriA protein requires cooperation of the N-terminal Dloop/arrested-fork binding and C-terminal helicase domains. J Biol Chem 2002, 277, (41), 38062-71. 46. Chen, H. W.; North, S. H.; Nakai, H., Properties of the PriA helicase domain and its role in binding PriA to specific DNA structures. J Biol Chem 2004, 279, (37), 38503-12. 47. Allen, G. C., Jr.; Kornberg, A., Assembly of the primosome of DNA replication in Escherichia coli. J Biol Chem 1993, 268, (26), 19204-9. 48. Liu, J.; Nurse, P.; Marians, K. J., The ordered assembly of the phiX174-type primosome. III. PriB facilitates complex formation between PriA and DnaT. J Biol Chem 1996, 271, (26), 15656-61.

174 49. Liu, J. H.; Chang, T. W.; Huang, C. Y.; Chen, S. U.; Wu, H. N.; Chang, M. C.; Hsiao, C. D., Crystal structure of PriB, a primosomal DNA replication protein of Escherichia coli. J Biol Chem 2004, 279, (48), 50465-71. 50. Lopper, M.; Holton, J. M.; Keck, J. L., Crystal structure of PriB, a component of the Escherichia coli replication restart primosome. Structure (Camb) 2004, 12, (11), 1967-75. 51. Shioi, S.; Ose, T.; Maenaka, K.; Shiroishi, M.; Abe, Y.; Kohda, D.; Katayama, T.; Ueda, T., Crystal structure of a biologically functional form of PriB from Escherichia coli reveals a potential single-stranded DNA-binding site. Biochem Biophys Res Commun 2005, 326, (4), 766-76. 52. Gill, S. C.; von Hippel, P. H., Calculation of protein extinction coefficients from amino acid sequence data. Anal Biochem 1989, 182, 319-26. 53. Antony, T.; Kumar, S.; Chauhan, M.; Atreyi, M.; Khatri, G. S., Effects of Mg2+ and denaturants on the unfolding pattern of DNA-T--a replication protein of E. coli. Int J Biol Macromol 1996, 19, (2), 91-7. 54. McPherson, A., Crystallization of Biological Macromolecules. Cold Spring Harbor Laboratory Press: new York, 1999. 55. Jancarik, J.; Kim, S. H., Sparse matrix sampling: a screening method for the crystallization of proteins. J. Appl. Cryst. 1991, 24, 409-11. 56. Carter, C. W., Jr.; Carter, C. W., Protein crystallization using incomplete factorial experiments. J Biol Chem 1979, 254, (23), 12219-23. 57. McPherson, A., Protein crystallization in the structural genomics era. J Struct Funct Genomics 2004, 5, (1-2), 3-12. 58. Yamazaki, I.; Yokota, K. N.; Shikama, K., Preparation Of Crystalline Oxymyoglobin From Horse Heart. J Biol Chem 1964, 239, 4151-3. 59. Miller, M.; Weinstein, J. N.; Wlodawer, A., Preliminary x-ray analysis of single crystals of ovalbumin and plakalbumin. J Biol Chem 1983, 258, (9), 5864-6. 60. McPherson, A., A comparison of salts for the crystallization of macromolecules. Protein Science 2001, 10, 418-422.

175

APPENDICES

Appendix 1: Maps of pET21a and pET28 vectors …………………………….. 176 Appendix 2: Buffers for Ni-NTA chromatography Appendix 3: Commonly used cryoprotectants

…………………………….. 177

………..……………………….. 177

Appendix 4: Secondary structure prediction for PriC (excerpt) ……………..... 178 Appendix 5: General index of trays

……………………………….....……..….. 180

176

Appendix 1: Maps of pET 21a and pET 28 vectors

177

Appendix 2: Buffers for Ni-NTA chromatography

Lysis buffer:

Wash buffer:

Adjust pH to 8.0 with NaOH

Adjust pH to 8.0 with NaOH

50 mM NaH2PO4

50 mM NaH2PO4

300 mM NaCl

300 mM NaCl

10 mM Imidazole

20 mM Imidazole

Elution buffer: Adjust pH to 8.0 with NaOH 50 mM NaH2PO4 300 mM NaCl 250 mM Imidazole

Appendix 3: Commonly used cryoprotectants

Cryoprotectant

Useful concentration range

Glycerol

15% - 50% (v/v)

Ethylene glycol

25% - 50% (v/v)

PEG-400

25% - 40% (w/v)

Xylitol

22% - 50% (w/v)

Glucose

25% - 50% (w/v)

MPD

20% - 30% (v/v)

1,2-propanediol

15% - 30% (v/v)

Li2SO4

> 2.5 M

178

Appendix 4: Secondary structure prediction for PriC (excerpt)

Appendix 4a: Overall secondary structure composition of PriC PROF predictions SYNOPSIS of prediction for query PROFsec summary overall your protein can be classified as: all-alpha given the following classes: 'all-alpha': %H > 45% AND %E < 5% 'all-beta': %H < 5% AND %E > 45% 'alpha-beta': %H > 30% AND %E > 20% 'mixed': all others Predicted secondary structure composition for your protein: sec str type HEL % in protein %H 86.21 %E 0 Other 13.79 About the alignment used: ali_orig/home/ppuser/server/work/predict_h21308.hsspPsiFil Residue composition for your protein: %A: 16.7%C: 1.7%D: 3.5%E: 9.8%F: 2.3 %G: 1.1%H: 3.5%I: 2.3%K: 2.9%L: 14.4 %M: 0.6%N: 1.1%P: 2.3%Q: 8.6%R: 14.4 %S: 2.9%T: 5.2%V: 4.0%W: 2.3%Y: 0.6 About the PROF methods used: prof_fparacc=/home/ppuser/server/pub/prof/net/PROFboth_best.par prof_nnetacc=6 Please quote: PROF: B Rost & C Sander (1993) J Mol Biol, 232:584-599 PROFhtm: B Rost, P Fariselli & R Casadio (1996) Prot Science, 7:1704-1718

 The calculated composition of PriC is 86.21% of α-helices, 0% of β-sheets and 13.79% of other structures (loops, coils) and the protein belongs to the “all-alpha” structural family.

Appendix 4b: Results of the secondary structure prediction based on the primary sequence of PriC ABBREVIATIONS used: AA : amino acid sequence OBS_sec: observed secondary structure: H=helix, E=extended (sheet), blank=other (loop) PROF_sec: PROF predicted secondary structure: H=helix, E=extended (sheet), blank=other (loop) PROF = PROF: Profile network prediction HeiDelberg Rel_sec: reliability index for PROFsec prediction (0=low to 9=high)

179 Note: for the brief presentation strong predictions marked by '*' SUB_sec: subset of the PROFsec prediction, for all residues with an expected average accuracy > 82% (tables in header) NOTE: for this subset the following symbols are used: L: is loop (for which above ' ' is used) .: means that no prediction is made for this residue, as the reliability is: Rel < 5

BODY with predictions for query PROF results (normal) ....,....1....,....2....,....3....,....4....,....5....,....6....,....7....,....8....,....9....,....10.1.,....11.1.,....12.1.,....13.1. ,....14.1.,....15.1.,....16.1.,....17.1. AA KTALLLEKLE GQLATLRQRC APVSQFATLS ARFDRHLFQT EAGDNLAALR HAVEQQQLPQ VAWLAEHLAA QLEAIAREAS AWSLREWDSA PPKIARWQRK RIQHQDFERR LREMVAERRA RLARVTDLVE QQTLHREVEA YEARLARCRH ALEKIENRLA RLTR OBS_sec PROF_sec ßHHHHHHHHH HHHHHHHHHß ßßßßßßßßßß HHHHHHHHHH HßHHHHHHHH HHHHHHHHHH HHHHHßßßßH HHHHHHHHHH HHHHHHHHHH HHHHHHHßßß ßßHHHHHHHH HHHHHHHHHH HHHHHHHHHH HHHHHHHHHH HHHHHHHHHH HHHHHHHHHH HHHHHHHHHH HHßß Rel_sec 9688898889 8898888612 6311567624 2455888864 3000037898 8888888888 8887258746 8888888899 8888887776 6866430578 8503368889 8867688898 9988898898 8865436788 8888888889 8877888898 8988898998 6229 SUB_sec LHHHHHHHHH HHHHHHHHßß LßßßLLLLßß ßßHHHHHHHß ßßßßßßHHHH HHHHHHHHHH HHHHßLLLßH HHHHHHHHHH HHHHHHHHHH HHHHßßßLLL LLßßßHHHHH HHHHHHHHHH HHHHHHHHHH HHHHßßHHHH HHHHHHHHHH HHHHHHHHHH HHHHHHHHHH HßßL

 The alignment shows that the vast majority of the residues were predicted to be part of α-helices (H above) with high reliability coefficients (7,8,9 out of 9). Appendix 4c: Prediction of the overall fold of PriC GLOBE prediction of globularity --- GLOBE: prediction of protein globularity --- nexp = 122 (number of predicted exposed residues) --- nfit = 80 (number of expected exposed residues --- diff = 42.00 (difference nexp-nfit) --- =====> your protein appears not to be globular ----- GLOBE: further explanations preliminaryily in: --http://cubic.bioc.columbia.edu/papers/1999_globe/paper.html ----- END of GLOBE

 PriC is predicted to be a non-globular protein.

180

Appendix 5: General index of trays

Protein

Tray #

Date

Temp.

Concentration

Drop size

Kits

Type of expansion

Number of Hits / Best conditions

PriA N / PriA

AI 001

5.14.03

RT

10mg/mL (both)

1+1

Wizard I+II

Screen

0

PriA N / PriA

AI 002

5.14.03

RT

10mg/mL (both)

1+1

Crystal Screen I+II

Screen

0

PriA N / PriA

AI 003

5.14.03

RT

10mg/mL (both)

1+1

Cryo I+II

Screen

0

PriA N / PriA

AI 004

5.14.03

RT

10mg/mL (both)

1+1

Natrix + Ion screen

Screen

0

PriA N / PriA

AI 005

5.20.03

4°C

10mg/mL (both)

1+1

Wizard I+II

Screen

0

PriA N / PriA

AI 006

5.20.03

4°C

10mg/mL (both)

1+1

Crystal Screen I+II

Screen

0

DnaT

AI 007

6.11.03

RT

10mg/mL

1+1

Wizard I+II

Screen

4

DnaT

AI 008

6.11.03

RT

10mg/mL

1+1

Cryo I+II

Screen

3

DnaT

AI 009

6.11.03

RT

10mg/mL

1+1

Crystal Screen I+II

Screen

9

DnaT

AI 010

6.11.03

RT

10mg/mL

1+1

Natrix + Ion screen

Screen

7

DnaT

AI 011

6.12.03

4°C

10mg/mL

1+1

Cryo I+II

Screen

8

DnaT

AI 012

6.12.03

4°C

10mg/mL

1+1

Crystal Screen I+II

Screen

4

DnaT

AI 013

6.12.03

4°C

10mg/mL

1+1

Wizard I+II

Screen

0

DnaT

AI 014

6.12.03

4°C

10mg/mL

1+1

Natrix + Ion screen

Screen

2

DnaT

AI 015

6.16.03

RT

10mg/mL

2+2

-

4x6

0

DnaT

AI 016

6.24.03

RT

10mg/mL

2+2

-

2x12

0

DnaT

AI 017

7.02.03

RT

17.7mg/mL

2+2

-

2x12

0

DnaT

AI 018

7.30.03

RT

21.5mg/mL

2+2

-

4x6

A1

DnaT

AI 019

7.29.03

RT

21.5mg/mL

2+2

-

4x6

B1,B2,B3

181

Protein

Tray #

Date

Temp.

Concentration

Drop size

Kits

Type of expansion

Number of Hits / Best conditions

DnaT

AI 020

7.30.03

RT

20mg/mL

2+2

-

4x6

A1,A2,A3, C1

DnaT

AI 021

8.27.03

RT

20mg/mL

2+2

-

2x12

C4,C5

DnaT

AI 022

8.27.03

RT

20mg/mL

2+2

-

2x12

A1

DnaT

AI 023

9.08.03

RT

20mg/mL

2+2

-

2x12

B5,D5

DnaT

AI 024

9.17.03

RT

20mg/mL

2+2

-

6x4 (columns)

C1, D5

DnaT

AI 025

9.24.03

RT

20mg/mL

2+2

-

2x12

C5

DnaT

AI 026

9.24.03

RT

20mg/mL

2+2

-

2x12

0

DnaT

AI 027

10.02.03

RT

20mg/mL

2+2 (+/microseeding)

-

2x12

B5,D5

DnaT

AI 028

10.09.03

RT

20mg/mL

2+2

-

1x24

D5

DnaT

AI 029

10.09.03

RT

20mg/mL

2+2 and 1+1

-

2x12

C2

DnaT

AI 030

10.09.03

RT

20mg/mL

2+2

-

2x12

0

DnaT

AI 031

10.13.03

RT

20mg/mL

2+2

-

2x12

A1,C2

DnaT

AI 032

10.16.03

RT

20mg/mL

2+2 and 1+1

-

2x12

B4-6, D1-3

DnaT

AI 033

10.21.03

RT

20mg/mL

2+2

-

2x12

0

PriA N

AI 034

10.28.03

RT

19mg/mL

1+1

Crystal Screen I+II

Screen

2

PriA N

AI 035

10.28.03

RT

19mg/mL

1+1

Cryo I+II

Screen

0

PriA N

AI 036

10.28.03

RT

19mg/mL

1+1

Wizard I+II

Screen

1

PriA N

AI 037

10.28.03

RT

19mg/mL

1+1

Natrix + Ion screen

Screen

2

DnaT

AI 038

10.25.03

RT

20-22mg/mL

2+2, 1+1+2, 2+2+4

-

2x12

B3?

DnaT

AI 039

10.28.03

RT

22mg/mL

2+2

-

2x12

0

182

Protein

Tray #

Date

Temp.

Concentration

Drop size

Kits

Type of expansion

Number of Hits / Best conditions

DnaT

AI 040

10.27.03

RT

22mg/mL

2+2, 1+1+2, 2+2+4

-

2x12

0

DnaT

AI 041

11.05.03

RT

18mg/mL

2+2

-

2x12

0

DnaT

AI 042

11.06.03

RT

18mg/mL

2+2

-

2x12

0

DnaT

AI 043

11.06.03

RT

18mg/mL

2+2

-

2x12

A1-A6

DnaT

AI 044

11.08.03

RT

18mg/mL

2+2

-

1x24

0

PriA N

AI 045

11.11.03

4°C

19mg/mL

1+1

Crystal Screen I+II

Screen

0

PriA N

AI 046

11.11.03

4°C

19mg/mL

1+1

Cryo I+II

Screen

1

PriA N

AI 047

11.11.03

4°C

19mg/mL

1+1

Wizard I+II

Screen

1

PriA N

AI 048

8.04.04

4°C

21mg/mL

0.5+0.5

Natrix + Ion screen

Screen

0

DnaT

AI 049

2.19.04

RT

18mg/mL

2+2

-

2x12

0

DnaT

AI 050

3.02.04

RT

20mg/mL

2+2

-

2x12

B3-B5

DnaT

AI 051

3.11.04

RT

20mg/mL

2+2

-

4x6

0

DnaT

AI 052

3.23.04

RT

18mg/mL

2+2

-

4x6

B5-B6

DnaT

AI 053

3.23.04

RT

18mg/mL

2+2

-

1x24

A6, C5

DnaT

AI 054

3.17.04

RT

20mg/mL

2+2

-

2x12

A1-A4

DnaT

AI 055

3.25.04

RT

18mg/mL

2+2

-

1x24

C6, D1

DnaT

AI 056

3.27.04

RT

18mg/mL

2+2

-

2x12

B3-B4, D4-D5

DnaT

AI 057

6.03.04

RT

20mg/mL

0.5+0.5

Index

Screen

5

DnaT

AI 058

6.03.05

RT

20mg/mL

0.5+0.5

Additive screen

Screen

0

DnaT

AI 059

6.03.06

RT

20mg/mL

0.5+0.5

Crystal Screen Cryo and MembFac

Screen

7

DnaT

AI 060

6.05.04

RT

18mg/mL

2+2

-

2x12 and 4x6

B4

183

Protein

Tray #

Date

Temp.

Concentration

Drop size

Kits

Type of expansion

Number of Hits / Best conditions

DnaT

AI 061

6.07.04

20°C

18mg/mL

2+2

-

2x12

B5-6, D6

DnaT

AI 062

6.07.04

RT

18mg/mL

2+2

-

2x12

B1

DnaT

AI 063

6.08.04

RT

18mg/mL

2+2

-

1x24

B2, B5, C5, D2-5

DnaT

AI 064

6.11.04

RT

18mg/mL

2+2

-

1x24

D2-3

DnaT

AI 065

6.14.04

RT

18mg/mL

2+2

-

1x24

C3,C6, D2-4

DnaT

AI 066

6.15.04

RT

18mg/mL

2+2

-

4x6

0

DnaT

AI 067

7.29.04

RT

18mg/mL

2+2

-

1x24

0

PriA

AI 068

8.04.04

RT

9mg/mL

0.5+0.5

Additive screen

Screen

6

PriA

AI 069

8.04.04

RT

9mg/mL

0.5+0.5

Index

Screen

0

PriA

AI 070

8.04.04

RT

9mg/mL

0.5+0.5

Crystal Screen Cryo and MembFac

Screen

1

PriA / PriA N

AI 071

8.04.04

4°C

9 and 16mg/mL

0.5+0.5

Additive screen

Screen

5

PriA / PriA N

AI 072

8.04.04

4°C

9 and 16mg/mL

0.5+0.5

Index

Screen

2

PriA / PriA N

AI 073

8.04.04

4°C

9 and 16mg/mL

0.5+0.5

Crystal Screen Cryo and MembFac

Screen

1

PriA / PriA N

AI 074

8.04.04

4°C

9 and 16mg/mL

0.5+0.5

Salt RX

Screen

1?

PriA

AI 075

8.09.04

RT

9mg/mL

2+2

-

4x6

Salt

PriA / PriA N

AI 076

8.12.04

4°C

9 and 16mg/mL

2+2

-

2x12

0

PriA N

AI 077

10.13.04

4°C

16mg/mL

2+2

-

4x6

0

PriA

AI 078

10.26.04

RT

9mg/mL

2+2

-

4x6

0

PriA+DNA

AI 079

12.13.04

RT

10mg/mL

0.5+0.5

Wizard I+II

Screen

?

PriA+DNA

AI 080

12.13.04

RT

10mg/mL

0.5+0.5

Crystal Screen I+II

Screen

?

PriA+DNA

AI 081

12.13.04

RT

10mg/mL

0.5+0.5

Index

Screen

?