UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Engineering proteins of novel mechanical properties : from single molecule to biomaterials Cao, Yi 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_fall_cao_yi.pdf [ 7.5MB ]
Metadata
JSON: 24-1.0061613.json
JSON-LD: 24-1.0061613-ld.json
RDF/XML (Pretty): 24-1.0061613-rdf.xml
RDF/JSON: 24-1.0061613-rdf.json
Turtle: 24-1.0061613-turtle.txt
N-Triples: 24-1.0061613-rdf-ntriples.txt
Original Record: 24-1.0061613-source.json
Full Text
24-1.0061613-fulltext.txt
Citation
24-1.0061613.ris

Full Text

ENGINEERING PROTEINS OF NOVEL MECHANICAL PROPERTIES: FROM SINGLE MOLECULE TO BIOMATERIALS  by  Yi Cao M. Sc., Nanjing University, China, 2004 B. Sc., Nanjing University, China, 2001  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in  THE FACULTY OF GRADUATE STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2009  © Yi Cao, 2009  Abstract Elastomeric proteins are an important class of mechanical proteins that take care of the strength, elasticity and extensibility of tissues and biological machineries. Elastomeric proteins are also essential building blocks of materials with outstanding mechanical properties. However, it was largely unknown how elastomeric proteins achieve the remarkable mechanical stability until the advent of single molecule force spectroscopy techniques, such as single molecule atomic force microscopy (AFM). Single molecule AFM has enabled the direct characterization of the mechanical properties of elastomeric proteins at the single molecule level and led to the new promising research area of single protein mechanics and engineering. Combined with molecular dynamics simulation and protein engineering, single molecule AFM has provided rich information about the mechanical design of elastomeric proteins. This dissertation focuses on engineering proteins with novel mechanical properties and makes use of this new technique. It is demonstrated how a non-mechanical protein GB1, the B1 immunoglobulin (IgG) binding domain of protein G from Streptococcus, shows superb mechanical stability. We also investigated the effect of denaturant, guanidium hydrochloride (GdmCl), on the mechanical stability of GB1. It was found that the mechanical stability of GB1 decreases with the increase of GdmCl concentration. Using GB1 as a model system, we demonstrated two ways to enhance the mechanical stability of proteins: by metal chelation and by stabilizing protein-protein interactions. It is revealed that preferentially stabilizing the native state over the mechanical unfolding transition state of proteins is the key to achieve enhanced mechanical stability.  ii  We also showed two applications of engineered elastomeric proteins. One is the design of an artificial elastomeric protein with dual mechanical stability that can be regulated reversibly by protein-protein interactions. We introduced proline mutations to GB1 to make it mechanically labile and behave as entropic springs. Upon binding of the Fc fragment of IgG, the proline mutants of GB1 switched into a state of significant mechanical stability and can serve as shock-absorbers. The other application is the engineering of the first tandem modular protein based thermo-reversible hydrogel, which paves the way for engineering hydrogels with much improved physical properties that can be used as artificial extracellular matrices and tissue engineering materials.  iii  Table of contents Abstract.............................................................................................................................. ii Table of contents .............................................................................................................. iv List of tables.................................................................................................................... viii List of figures.................................................................................................................... ix List of symbols and abbreviations ................................................................................ xiv Acknowledgements ........................................................................................................ xiv Dedication ....................................................................................................................... xvi Co-authorship statement .............................................................................................. xvii Chapter 1: Introduction ................................................................................................... 1 1.1 Investigating the mechanical properties of proteins using single molecule atomic force microscopy (AFM) .................................................................................... 3 1.1.1 Single molecule techniques and the principles of single molecule atomic force microscopy ...................................................................................................... 3 1.1.2 Two operation modes and force spectroscopy of single protein molecules ...... 6 1.1.3 The kinetics and energetics of protein folding and unfolding ......................... 11 1.1.3.1 The two-state model for protein folding and unfolding............................ 11 1.1.3.2 The effect of force on protein folding and unfolding ............................... 13 1.1.4 Mechanical stability is a kinetic stability......................................................... 15 1.1.5 Mapping the free energy landscape of proteins using single molecule AFM.. 16 1.1.5.1 Extracting the kinetic parameters from the unfolding force distribution.. 16 1.1.5.2 Extracting the kinetic parameters from a pulling speed dependent experiment................................................................................................. 17 1.2 The history and the achievements of “protein mechanics”..................................... 19 1.2.1 The study of naturally occurring elastomeric proteins..................................... 23 1.2.2 Polyprotein engineering allows the study of individual protein domains in detail.............................................................................................................. 24 1.2.3 Detecting rare misfolding events ..................................................................... 25 1.2.4 Studying the mechanical unfolding pathway and intermediate state by glycine insertion and disulfide crosslinking mutations. ............................................ 26 1.2.5 Scaling up single molecule AFM results explains the macroscopic elasticity of muscle myofibrils.......................................................................................... 27 1.2.6 Mechanical stability of proteins depends on the direction of force ................. 28 1.2.7 The folding trajectory of proteins studied by force-clamp experiment ........... 29 1.2.8 Single molecule mechanics on membrane proteins ......................................... 29 1.3 Engineering proteins with novel mechanical stability............................................ 30 1.3.1 Experimental efforts on the engineering of mechanical proteins .................... 30 1.3.2 Investigation of the mechanical unfolding of proteins in silico....................... 33 1.3.3 The secondary structure, topology and mechanical stability of proteins......... 35 1.4 Aim of this dissertation........................................................................................... 37 1.4.1 Searching for mechanically stable non-mechanical proteins........................... 38 1.4.2 Rational tuning of the mechanical stability of proteins ................................... 38 1.4.3 Bottom-up construction of elastomeric protein based materials...................... 40 1.5 References............................................................................................................... 42 Chapter 2: Expanding the toolbox of elastomeric proteins to include non-mechanical proteins................................................................................................................. 47  iv  2.1 Introduction............................................................................................................. 47 2.2 Results..................................................................................................................... 49 2.2.1 GB1 exhibits significant mechanical stability ................................................. 49 2.2.2 Mechanical unfolding of GB1 is a non-equilibrium process .......................... 54 2.2.3 The fast mechanical folding kinetics of GB1 ................................................. 56 2.2.4 GB1 can fold under residual forces ................................................................ 59 2.2.5 Free energy diagram for the mechanical unfolding and refolding of GB1..... 61 2.2.5 Polyprotein GB1 does not show noticeable mechanical fatigue..................... 62 2.2.7 High fidelity in the refolding of GB1 ............................................................. 66 2.3 Discussion............................................................................................................... 67 2.4 Conclusion ............................................................................................................. 68 2.5 Materials and methods ........................................................................................... 69 2.5.1 Protein engineering ......................................................................................... 69 2.5.2 Single molecule atomic force microscopy ...................................................... 70 2.5.3 Monte Carlo simulations................................................................................. 70 2.6 References.............................................................................................................. 73 Chapter 3: Tuning the mechanical stability of proteins by varying solvent conditions ............................................................................................................................... 75 3.1 Introduction............................................................................................................. 75 3.2 Results..................................................................................................................... 77 3.2.1 GB1 is mechanically weakened by denaturant ................................................ 77 3.2.2 The mechanical unfolding distance is unaltered by denaturant ....................... 80 3.2.3 Chemical denaturants speed up the mechanical unfolding event. ................... 82 3.2.4 Chemical denaturant slows down the folding reaction of GB1. ...................... 83 3.3 Discussion............................................................................................................... 85 3.3.1 A mechanical chevron plot quantitatively describes the effects of chemical denaturants on the mechanical unfolding and folding reactions................... 85 3.3.2 The mechanical unfolding kinetics are affected by chemical denaturants in a similar way to chemical unfolding................................................................ 87 3.3.3 The mechanical unfolding transition state of GB1 is more solvent exposed than the native state............................................................................................... 91 3.3.4 The effect of chemical denaturant on mechanical refolding............................ 92 3.4 Conclusion .............................................................................................................. 93 3.5 Materials and methods ............................................................................................ 94 3.5.1 Protein engineering and expression ................................................................. 94 3.5.2 Force spectroscopy of single proteins.............................................................. 94 3.5.3 Monte Carlo simulations.................................................................................. 95 3.6 References............................................................................................................... 96 Chapter 4: Enhancing the mechanical stability of proteins by engineered metal chelation ............................................................................................................... 98 4.1 Introduction............................................................................................................. 98 4.2 Results................................................................................................................... 100 4.2.1 Rationale for enhancing the mechanical stability of proteins ........................ 100 4.2.2 Binding of metal ions significantly enhances the mechanical stability of GB1 bi-His mutants. ............................................................................................ 104  v  4.2.3 The enhancement of mechanical stability by metal ion binding is fully reversible..................................................................................................... 106 4.2.4 The enhancement in mechanical stability by binding of metal ion is context dependent. ................................................................................................... 108 4.2.5 The metal chelation sites engineered outside the force-bearing region do not affect the mechanical stability of GB1........................................................ 109 4.2.6 Enhancing the mechanical stability by increasing the free energy barrier .... 111 4.3 Discussion............................................................................................................. 114 4.4 Materials and methods .......................................................................................... 117 4.4.1 Protein engineering ........................................................................................ 117 4.4.2 Single molecule AFM experiment ................................................................. 117 4.5 References............................................................................................................. 119 Chapter 5: Enhancing the mechanical stability of proteins by protein-protein interactions ........................................................................................................ 121 5.1 Introduction........................................................................................................... 121 5.2 Results................................................................................................................... 123 5.2.1 Binding of IgG fragments can significantly enhance the mechanical stability of GB1. ............................................................................................................ 123 5.2.2 The binding of the IgG fragments does not shift the mechanical unfolding transition state for GB1. .............................................................................. 132 5.2.3 The amplitude of the mechanical stability enhancement does not correlate with the binding affinity...................................................................................... 133 5.2.4 The enhancement of the mechanical stability of GB1 by IgG fragment is through long range coupling ....................................................................... 137 5.2.5 The long range coupling regulation of the mechanical stability of GB1 is robust........................................................................................................... 139 5.3 Discussion............................................................................................................. 140 5.4 Materials and methods .......................................................................................... 141 5.4.1 Protein engineering ........................................................................................ 141 5.4.2 Single molecule AFM experiment ................................................................. 142 5.5 References............................................................................................................. 143 Chapter 6: Engineered elastomeric proteins with dual elasticity can be controlled by a molecular regulator ....................................................................................... 145 6.1 Introduction........................................................................................................... 145 6.2 Results................................................................................................................... 148 6.3 Discussion............................................................................................................. 154 6.4 Experimental section............................................................................................. 155 6.4.1 Protein engineering ........................................................................................ 155 6.4.2 Single-molecule AFM experiment................................................................. 156 6.4.3 Circular dichoism (CD) and Surface plasmon resonance (SPR) measurements ..................................................................................................................... 157 6.5 References............................................................................................................. 158 Chapter 7: Engineering tandem modular protein-based reversible hydrogels ..... 160 7.1 Introduction........................................................................................................... 160 7.2 Results................................................................................................................... 161 7.3 Conclusion ............................................................................................................ 169  vi  7.4 Experimental section............................................................................................. 169 7.4.1 Protein engineering and hydrogel formation ................................................. 169 7.4.2 Circular dichroism measurement (CD) .......................................................... 171 7.4.3 Scanning electron microscopy (SEM) ........................................................... 171 7.4.4 Erosion measurement..................................................................................... 171 7.5 References............................................................................................................. 173 Chapter 8: Summary and future directions .............................................................. 174 8.1 Summary............................................................................................................... 174 8.2 Future directions ................................................................................................... 179 8.2.1 Can non-mechanical proteins with other folds have significant mechanical stability? ...................................................................................................... 179 8.2.2 Rational tuning the mechanical stability of proteins...................................... 180 8.2.3 Beyond single protein domains...................................................................... 183 8.3 References............................................................................................................. 184 Appendix A: Polyprotein engineering......................................................................... 185 A1. Engineering of plasmids containing octameric genes encoding proteins of interest ........................................................................................................................ 185 A2. Polyprotein expression......................................................................................... 188 A3. Sequences of proteins and the encoding cDNAs ................................................. 190 A3.1 wild type GB1 ................................................................................................ 190 A3.2 G6-53 ............................................................................................................ 190 A3.3 G4-51 ............................................................................................................ 190 A3.4 G8-55 ............................................................................................................ 190 A3.5 G4-6 .............................................................................................................. 191 A3.6 G32-36 .......................................................................................................... 191 A3.7 GT25A .......................................................................................................... 191 A3.8 GK28A .......................................................................................................... 191 A3.9 GK31A ........................................................................................................ 192 A3.10 GN35A ........................................................................................................ 192 A3.11 NuG2 ........................................................................................................... 192 A3.12 Gc3b4 .......................................................................................................... 192 A3.13 GT18P ......................................................................................................... 193 A3.14 GV54P......................................................................................................... 193 A3.15 Leucine zipper A ......................................................................................... 193 A3.16 Leucine zipper P.......................................................................................... 194  vii  List of tables Table 1.1 Table 1.2 Table 4.1 Table 4.2  Table A1  Comparison of single molecule force spectroscopy techniques ……… 4 Mechanical properties of proteins studied by AFM…………………... 19 The free energy changes upon metal chelation of GB1 bi-His mutants…………………………………………………………………. 104 The summary of unfolding force, unfolding distance ( Δ x u ) and spontaneous unfolding rate constant α0 at zero force for bi-His mutants…………………………………………………………………. 109 The restriction sites typically used for protein engineering…………… 185  viii  List of figures Figure 1.1 Figure 1.2 Figure 1.3 Figure 1.4 Figure 1.5 Figure 1.6 Figure 1.7 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5  Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4  Figure 4.5 Figure 4.6 Figure 5.1 Figure 5.2 Figure 5.3  The force-spectroscopy mode of atomic force microscopy (AFM)… Using single molecule atomic force microscopy to probe the mechanical properties of single proteins……………………………. The contour length increment of unfolding a protein domain………. The energetics of protein folding and unfolding in a two-state model The effect of force on the free energy of mechanical folding and unfolding…………………………………………………………….. The achievements in the field of protein mechanics………………... The secondary structure and topology of proteins that have been studied by single molecule AFM……………………………………. Polyprotein (GB1)8 has significant mechanical stability……………. Mechanical unfolding of GB1 is a non-equilibrium process……….. The fast folding kinetics of GB1……………………………………. GB1 can fold in the presence of residual forces…………………….. Schematic drawing of the free energy diagram for the mechanical unfolding and refolding of B1……………………………………..... Polyprotein (GB1)8 does not show noticeable mechanical fatigue…. Mechanical stability of GB1 decreases with the increase of denaturant concentration…………………………………………….. The relationship of mechanical stability of GB1 and the denaturant concentration………………………………………………………… The mechanical unfolding distance is unaltered by denaturant……... The refolding kinetics of GB1 at different GdmCl concentrations... A mechanical chevron plot quantitatively describes the effect of a chemical denaturant on the mechanical unfolding and folding kinetics………………………………………………………………. Rationale for enhancing the mechanical stability of proteins……….. The thermodynamic stability of bi-His mutants of GB1 in the absence and presence of 14.3mM of NiCl2…………………………. The mechanical stability of GB1 bi-His mutants is enhanced by the binding of Ni2+………………………………………………………. The mechanical stability of G6-53 can be regulated reversibly by the binding of Ni2+ ions as well as its competitive binding reagent imidazole…………………………………………………………….. Metal chelation site outside the force-bearing region does not enhance the mechanical stability of biHis mutants…………………. The pulling speed dependence of the mechanical unfolding forces of bi-His mutants in the absence and presence of 4mM Ni2+………….. hFc and hFab bind GB1 in the different regions of GB1…………… The mechanical stability of protein GB1 is significantly enhanced by the binding of human IgG fragments hFc and hFab…………….. The mechanical stability of GB1/hFc complex is independent on the  5 10 11 13 14 21 37 51 53 55 58 62 65 77 78 80 83  84 102 103 105  107 111 113 124 125  ix  Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4 Figure 6.5 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 7.6 Figure 7.7 Figure 7.8 Figure 8.1  Figure 8.2 Figure A1 Figure A2  Figure A3  concentration of hFc………………………………………………… Enhancement of mechanical stability by the binding of hFc is not an artifact due to the polymerized form of GB1………………………... The unfolding force increment of GB1 upon binding of hFc is independent on the pulling speed…………………………………… The mechanical stability of GB1/hFc complex does not correlate with the binding affinity…………………………………………….. The long distance coupling of mechanical stability is retained in GB1 mutants with large structural perturbation…………………….. Designing elastomeric chameleon proteins using proline mutagenesis………………………………………………………….. Elastomeric chameleon proteins show dual mechanical elasticity. SPR sensorgrams for the binding and dissociation of monomeric chameleon proteins………………………………………………….. Reversible elastic behaviors of polyprotein (GT18P)8…………….... Schematic illustration of the general concept of elastomeric chameleon proteins………………………………………………….. Schematic of the artificial protein A(G)8A and its SDS-PAGE picture……………………………………………………………….. CD spectra of (G)8 and A(G)8A……………………………………. Optical and SEM image of A(G)8A hydrogel………………………. Hydrogel of A(G)8A at different temperatures…………………… Solutions of A(G)8A at different concentrations……………………. Schematic drawing of the A(G)8A hydrogel………………………... Erosion profile of A(G)8A hydrogel……………………………….. Protein sequence of A(G)8A………………………………………... Tuning the mechanical stability of proteins by modulating the free energy difference between the native state and the transition state (ΔΔGN-T)…………………………………………………………….. Summary of currently available methods to tune the mechanical stability of proteins………………………………………………… General procedure for polyprotein gene engineering………………. The representative agarose gel DNA electrophoresis pictures for plasmids containing monomer, dimer, tetramer and octamer of the gene encoding the protein of interest……………………………….. A representative SDS-PAGE gel for a purified protein (GB1)8…  128 130 133 136 138 147 150 151 153 154 162 163 164 165 165 166 167 170  177 182 187  188 189  x  List of symbols and abbreviations a  loading rate of force  AFM  atomic force microscopy (microscope)  CD  circular dichroism  DTT  dithiothreitol  E. Coli  Escherichia coli  Ex  extension of the molecule  F  force  Fab  the Fab fragment of Immunoglobulin G antibody  Fc  the Fc fragment of Immunoglobulin G antibody  FnIII  fibronectin type 3  Fu  unfolding force  GB1  the B1 binding domain of protein G from Streptococcus  Gc3b4  a computationally designed hyperthermophilic variant of GB1  GdmCl  guanidinium chloride  Ig  Immunoglobulin  IgG  Immunoglobulin G  IPTG  isopropyl-1-β-D-thiogalactoside  kB  Boltzmann constant  kc  spring constant of the cantilever  Kd  dissociation constant  LB  Luria-Bertani broth  xi  Lc  Contour length  M  molar  N  newton  nm  nanometer  NuG2  a de novo designed version of GB1 in which 11 residues in the first b hairpin were mutated to increase its folding rate  OD  optical density  P  persistence length  PBS  phosphate buffer saline  PID  proportional, integral and differential amplifier  Pu(F)  unfolding probability at force F  R  gas constant  RU  resonance units  s  second  S.D.  standard deviation  SASA  solvent accessible surface area  SCOP  structural classification of proteins  SDS-PAGE  sodium dodecyl sulfate polyacrylamide gel electrophoresis  SMD  steer molecular dynamics  SPR  surface plasmon resonance  T  temperature  TNfn3  the third fibronectin type 3 domain of human tenascin C  v  pulling speed  xii  WLC  worm-like chain  x  extension  ΔGT-N  free energy difference between the transition state and the native state  ΔGU-N  free energy difference between the unfolded state and the native state  ΔLc  contour length increment  Δxf  folding distance, the distance between the unfolded and the transition state  Δxu  unfolding distance, the distance between the native and the transition state  Δzc  displacement of the cantilever  Δzp  movement of the piezoelectric positioner  α0  unfolding rate at zero force  β0  folding rate at zero force  xiii  Acknowledgements I would like to express my cordial gratitude to Dr. Hongbin Li for his supervision, guidance, and support throughout this work. Hongbin is an exceptional teacher and a great friend. His dedications to science and brilliant intellectual insights have inspired me through all these years. I have also been deeply impressed by his enthusiasm for his assisting and supporting students in many ways, from improving scientific writing skills to counseling career choices. His attitude towards science and people will have a profound influence on me in my future.  I would also like to thank all of the members of the Li’s laboratory that I have worked with in the past and present years, for both technical and intellectual contributions to my work, and for friendships that have enriched my experience at UBC. In particular, thank you to Canaan Lam and Deepak Sharma for teaching me basic molecular biology techniques when I first started my work. Thank you to Qing Peng, Shunlin Zhuang, Meijia Wang, Ying Guo, Ashlee Jollymore, M.M. Balamurali, Eileen Wang, Peng Zheng, Shanshan Lv, Teri Yoo, Kai Shih Er and Rakesh Parhar for your invaluable friendship that have supported me in completing this endeavor.  Thank you to Dr. Suzana K. Straus, Dr. Pierre Kennepohl and Dr. Elliott Burnell for providing technical and intellectual guidance as my supervision committee. I especially thank Dr. Suzana K. Straus for her critical evaluation of this manuscript.  xiv  I would like to thank the whole UBC community for providing such a close and friendly environment. In particular, thank you to Hui Wang, for your invaluable friendship and help during my stay at UBC.  Last but not least, I would like to thank my parents for their deep love and greatest emotional support through all these years and my entire life.  xv  Dedication This dissertation is dedicated to my wife, Ying Li. Without her love, support, and patience, I probably would not be where I am now.  xvi  Co-authorship statement The work presented in this dissertation is the result of collaborative projects. The contributions of other scientists are summarized below.  Chapter 2 has been published as two papers: “Cao, Y.; Lam, C.; Wang, M.; Li, H., Nonmechanical protein can have significant mechanical stability. Angew Chem Int Ed Engl 2006, 45, (4), 642-5.” and “Cao, Y.; Li, H., Polyprotein of GB1 is an ideal artificial elastomeric protein. Nat Mater 2007, 6, (2), 109-14.”. My supervisor, Hongbin Li, designed the experiments. Canaan Lam and Meijia Wang helped me in polyprotein engineering. I performed all the other experiments and analyzed the data. I wrote the manuscript together with Hongbin Li.  Chapter 3 has been published as: “Cao, Y.; Li, H., How do chemical denaturants affect the mechanical folding and unfolding of proteins? J Mol Biol 2008, 375, (1), 31624.”. Hongbin Li designed the research. I performed all experiments and analyzed the data. I wrote the manuscript together with Hongbin Li.  Chapter 4 has been published as: “Cao, Y.; Yoo, T.; Li, H., Single molecule force spectroscopy reveals engineered metal chelation is a general approach to enhance mechanical stability of proteins. Proc Natl Acad Sci U S A 2008, 105, (32), 11152-7.”. Hongbin Li and I conceived the experiments. Teri Yoo engineered 3 polyproteins of bi-  xvii  histidine mutants of GB1. I conducted all the other experiments in this chapter. I analyzed the data. I wrote the manuscript together with Hongbin Li.  Chapter 5 has been published as: “Cao, Y.; Balamurali, M. M.; Sharma, D.; Li, H., A functional single-molecule binding assay via force spectroscopy. Proc Natl Acad Sci U S A 2007, 104, (40), 15677-81.” and “Cao, Y.; Yoo, T.; Zhuang, S.; Li, H., Proteinprotein interaction regulates proteins' mechanical stability. J Mol Biol 2008, 378, (5), 1132-41.”. Hongbin Li and I designed the experiments. M.M. Balamurali and Deepak Sharma helped me with surface plasmon resonance (SPR) experiments. Teri Yoo and Shunlin Zhuang helped me with protein engineering. I am responsible for the remaining experiments. I analyzed all the data. I wrote the manuscripts together with Hongbin Li.  Chapter 6 has been published as “Cao, Y.; Li, H., Engineered elastomeric proteins with dual elasticity can be controlled by a molecular regulator. Nat Nanotechnol 2008, 3, (8), 512-6.”. Hongbin Li and I conceived the experiments. I conducted all the experiments and the data analysis. I wrote the paper together with Hongbin Li.  Chapter 7 has been published as “Cao, Y.; Li, H., Engineering tandem modular protein based reversible hydrogels. Chem Commun (Camb) 2008, (35), 4144-6.”. Hongbin Li and I designed the research. I performed the experiments and analyzed the data. I wrote the paper together with Hongbin Li.  xviii  Chapter 1: Introduction  Many biological processes require the conversion of chemical energy to mechanical work, such as muscle contraction, membrane fusion, cell division, cell crawling, cell adhesion, organelle transport, protein translocation, protein degradation, protein folding and unfolding (1, 2). There are a specific number of proteins that mediate this conversion and they do so via their conformational changes. These molecules are called mechanical proteins. There are two important classes of mechanical proteins: one is the motor proteins which directly generate mechanical work, such as ATP synthase, myosin, kinesin, dynamin and polymerases (3-13); the other is the elastomeric class of proteins that provide elasticity, extensibility and necessary strength for cells and tissues such as titin, fibronectin, tenascin, spectrin and elastin (14-19). Elastomeric proteins can withstand significant deformation without rupture and can return to their original state when the stretching force is removed. Besides their important biological function, elastomeric proteins are also biomaterials of extraordinary mechanical properties (21). For example, spider dragline silks are highly elastic and tear-proof fibers that outperform any artificial fibrous materials (22, 23). Resilin, an elastomeric protein found in specialized regions of the cuticle of most insects, has high rubber efficiency (resilience) and plays a role in energy storage and repetitive movement necessary for flight and locomotion of insects (24). The mechanical properties of these amazing elastomeric proteins are “encoded” in their primary sequence as well as their three dimensional arrangements of single domains and the whole complex. Thanks to recently developed single molecule force spectroscopy techniques, extensive experimental efforts at the 1  single molecule level have been well under way to decipher the mystery that underlies the unique mechanical properties of elastomeric proteins (25-28). Single molecule force spectroscopy techniques has made it possible to stretch individual elastomeric proteins and to study their mechanical properties and structure-function relationship one molecule at a time. Understanding the molecular details of the design of elastomeric proteins is not only important for elucidating biophysical principles underlying a wide variety of biological processes, but may also reveal new design principles for biomaterials that could pave the way to design novel elastomeric proteins with well-defined mechanical properties using bottom-up approaches. These efforts will also lead to use of these novel elastomeric proteins for nanobiotechnological applications. In this chapter, I will provide a review on this new and exciting field. It includes 4 sections: 1. Introduction of the techniques that are used to study the mechanical properties of elastomeric proteins, with the focus of the use of single molecule atomic force microscopy (AFM). The general experimental procedures, the interpretation of data and the models for unfolding of elastomeric proteins are presented. 2. Review of the history and achievements of the field of “protein mechanics”. The proteins that have been studied by AFM so far are summarized and the important advances are highlighted. 3. Summary of the achievements and challenges in the engineering of proteins with novel mechanical properties. 4. Aim of this dissertation.  2  1.1 Investigating the mechanical properties of proteins using single molecule atomic force microscopy (AFM) 1.1.1  Single molecule techniques and the principles of single molecule atomic force microscopy Despite the significance of mechanical properties of protein based materials, it  was not possible to measure the mechanical properties of proteins at the single molecule level until the invention of force spectroscopic techniques. These methods include optical tweezers (29-31), magnetic tweezers (32), atomic force microscopy (33, 34), microneedle manipulation (35), biomembrane force probe (36) and flow-induced stretching (37, 38). The first three are the most widely used and a summary of their technical features are listed in table 1.1. The single molecule force spectroscopy techniques make it possible to relate the mechanical properties of individual proteins to their structures and conformations. Of the many single molecule force spectroscopy techniques, single molecule AFM is particularly valuable for measuring mechanical properties of proteins due to its excellent spatial resolution (Å), high force sensitivity (~10 pN), and large force range (from ~5 pN to ~10 nN). The AFM apparatus, invented in 1986 by Binnig and coworkers, was a relatively simple apparatus made of a soft cantilever based force sensor, a high resolution piezoelectric positioner and other regular electronic devices (39, 40). As shown in Fig. 1.1, when a polymer molecule is picked up by the cantilever from the surface of the substrate (typically glass coverslips), a molecular mechanical linkage is formed between the cantilever and the surface. The force applied to the molecule can be monitored by the deflection of the cantilever.  3  Table 1.1 Comparison of single molecule force spectroscopy techniques (41, 42) Optical tweezers Spatial resolution (nm) Measurement time scale (s) Stiffness (pN nm-1) Force range (pN) Displacement range (nm) Probe size (µ µm) Typical applications Typical attachment chemistries Features  Limitations  Magnetic tweezers  AFM  0.1-2  5-10  0.5-1  10-4-103  10-3-105  10-3-102  0.005-1  10-3-10-6  1-105  0.1-100  10-3-102  2-104  0.1-105  5-104  0.5-104  0.25-5  0.5-5  100-250  3D manipulation Tethered assay Interaction assay  Tethered assay DNA topology  Covalent, or specific noncovalent Low-noise and lowdrift dumbbell geometry Active or passive force clamp  Covalent, or specific noncovalent  Nonspecific  Force clamp Bead rotation Specific interactions  High-resolution imaging High force range Active force clamp  Difficult to manipulate molecule Large particles Bulky handles  Large high stiffness probe Large minimal force Polyproteins or chimeric constructs Random attachment geometries  Photodamage Sample heating Bulky handles  High-force pulling and interaction assays  4  A B  Laser  Photodiode Cantilever Si3N4 tip Glass substrate  Piezoelectric positioner  Figure 1.1 The force-spectroscopy mode of atomic force microscopy (AFM). The molecule of interest is picked up by the tip of the cantilever from the substrate. The force exerted on the molecule is measured by the deflection of the cantilever and the extension of the molecule is measured by the movement of the piezoelectric positioner. The cantilever of an AFM behaves as a Hookean spring. The deflection of the cantilever can be measured by a laser beam bounced back from the back of the cantilever to a split photodiode detector. The force acting on the cantilever is given by F = k c ∆z c  (1.1)  where F is the force, kc is the spring constant of the cantilever and ∆zc is the displacement of the cantilever along the z-axis (Fig 1.2). The spring constant, kc, of the cantilever can be measured from the fluctuation of the cantilever driven by thermal motion. The  5  cantilever is modeled as a one dimensional oscillator and its energy can be calculated using the equipartition theorem: 1 1 2 k c < ∆z c >= k B T 2 2  (1.2)  where kc and ∆zc are as previously defined, kB is the Boltzmann constant and T is the temperature in Kelvin. The average “thermal energy” of cantilever,  1 2 kc < ∆zc > , is 2  typically not evaluated in the time domain but is converted to the frequency domain using a Fourier transformation. It is of note that the spring constant of cantilevers calibrated by the thermal method may have errors of 10-20% (43). The extension of the molecule can be calculated from the movement of the piezoelectric positioner and the deflection of the cantilever. E x = ∆z p − ∆ z c  (1.3)  where Ex is the extension of the molecule, ∆zp is the movement of the piezoelectric positioner (Fig. 1.2) and ∆zc is the deflection of the cantilever (Fig. 1.2).  1.1.2  Two operation modes and force spectroscopy of single protein molecules There are two typical operation modes of AFM: distance-ramp mode (16, 34, 44)  and force-ramp mode (45, 46). In distance-ramp mode, the protein molecules are stretched with a constant pulling velocity; whereas in the force-ramp mode, the force applied to the molecule is controlled at a constant ramp rate through an active force feedback system. In distance-ramp mode, the protein or polymer samples which are deposited on the substrate are first brought into contact with the cantilever to establish the single  6  molecule connection between cantilever tip and substrate. Then, the substrate is driven away from the cantilever by the piezoelectric positioner at a constant velocity to a set distance. The force and extension of the molecule are monitored simultaneously. A plot of the force-extension curve illustrates the force spectra of molecules, which directly reports the mechanical property of polymers at the single molecule level. Fig. 1.2A shows a schematic of distance-ramp experiments. Unfolding of an artificial elastomeric protein that is made of tandem repeats of individually folded GB1, the first IgG binding domain of protein G from group G streptococcus, shows a complicated periodic pattern. When the polyprotein of GB1 is stretched, the force will increase nonlinearly with the elongation of the polyprotein molecule (Fig. 1.2A, from stage 1 to stage 2). The relationship between force and extension can be well described using the worm-like chain (WLC) model of polymer elasticity (1.4):  F ( x) =  1 1 x k BT 1 ( ⋅ − + ) p 4 (1 − x ) 2 4 Lc Lc  (1.4)  where F(x) is the force at extension x, kB is Boltzmann constant, T is absolute temperature, p is the persistence length and Lc is the contour length of the polymer. In WLC model, the polymer is described as an isotropic chain that is continuously flexible. The persistence length, p, defines the rigidity of the chain. The entropic restoration force of the polymer can be calculated at any given extension of the molecule. The individually folded domains are not extensible before they are unfolded. Force will increase the unfolding probability of protein domains and trigger the unfolding of protein domains in the polyprotein chain. If one of the domains unfolds and becomes extensible, the total contour length of the polyprotein increases and the force acting on the polyprotein drops  7  back to a low value (Fig. 1.2A, from stage 2 to stage 3). The force increases back again following the WLC model of polymer elasticity until the unfolding of another domain occurs in the polyprotein (Fig. 1.2A, from stage 3 to stage 4). The sequential unfolding of protein domains in the polyprotein gives rise to a saw-tooth like pattern of unfolding of elastomeric proteins made of tandem repeats of individually folded proteins. Each peak in the force-extension curve corresponds to the unfolding of a GB1 domain in the polyprotein chain. In the force-ramp mode, the force can be controlled through an additional feedback loop. The deflection of the cantilever is compared with a computer-controlled set point to generate an error signal. Then the error signal is amplified through a proportional, integral, and differential amplifier (PID) and fed back to the piezoelectric positioner. The force exerted onto the molecule is kept the same as the set number by controlling the extension of the molecule through the piezoelectric positioner. Comparing to the complex non-linear force-extension relationship obtained in distance-ramp mode, the kinetic data of the molecule can be obtained in a model-free manner in force-ramp mode and analyzed in a more straightforward fashion. A typical force ramp experiment of polyprotein (GB1)8 is shown schematically in Fig. 1.2B. The force increases linearly with time and the length of the molecule increases in discreet stair case-like stages of similar height corresponding to the sequential unfolding of protein domains in the polyprotein. In each stage, the length of molecule increases non-linearly with time (time is proportional to force in force ramp experiments), following the WLC model of polymer elasticity (dotted line). The main advantage of force-ramp experiments is that the kinetic parameters of mechanical unfolding can be directly extracted by fitting the experimental  8  data, as will be discussed in section 1.1.3. However, the drift of the cantilever and the piezoelectric positioner make the force-ramp experiments challenging. In most of the experiments in this thesis, we employed the distance-ramp mode of AFM. The force spectra of a single protein molecule contain two kinds of important information: firstly, the structural architecture of the protein domain and secondly, the kinetics of the mechanical unfolding reaction. In the force-extension curve in distanceramp mode, we are able to measure the contour length increment upon unfolding of one protein domain. As shown in Fig. 1.3, the unfolded length of a protein can be calculated from the number of amino acids in the polypeptide chain and the distance between the Nand C- termini of the folded protein can be estimated from the crystal structure. Therefore the contour length increment observed upon unfolding of a single protein domain can be directly obtained and compared with the number obtained from the single molecule AFM experiments. The discrepancy of these may indicate the existence of unfolding intermediate states (47). When a protein with unknown structure is stretched from different positions, it is possible to find the distance between the two amino acids in the folded structure where the force is applied by measuring the contour length increment (if the unfolding transition state is highly native like) (48). On the other hand, the kinetics of the mechanical unfolding reaction can be deduced from the amplitude of each unfolding peak. The average unfolding force at a given pulling speed defines the mechanical stability of the protein. The force that is required to unfold a protein domain depends on the 3-dimensional structure of the protein. The amplitude and the distribution of the unfolding force are determined by the free energy landscape of mechanical unfolding. This part will be discussed in detail in the following sections.  9  Figure 1.2 Using single molecule atomic force microscopy to probe the mechanical properties of single proteins. In a typical single molecule AFM experiment, a tandem modular protein, that is adsorbed on a glass coverslip, is picked up randomly by the tip of AFM cantilever and stretched from the glass substrate, which is mounted on top of a high piezoelectric positioner. The single molecule AFM experiments can be operated in two different modes: distance-ramp mode (A) and force-ramp mode (B). A) A shows the schematic of a distance-ramp measurement on a polyprotein. Stretching a tandem modular protein in the distance-ramp mode results in force-extension curves of the characteristic saw-tooth pattern appearance. Each peak corresponds to the unfolding of a protein domain in the tandem modular protein. The last peak in the force-extension curve corresponds to the full extension of the unfolded polyprotein and the detachment of the molecule from either the AFM tip or the glass substrate. B) A schematic of a single molecule force-ramp experiment on a polyprotein. The stretching force F is set as a linear function of time (F=at, where a is the ramp rate and t is time), and the time evolution of the length of the polyprotein is monitored as shown in the middle panel. The mechanical unfolding of individual domains gives rise to the staircase appearance of the resultant length-time traces. Each stair case corresponds to the unfolding of a protein domain in the polyprotein. The measured force signal as a function of time is shown at the bottom panel. Due to the finite response frequency, the unfolding of a protein domain will suddenly increase the contour length of the polyprotein and the force is intransient relaxed to a low value, resulting in a spike in the force trace.  10  Lu=0.36nm/aa ×Naa F  Lf  ∆Lc=Lu-Lf ∆Lc F  Lu Figure 1.3 The contour length increment (∆Lc) of unfolding a protein domain depends on the number of amino acids (Naa) and the folded length of the protein (Lf). Lu is the contour length of unfolded protein, which equals the length of an extended amino acid (0.36 nm) times the number of amino acids in the protein. 1.1.3  The kinetics and energetics of protein folding and unfolding Since the mechanical stability of a protein is largely related to its intrinsic  folding/unfolding free energy landscape, it is thereby critical to understand the kinetics and energetics of the folding and unfolding of proteins under force. These will also serve as the base for finding proteins with high mechanical stability and for rationally tuning the mechanical stability of proteins.  1.1.3.1 The two-state model for protein folding and unfolding Protein folding/unfolding is often interpreted as a two-state process by conventional transition state theory (49). Only two populations are dominant in the free energy landscape: folded and unfolded state. The interconversion between the two states requires the crossing of a high-energy barrier. The saddle point of the energy barrier is  11  called the transition state (Fig.1.4). The folding and unfolding rates are given in eq. 1.5 and 1.6. kf =κ  ∆GT − D k BT exp( − ) h k BT  (1.5)  ku = κ  k BT ∆GT − N exp( − ) h k BT  (1.6)  Here, kf and ku are the folding and unfolding rate constants of protein, respectively. The energy barriers for folding and unfolding are denoted as ∆GT-D and ∆GTN,  respectively. κ is the transmission coefficient, kB is the Boltzmann constant, T is the  absolute temperature and h is the Planck constant.  12  Figure 1.4 The energetics of protein folding and unfolding in a two-state model. N, T and U denote the native state, the transition state and the unfolded state of protein, respectively. ∆xu is the distance between the native and transition states along the reaction coordinate. ∆xf is the distance between the unfolded and transition states along the reaction coordinate. ∆GT-N and ∆GT-u define the free energy barriers for unfolding and folding, respectively. The thermodynamic stability is determined by the free energy difference between the native and unfolded states, ∆GU-N.  1.1.3.2 The effect of force on protein folding and unfolding Proteins will favor the native state in the absence of force, as the native state is thermodynamically more stable than the denatured state. In addition the high transition state barrier between the native state and the transition state makes it impossible to observe the spontaneous unfolding of a protein without force within the experimental time window of single molecule AFM experiments (Table 1.1) (Fig. 1.4). However, when force is applied to the protein, it will tilt the free energy landscape. If the force is 13  high enough, the denatured state may become more stable than the native state and the transition state barrier will be sufficiently low such that protein unfolding will be observed within the experimental time scale, as shown in Fig. 1.5. The effect of force on the folding and unfolding rate constant can be described using the Bell-Evans model (5052). k f (F ) = κ  ∆GT −U + F∆x f F∆x f k BT exp( − ) = β 0 exp( − ) h k BT k BT  (1.7)  ku ( F ) = κ  k BT ∆GT − N − F∆xu F∆xu exp( − ) = α 0 exp( ) h k BT k BT  (1.8)  Figure 1.5. The effect of force on the free energy of mechanical folding and unfolding. The top dashed blue line indicates the effect of force on the free energy. The red dashed curve describes the free energy diagram of mechanical folding and unfolding without force and the blue curve indicate the free energy diagram under force. Force destabilizes the native state (N), transition state (T) and unfolded state (U). However, the unfolded barrier becomes lower under force, because the native state is destabilized more than the transition state. The folding barrier becomes higher, because the unfolded state is destabilized less than the transition state. 14  Here, ∆xf is the distance between the denatured state and the transition state (folding distance) and ∆xu is the distance between the native state and the transition state (unfolding distance) (Fig. 1.5). The mechanical folding and unfolding rate constants at zero force are denoted as β0 and α0, respectively. It is worth noting that in the Bell-Evans model, the folding and unfolding distances are independent of force. If the free energy profile for protein unfolding is curved, this assumption may not hold true and other models, such as Kramer’s diffusion model, may be required to resolve the shape of the energy landscape (53, 54). It is also important to note that mechanical unfolding of proteins using AFM is typically a non-equilibrium process. This is mainly due to the asymmetry between the folding distance and the unfolding distance. The unfolding distance (∆xu) is much shorter than the folding distance (∆xf). Therefore, at the force where the unfolding is observed, the refolding is completely inhibited by the force; whereas, at the force where folding is still observable, the unfolding cannot be detected within the timeframe of the experiment (55).  1.1.4  Mechanical stability is a kinetic stability We define the average unfolding force of a protein at a given pulling speed as the  measure of the mechanical stability of a protein. As shown in Fig. 1.5, the mechanical stability of a protein is determined by the free energy barrier (∆GT-N) along the mechanical unfolding pathway defined by the stretching force and the distance between the mechanical unfolding transition state and the native state (∆xu) (26). The mechanical stability is obviously not directly correlated with the thermodynamic stability of proteins,  15  which is the free energy difference between the unfolded state and the native state (∆GU-N). Since the mechanical unfolding pathway is different in different pulling directions and is different from the unfolding pathway measured using chemical methods, generally there is no correlation between mechanical stability and chemical kinetic stability (25-28, 56, 57). It is evident that, although for some proteins the mechanical unfolding and chemical unfolding rate constants are close to one another, such coincidence is not generally valid for all proteins, even for their point mutants (58, 59). It is therefore not possible to directly predict a priori the mechanical stability based on available kinetic and thermodynamic data of proteins measured by ensemble methods. It is evident that ∆GT-N (or α0) and ∆xu influence the mechanical stability of proteins in a combined fashion. It can be shown that a smaller spontaneous unfolding rate constant α0 (or larger unfolding free energy barrier) and a smaller unfolding distance ∆xu favor a higher unfolding force. Thus, upon mutation, the change in the unfolding force of the protein cannot be predicted from ∆GT-N (or α0) or ∆xu alone. For example, an increased mechanical unfolding energy barrier ∆GT-N upon mutation does not necessarily lead to higher mechanical stability as the information about ∆xu is also required. In contrast, the loading rate a influences the mechanical unfolding force of proteins in a straightforward way: the higher the loading rate, the higher the unfolding force.  1.1.5  Mapping the free energy landscape of proteins using single molecule AFM  1.1.5.1 Extracting the kinetic parameters from the unfolding force distribution Protein unfolding is a stochastic (probabilistic) process and the distribution of the unfolding forces reveals the probability density (dPu(F)/dF) of the unfolding at a given  16  force, F. It is worth noting that the broad distribution of unfolding forces is not due to experimental errors (e.g. the error in the calibration of the spring constant of the cantilever or the drift of the baseline of the force-extension curves), but reflects the fluctuation of the energy of proteins in the native state and the transition state along the reaction coordinate. Instead of showing a Gaussian distribution, it shows an asymmetric distribution with a tail in the low unfolding force region. The width of the population directly relates to the unfolding distance, xu, of the protein. The broader the distribution, the smaller the ∆xu. The average unfolding force of the distribution is also related to ∆xu and the unfolding rate at zero force, α0. The higher the unfolding force, the smaller the  ∆xu and α0. Monte Carlo simulation or numeric fitting of the unfolding force distribution is widely used to estimate the ∆xu and α0 of the free energy landscape of mechanical unfolding of proteins (44, 59-63). In practice, however, the calibration error of the cantilever, the drift of the baseline of force-extension curves and the number of domains or the spacers in the polyprotein may broaden the distribution of unfolding forces. Moreover, the ∆xu and α0 are correlated with each other. Simply using the unfolding force distribution at one pulling speed may introduce some degeneracy to the estimated ∆xu and  α0 .  1.1.5.2 Extracting the kinetic parameters from a pulling speed dependent experiment As shown in Fig. 1.2B, the mechanical unfolding of a protein is typically a nonequilibrium process. As predicted by Evans (51, 52) and verified by experimental work (16, 33), for non-equilibrium mechanical unfolding, the unfolding force depends on the  17  effective loading rate of the force. If the force increases in a linear fashion with a loading rate of a (F=a⋅t, t is time), the most probable unfolding force Fu is (51):  Fu =   a ⋅ ∆xu  k BT  ⋅ ln ∆xu  α 0 ⋅ k BT   (1.9)  where ∆xu is the unfolding distance, α0 is the unfolding rate without force, and kBT is the thermal energy. Therefore by analyzing mechanical unfolding data at different loading rates, one can directly extract the unfolding distance and the spontaneous unfolding rate of a protein, the two most important kinetic parameters defining the mechanical unfolding free energy landscape. However most single molecule AFM experiments are performed in distance ramp mode, in which the end-to-end distance of a protein is increased linearly with time by moving the piezoelectric positioner away at a constant velocity. The force changes in a non-linear fashion with the extension of the protein, in accordance with the WLC model and the loading rate is not a constant number, as given in eq. 1.10. a=  dF dF dx k B T 1 = ⋅ = ( + 1)v dt dx dt pLc 2(1 − vt ) 3 Lc  (1.10)  where kB, p, x and Lc are defined in eq. 1.4. a is the loading rate, v is the pulling speed and t is the time. However, increasing v will increase the loading rate a, according to eq. 1.10, and ultimately will increase the unfolding force of proteins. The unfolding force of proteins in force-extension measurements is therefore dependent upon the pulling speed: the faster the pulling speed is, the higher the unfolding force is. It is worth noting that although eq. 1.9 is universal in describing the mechanical unfolding of proteins, due to the non-linear fashion of the loading rate in a distance-ramp experiment, the most  18  probable unfolding force (Fu) measured from force-extension curves cannot be predicted by a simple analytical formula, which is in sharp contrast with force-ramp experiments. Typically Monte Carlo simulation (16, 34, 55) or numerical fitting (59, 62) is used to extract the kinetic parameters from pulling speed dependent experiments.  1.2 The history and the achievements of “protein mechanics”. Since the first single molecule AFM experiment on elasomeric protein 12 years ago (33), many proteins, including naturally occurring ones and non-mechanical ones, have been studied (see Table 1.2). For many naturally occurring elastomeric proteins, the mechanical properties have been investigated at both the entire molecule level and individual domain level. A few important achievements in this burgeoning field are summarized below.  Table 1.2 Mechanical properties of proteins studied by AFM Protein*  Construct*  SCOP Class¶  SCOP Fold¶  Force in pN (speed in nms-1)  Reference  Calmodulin  (Cam)4  all α  EF Hand-like  <20 (600)  (64)  Im9  (I27-Im9)3-I27  all α  <20 (700)  (65)  (I27-R16)4  all α  ~30 (600)  (66)  (I27-R17)4  all α  ~25 (600)  (66)  T4-lysozyme  (GB1)4-lysozyme-(GB1)4  α+β  50 (400)  (67)  Barnase  (I27)5(Ba)3  α+β  Ubiquitin Protein L GB1 Top7  (Ub)9 (Protein L)5 (GB1)8 (GB1)4-(Top7)2-(GB1)4 (GB1)4-(Barstar)2(GB1)4  R16 (Spectrin) R17 (Spectrin)  Barstar  Acyl carrier protein-like Spectrin repeatlike Spectrin repeatlike Lysozyme-like  70 (300)  (68)  α+β α+β α+β α+β  Microbial Ribonuclease β-grasp β-grasp β-grasp Top7  203 (400) 136 (400) 184 (400) 155 (400)  (69) (62) (63, 70) (71)  α+β  Barstar-like  <50 (400)  (72)  19  Protein*  Construct*  SCOP Class¶  SCOP Fold¶  Force in pN (speed in nms-1)  Reference  α+β  GFP-like  104 (3000)  (47)  C2A  (Ig)4GFP(Ig)4 or (DdFLN)3GFP(DdFLN)2 (C2A)9  all β  60 (600)  (64)  E2lip3  (I27)4E2lip3  all β  <20 (600)  (73)  FLN4  (I27–30)FLN4(I31–34)  all β  63 and 53 (250– 350)  (74)  (1FNIII-2FNIII)6  all β  220 (600)  (75)  (10FNIII-I27)4  all β  100 (400)  (76)  (12FNIII-13FNIII)5  all β  Ferredoxin-like Barrel-sandwich like Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich  125 (400)  (75)  (I27-13FNIII)8  all β  89 (400)  (75)  FNIII (tenascin)  (3FNIII)8  all β  120-130 (400)  (77)  I1  (I27-I1)4  all β  127 (600)  (78)  I4  (I4)8  all β  171 (400)  (79)  I5  (I5)8  all β  155 (400)  (79)  I27  (I27)8  all β  204 (400)  (44)  I28  (I28)8  all β  257 (400)  (80)  I32  (I32)8  all β  298 (400)  (79)  I34  (I34)8  all β  281 (400)  (79)  (I27-PKDd1)3-I27  all β  ~200 (1000)  (81)  GFP  1  FNIII (fibronectin) 10 FNIII (fibronectin) 12 FNIII (fibronectin) 13  FNIII (fibronectin) 3  PKDd1 Polycystin-1  Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike b-sandwich Immunoglobulinlike β-sandwich Immunoglobulinlike β-sandwich  * The names of all proteins and constructs are described in detail in the corresponding references. ¶ SCOP: structural classification of proteins (20) http://scop.mrc-lmb.cam.ac.uk/scop/  20  Figure 1.6 The achievements in the field of protein mechanics. A) Force-extension curves of stretching native titin shows a sawtooth pattern with as many as 20 force peaks  21  that varied between 150 and 300 pN and were spaced between 25 and 28 nm. Preceding the sawtooth pattern, a long spacer region was observed in most traces, which corresponds the strengthening of the regions of folded Ig and FnIII domains and the elongation of the unstructured region in titin. Inset is the schematic of titin structure. Adapted from (33) B) Engineering polyproteins by directional DNA concatemerization. The nonpalindromic AvaI restriction site (CTCGGG) was used for self-ligation of monomers. The agarose gel picture shows the I27 monomer (lane a) and a ladder of concatemers with various numbers of I27 monomers (lane b). Right panel shows the map of modified pET expression vector designed for single-step cloning of the Ava I concatemers. The sequence of the polyprotein engineered by this method is shown below. Adapted from (64). C) Misfolding events were observed in polyprotein (I27)8. The contour length increment for unfolding of misfolded protein domains (∆skip) is more than that for unfolding of two I27 domains (∆2x). The possible mechanism for such misfolding events is shown on the right panel. Adapted from (82). D) Glycine insertion can be used to map the key regions for the mechanical stability of I27. 5 glycines were inserted to different regions of I27. However, only at position 75, the inserted glycine residues increased the contour length increment of I27, which indicated that position 75 is in the folded structure of I27. Adapted from (83) E) The unfolding intermediate state of I27 showed as a hump preceding the main unfolding peak, which makes the forceextension curve deviates from the WLC fitting. The conversion between the native state and the intermediate state is reversible in the mechanical unfolding experiment. The hump is more obvious in the first peak of the spectrum, because it contains the contribution of all the events from the stretched domains (Right panel). After the domains are completely unfolded, they can not go back to the intermediate state in experimental time window and the hump gets less when more domains in the protein are unfolded. Adapted from (84). F) The elasticity of muscle myofibrils can be explained at the single molecule level. The red line plots the calculated length-stretching force relationship of I-band titin. The symbols show the force–extension relationship of nonactivated rabbit cardiac myofibrils. To compare with single molecule data, the values of the measured force were scaled by assuming 6 109 titin molecules per mm2 of crosssectional area. The single-molecule data can fully explain the force-extension relationship of myofibrils in and out of the physiological condition (pink box). Adapted from (79). G) The mechanical response of GFP depends on the direction of applied force. The left panel shows the location of cysteine mutations and the direction of force applied to GFP. The right panel shows representative force-extension curves of unfolding disulfide bond linked polyprotein of GFP. The mechanical stability of GFP depends on the pulling direction. Adapted from (61). H) Refolding of polyubiquitin under force monitored by force-clamp AFM. The force applied to the molecule is shown on the bottom panel. The force is force set to 122 pN to unfold ubiquitin domains. Then, the force is quenched to 15 pN to allow ubiquitin domains to fold. The folding trajectory of ubiquitin is shown on the top panel. The inset is the schematic of the snapshots of the force clamp experiments. In the end, the force is set back to 122 pN to check whether the protein is completely folded. Adapted from (46).  22  1.2.1  The study of naturally occurring elastomeric proteins Many naturally occurring elastomeric proteins have been studied by single  molecule atomic force microscopy, including titin (33), tenascin (16), fibronectin (75), spectrin (17) and ankyrin (85). They have well defined mechanical functions in their biological settings and diverse mechanical responses upon stretching. For example, the gigantic muscle protein titin is made of about 300 domains and a few random coil segments. It takes care of the passive elasticity of muscles. Like other mechanical proteins, an important feature of titin is that it must be able to extend to several times of its resting length without breaking in response to high stretching forces. The design principles of titin were little known until single molecule AFM experiments were reported (33). It was discovered that titin shows two distinct elastic features. At low force, the extension of the titin molecule is mainly due to the elongation of the unstructured region, such as the PEVK and N2B sequences. In this case the elasticity is entropic. The titin molecule has an intelligent mechanism to deal with high stretching force that may potentially cause severe damage to the whole sarcomere. At a high stretching force, the individually folded immunoglobulin (Ig) and fibronectin type 3 (FnIII) domains are not fall apart at once. Instead, they can unfold sequentially (Fig. 1.6A) to increase the total contour length of titin and release the force that acts on it. Such “modular” unfolding prevents the high force damage to the whole protein by sacrificially unfolding of the individual domains one by one. This unique mechanism to dissipate energy conveys high toughness to elastomeric proteins and makes them perfect shock-absorbers (23, 86). The unfolding and refolding of individually folded protein domains have been found to be reversible, thereby providing a self-healing property to the elastomeric protein based  23  material (23, 86). The elastic properties of individual protein domains are combined to determine the overall mechanical properties of elastomeric proteins. A recent single molecule AFM experiment on ankyrin highlights the importance of quaternary structure to the mechanical stability of elastomeric proteins (85). Unlike titin, ankyrin is a superhelical spiral formed by stacking different number of repeats of antiparallel alpha-helices. Stretching ankyrin results in Hookean spring behavior before the breaking of the spiral structure, which is in contrast with many other elastomeric proteins (16, 17, 33, 75).  1.2.2  Polyprotein engineering allows the study of individual protein domains in detail In pioneering experiments, the sawtooth pattern resulting from the naturally  occurring elastomeric proteins that are made of domains of similar length not only yielded a wealth of information about the mechanical design of these mechanical proteins, but also served as a characteristic feature to identify single molecule stretching events. As shown in Figure 1.3, the contour length increment of unfolding of each domain can be calculated directly using the number of amino acids and the folded length of individual domains. However, since all domains in naturally occurring modular proteins are different and the unfolding of these domains are stochastic, it is not possible to study the properties of individual protein domains in their native form. Inspired by the naturally occurring elastomeric proteins, polyproteins made of identical repeats of the same domain have been engineered to study the mechanical property of the protein of interest (44). These artificial polyproteins have a perfect repetitive structure and display sawtooth  24  like force-extension curves with uniform spacings. Such characteristic sawtooth patterns can be used to accurately identify single molecule events from a big data pool in which non-specific interactions and stretching of multiple molecules may also exist. The polyprotein strategy has been extensively used in the study of protein mechanics. As different proteins have their unique fingerprint in force-extension curves, this fingerprint technique has been expanded to hetero-polyproteins made of a combination of a protein of known unfolding pattern and a protein of interest (80) (87). Unfolding of such polyproteins shows force-extension traces consisting of mixed patterns with the pattern from the known proteins serving as the fingerprint of single molecule stretching. The hetero-polyprotein strategy is extremely useful to study proteins with complicated unfolding behaviors, such as unfolding intermediates and multiple unfolding pathways, or proteins that cannot be expressed as homo-polyproteins (65, 67, 68, 71, 75, 76, 78, 80, 87, 88). There are two major methods to engineer polyproteins. The first one is through recombinant DNA technique (44) (Fig. 1.6B) (Appendix A). Polyprotein genes are engineered through multiple cloning and expressed directly. The second one is a chemical method in which the oxidation of cysteines in monomeric proteins are used to form intermolecular disulfide bonds (48, 61, 89).  1.2.3  Detecting rare misfolding events By repeated stretching and relaxation of polyprotein I27 and native tenascin,  Fernandez and coworkers observed rare misfolding events in single-molecule AFM experiments (82). They found that neighboring I27 domains can misfold into a bigger  25  fold with a contour length increment of more than twice of single I27 domains (Fig. 1.6C). Such misfolding events only account for 2% of the total unfolding traces, and are impossible to detect at macroscopic level due to ensemble averaging. This finding also has an important impact on the design of tandem modular elastomeric proteins. The sequence diversity of different naturally occurring elastomeric proteins may be the direct consequence of evolutionary pressure to avoid misfolding.  1.2.4  Studying the mechanical unfolding pathway and intermediate state by glycine insertion and disulfide crosslinking mutations. The contour length increment (∆Lc) of unfolding a protein domain is proportional  to the number of amino acids in the folded structure. Therefore, by inserting glycine residues in the loop of a protein or using bi-cysteine mutations to shorten the extensible segment of a protein in AFM experiments, the ∆Lc for protein unfolding can be used to locate the key regions for the mechanical stability of proteins and the structure of the intermediates (Fig. 1.6D). If the loop that is elongated by glycine insertion is inside the protein fold but not inside the intermediate state, the extensible segment of the intermediate state remains the same, but the number of extensible residues of the entire protein becomes larger. Therefore, the ∆Lc for unfolding of the intermediate state remains the same, but the ∆Lc from the unfolding of the native state increases. Similarly, if the inserted glycines are in the structure of the intermediate state, the ∆Lc for the unfolding of intermediate state will increase, whereas the ∆Lc from the native to the intermediate state will remain the same. The disulfide crosslinking mutants will have opposite effects as those due to glycine insertion. Therefore, by monitoring the length phenotype of  26  mechanical unfolding upon glycine insertion or bi-cysteine mutations, we can obtain the mechanical unfolding pathway in detail. Such a strategy was first used in the study of mechanical unfolding of I27 by Fernandez and coworkers (83) and then extended to other proteins including ddFLN (the immunoglobulin rod domains of filamin from Dictyostelium discoideum) (74), GFP (green fluorescence protein) (90), and MBP (Maltose Binding Protein) (91).  1.2.5  Scaling up single molecule AFM results explains the macroscopic elasticity of muscle myofibrils An important achievement of single molecule force spectroscopy is that the  passive elasticity of intact myofibrils can be explained by simply scaling up the mechanical properties of single titin molecules (Fig. 1.6F). The mechanical property of titin is reconstructed by summing up the mechanical properties of different elastic regions of titin, including N2B, PEVK, and the proximal and distal Ig regions. It was revealed that upon stretching, titin behaves as a non-linear spring and different regions have different mechanical responses. At low forces, the straightening of folded Ig domains and the extension of N2B and PEVK random coiled regions dominate the elongation of the titin molecule. Whereas under a high non-physiological force (around hundreds of pN), the entire titin serves as a shock absorber, and the Ig domains can be unfolded to avoid severe damage to the sarcomere of muscle.  27  1.2.6  Mechanical stability of proteins depends on the direction of force Different from thermodynamic stability, mechanical stability of proteins is an  anisotropic property. Independent studies by Fernandez’s group (69) and Radford’s group (73) demonstrated that the mechanical stability of the same protein depends on the pulling directions: the same protein can exhibit drastically different mechanical stability if the protein is pulled from different directions via different pairs of residues. For example, ubiquitin unfolds at ~200 pN when it is pulled from its N- and C-termini. In contrast, the same protein ubiquitin will unfold at much lower force of ~80 pN when pulled from its C-terminus and residue Lys48. The anisotropic nature of the mechanical response of the protein provides unqiue possibilities to explore the free energy landscape for mechanical unfolding of a protein in different pulling directions, which has been exploited systematically by Rief and coworkers using green fluorescent protein (GFP) as a model system (Fig. 1.6G) (61). They substitute a pair of residues in GFP at selected locations to cysteine residues and used them to connect several GFP into a polyprotein via oxidizing cysteine residues. The disulfide linkage established upon oxidation allows the stretching force to be applied to GFP along the direction pre-determined by the two cysteine residues. As shown in Fig. 1.6G, the mechanical unfolding force of GFP exhibits great diversity and anisotropy depending on the pulling direction. The unfolding forces of GFP show a broad distribution with average forces ranging from ~100 pN to more than 600 pN. In addition, each individual GFP polyprotein also has a different spring constant. This study elegantly demonstrates that one protein building block can be used as a multifunctional nanomechanical protein building block with different mechanical stability in different directions.  28  1.2.7  The folding trajectory of proteins studied by force-clamp experiment The force-clamp technique enabled the first study of the folding of a protein under  a low constant force (Fig. 1.6H) (46). In the experiment, a polyubiquitin chain was first subjected to a high, constant force to unfold all the protein domains that have been stretched. Then the force was quickly quenched to a low constant value for a given time to allow the unfolded protein domains to fold. The force was again increased to a high value again to confirm the folding of protein domains in the given time. The folding time depends on the force applied to the protein. The larger the force is, the longer the time required to fold ubiquitin. Instead of observing a stepwise folding of individual domains in the polyubiquitin, Fernandez and Li found that all the domains were folded in a continuous fashion. This observation is in contrast with traditional views of protein folding and requires new physical models to explain it.  1.2.8  Single molecule mechanics on membrane proteins Because membrane proteins are highly hydrophobic, difficult to express in large  quantities and form large complexes, it is hard to study them using standard structural methods. AFM can study membrane proteins in their native form, deciphering the subunit structure and supermolecular arrangement by directly imaging the structure of the membrane surface (92) (93). Recently, the single molecule force spectroscopy has also been applied to address the structural arrangement and folding/unfolding of single membrane proteins. For example, the unfolding pathway of individual bacteriorhodopsin (BR) molecules has been investigated by AFM using force spectroscopy combined with direct imaging (94). The unfolding pathway of BR was found to be temperature  29  dependent (95). The free energy of BR unfolding has been calculated from its unfolding traces (96). Single molecule AFM also enables direct observation of BR folding under an applied external force (97), which is in contrast to many other proteins. These studies provide invaluable information about the structure and assembly of membrane proteins.  1.3 Engineering proteins with novel mechanical stability Although it has been more than 12 years since the first single molecule AFM experiment on the mechanical unfolding of a protein has been performed (33), the field of single molecule mechanics is still in its infancy. Only about 60 mechanical and nonmechanical proteins have been studied by single molecule force spectroscopy. Such a small pool of elastomeric proteins may be limited, since elastomeric proteins are in high demand for nano-mechanical applications. Therefore, it is of great importance and general interest to engineer proteins with novel mechanical properties. Extensive single molecule AFM studies and molecular dynamics simulations have been carried out to quantify the mechanical properties of a large spectrum of proteins and to understand the underlying mechanical design principles of these proteins (28, 57, 64, 98-104).  1.3.1  Experimental efforts on the engineering of mechanical proteins Significant mechanical stability has been found to be a common feature of  naturally occurring elastomeric proteins. It is natural to ask whether non-mechanical proteins can have significant mechanical stability. Early single molecule experiments on barnase showed that it is only marginally stable under an applied stretching force (68). However, close examinination of naturally occurring elastomeric proteins revealed that  30  they have neither special amino acid residues in the sequence nor unique structural motifs. Therefore, the efficient arrangement of the non-covalent interactions is the key to achieve high mechanical stability. This also implies that non-mechanical proteins that share similar structural motifs as mechanical proteins may achieve high mechanical stability. The search for mechanically stable non-mechanical proteins is well under way and is one topic in this dissertation (47, 62, 63). Another important effort in engineering the mechanical properties of proteins is trying to find the correlation between mechanical stability and other biophysical properties of proteins. Since single molecule AFM experiments typically require tedious protein engineering steps and it is still impossible to screen the proteins with unique mechanical stability from a large protein library, correlation of the mechanical stability with other properties, such as thermodynamic stability or chemical unfolding kinetics may provide an efficient way to screen proteins for mechanical applications (71, 105). It was found that the unfolding rate for I27 at zero molar of denaturant, guanidium hydrochloride (GdmCl), is similar to that at zero unfolding force (44). Therefore, based on the chemical unfolding kinetics, we may be able to predict the mechanical stability of the mutants of I27. However, further studies indicated the mutants of I27 did not follow this correlation (106). There are still no direct methods to predict the mechanical stability of a protein simply from its other biophysical properties. There are also considerable efforts being made in finding which types of interactions are important to the mechanical stability of a protein. A pioneering work on I27 indicated that the hydrogen bonding is critical to the mechanical stability (58). By deleting the backbone hydrogen bonds through proline mutation, the mechanical stability  31  of I27 is reduced significantly. A follow up experiment on the third fibornectin type III (FnIII) domain of tenascin indicated that the hydrophobic interactions are important to the mechanical stability of FnIII (77). However, the widely accepted view is that all kinds of different interactions including hydrogen bonding, hydrophobic interaction, chargecharge interaction and salt bridge can contribute to the mechanical stability of a protein. Their relative contribution largely depends on the structure of the protein and the position of the interaction, which cannot be predicted a priori. Many efforts in the engineering of mechanical proteins have been focused on tuning the mechanical stability of proteins. It was found that single proline mutations on the key regions for the mechanical unfolding of I27 can cause phenotypical effects on its mechanical stability, leading to a wide spectrum of different unfolding forces (58, 106). Such findings are also valid for other proteins, such as FnIII, with either alanine or proline mutations (77). However, these point mutations mainly produce mechanically weaker mutants. Although in a few cases, the resultant protein is mechanically more stable, the mechanism for increasing mechanical stability is still unknown. Recently, there were a few reported successes in the rational enhancement of the mechanical stability of proteins. For example, Li and coworkers were able to increase the mechanical stability of Top7 by changing its unfolding pathway through disulfide bond engineering (71). Clarke and coworkers showed that the mechanical stability of a FnIII domain from fibronectin can be enhanced by transplanting the hydrophobic core from a mechanically more stable FnIII domain from tenascin (107). However, general approaches to rationally enhance the mechanical stability of proteins have not yet been established.  32  1.3.2  Investigation of the mechanical unfolding of proteins in silico Single-molecule AFM experiments directly characterize the mechanical  properties of elastomeric proteins. However, the molecular scale mechanisms underlying these mechanical properties cannot be revealed. Molecular dynamics simulations offer an atomic scale picture of mechanical unfolding, provide insights into the mechanical unfolding intermediate state and guide single molecule experiments. Many molecular dynamics simulation methods have been established to study the mechanical unfolding of proteins (108, 109) (102, 103, 110-124). Among them, steered molecular dynamics simulation (SMD), in which external forces are applied to the macromolecules to investigate their mechanical responses, has become an important tool in the study of mechanical properties of elastomeric proteins. For example, the mechanical unfolding of I27 in the single molecule AFM experiment can be well explained by SMD simulations. The main unfolding barrier of I27 was found to be the breaking of 6 hydrogen bonds between strand A’ and G. Once this key region is unraveled, the rest of the structure can be unfolded by unzipping hydrogen bonds one by one. Moreover, the intermediate state of unfolding I27, which shows up as a hump in the force-extension curve before the main unfolding peak (Fig 1.6E), was identified as the breaking of 3 hydrogen bonds between the A and B strands. This event occurs a bit earlier than the main unfolding event. SMD results suggested that introducing mutants to disrupt the A-B interstrand hydrogen bonds would abolish the intermediate state. Indeed, the hump in the unfolding of I27 was abolished by a proline mutation at position 6, corroborating the scenario depicted in simulations (84). A recent simulation and experiment of mechanical unfolding of I27 also highlighted the importance of the solvent conditions used to the mechanical unfolding  33  (125, 126). By replacing water with glycerol or deuterium oxide, the unfolding barriers as well as the unfolding distance of I27 are changed. The experimental findings can be adequately described by SMD simulation. Molecular dynamics simulation also found that the mechanical stability of a protein is mainly determined by the interactions in the key regions for the mechanical unfolding of proteins (109, 121). Similar to the active site in enzymes, the key region for mechanical stability of a protein can be looked at as a “mechanical active site”. Before unraveling the interactions in the key regions, all the other interactions are not exposed to force. After the disruption of this region, the unfolding becomes barrier-less. It is worth noting that despite the significance of the mechanical active site for the mechanical stability of a protein, the overall structure of a protein also has a collective contribution to the mechanical stability. Therefore, engineering of proteins with novel mechanical stability should focus on but not be limited to the reengineering of the mechanical active site of a protein. Molecular dynamics simulation has also been used to predict the mechanical stability of proteins. Makarov and coworkers predicted that the ubiquitin-like proteins can have significant mechanical stability. (122, 127) Recent ambitious work by Cieplak and coworkers has attempted to predict the mechanical stabilities of all the proteins that have known three dimensional structure in protein data bank (PDB) using a Go-like coarsegrained models (103). Although Go-like model was disputed in studies of folding for oversimplifying the protein structure, it can be successfully used to predict the mechanical unfolding of proteins. The authors observed a good correlation (the Pearson  34  coefficient is 0.89) between the predicted unfolding forces and the experimentally obtained values.  1.3.3  The secondary structure, the topology and the mechanical stability of proteins The structure of proteins can be identified at different levels of detail. The  primary structure of a protein describes the amino-acid components and their order in the protein chain. The secondary structure describes how the amino acids form local threedimensional segments, such as alpha helices and beta sheets. The tertiary structure is the three dimensional framework of the whole protein that is defined by the atomic coordinates. “Protein topology” is a term used to specify how secondary structure of a protein is arranged in the three dimensional space, which denotes an intermediate level between secondary and tertiary structure of a protein. Both SMD simulations and single molecule AFM experiments of the mechanical unfolding of natural mechanical proteins revealed that the secondary structure and the topology of a protein play important roles in determining the mechanical stability (102-104, 108, 109, 121, 128). As we can see from Table 1.2 and Fig. 1.7, the alpha helical proteins are generally less stable than beta sheet ones and the proteins with shear topology are more stable than the proteins without (17, 44, 47, 62-64, 66-72, 74, 75, 77, 79, 88, 89, 129, 130). The secondary structure and the topology are very important for proteins to respond to the stretching force. Unlike chemical denaturation experiments on proteins, in which all the surface residues are exposed to denaturants, in the mechanical unfolding process, the force can only be felt by a small region of proteins between their N- and C-termini. Such a local region serves as a  35  mechanical clamp to protect the unraveling of the whole structure of proteins. If the secondary structure elements of this region are not stable enough under force, the protein will be easily unraveled. Although alpha helices are thermodynamically more stable than beta sheets, it is mechanically less stable. This is mainly due to different hydrogen bond patterns between alpha helix and beta sheets. In the alpha helix, hydrogen bonds are arranged in series along the force axis and can be unraveled sequentially where only a small force is required to unravel the whole structure of the protein. For beta sheet structure, depending on the topology of the protein, the force can be applied parallel to the beta sheet, resulting in a shearing force, or be applied perpendicular to the beta sheet, resulting in an unzipping force. If the force is parallel to the beta sheet, the hydrogen bonds in the beta sheet will break more-or-less concurrently, giving rise to the high mechanical strength of these proteins. However, if the beta sheet is unzipped by the force, the hydrogen bonds can still be unraveled sequentially, resulting in low mechanical stability. As shown in Fig. 1.7, a common feature of a mechanically stable protein is that the two terminal force-bearing β strands are arranged in anti-parallel. Such an arrangement of the force-bearing β strands constitutes a shear topology. The arrangement of A’ and G β strands in I27 is a typical example of the shear topology (Fig. 1.7). When the protein with shear topology is extended, the interactions (backbone hydrogen bonding and hydrophobic interactions) between the two force-bearing strands have to be unraveled synchronously. Therefore, these interactions serve as a mechanical clamp to resist mechanical force and form the molecular basis for the mechanical stability of the protein.  36  Figure 1.7 The secondary structure and topology of proteins that have been studied by single molecule AFM (The references refer to table 1.2).  1.4 Aim of this dissertation Although there have been great advances in the engineering of mechanical stability of proteins, the molecular determinants of mechanical stability are still largely unknown. It is still impossible to directly predict the mechanical stability of a protein from its three dimensional structure or even its sequence. It is also difficult to tune the mechanical stability of a protein in a rational fashion. Moreover, the application of elastomeric proteins as building blocks for biomaterials remains to be demonstrated. It is still challenging to engineer elastomeric protein based materials for biomedical applications.  37  In this dissertation I will mainly address the following 3 aspects of single molecule protein mechanics.  1.4.1  Searching for mechanically stable non-mechanical proteins Single molecule AFM combined with SMD has revealed that many mechanical  proteins share high mechanical stability, and those non-mechanical proteins having similar structural motifs as mechanical protein may also be mechanically stable. However, only a few non-mechanical proteins have been studied by single molecule AFM (Table 1.2). Mechanically stable non-mechanical proteins have only been reported in isolated cases. Have all the mechanically stable folds been sampled by naturally occurring mechanical proteins? Can proteins with novel folds that do not overlap with any known mechanical proteins be mechanically stable? Moreover mechanically stable nonmechanical proteins may impart novel functionality to elastomeric proteins, and enable applications such as force sensors (47) and switches.  1.4.2  Rational tuning of the mechanical stability of proteins Tuning the mechanical stability of proteins rationally is an important test for our  understanding of the molecular determinants of mechanical stability, and is also key toward designing elastomeric proteins for nanomechanical, material and biomedical applications. Mechanical stability is determined by the unfolding distance ∆xu as well as the free energy barrier, ∆GT-N, between the mechanical unfolding transition state and native state. To tune the mechanical stability of proteins, it is necessary to change the relative energetics of the native state and mechanical unfolding transition state, which has  38  proven challenging. Without clear knowledge about the molecular determinants of mechanical stability, the changes of unfolding force of proteins can only be gleaned in a trial-and-error fashion. As demonstrated in the pioneering work on I27, the proline mutations introduced using site directed mutagenesis were expected to disrupt the hydrogen bonding between A’ and G strands of I27 and lower the energy barrier of the transition state. However, single molecule AFM revealed that the unfolding distance of these mutants are different from the wild type I27 and the free energy barriers for unfolding are increased instead of being lowered in some of the mutants (58). The changes in both unfolding distance and the spontaneous unfolding rate made the unfolding force of I27 proline mutants unpredictable. Similar results have been shown on recombinants of I27 and I32 (131). It is clear that proteins will show different mechanical phenotypes upon mutations. To increase the mechanical stability of a protein using random mutations requires a big pool of protein mutants and an efficient screening strategy. These are beyond current experimental efforts. As shown in the free energy diagram of protein unfolding (Fig. 1.4), there are two ways to rationally tune the mechanically stability of a protein: 1) change the unfolding pathway to shift the position of unfolding transition state or 2) change the relative energy of the native state and transition state. There are a few rational ways to change the unfolding pathways. The first one is to change the direction of force applied to the protein. This has been demonstrated by a few groups (61, 69, 73). The second one is to shift the unfolding transition state along the force direction. For example, Li and coworkers have successfully tuned the mechanical unfolding pathway of a protein by introducing a disulphide bond to block the original unfolding pathway and force the protein to unfold  39  through a transition state with higher free energy (71). However, rationally altering the mechanical stability of a protein by changing the unfolding pathway requires preknowledge about the energetics of different unfolding pathways from either simulation or experimental work. In this thesis, I mainly focus on tuning the mechanical stability of a protein by changing its mechanical unfolding barrier. I demonstrate the modulation of mechanical stability of GB1 by chemical denaturant in Chapter 3. I also show two successful examples of enhancing the mechanical stability of GB1 by engineered metal ion chelation in Chapter 4, and by protein-protein interactions in Chapter 5. Based on the finding that the mechanical stability of GB1 can be tuned by protein-protein interaction, I demonstrate the engineering of elastomeric protein of dual mechanical stability that can be modulated reversibly by protein-protein interactions in Chapter 6.  1.4.3  Bottom-up construction of elastomeric protein based materials Single molecule AFM experiments not only provide new insight into the  mechanical design of proteins at the single protein domain level, but also shed light on the molecular basis of interactions that are important for the association and crosslinking of individual protein chains. For example, protein-ligand or protein-protein interactions have been characterized extensively using single molecule AFM under non-equilibrium and equilibrium conditions (132-135). The reduction of disulfide bonds under force has been quantitatively addressed by force-clamp force spectroscopy techniques (136). The direct relationship between the mechanical properties of single protein domains and inter domain interactions to the mechanical properties of the protein materials composed of these domain building blocks are to be established. It has been reported recently that a  40  titin mimic polymer that performs like a shock absorber at the single molecule level increases mechanical strength at the materials level (137). Therefore, the understanding of the mechanical architecture of elastomeric proteins and engineering novel artificial elastomeric proteins with tunable mechanical properties will enable the bottom-up construction of a new generation of protein-based biomaterials with well-defined mechanical properties. However, it is challenging to efficiently program inter-molecule interaction to efficiently assemble individual protein molecules into well-organized protein materials in a controllable and predictable fashion. As the first attempt to build macroscopic materials with elastomeric proteins, we have constructed novel hydrogels using a well-characterized artificial polyprotein (GB1)8 (138). Since we have discovered many means to tune the mechanical stability of GB1 at single molecule level, it will be interesting to further explore the correlation between single molecule mechanics with macroscopic mechanical performance. We anticipate that the concept of “mechanical engineering” will be revolutionary for the design and engineering of biomaterials with well-defined mechanical stability. More exciting discoveries and new surprises are yet to appear!  41  1.5 References 1. Howard J (2001) Mechanics of motor proteins and the cytoskeleton (Sinauer Associates, Publishers, Sunderland, Mass.). 2. Bustamante C, Chemla YR, Forde NR, & Izhaky D (2004) Annu Rev Biochem 73, 705-748. 3. Block SM (2007) Biophys J 92, 2986-2995. 4. Burgess SA & Knight PJ (2004) Curr Opin Struct Biol 14, 138-146. 5. Grabe M, Wang H, & Oster G (2000) Biophys J 78, 2798-2813. 6. Hopfner KP & Michaelis J (2007) Curr Opin Struct Biol 17, 87-95. 7. Lindemann CB (2003) Biophys J 84, 4115-4126. 8. Mehta AD, Rief M, Spudich JA, Smith DA, & Simmons RM (1999) Science 283, 1689-1695. 9. Wickner W & Schekman R (2005) Science 310, 1452-1456. 10. Xu L & Massague J (2004) Nat Rev Mol Cell Biol 5, 209-219. 11. Yoshida M, Muneyuki E, & Hisabori T (2001) Nat Rev Mol Cell Biol 2, 669-677. 12. Itoh H, Takahashi A, Adachi K, Noji H, Yasuda R, Yoshida M, & Kinosita K (2004) Nature 427, 465-468. 13. Berg HC (2003) Annu Rev Biochem 72, 19-54. 14. Labeit S & Kolmerer B (1995) Science 270, 293-296. 15. Hynes RO (1999) Proc Natl Acad Sci U S A 96, 2588-2590. 16. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. 17. Rief M, Pascual J, Saraste M, & Gaub HE (1999) J Mol Biol 286, 553-561. 18. Tatham AS & Shewry PR (2000) Trends Biochem Sci 25, 567-571. 19. Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, & Savage K (2002) Philos Trans R Soc Lond B Biol Sci 357, 121-132. 20. Murzin AG, Brenner SE, Hubbard T, & Chothia C (1995) J Mol Biol 247, 536540. 21. Bao G & Suresh S (2003) Nat Mater 2, 715-725. 22. Guerette PA, Ginzinger DG, Weber BH, & Gosline JM (1996) Science 272, 112115. 23. Becker N, Oroudjev E, Mutz S, Cleveland JP, Hansma PK, Hayashi CY, Makarov DE, & Hansma HG (2003) Nat Mater 2, 278-283. 24. Elvin CM, Carr AG, Huson MG, Maxwell JM, Pearson RD, Vuocolo T, Liyou NE, Wong DC, Merritt DJ, & Dixon NE (2005) Nature 437, 999-1002. 25. Li H (2007) Organic & Biomolecular Chemistry 5, 3399-3406. 26. Li HB (2008) Advanced Functional Materials 18, 2643-2657. 27. Oberhauser AF & Carrion-Vazquez M (2008) J Biol Chem 283, 6617-6621. 28. Brockwell DJ (2007) Current Nanoscience 3, 3-15. 29. Smith SB, Cui Y, & Bustamante C (1996) Science 271, 795-799. 30. Tskhovrebova L, Trinick J, Sleep JA, & Simmons RM (1997) Nature 387, 308312. 31. Kellermayer MS, Smith SB, Granzier HL, & Bustamante C (1997) Science 276, 1112-1116. 32. Strick TR, Croquette V, & Bensimon D (2000) Nature 404, 901-904.  42  33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.  Rief M, Gautel M, Oesterhelt F, Fernandez JM, & Gaub HE (1997) Science 276, 1109-1112. Rief M, Oesterhelt F, Heymann B, & Gaub HE (1997) Science 275, 1295-1297. Cluzel P, Lebrun A, Heller C, Lavery R, Viovy JL, Chatenay D, & Caron F (1996) Science 271, 792-794. Evans E, Ritchie K, & Merkel R (1995) Biophys J 68, 2580-2587. Smith SB, Finzi L, & Bustamante C (1992) Science 258, 1122-1126. Kim S, Blainey PC, Schroeder CM, & Xie XS (2007) Nat Methods 4, 397-399. Binnig G, Frank KH, Fuchs H, Garcia N, Reihl B, Rohrer H, Salvan F, & Williams AR (1985) Phys Rev Lett 55, 991-994. Binnig G & Rohrer H (1986) Ibm J Res Dev 30, 355-369. Greenleaf WJ, Woodside MT, & Block SM (2007) Annu Rev Biophys Biomol Struct 36, 171-190. Neuman KC & Nagy A (2008) Nat Methods 5, 491-505. Florin EL, Rief M, Lehmann H, Ludwig M, Dornmair C, Moy VT, & Gaub HE (1995) Biosensors & Bioelectronics 10, 895-901. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Oberhauser AF, Hansma PK, Carrion-Vazquez M, & Fernandez JM (2001) Proc Natl Acad Sci U S A 98, 468-472. Fernandez JM & Li H (2004) Science 303, 1674-1678. Dietz H & Rief M (2004) Proc Natl Acad Sci U S A 101, 16192-16197. Dietz H & Rief M (2006) Proc Natl Acad Sci U S A 103, 1244-1247. Fersht A (1999) Structure and mechanism in protein science : a guide to enzyme catalysis and protein folding (W.H. Freeman, New York). Bell GI (1978) Science 200, 618-627. Evans E & Ritchie K (1997) Biophys J 72, 1541-1555. Evans E & Ritchie K (1999) Biophys J 76, 2439-2447. Schlierf M & Rief M (2006) Biophys J 90, L33-35. Dudko OK, Hummer G, & Szabo A (2008) Proc Natl Acad Sci U S A 105, 1575515760. Rief M, Fernandez JM, & Gaub HE (1998) Phys Rev Lett 81, 4764-4767. Carrion-Vazquez M, Oberhauser A, Diez H, Hervas R, Oroz J, Fernandez J, & Martinez-Martin A (2006) in Advanced Techniques in Biophysics, eds. Arrondo J & Alonso A (Springer), pp. 163-236. Borgia A, Williams PM, & Clarke J (2008) Annu Rev Biochem 77, 101-125. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Williams PM, Fowler SB, Best RB, Toca-Herrera JL, Scott KA, Steward A, & Clarke J (2003) Nature 422, 446-449. Schlierf M & Rief M (2005) J Mol Biol 354, 497-503. Dietz H, Berkemeier F, Bertz M, & Rief M (2006) Proc Natl Acad Sci U S A 103, 12724-12728. Brockwell DJ, Beddard GS, Paci E, West DK, Olmsted PD, Smith DA, & Radford SE (2005) Biophys J 89, 506-519. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645.  43  64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.  Carrion-Vazquez M, Oberhauser AF, Fisher TE, Marszalek PE, Li H, & Fernandez JM (2000) Prog Biophys Mol Biol 74, 63-91. Hann E, Kirkpatrick N, Kleanthous C, Smith DA, Radford SE, & Brockwell DJ (2007) Biophys J 92, L79-81. Randles LG, Rounsevell RW, & Clarke J (2007) Biophys J 92, 571-577. Peng Q & Li H (2008) Proc Natl Acad Sci U S A 105, 1885-1890. Best RB, Li B, Steward A, Daggett V, & Clarke J (2001) Biophys J 81, 23442356. Carrion-Vazquez M, Li H, Lu H, Marszalek PE, Oberhauser AF, & Fernandez JM (2003) Nat Struct Biol 10, 738-743. Cao Y & Li H (2007) Nat Mater 6, 109-114. Sharma D, Perisic O, Peng Q, Cao Y, Lam C, Lu H, & Li H (2007) Proc Natl Acad Sci U S A 104, 9278-9283. Sharma D, Feng G, Khor D, Genchev GZ, Lu H, & Li H (2008) Biophys J 95, 3935-3942. Brockwell DJ, Paci E, Zinober RC, Beddard GS, Olmsted PD, Smith DA, Perham RN, & Radford SE (2003) Nat Struct Biol 10, 731-737. Schwaiger I, Kardinal A, Schleicher M, Noegel AA, & Rief M (2004) Nat Struct Mol Biol 11, 81-85. Oberhauser AF, Badilla-Fernandez C, Carrion-Vazquez M, & Fernandez JM (2002) J Mol Biol 319, 433-447. Li L, Huang HH, Badilla CL, & Fernandez JM (2005) J Mol Biol 345, 817-826. Ng SP, Rounsevell RW, Steward A, Geierhaas CD, Williams PM, Paci E, & Clarke J (2005) J Mol Biol 350, 776-789. Li H & Fernandez JM (2003) J Mol Biol 334, 75-86. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Li H, Oberhauser AF, Fowler SB, Clarke J, & Fernandez JM (2000) Proc Natl Acad Sci U S A 97, 6527-6531. Forman JR, Qamar S, Paci E, Sandford RN, & Clarke J (2005) J Mol Biol 349, 861-871. Oberhauser AF, Marszalek PE, Carrion-Vazquez M, & Fernandez JM (1999) Nat Struct Biol 6, 1025-1028. Carrion-Vazquez M, Marszalek PE, Oberhauser AF, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 11288-11292. Marszalek PE, Lu H, Li H, Carrion-Vazquez M, Oberhauser AF, Schulten K, & Fernandez JM (1999) Nature 402, 100-103. Lee G, Abdi K, Jiang Y, Michaely P, Bennett V, & Marszalek PE (2006) Nature 440, 246-249. Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, & Hansma PK (1999) Nature 399, 761-763. Li H, Oberhauser AF, Redick SD, Carrion-Vazquez M, Erickson HP, & Fernandez JM (2001) Proc Natl Acad Sci U S A 98, 10682-10686. Ainavarapu SR, Li L, Badilla CL, & Fernandez JM (2005) Biophys J 89, 33373344.  44  89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118.  Yang G, Cecconi C, Baase WA, Vetter IR, Breyer WA, Haack JA, Matthews BW, Dahlquist FW, & Bustamante C (2000) Proc Natl Acad Sci U S A 97, 139-144. Mickler M, Dima RI, Dietz H, Hyeon C, Thirumalai D, & Rief M (2007) Proc Natl Acad Sci U S A 104, 20268-20273. Bertz M & Rief M (2008) J Mol Biol 378, 447-458. Kedrov A, Janovjak H, Sapra KT, & Muller DJ (2007) Annu Rev Biophys Biomol Struct 36, 233-260. Engel A & Gaub HE (2008) Annu Rev Biochem 77, 127-148. Oesterhelt F, Oesterhelt D, Pfeiffer M, Engel A, Gaub HE, & Muller DJ (2000) Science 288, 143-146. Janovjak H, Kessler M, Oesterhelt D, Gaub H, & Muller DJ (2003) Embo J 22, 5220-5229. Preiner J, Janovjak H, Rankl C, Knaus H, Cisneros DA, Kedrov A, Kienberger F, Muller DJ, & Hinterdorfer P (2007) Biophys J 93, 930-937. Kessler M, Gottschalk KE, Janovjak H, Muller DJ, & Gaub HE (2006) J Mol Biol 357, 644-654. Brockwell DJ (2007) Biochem Soc Trans 35, 1564-1568. Rounsevell R, Forman JR, & Clarke J (2004) Methods 34, 100-111. Fisher TE, Oberhauser AF, Carrion-Vazquez M, Marszalek PE, & Fernandez JM (1999) Trends Biochem Sci 24, 379-384. Zhuang XW & Rief M (2003) Current Opinion in Structural Biology 13, 88-97. Paci E & Karplus M (2000) Proc Natl Acad Sci U S A 97, 6521-6526. Sulkowska JI & Cieplak M (2007) J Phys-Condens Mat 19, -. Sotomayor M & Schulten K (2007) Science 316, 1144-1148. Balamurali MM, Sharma D, Chang A, Khor D, Chu R, & Li H (2008) Protein Sci 17, 1815-1826. Fowler SB, Best RB, Toca Herrera JL, Rutherford TJ, Steward A, Paci E, Karplus M, & Clarke J (2002) J Mol Biol 322, 841-849. Ng SP, Billings KS, Ohashi T, Allen MD, Best RB, Randles LG, Erickson HP, & Clarke J (2007) Proc Natl Acad Sci U S A 104, 9633-9637. Lu H, Isralewitz B, Krammer A, Vogel V, & Schulten K (1998) Biophys J 75, 662-671. Lu H & Schulten K (2000) Biophys J 79, 51-65. Lu H, Krammer A, Isralewitz B, Vogel V, & Schulten K (2000) Adv Exp Med Biol 481, 143-160; discussion 161-142. Lu H & Schulten K (1999) Proteins 35, 453-463. Lee EH, Gao M, Pinotsis N, Wilmanns M, & Schulten K (2006) Structure 14, 497-509. Craig D, Gao M, Schulten K, & Vogel V (2004) Structure 12, 21-30. Gao M, Craig D, Lequin O, Campbell ID, Vogel V, & Schulten K (2003) Proc Natl Acad Sci U S A 100, 14784-14789. Tajkhorshid E, Aksimentiev A, Balabin I, Gao M, Isralewitz B, Phillips JC, Zhu F, & Schulten K (2003) Adv Protein Chem 66, 195-247. Gao M, Lu H, & Schulten K (2002) J Muscle Res Cell Motil 23, 513-521. Gao M, Wilmanns M, & Schulten K (2002) Biophys J 83, 3435-3445. Gao M, Craig D, Vogel V, & Schulten K (2002) J Mol Biol 323, 939-950.  45  119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138.  Gao M, Lu H, & Schulten K (2001) Biophys J 81, 2268-2277. Isralewitz B, Gao M, & Schulten K (2001) Curr Opin Struct Biol 11, 224-230. West DK, Brockwell DJ, Olmsted PD, Radford SE, & Paci E (2006) Biophys J 90, 287-297. Li PC & Makarov DE (2004) Journal of Physical Chemistry B 108, 745-749. Huang L, Kirmizialtin S, & Makarov DE (2005) J Chem Phys 123, 124903. Kirmizialtin S, Huang L, & Makarov DE (2005) J Chem Phys 122, 234915. Dougan L, Koti AS, Genchev G, Lu H, & Fernandez JM (2008) Chemphyschem 9, 2836-2847. Dougan L, Feng G, Lu H, & Fernandez JM (2008) Proc Natl Acad Sci U S A 105, 3185-3190. Li PC, Huang L, & Makarov DE (2006) J Phys Chem B 110, 14469-14474. Klimov DK & Thirumalai D (2000) Proc Natl Acad Sci U S A 97, 7254-7259. Junker JP, Hell K, Schlierf M, Neupert W, & Rief M (2005) Biophys J 89, L46-48. Wilcox AJ, Choy J, Bustamante C, & Matouschek A (2005) Proc Natl Acad Sci U S A 102, 15435-15440. Sharma D, Cao Y, & Li H (2006) Angew Chem Int Ed Engl 45, 5633-5638. Florin EL, Moy VT, & Gaub HE (1994) Science 264, 415-417. Baumgartner W, Hinterdorfer P, Ness W, Raab A, Vestweber D, Schindler H, & Drenckhahn D (2000) Proc Natl Acad Sci U S A 97, 4005-4010. Nevo R, Stroh C, Kienberger F, Kaftan D, Brumfeld V, Elbaum M, Reich Z, & Hinterdorfer P (2003) Nat Struct Biol 10, 553-557. Cao Y, Balamurali MM, Sharma D, & Li H (2007) Proc Natl Acad Sci U S A 104, 15677-15681. Wiita AP, Perez-Jimenez R, Walther KA, Grater F, Berne BJ, Holmgren A, Sanchez-Ruiz JM, & Fernandez JM (2007) Nature 450, 124-127. Guan Z, Roland JT, Bai JZ, Ma SX, McIntire TM, & Nguyen M (2004) J Am Chem Soc 126, 2058-2065. Cao Y & Li H (2008) Chem Commun (Camb), 4144-4146.  46  Chapter 2: Expanding the toolbox of elastomeric proteins to include non-mechanical proteins†  2.1 Introduction Elastomeric proteins are passive elastic elements and serve as molecular springs inside cells as well as in extracellular matrix space to establish elastic connections and provide mechanical strength and extensibility (1, 2). Elastomeric proteins can also function as the building blocks for materials of superb mechanical properties, such as spider dragline silk protein (3-5). Constantly subject to repeated stretching-relaxation cycles, many elastomeric proteins demonstrate remarkable consistency and reliability in their mechanical performance. One of the common features of the elastomeric proteins is their tandem modular construction (1, 2). For example, the giant muscle protein titin is composed of hundreds of individually folded immunoglobulin-like (Ig) domains and fibronectin type III (FnIII) domains, interspersed by random coil-like unique sequences (6, 7). It was discovered that a stretching force can trigger sequential mechanical unfolding of individually folded domains (8-10). Such a “modular” unfolding mechanism conveys high toughness to elastomeric proteins and makes them perfect shock-absorbers. Upon relaxation, the unfolded domains can refold to their original folded state and recover their mechanical resistance. This mechanism has been exploited extensively in nature and can be found in a wide variety of materials, ranging from muscle fibers (6), to †  A version of this chapter has been published as two papers: “Cao, Y.; Lam, C.; Wang, M.; Li, H., Nonmechanical protein can have significant mechanical stability. Angew Chem Int Ed Engl 2006, 45, (4), 642-5.” and “Cao, Y.; Li, H., Polyprotein of GB1 is an ideal artificial elastomeric protein. Nat Mater 2007, 6, (2), 109-14.”.  47  spider silk fibers (4) to biological adhesives in abalones (11). Inspired by the development in nanotechnology as well as the naturally evolved elastomeric proteins, researchers have recently started to explore the potential unique mechanical features that non-mechanical proteins may offer (12-15). Although nonmechanical proteins do not have known mechanical function and are not evolved in nature for mechanical purposes, they may carry unique traits that nature does not select for but can be used for mechanical purposes. This will open up new avenues in the design of artificial elastomeric proteins for novel nanomechanical applications. For example, the mechanical unfolding of green fluorescent protein (GFP) was investigated in detail by single molecule AFM, with the ultimate goal of correlating the mechanical unfolding with the fluorescence change of GFP and employing GFP as a molecular force sensor (12). Along a similar line, experimental efforts are underway to explore non-mechanical proteins with significant mechanical stability to expand the toolbox of elastomeric proteins and develop artificial protein-based molecular springs. As discussed in chapter 1, mechanically stable proteins share a common shear topology. Using the shear topology as a search criterion, we identified that the B1 immunoglobulin-binding domain of protein G from Streptococcus (referred to as GB1) can have significant mechanical stability (16). Here we characterize the mechanical performance of GB1 in the form of a polyprotein (GB1)8 that is made of eight direct tandem repeats of GB1 domains to fully explore the potential of these non-mechanical proteins as elastomeric proteins. It was found that GB1 has significant mechanical stability which is comparable with that of I27 of the muscle protein titin. In addition to its mechanical stability, we show that the GB1 polyprotein exhibits a unique combination of  48  mechanical features, including the fastest folding kinetics measured to date for a tethered protein, high folding fidelity, and low mechanical fatigue during repeated stretchingrelaxation cycles and ability to fold against residual forces. These features make GB1 polyprotein an ideal artificial protein-based molecular spring that could function in a challenging working environment requiring repeated stretching-relaxation. This study represents a key step towards engineering artificial molecular springs with tailored nanomechanical properties for bottom-up construction of new devices and materials.  2.2 Results 2.2.1 GB1 exhibits significant mechanical stability GB1 is a small α+β protein with only 56 amino acid residues and consists of a four stranded β sheet packed against a long α helix (16). The two terminal β strands of GB1 are arranged in parallel and have the characteristic shear topology (Fig. 2.1A). To mimic the tandem modular nature of naturally occurring elastomeric proteins, we used protein engineering techniques to construct polyprotein (GB1)8 consisting of eight identical tandem repeats of GB1 domains, and used single molecule AFM to stretch the constructed polyprotein to measure its mechanical properties (Fig. 2.1B). Stretching polyprotein (GB1)8 results in force-extension curves of characteristic saw-tooth pattern appearance, with as many as eight force peaks. Typical force-extension curves are shown in Fig. 2.1C. Similar to naturally occurring molecular nanosprings, (GB1)8 appears to be a nonlinear spring for which the force increases nonlinearly as a function of extension, in contrast with the recently discovered Hookean or linear spring elastic behavior for ankyrin (17). This nonlinear entropic elastic behavior can be well described by the Worm-  49  like-chain (WLC) model of polymer elasticity (Fig. 2.1C) (18). The saw-tooth force peaks arise from the sequential unfolding of the individual GB1 domains in the polyprotein chain and are equally spaced by an average peak-to-peak distance of 15.7±0.9 nm (n=910). The WLC model fits the force-extension curves well and no unfolding intermediate state is detectable, suggesting that the mechanical unfolding of GB1 is an apparent two-state process. WLC fits to the consecutive force peaks yield a contour length increment of 18.0±0.5 nm (n=472), indicating that each GB1 will elongate by 18 nm upon its mechanical unfolding. A GB1 domain is 20.2 nm (56aa×0.36 nm/aa) long when it is fully stretched, while the length of the folded GB1 domain (the distance between its N- and C- termini) is 2.6 nm (16). Hence, the mechanical unfolding of a GB1 domain will lengthen the molecule by 17.6 nm, which is in good agreement with the contour length increment between consecutive peaks measured from the force-extension curves.  50  Figure 2.1 Polyprotein (GB1)8 has significant mechanical stability. A) The three dimensional structure of GB1 displays a shear topology. The two terminal strands (strands 1 and 4) are parallel and are bonded by a series of backbone hydrogen bonds (indicated by lines). Upon stretching (indicated by arrows), the backbone hydrogen bonds between these two strands are subject to a shear force and may form the mechanical resistance to unfolding. B) The schematic diagram of polyprotein (GB1)8. Eight identical GB1 monomers are joined in tandem by connecting the N- and C- termini. C) Typical force-extension curves of (GB1)8 polyproteins. These force-extension curves show characteristic saw-tooth pattern, with equally spaced force peaks, which arise from the mechanical unraveling of each individual GB1 domains in the polyprotein chain. The last peak in each force-extension curve corresponds to the detachment of the protein from either the AFM tip or substrate. The red lines are WLC fits to the individual unfolding force peaks in a representative force-extension curve. D) Histogram of unfolding forces for GB1 protein. The gray bars are experimental data measured from ~1200 forceextension curves (as those shown in C) with a total of 6691 unfolding events, at a pulling speed of 400nm/s. The unfolding force histogram for GB1 peaks at ~ 180pN and shows a broad distribution spanning a range of ~150pN. The solid red line is a Monte Carlo simulation of the mechanical unfolding of (GB1)8 polyprotein by using a spontaneous unfolding rate constant, α0=0.039s-1, an unfolding distance, ∆xu=0.17nm and a pulling velocity of 400nm/s. Despite the fact that (GB1)8 consists of eight identical tandem repeats of GB1, the amplitude of each individual GB1 unfolding force peak in the force-extension curves varies randomly around an average value of approximately 180 pN, reflecting the  51  stochastic nature of mechanical unfolding of proteins. The histogram of the unfolding forces measured at a pulling velocity of 400nm/s shows a broad distribution, spanning a range from 50 pN to 300 pN (Fig. 2.1D). The average unfolding force of GB1 is 184±41 pN (average±S.D., n=6991), similar to that of I27, which is a well-characterized and naturally occurring mechanical protein from the giant muscle protein titin (19). The amplitude of unfolding force peaks is a direct measure of the mechanical stability of a protein. Therefore, the mechanical stability of GB1 is comparable to that of the I27 domain. This result demonstrates that, although GB1 is not naturally evolved for mechanical purposes and does not have known mechanical functions, GB1 has significant mechanical stability. Such high mechanical stability offers the potential for GB1 to be used in mechanical applications. It is worth noting that the unfolding force distribution for GB1 is very broad. The standard deviation of the unfolding forces for GB1 is 41 pN and is significantly larger than that of any other mechanical protein studied so far by single molecule AFM. The standard deviation of unfolding force is governed by the thermal force (kBT/xu) instead of experimental uncertainty. Therefore, the broad distribution in the unfolding force for GB1 reflects the intrinsic properties of the underlying mechanical unfolding energy landscape.  52  Figure 2.2 Mechanical unfolding of GB1 is a non-equilibrium process. A) A pair of typical stretching (black line) and relaxation (red line) curves of a polyprotein GB1 at a pulling speed of 400nm/s. The stretching curve shows a saw-tooth pattern with 7 GB1 unfolding peaks. In contrast, the relaxation curve shows only a non-linear entropic elastic behaviour. The shadowed area reflects the energy dissipated during stretching. B) Unfolding kinetics of GB1 protein. The unfolding forces of GB1 (squares) strongly depend on the pulling speed at which the polyprotein was stretched and unraveled. The measured unfolding forces at high pulling speeds were corrected for the hydrodynamic drag force acting directly on the AFM cantilever. The unfolding kinetics of GB1 can be reproduced adequately by Monte Carlo simulation using α0=0.039s-1 and ∆xu=0.17nm (solid line).  53  2.2.2 Mechanical unfolding of GB1 is a non-equilibrium process. To check the reversibility of the mechanical unfolding of GB1, we first extended polyprotein (GB1)8 to unravel all the GB1 domains in the polyprotein chain (black trace, Fig. 2.2A) and then relaxed the unfolded polypeptide chain to measure its force-extension relationship during relaxation (red trace, Fig. 2.2A). The force-extension relationship of the unfolded (GB1)8 during relaxation shows a nonlinear entropic elastic behavior, which is distinct from the saw-tooth force-extension curve of (GB1)8 during stretching. This result indicated that much of the energy invested during stretching (shadowed area) is dissipated as heat in the mechanical unfolding of GB1 domains, and suggested that the mechanical unfolding of GB1 is a non-equilibrium process. The non-equilibrium nature of the mechanical unfolding of GB1 is similar to many naturally occurring mechanical proteins. Since the mechanical unfolding of GB1 is a non-equilibrium process, it is expected that the unfolding force of GB1 will depend upon the velocity at which polyprotein (GB1)8 is extended and unraveled. Indeed, as shown in Fig. 2.2B, the higher the pulling velocity is, the higher the force is required to unravel GB1. Such velocitydependence of the mechanical stability is a general feature of non-equilibrium bond rupture-like processes (20). Since there is no detectable unfolding intermediate state in the force-extension curves, we can model the mechanical unfolding reaction of GB1 as a two-state unfolding process with a force-dependent unfolding rate constant. Using a standard Monte Carlo simulation procedure, we reproduced the force-extension curves of polyprotein (GB1)8 as those shown in Fig. 2.1C to estimate the unfolding rate constant α0 and the unfolding  54  distance between the folded state and transition state xu. We found that, using a α0 of 0.039 s-1 and a xu of 0.17 nm, we can adequately reproduce the unfolding force distribution (Fig. 2.1D), as well as the velocity dependence of the average unfolding force (Fig. 2.2B). The small unfolding distance indicates that the transition state for mechanical unfolding for GB1 is close to the native state. The estimated xu for GB1 is smaller than most mechanical proteins, and underlies the broad distribution of unfolding forces and the stronger dependence of the average unfolding force on the pulling velocity.  Figure 2.3 The fast folding kinetics of GB1. A) The folding kinetics of GB1 is probed by AFM using a double-pulse protocol (inset). The polyprotein is first stretched to unfold all the GB1 domains in the chain and count the total number of domains that are available in the polyprotein chain, Ntotal, (upper traces), and then the unfolded polyprotein is quickly relaxed back to its original length within 2ms. After a relaxation time t, the protein is stretched again to count the number of domains that refolded during the relaxation time, Nrefolded (lower traces). B) Plot of the refolding probability, Nrefolded/Ntotal, versus t. Since the unfolding rate constant is negligible at zero force, the folding kinetics can be described adequately using a simple first-order kinetic equation, Nrefolded/Ntotal (t) =1exp(-β0⋅ t) using β0= 720±120 s-1(black solid line). The data point at zero waiting time was added to facilitate the fitting.. For comparison, we also plotted equation (1) using β0 of 2000 s-1 (grey solid line) and 200 s-1 (grey dotted line), the fast phase folding rate constant of GB1 in water and the fastest folding rate constant reported so far for tethered proteins. 55  2.2.3 The fast mechanical folding kinetics of GB1 The capability of a protein to fold fast with high fidelity is an important requisite for a candidate protein to be used for nanomechanical applications. Using ultrafast mixing, it was shown that GB1 is a fast folder via biphasic kinetics: a fast phase with a folding rate constant of ~2000s-1 (minor phase) and a slow phase with a folding rate constant of ~700 s-1 (major phase) in water (21). However, upon stretching, both ends of the protein are tethered and its folding kinetics may differ from that of the protein when the protein is free in solution. Compared with proteins free in solution, tethering may result in a restriction in the degrees of freedom for the protein, which might influence the folding kinetics of the protein. It has been shown that for some mechanical proteins (19, 22, 23), tethering both ends of the protein slowed down the folding kinetics significantly. For example, the folding rate constant of I27 at zero force was almost 30 times slower than that of I27 free in solution (19). However, it remains to be established whether the slowing down effect observed in these proteins is intrinsically due to the tethering and ubiquitous among proteins. Here we use single molecule AFM to measure the folding kinetics of GB1 when it is tethered at both ends. We used a double pulse protocol to probe the mechanical folding kinetics of GB1 at zero force (Fig. 2.3A inset). First, (GB1)8 polyprotein was extended to unfold all the GB1 domains in the polyprotein chain being picked up and stretched by the AFM tip. This allows us to count the total number of GB1 domains, Ntotal, that are available in the given polyprotein chain. Since (GB1)8 polyprotein is picked up randomly along its contour by the AFM tip, Ntotal, is typically less than the maximum number of 8. After the first pulse, the unfolded polyprotein chain was relaxed quickly to zero extension (usually within 2 ms) before it detached from either  56  the AFM tip or substrate. After the polyprotein relaxed at zero extension for a variable period of time, t, (from 0 to 50 ms), it was stretched again by the second pulse, resulting in a force-extension curve of sawtooth appearance. Since the mechanical stability is a unique property of the folded GB1 domains, we interpret the sawtooth pattern as evidence that some of the GB1 domains that unfolded in the first pulse refolded spontaneously to their native state and regained their mechanical stability upon relaxation. The number of GB1 domains that refolded (Nrefold) was observed to depend exponentially upon the length of the relaxation time t. It is evident that almost 100% of GB1 domains will refold when the relaxation time is longer than 10 ms. Since the unfolding rate constant α0 at zero force is significantly smaller than the folding rate constant β0, we can treat the folding event of GB1 at zero force using first-order kinetics, Nrefold/Ntotal=1exp(β0×t), where β0 is the folding rate constant of GB1 at zero force. Fitting this function to our experimental data yields a folding rate constant of β0=720±120 s-1. A folding rate constant of ~700 s-1 is the fastest folding rate constant reported so far for proteins tethered at both termini: 600 times faster than that of I27 and more than 3 times faster than the fast folder filamin (24). Such folding rate constant makes GB1 the fastest folder under tethering reported so far. This result indicates that it is possible for a protein to fold quickly when tethered at both ends, making polyprotein (GB1)8 a perfect candidate for artificial nanomechanical springs, in that it can regain its mechanical stability much more quickly. The folding rate constant measured at zero force for GB1 is very similar to the rate constant of the major phase of GB1 free in solution, indicating that mechanical tethering at both ends of GB1 does not impose significant constraints on the folding of  57  GB1. This observation is in sharp contrast to the slowing down effect of tethering on the protein folding kinetics previously observed on a range of proteins, such as I27 from titin(19, 23) and ubiquitin (22).  Figure 2.4 GB1 can fold in the presence of residual forces. A) Folding kinetics of GB1 under different force is probed by a modified double pulse protocol (inset). First, the protein is stretched to a certain length (L0) to unfold all the GB1 domains in the polyprotein chain. Then, it is rapidly relaxed to a shorter length (x) and held for a fixed time period (10ms). A second pull of the protein measures the number of domains refolded (Nrefolded) during the waiting time under the force. The total number of peaks in the first pulling curve within the length x to L0 is counted as Ntotal. B) Plot of Nrefolded/Ntotal vs. residual force. Black squares correspond to the data obtained with t=10 ms, while grey squares correspond to the data obtained with t=1 s. The residual force acting on the unfolded polyprotein chain slowed down the folding process of GB1 significantly. Solid lines correspond to the fits, defined as Pƒ(F)= Nrefolded/Ntotal =1- exp(-t β0·exp(-F·∆xf/kBT), where t is the relaxation time, F is the force acting on the protein, ∆xf is the folding distance, kB is Boltzmann Constant, T is absolute temperature and β0 is the folding rate in the absence of force. The red line was generated using ∆xf=2.1 nm, β0 = 720 s-1, and t=10 ms; while the light orange line was generated using ∆xf = 2.1 nm, β0 = 720 s-1, and t = 1 s.  58  2.2.4 GB1 can fold under residual forces. Upon stretching, the folding rate constant of GB1 is also affected by the stretching force acting on the polyprotein chain, which can be described as follows: β(F)= β0 ·exp(F·∆xf/kBT), where, β(F) is the folding rate constant in the presence of force F, β0 is the folding rate constant at zero force; ∆xf is the folding distance between the folded state and the unfolded state along the reaction coordinate defined by the stretching force, kB is the Boltzmann constant, and T is absolute temperature in Kelvin. We used a modified double pulse protocol to estimate the folding distance of GB1 by measuring the folding rate constant in the presence of residual forces. As shown in Fig. 2.4A inset, the protein was first extended to a certain length (L0) to unravel all the GB1 domains in the polyprotein chain, allowing us to count the total number of GB1 domains in the given polyprotein chain. Then the unfolded polyprotein chain was relaxed partially to a shorter extension (x) and allowed to refold for a fixed period of time (10 ms). For any x that is larger than zero, a residual force will be acting on the polyprotein chain due to the entropic elasticity of polymer chain, which can be calculated using the WLC model of polymer elasticity. During the second pulse, the polyprotein chain was extended again to L0 to measure the total number of GB1 domains that refolded (Nrefolded) at an extension of x in 10 ms, which manifested themselves in force-extension curves as sawtooth peaks. One such experiment is illustrated in Fig. 2.4A. A polyprotein fragment containing 7 GB1 domains was picked up and extended in the first pulse (Fig. 2.4A), resulting in force-extension curves with 7 unfolding force peaks. When the molecule was relaxed to zero extension for 10 ms, all  59  the unfolded GB1 domains refolded. When the polyprotein was only relaxed partially (x>0), the number of domains that refolded during 10 ms relaxation time decreased sharply as x increased. Strictly speaking, the residual force acting on the protein does not remain constant but increases when one domain manages to refold. However, the force increase is very small (a few pN). Hence, as an approximation, we assumed that the force remained constant. The curve of probability of refolding (Nrefolded/Ntotal) versus the residual force (F), which is calculated from x using the WLC model, shows a reverse sigmoid shape (Fig. 2.4B). It is of note that the folding for GB1 is almost completely inhibited in a time window of 10 ms by a force of ~12 pN. Assuming the folding reaction under force is a two-state reaction, we can use the following equation to fit the relationship of Nrefolded/Ntotal versus F: Nrefolded/Ntotal (F) =1- exp(-t·β0·exp(-F·∆xf/kBT)  (2.1)  where t is the folding time of 10 ms and β0 is the folding rate constant of GB1 at zero force of 720 s-1. The experimental data can be described adequately by equation (2.1) using ∆xf=2.1 nm. This result suggested that the folding of GB1 involves a contraction of ~2.1 nm along the direction of the applied force. It is also evident that the probability of refolding at a given force F will increase upon increasing the observation time window t. Indeed, when the observation time window increased to 1s, the observed Nrefolded/Ntotal vs. F curve shifted towards higher forces and the folding probability at the same force increased considerably. It is evident that GB1 can fold at a considerable rate even at a residual force up to 15 pN.  60  2.2.5 Free energy diagram for the mechanical unfolding and refolding of GB1. Combining the unfolding and refolding experiments on polyprotein (GB1)8, we reconstructed the free energy diagram characterizing the mechanical unfolding and refolding reaction of GB1 (Fig. 2.5). The free energy diagram is characterized by its asymmetry between the unfolding and refolding reactions: the distance between the folded and transition state is only 0.17 nm, while the folding distance 2.1 nm is more than 10 times the unfolding distance. This asymmetry is similar to that of the mechanical protein I27, raising the question whether such asymmetry is an essential feature of mechanically stable proteins. The small unfolding distance is essential for conferring the high mechanical stability to GB1. Such an asymmetry also results in the different responses of unfolding and folding rate constants to the stretching force: the folding rate constant is much more sensitive to force than the unfolding rate constant. Although the GB1 folds really fast with a folding rate constant of ~700 s-1 at zero force, the folding reaction of GB1 does not generate noticeable forces (>20 pN) at a pulling speed of 400 nm/s, as no force peaks can be observed in the relaxation curve of polyprotein (GB1)8. This is in contrast to ankyrin which has the ability to generate force during folding (17). More accurate force measurements are required to precisely determine the folding force of GB1 (25-27). It is of note that, although GB1 is mechanically very stable, the unfolding rate constant for GB1 is significantly faster than that of typical mechanical proteins studied so far (9, 19, 28-30), indicating that the free energy barrier for the mechanical unfolding of GB1 is much lower than that of known mechanical proteins. This observation corroborates that mechanical stability is not only correlated with the free energy barrier  61  for unfolding, but also depends upon the width of the potential well (31). A steep potential well of protein GB1 could yield higher unfolding forces than a shallow potential well of similar depth. -1  720s  -1  0.039s  Extended Unfolded  Free Energy  Unfolded  Folded  2.1nm 0.17nm 18.0nm  End to End Distance Figure 2.5 Schematic drawing of the free energy diagram for the mechanical unfolding and refolding of GB1.  2.2.6 Polyprotein GB1 does not show noticeable mechanical fatigue. Force induced unfolding-refolding cycles may be part of the natural life of natural elastomeric proteins. Some mechanical fatigue may occur for mechanical proteins, as it was observed for native titin molecules at low force. In fact, titin exhibited a progressive lengthening of the contour length during repetitive stretching/relaxation cycles. This phenomenon is believed to be the origin of the mechanical fatigue observed on relaxed striated muscle cells. In the previous sections, we have shown that GB1 refolds fast and that it can fold under residual forces. Can a molecular spring made from the nonmechanical protein GB1 function in the challenging working environment requiring  62  repetitive stretching/relaxation cycles display similar or even less mechanical fatigue than its natural counterparts? To address this question, we subject a single (GB1)8 molecule to repetitive stretching/relaxation for as many cycles as possible before the protein detaches from either the AFM tip or substrate. Fig. 2.6A shows the force-extension curves after a number of stretching/relaxation cycles from such an experiment in which the protein survived a total of 276 cycles before it detached. Between consecutive cycles, the protein was relaxed at zero extension for 15 ms. The folding probability of GB1 remained constant throughout the experiment (Fig. 2.6B, top panel, blue symbols), indicating that almost all the GB1 domains in the chain can regain their mechanical stability after relaxation, regardless of the number of undergone stretching/relaxation cycles. This result demonstrates that there is no noticeable mechanical fatigue preventing GB1 domains from folding. Furthermore, GB1 did not show significant fatigue in the form of reduced mechanical stability. As shown in Fig. 2.6B (top panel, red symbols), the average unfolding forces of GB1 remained constant (~180 pN) throughout the experiment. Similar results were observed on additional polyprotein GB1 molecules which were subject to repetitive stretching/relaxation cycles (Fig. 2.6B, middle and bottom panels). The overall unfolding force histogram of GB1 compiled from repetitive stretching/relaxation cycles (Fig. 2.6C, red) is almost identical to that obtained by stretching individual (GB1)8 (Fig. 2.6C, grey). These results indicate that GB1 retained its mechanical stability during repeated stretching-relaxation cycles and that no mechanical fatigue is present, at both the individual molecule and the ensemble levels. This stability is similar to that of projectin (an insect flight muscle protein) (32), and clearly outperformed mammalian titin (33, 34), at least at the single molecule level.  63  This observation also serves as a good example in demonstrating the validity of the ergodic hypothesis implicitly assumed when chemical thermodynamics is applied to single molecule studies. The hypothesis of ergodicity states that the properties of a single molecule averaged over time are equivalent to the average properties of many molecules at one instant in time (35). The unfolding force histogram obtained on a single molecule during repeated stretching-relaxation cycles is the time-average of a single molecule, while the unfolding force histogram obtained by stretching different individual molecules represents the average over many different molecules (ensemble average). The close agreement between these two unfolding force histogram indicates that the ergodicity is valid.  64  Figure 2.6 Polyprotein (GB1)8 does not show noticeable mechanical fatigue. A) Forceextension curves of a polyprotein (GB1)8 during repeated stretching-relaxation experiments. The number above each curve indicates the number of cycles that the polyprotein has been subject to. B) The unfolding force and folding probability of GB1 remain unchanged during repetitive stretching/relaxation cycles. The top panel was measured from the same molecule shown in Fig. 2.6A. The middle and bottom ones were measured from two additional molecules that were subjected to repeated stretchingrelaxation. The blue diamonds and red squares correspond to the average folding probability and unfolding force of GB1 in 10 consecutive cycles, respectively. The lines indicate linear fitting to experimental data. C) Repeated stretching-relaxation cycles do not weaken the mechanical stability of GB1. The unfolding force histogram compiled from the unfolding events of GB1 during the repeated stretching/relaxation cycles for the same polyprotein (red bars) is indistinguishable from the unfolding force histogram obtained by stretching different individual (GB1)8 polyproteins (grey bars) as those shown in Fig. 2.1C.  65  2.2.7 High fidelity in the refolding of GB1 Misfolding can occur for elastomeric proteins and will result in an altered mechanical response (36). A recent ensemble chemical unfolding/folding study suggested that tandem modular proteins with high sequence identity are prone to misfolding and aggregation due to the effective high local protein concentration in the vicinity of constituting domains in the tandem modular proteins (37). An implication of this study is that homopolyproteins, such as (GB1)8, are potentially prone to misfolding and as such their mechanical properties will be compromised. Indeed, misfolding events were reported for tandem modular proteins with varied frequency (2% for polyprotein (I27)12 and 4% for a recombinant fragment of tenascin (36)). To explore whether misfolding of GB1 in the polyprotein could potentially jeopardize the mechanical performance of GB1, we monitored the folding fidelity of (GB1)8 using repeated stretching/relaxation protocols. The contour length increment upon domain unfolding (∆Lc) is sensitive to misfolding and has been used to monitor the formation of misfolded states (23, 36). GB1 has a ∆Lc of 18.0±0.5 nm. We found that more than 99.8% of the unfolding events of GB1 domains in repetitive stretching/relaxation cycles show ∆Lc of ~18nm, identical to those in single pulling experiments (Fig. 2.1C). And the average unfolding forces of GB1 domains in repetitive stretching/relaxation cycles remained unchanged (Fig. 2.4B). These results strongly indicated that the folding of GB1 proceeded with exceptionally high fidelity, which is even superior to some natural elastomeric proteins (36).  66  2.3 Discussion Our results demonstrated that the protein GB1 is mechanically stable and that polyprotein (GB1)8 has ideal features for being an artificial molecular nanospring. What are the molecular determinants for these desired mechanical properties and what can we learn from GB1 to help us design proteins with tailored nanomechanical properties? Molecular dynamics simulations of protein GB1 homolog protein L (13) and the third IgG binding domain of protein G (14, 15) showed that the main unfolding event involves the separation of the two parallel terminal β strands, e. g. simultaneous disruption of the backbone hydrogen bonds linking β strands 1 and 4, which is similar to the mechanical unfolding of I27 (38). Considering the intimate interaction of β strands 1 and 4 with the α-helix as well as with β strands 2 and 3, it is also expected that unfolding will also involve the disruption of those interactions, at least partially. Although protein GB1 and protein L share high structural homology and have similar hydrogen bonding patterns in their respective β sheets, GB1 is considerably more stable than protein L under similar pulling velocity (180 pN vs. ~130 pN). It is still challenging to quantitatively account for the interactions that are critical for the mechanical stability of proteins. The exact connection between shear topology and mechanical stability remains to be established, due to the limited number of proteins studied so far by single molecule AFM and the lack of detailed three dimensional structural information for the proteins being studied. As such, for designing protein-based artificial molecular springs with tailored nanomechanical properties, it is critical to build an expanded library of elastomeric proteins with diverse mechanical stability as well as diverse structures. Such a library will enable one to combine protein engineering, molecular dynamics simulation and single  67  molecule force spectroscopy to delineate the molecular interactions that are critical for the mechanical stability, laying the foundation for the ultimate de novo design of elastomeric proteins. Recent de novo design of proteins has made tremendous progress, and proteins with novel sequence and fold have been designed with atomic-level accuracy (39). The de novo design of enzymes with well defined functions is underway (40, 41). We anticipate that de novo design of elastomeric proteins with well-defined topology and tailored interactions should be within the reach of current computational biology methodology. In fact, the computationally designed artificial protein Top7 has been proven to have significant mechanical stability (42). Such efforts will lead to the exciting possibilities in designing proteins with tailored mechanical stability, paving the way for designing proteins with more complex structure and properties.  2.4 Conclusion In conclusion, we demonstrate that artificial polyprotein (GB1)8 exhibits a unique combination of mechanical features that enables it to function as a molecular spring under challenging conditions of continuous stretching/relaxation cycles. These mechanical properties, including high mechanical stability, fast and high fidelity folding kinetics, low mechanical fatigue and the ability to fold against residual force, allow the artificial GB1 polyprotein to recover its mechanical stability more efficiently and help to reduce mechanical fatigue over long periods of continuous stretching-relaxation cycles. These mechanical features are either comparable or superior to those of naturally occurring elastomeric proteins, making GB1 polyprotein an ideal candidate as an artificial elastomeric protein. Since GB1 is not naturally evolved for mechanical function,  68  the superior mechanical properties displayed by the GB1 polyprotein reveal promising prospect for engineering elastomeric proteins using non-mechanical proteins. It is anticipated that the mechanical properties of GB1 can be further finely tuned using protein engineering techniques, an important step towards tailoring the mechanical properties of elastomeric proteins to meet the requirements of different working environments and integrating artificial elastomeric proteins into nanomechanical devices and/or constructing materials (such as hydrogels) of superior mechanical properties.  2.5 Materials and Methods 2.5.1 Protein engineering The plasmid encoding GB1 protein was generously provided by David Baker of University of Washington. GB1 monomer, flanked with a 5’ BamHI restriction site and 3’ BglII, KpnI restriction sites, was amplified by polymerase chain reaction and subcloned into a pQE80L expression vector. The (GB1)8 polyprotein gene was constructed using a previously described iterative method based on the identity of the sticky ends generated by BamHI and BglII restriction enzymes (19). The sequence of the polyprotein (GB1)8 is: MetArgGlySer(His)6-GlySer(GB1-ArgSer)8-CysCys, where the linker sequence ArgSer between GB1 domains comes from the hybrid sites of BamHI and BglII. The polyprotein (GB1)8, carrying two cysteine residues at the C-terminus, was overexpressed in the DH5α strain of E. Coli and purified by Ni2+-affinity chromatography. The purified polyprotein sample was at a final concentration of ~740µg/mL, and was kept at 4 °C in PBS buffer with 5 mM dithiothreitol (DTT) to prevent the dimerization of (GB1)8 via the two C-terminal cysteine residues.  69  2.5.2 Single molecule atomic force microscopy Single-molecule AFM experiments were carried out on a custom-built AFM. In our AFM, we used a high-speed, high performance PicoCube XYZ piezo stage (P-363) from Physik Instrumente (Karlsruhe, Germany). This actuator is equipped with capacitive sensors for all three axes and has a high resonant frequency along the z axis (9.8 kHz). All the force-extension measurements were done in phosphate buffer saline (PBS) at room temperature. In a typical experiment, 1 µL of (GB1)8 protein solution was deposited onto a clean glass cover slip covered by ~50 µL PBS buffer, and was allowed to adsorb for ~5 minutes before pulling experiments proceeded. The spring constant of each individual cantilever (Si3N4 cantilevers from Vecco, with a typical spring constant of 40 pN/nm) was calibrated in solution using the equipartition theorem before and after each experiment. During the AFM experiment, the glass cover slip was brought into contact with the AFM tip by the piezoelectric actuator and then pulled away at a constant velocity. During the contact with the AFM tip, polyprotein molecules can occasionally adsorb onto the tip due to non-specific physical adsorption, allowing for the polyprotein to be stretched between the AFM tip and the solid substrate. The contact force between the AFM tip and the substrate was about 1-2 nN, and the pulling velocity ranged from 5nm/s to 5000nm/s. The measured unfolding forces at high pulling speed were corrected for the hydrodynamic drag force acting directly on the AFM cantilever.  70  2.5.3 Monte Carlo simulations The mechanical unfolding of protein GB1 can be modeled as an all-or-none process with force dependent rate constants, in which only the folded and unfolded states will be populated during the course of the folding/unfolding events. The force dependent unfolding and folding rate constants can be described as: α(F)=α0exp(F∆xu/kBT),  (2.2)  β(F)=β0exp(-F∆xf/kBT),  (2.3)  where, kB is the Boltzmann constant, T is the temperature in Kelvin, α(F) and β(F) are the unfolding and folding rate constants at a stretching force of F, α0 and β0 correspond to the unfolding and folding rate constants at zero force, ∆xu is the distance between the folded state and the transition state, and finally ∆xf is the folding distance between the unfolded state and the transition state. Since there are no analytical expressions to fit the unfolding force distribution, Monte Carlo simulation is widely used to estimate ∆xu and α0 from experimental unfolding force distribution or the relationship between unfolding force and pulling speed (8, 9, 19). The Monte Carlo simulation algorithm is used to compute whether a protein module has unfolded at a given extension or force. The procedure is as follows. We assume that the polyprotein is elongated in discrete time steps of δt (typically set to 10-6 s or even smaller). The length of the protein can be calculated from the number of folded domains, Nf, the number of unfolded domains, Nu and the spacer length Ls. Lc = Ls + N f ⋅ L f + N u ⋅ Lu  (2.4)  Here Lc is the contour length of the molecule, Lf is the folded length of protein and Lu is the unfolded length of protein. 71  The extension of the protein can be calculated as v·t, where v is the pulling speed and t is the time. Therefore, the force can be calculated using the worm like chain model: F=  k BT 1 ( ⋅ p 4 (1 −  1 1 v⋅t − + ) v⋅t 4 L + N ⋅ L + N ⋅ L 2 s f f u u ) Ls + N f ⋅ L f + N u ⋅ Lu  (2.5)  The unfolding probability is calculated using eq. 2.6:  Pu (δt ) = k u (δt ) ⋅ δt = N f ⋅ α 0 exp(  F∆xu ) ⋅ δt k BT  (2.6)  The Pu(δt) is then compared with a random number between 0 and 1. If Pu(δt) is bigger than the random number, we consider one of the folded domains will unfold at time t and the force value at the same t will be the unfolding force. After one domain unfolds, Nf = Nf -1 and Nu = Nu +1. The procedure will be repeated until all the domains in the protein are unfolded (Nf=0). The whole process will be repeated for a minimum of times to generate a statistically significant number of unfolding events.  72  2.6 References 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.  Tatham AS & Shewry PR (2000) Trends Biochem Sci 25, 567-571. Vogel V (2006) Annu Rev Biophys Biomol Struct 35, 459-488. Guerette PA, Ginzinger DG, Weber BH, & Gosline JM (1996) Science 272, 112115. Becker N, Oroudjev E, Mutz S, Cleveland JP, Hansma PK, Hayashi CY, Makarov DE, & Hansma HG (2003) Nat Mater 2, 278-283. van Hest JC & Tirrell DA (2001) Chem Commun (Camb), 1897-1904. Labeit S & Kolmerer B (1995) Science 270, 293-296. Freiburg A, Trombitas K, Hell W, Cazorla O, Fougerousse F, Centner T, Kolmerer B, Witt C, Beckmann JS, Gregorio CC, et al. (2000) Circ Res 86, 1114-1121. Rief M, Gautel M, Oesterhelt F, Fernandez JM, & Gaub HE (1997) Science 276, 1109-1112. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, & Hansma PK (1999) Nature 399, 761-763. Dietz H & Rief M (2004) Proc Natl Acad Sci U S A 101, 16192-16197. Brockwell DJ, Beddard GS, Paci E, West DK, Olmsted PD, Smith DA, & Radford SE (2005) Biophys J 89, 506-519. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Cao Y & Li H (2007) Nat Mater 6, 109-114. Gronenborn AM, Filpula DR, Essig NZ, Achari A, Whitlow M, Wingfield PT, & Clore GM (1991) Science 253, 657-661. Lee G, Abdi K, Jiang Y, Michaely P, Bennett V, & Marszalek PE (2006) Nature 440, 246-249. Marko JF & Siggia ED (1995) Macromolecules 28, 8759-8770. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Evans E & Ritchie K (1997) Biophys J 72, 1541-1555. Park SH, Shastry MC, & Roder H (1999) Nat Struct Biol 6, 943-947. Fernandez JM & Li H (2004) Science 303, 1674-1678. Dougan L & Fernandez JM (2007) J Phys Chem A 111, 12402-12408. Schwaiger I, Schleicher M, Noegel AA, & Rief M (2005) EMBO Rep 6, 46-51. Cecconi C, Shank EA, Bustamante C, & Marqusee S (2005) Science 309, 20572060. Schlierf M & Rief M (2009) Angew Chem Int Ed Engl. 48, 820-822. Schlierf M, Berkemeier F, & Rief M (2007) Biophys J 93, 3989-3998. Ng SP, Rounsevell RW, Steward A, Geierhaas CD, Williams PM, Paci E, & Clarke J (2005) J Mol Biol 350, 776-789. Carrion-Vazquez M, Li H, Lu H, Marszalek PE, Oberhauser AF, & Fernandez JM (2003) Nat Struct Biol 10, 738-743. Oberhauser AF, Badilla-Fernandez C, Carrion-Vazquez M, & Fernandez JM  73  31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.  (2002) J Mol Biol 319, 433-447. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Bullard B, Garcia T, Benes V, Leake MC, Linke WA, & Oberhauser AF (2006) Proc Natl Acad Sci U S A 103, 4451-4456. Kellermayer MS, Smith SB, Granzier HL, & Bustamante C (1997) Science 276, 1112-1116. Kellermayer MS, Smith SB, Bustamante C, & Granzier HL (2001) Biophys J 80, 852-863. Tinoco I, Jr. & Bustamante C (2002) Biophys Chem 101-102, 513-533. Oberhauser AF, Marszalek PE, Carrion-Vazquez M, & Fernandez JM (1999) Nat Struct Biol 6, 1025-1028. Wright CF, Teichmann SA, Clarke J, & Dobson CM (2005) Nature 438, 878-881. Lu H & Schulten K (2000) Biophys J 79, 51-65. Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard BL, & Baker D (2003) Science 302, 1364-1368. Rothlisberger D, Khersonsky O, Wollacott AM, Jiang L, DeChancie J, Betker J, Gallaher JL, Althoff EA, Zanghellini A, Dym O, et al. (2008) Nature 453, 190195. Jiang L, Althoff EA, Clemente FR, Doyle L, Rothlisberger D, Zanghellini A, Gallaher JL, Betker JL, Tanaka F, Barbas CF, 3rd, et al. (2008) Science 319, 13871391. Sharma D, Perisic O, Peng Q, Cao Y, Lam C, Lu H, & Li H (2007) Proc Natl Acad Sci U S A 104, 9278-9283.  74  Chapter 3: Tuning the mechanical stability of proteins by varying solvent conditions† 3.1 Introduction Solvent condition and temperature can affect the thermodynamics of proteins based on different mechanisms (1). However, the mechanical stability of a protein is kinetic stability and as discussed previously, it does not correlate with the thermodynamic stability and kinetic stability obtained from chemical denaturation experiments (2). The influence of a chemical denaturant on the mechanical folding and unfolding is largely unexplored. Here we combine single molecule AFM with chemical denaturation methodology to investigate how the mechanical folding and unfolding pathways are modulated by the presence of the chemical denaturant, guanidinium chloride (GdmCl). The results obtained will provide valuable information about the structure of the mechanical unfolding/folding transition state and the mechanical unfolding/folding mechanism. In single molecule AFM experiments, a stretching force is used as a perturbation to trigger a mechanical unfolding event. Proteins are forced to undergo unfolding reactions along a pre-defined reaction coordinate determined by the stretching force acting on the protein of interest. Mechanical stability of a protein provides invaluable information about the energy landscape underlying the folding/unfolding process. This unique experimental setting has revealed novel information about protein (un)folding dynamics, which would otherwise not be possible to observe. For example, single †  A version of this chapter has been published as: “Cao, Y.; Li, H., How do chemical denaturants affect the mechanical folding and unfolding of proteins? J Mol Biol 2008, 375, (1), 316-24.”.  75  molecule AFM experiments have revealed complex continuous folding trajectories for ubiquitin under a small but constant stretching force (3). Stretching proteins from different directions allows one to map the folding/unfolding energy landscape in an unrivaled level of detail and dimensions (4, 5). A number of theoretical and experimental studies have revealed that a change in temperature and pH will cause a dramatic alternation of the mechanical stability of a protein. Recently, single molecule AFM has been used to investigate how thermal perturbation affects the mechanical unfolding pathway of proteins (6-9). Chemical denaturation is also a widely used method to study protein folding/unfolding kinetics. Surprisingly, it remains unexplored how chemical denaturants will weaken the mechanical stability of proteins and how the chemical denaturant affects the mechanical (un)folding pathway along the direction defined by the stretching force. Here we use a small protein, the B1 immunoglobulin-binding domain of protein G from Streptococcus (referred to as GB1 hereafter), as a model system to quantitatively examine the effects of the chemical denaturants GdmCl on the mechanical (un)folding pathways and kinetics. We found that the mechanical stability of GB1 is weakened by denaturant. The mechanical unfolding rate is sped up by the presence of denaturant while the position of the mechanical unfolding transition states remains unaltered. The mechanical folding of GB1 is also slowed down by GdmCl. Based on our mechanical folding/unfolding results, we developed a “mechanical chevron plot”, analogous to the chevron plot used in chemical denaturation studies, to quantify the effect of GdmCl on the folding and unfolding kinetics and energetics.  76  Figure 3.1 The mechanical stability of GB1 decreases with the increase of denaturant concentration. A) Three dimensional structure of GB1. Grey bars indicate the backbone hydrogen bonds connecting the force-bearing β strands 1 and 4. B) Representative forceextension curves of (GB1)8 polyproteins at 0 and 2.0 M of GdmCl show the characteristic saw-tooth patterns with equally spaced force peaks that correspond the mechanical unfolding of GB1. Fits obtained using the worm-like chain (WLC) (grey lines) are used to measure the contour length increment (ΔL) of around 18 nm for the unfolding of each individual GB1 domains in both conditions. Despite the similar contour length increment, the unfolding force of GB1 decreases from ~180 pN at 0 M GdmCl to ~ 110 pN at 2.0M GdmCl. C) Force-extension curves of (GB1)8 polyproteins under different GdmCl concentrations. The GdmCl concentrations for the curves (from left to right) are 0, 0.6, 0.8, 1.0, 1.25, 1.5, 2.0, 2.25 M, respectively. The average unfolding force of GB1 decreases from ~180 pN at 0 M GdmCl to ~100 pN at 2.25 M.  3.2 Results 3.2.1 GB1 is mechanically weakened by denaturant GB1 is a classical paradigm for protein folding/unfolding studies. It is a small α+β protein with only 56 amino acid residues, composed of a four stranded β sheet packed against a long α helix (Fig. 3.1A) (10). The folding/unfolding kinetics of GB1 77  has been well characterized by both thermal and chemical denaturation methods (11-15). In order to study the mechanical stability of GB1, we constructed a polyprotein of (GB1)8 that is composed of eight identical tandem repeats of GB1. As we showed in Chapter 2, stretching polyprotein (GB1)8 resulted in force-extension curves of the characteristic sawtooth pattern appearance (Fig.3.1B, C), in which the individual force peaks correspond to the sequential mechanical unfolding event of each individual GB1 domains in the polyprotein chain(16, 17). The last peak corresponds to the stretching and the subsequent detachment of the completely unfolded polyprotein chain from either the cantilever tip or the glass substrate. In PBS, GB1 unfolds at an average force of 178 pN at a pulling speed of 400 nm/s.  Figure 3.2 The relationship between the mechanical stability of GB1 and the denaturant concentration. A) The unfolding force histograms of GB1 protein at different denaturant concentration. From bottom to top, the concentrations of GdmCl are 0, 0.6, 0.8, 1.0, 1.25, 1.5, 2.0, 2.25 M, respectively. The unfolding forces of GB1 spans a range of approximately 150 pN for all the GdmCl concentrations, while the average unfolding force of GB1 is deceased with the increase of GdmCl concentration. The solid lines refer to the Gaussian fit to experimental data. The numbers of events in each histogram from top to bottom is 3190, 1785, 767, 1428, 1304, 1032, 1133 and 136, respectively. B) The average unfolding force of GB1 decreases with increasing GdmCl concentration.  78  To study how chemical denaturation alters the mechanical stability of GB1 quantitatively, we carried out mechanical unfolding experiments of GB1 in the presence of different concentrations of GdmCl. As shown in Fig. 3.1C, the mechanical unfolding of (GB1)8 in the presence of GdmCl results in sawtooth-like force-extension curves. Worm-like chain model of polymer elasticity fits (18) to consecutive unfolding events yielded an identical contour length increment as those in the absence of denaturant (17), indicating that chemical denaturants does not affect the contour length increment of GB1. This result suggests that the chemical denaturants do not change the nature of the twostate unfolding of GB1, and the unfolding events of GB1 observed in the sawtooth pattern correspond to the complete unfolding of GB1. Despite the identical contour length increment, the mechanical forces required to unfold GB1 decreased with increasing of GdmCl concentration (Fig. 3.1C). The unfolding force histograms of GB1 under 7 different GdmCl concentrations (0, 0.6 M, 0.8 M, 1 M, 1.25 M, 1.5 M, 2 M and 2.25 M) are shown in Fig. 3.2A. All the unfolding force histograms showed unimodal distributions with a similar width of ~55 pN. We observed a systematic shift of the average unfolding force towards a lower value when the GdmCl concentration is increased: the average unfolding forces of GB1 decreased from 178 pN in PBS to 97 pN at a GdmCl concentration of 2.25 M (Fig. 3.2B). Since the unfolding force is a direct measure of the mechanical stability of a protein, these results indicate that chemical denaturants weaken the mechanical stability of GB1 domains in a concentrationdependent fashion.  79  Figure 3.3 The mechanical unfolding distance is unaltered by denaturant. The unfolding force of GB1 at GdmCl concentration of 1.0 M is plotted versus the pulling speed (black squares). The number of events for each point from left to right is 50, 93, 81, 168, 148, 180, 136, 160, 137, 169, 133, 97 and 204, respectively. The error bars indicate the standard deviation of the unfolding forces. The unfolding kinetics of GB1 in 1.0 M GdmCl can be reproduced adequately by Monte Carlo simulations with an unfolding distance of 0.17 nm, the same as that for GB1 in PBS, and a spontaneous unfolding rate constant α0 of 0.1 s-1. For comparison, the unfolding forces of GB1 in PBS at different pulling speed, as well as its Monte Carlo fit, are also plotted (grey triangle). The two curves are parallel, indicating that the mechanical unfolding distance is the same under both conditions.  3.2.2 The mechanical unfolding distance is unaltered by denaturant The mechanical unfolding of proteins is determined by the underlying energy landscape. The mechanical unfolding energy landscape is characterized by two important parameters: the energy barrier for mechanical unfolding, as well as the distance between the folded state and the transition state (Δxu). The mechanical unfolding rate constant at a given force F follows the following relationship (19, 20):  80  ⎛ ΔGu − F ⋅ Δx u k BT ⎝  α ( F ) = ν ⋅ exp⎜⎜ −  ⎞ ⎛ F ⋅ Δx u ⎟⎟ = α 0 ⋅ exp⎜⎜ ⎠ ⎝ k BT  ⎞ ⎟⎟ ⎠  (3.1)  where ν is the prefactor, ΔGu is the unfolding free energy barrier and α0 is the spontaneous unfolding rate constant along the reaction coordinate defined by the stretching force, kB is the Boltzmann constant and T is temperature. To quantitatively describe the effect of chemical denaturants on the mechanical unfolding, we determine the Δxu and α0 in the presence of chemical denaturants. Previous single molecule AFM experiments on proteins have demonstrated that the analysis of force distributions provides information about the underlying free energy landscape (4, 6, 21). The width of the force histogram is directly related to the distance between the folded state and the transition state (Δxu) along the mechanical unfolding reaction coordinate (20). As shown in Fig. 3.2 A, the width of unfolding force histograms at different GdmCl concentration is around 55pN, which is very close to that for GB1 in PBS, indicating that the distance between the folded state and the transition state is unaltered by chemical denaturants. To further confirm this observation, we also measured the pulling speed dependence of unfolding forces of GB1 in the presence of GdmCl. Fig. 3.3 shows the average unfolding forces of GB1 as a function of pulling speeds (ranging from 47 nm/s to 5000 nm/s) at a GdmCl concentration of 1.0 M. For comparison, the pulling speed dependency of the average unfolding forces of GB1 in PBS is also shown. As expected, the two curves are parallel, which further confirmed that the unfolding distance of GB1 at different GdmCl concentrations remained the same as that when [GdmCl] = 0. Using a standard Monte Carlo simulation procedure and modeling the mechanical unfolding of GB1 as a two-state process, we reproduced the average unfolding forces at different pulling speeds using an unfolding distance of 0.17 nm, the 81  same as that of GB1 in PBS. Note that in the presence of GdmCl, a spontaneous unfolding rate constant of 0.1 s-1 is found, which is 2.5 times faster than that in PBS.  3.2.3 Chemical denaturants speed up the mechanical unfolding event. Since the unfolding distance at different GdmCl concentration remained unchanged as compared with that for GB1 in PBS, we extracted the unfolding rate constant at zero force for GB1 at different GdmCl concentrations using Monte Carlo simulation. We found that the spontaneous unfolding rate constant α0 of GB1 increases with the increase of GdmCl concentration, from 0.039 s-1 in PBS to 0.42 s-1 in 2.25M GdmCl. It is evident that the chemical denaturant facilitates the mechanical unfolding reaction by speeding up the unfolding rate constant, and has little effect on the unfolding distance. This result suggests that chemical denaturants do not alter the mechanical unfolding pathway for GB1.  82  Figure 3.4 The refolding kinetics of GB1 at different GdmCl concentrations. The folding probability of GB1 is plotted versus the relaxation time, Δt. The folding probability of GB1 increases exponentially with the increase of relaxation time at all GdmCl concentrations. The open and filled symbols represent the data from 2 separated experiments for each GdmCl concentration. Fitting of the data to the function, P(t)=1exp(-β0·t), yields the folding rate constant of GB1 at different GdmCl concentration (solid line): 425±29 s-1 at 0.6 M, 36.3±6.9 s-1 at 0.8 M,5.6±1.1 s-1 at 1.0 M and 3.3±0.4 s-1 at 1.25 M. The folding rate constant of GB1 decreases with the increase of GdmCl concentration.  3.2.4 Chemical denaturant slows down the folding reaction of GB1. In ensemble denaturation studies, the folding rate constant of protein is demonstrated to be slowed down upon addition of denaturant. However, in single molecule AFM folding experiment, the folding of protein starts from an extended unfolded state other than a compacted unfolded state as in ensemble studies (3). The effect of chemical denaturants on the folding of tethered proteins may be different (22).  83  Despite its importance, this question has not been probed before. We used a double-pulse protocol to measure the mechanical folding kinetics of GB1 at zero force at different GdmCl concentrations. First, a polyprotein of GB1was extended to count the total number of GB1 domains, Ntotal, in the polyprotein chain being picked up and stretched by the AFM tip. Then, the unfolded polyprotein chain was quickly relaxed to zero extension (usually within 2 ms) before it detached from either the AFM tip or substrate. The dead time of our mechanical folding experiment depends on how fast we can relax the molecule to zero extension, which is analogue to the mixing time in stop-flow experiments. After the polyprotein had been relaxed at zero extension for a variable period of time, t, it was stretched again to count the number of domains folded (Nrefold) during the waiting time t. Nrefold was observed to depend exponentially upon the length of the relaxation time t at all the GdmCl concentrations, despite the fact that the folding was slowed down dramatically with the increase of GdmCl concentration. As shown in Fig. 3.4, the probability that GB1 folds in 15 ms is decreased from ~100% at GdmCl concentration of 0.6 M to ~30% at GdmCl concentration of 1.25 M. Since the unfolding rate constant α0 at zero force is significantly smaller than the folding rate constant β0, we treated the folding reaction of GB1 as a first-order kinetics, Nrefold/Ntotal=1-exp(β0·t), where β0 is the folding rate constant of GB1 at zero force. Fitting this function to our measured folding kinetics data measures the folding rate constants of GB1 at different GdmCl concentrations. The folding rate constant decreases from 425±29 s-1 at a GdmCl concentration of 0.6 M to 3.3±0.4 s-1 at a GdmCl concentration of 1.25 M. It is clear that the chemical denaturants significantly slowed down the folding reaction of GB1 tethered between the AFM tip and the substrate.  84  3.3 Discussion 3.3.1 A mechanical chevron plot quantitatively describes the effects of chemical denaturants on the mechanical unfolding and folding reactions. In order to reveal the effect of denaturant on the mechanical folding-unfolding reactions in a systematic fashion, we plot the measured mechanical folding and unfolding rate constants as a function of the concentration of chemical denaturant GdmCl. As shown in Fig. 3.5, both the logarithm of the spontaneous mechanical unfolding rate constant α0 and that of the mechanical folding rate constant β0 at zero force depend on the denaturant concentration in a linear fashion: logβ0 decreases linearly as GdmCl concentration increases following the relationship of lnβ(denaturant) = lnβ0(PBS)  − mf[GdmCl], while logα0 increases linearly when GdmCl concentration increases following a similar relationship of lnα(denaturant) = lnα0 + mu[GdmCl]. These results indicated that the free energy barriers for the mechanical unfolding and folding, ΔGu and  ΔGf, change linearly with respect to GdmCl concentration. The resulting dependence of mechanical unfolding/folding kinetics on GdmCl concentration has the shape of a right angle bracket. Since this plot is a mechanical analogue to the classical chevron plot in protein folding/unfolding studies (by chemical denaturation), we term this plot as a “mechanical chevron plot”. As in a chemical chevron plot, we define the mechanical m value: m = ΔΔG/[denaturant], where ΔΔG is the free energy difference between native and unfolded states and [denaturant] is the concentration of denaturant. mu and mf are the m values for mechanical unfolding and folding, respectively. Mechanical m values provide a quantitative measure how the free energy barrier of mechanical folding and  85  unfolding reactions are changed by chemical denaturant. The detailed physical meaning of m value will be discussed in section 3.3.3. Fitting the two arms of mechanical chevron plot, we obtain a mu of 0.53±0.06 kcal mol-1 M-1 and a mf of 2.79±0.71 kcal mol-1 M-1.  Figure 3.5 A mechanical chevron plot quantitatively describes the effect of a chemical denaturant on the mechanical unfolding and folding kinetics. The natural logarithm of mechanical folding and unfolding rate constants at different GdmCl concentrations are plotted versus the concentration of GdmCl. For comparison, the folding rate constants of GB1 monomer (in grey) and (GB1)8 polyprotein (open triangle) measured by a stop-flow experiment are also plotted (taken from ref.(14)). The mechanical unfolding rate constant increases with increased GdmCl concentration, while the mechanical folding rate constant decreases with increased GdmCl concentration.  Despite the similarity between chemical and mechanical chevron plots, it is of importance to note the significant difference between the two. First, a mechanical chevron plot probes a region that is not accessible to a chemical chevron plot. At low  86  denaturant condition, a mechanical chevron plot directly measures the mechanical unfolding rate constant, as well as the folding rate constant. In contrast, the information about the unfolding rate constant is not directly visible in a chemical chevron plot at low denaturant concentrations, as the folding reaction dominates under such experimental conditions. Thus, the mechanical chevron plot provides unique information pertaining to the folding and unfolding kinetics at low denaturant concentrations, such as deviation of two-state unfolding behavior at low denaturant concentration. Such information cannot be directly probed in a chemical chevron plot. Second, in a chemical chevron plot, the rate constant plotted is kobs, which is the sum of ku and kf at a given denaturant concentration (kobs= ku + kf). In contrast, in a mechanical chevron plot, α0(GdmCl) and β0(GdmCl) are plotted. This is because the mechanical unfolding experiments are carried out under non-equilibrium conditions and no reverse reaction is possible. Third, single molecule AFM experiments use unfolding force as a probe to measure folding and unfolding kinetics. Hence, a mechanical chevron plot only covers low chemical denaturant concentration range, while a chemical chevron plot covers a full range of GdmCl concentrations.  3.3.2 The mechanical unfolding kinetics are affected by chemical denaturants in a similar way to chemical unfolding. To compare the effects of chemical denaturants on mechanical and chemical unfolding pathways, we also plotted a chemical chevron plot of GB1 measured using stopped-flow spectrofluoremetry (taken from ref.(14)) It is surprising to find that the extrapolation of the unfolding arm of the chemical chevron plot superimpose with the  87  unfolding arm of mechanical chevron plot. This result suggests that the chemical denaturant speeds up mechanical unfolding in a similar way to the chemical unfolding events. The measured spontaneous unfolding rate constant for mechanical and chemical unfolding are indistinguishable from each other, so are mu values measured for mechanical and chemical unfolding of GB1. These results indicate that chemical denaturant softens the mechanical unfolding barrier of GB1 in a same scale as it does on the chemical unfolding barrier, and the mechanical unfolding pathway of GB1 is likely to coincide with or be part of the ensemble of chemical unfolding pathways. Chemical unfolding experiments make use of the change in tryptophan fluorescence, upon disruption of the hydrophobic core of proteins, or in the CD signal upon the change of secondary structure of proteins, which are global in nature. In contrast, the mechanical unfolding uses a directional unfolding pathway defined by the stretching force, thus it is local in nature. Due to the different nature of the two unfolding pathways, generally the two are not necessarily the same. The finding that the mechanical unfolding pathway of GB1 coincides with the chemical one represents a unique case in which mechanical and chemical unfolding experiments share common kinetic features. Similar findings have been reported on wild type Ig domains from titin (21, 23, 24). A common feature of Ig domains of titin and GB1 is their highly native-like transition states for both mechanical and chemical unfolding pathways. It is unknown whether such coincidences for mechanical and chemical unfolding pathways can be extended to other GB1-like domains, such as protein L. Previous single molecule AFM studies revealed that the coincidence is only present for wild type Ig domains (21, 23). In point mutants of Ig domains, the  88  coincidence between the two unfolding pathways was observed to disappear (25, 26). It remains to be seen whether a similar breaking down of such coincidence will also occur for mutants of GB1. Moreover, the softening effect we observed in our experiments indicated that the unfolding energetics in mechanical unfolding experiments are also affected by the denaturing condition. As demonstrated before, mechanical unfolding is anisotropic (4). The mechanical unfolding kinetics of the same protein can show dramatically different behaviors when the protein is pulled and unfolded from different directions. Hence, the softening effect we reported here for stretching GB1 from its N- and C-termini may not be extrapolated to other stretching and unfolding directions. The effect of chemical denaturant on individual mechanical unfolding pathways will have to be examined on an individual basis. Similar to our chemical softening of the mechanical stability of a protein, thermal softening of mechanical stability of proteins was investigated by various groups (6-9). It was observed that increasing temperature reduced the mechanical stability of the proteins under investigation. Moreover, Rief and co-workers also observed that increasing the temperature led to a gradual shift of the mechanical unfolding transition state away from the folded state on ddFFL4, resulting in an increase of unfolding distance (6). They suggested that the increase of unfolding distance of ddFLN4 from a temperature of 5 oC to 37oC reflects a shift of the mechanical resistance of the protein from hydrogen bonding dominated interactions to hydrophobic interactions. In contrast, such effects were not found in our mechanical unfolding experiment of GB1 in GdmCl solution. Instead, we found that the unfolding distance remains constant at all GdmCl concentrations. We  89  propose that the interactions that are crucial to mechanical unfolding of GB1 remain unchanged under denaturing conditions, while their strength is weakened by chemical denaturant. The mechanism of how chemical denaturants, such as GdmCl, unfold proteins remains controversial because of the difficulty in directly probing the interactions between denaturants and proteins experimentally. There are two distinct views on the interaction mechanism. One is the direct binding model (27), which suggests that urea and GdmCl interact with the hydrophilic polar side chains and the backbone by hydrogen bonding (28, 29). Another model, proposed by Tanford et al. (30, 31), indicates that urea and GdmCl denature proteins by altering the strength of the hydrophobic interactions (32). Molecular dynamics simulations showed that the mechanical resistance of GB1 mainly comes from the hydrogen bonding between the two terminal force-bearing strands of GB1(33). Our results indicated that these interactions are weakened in the presence of GdmCl, which may serve as experimental evidence of the direct binding model mentioned above. This also qualitatively explains why the mechanical and chemical unfolding kinetics show similar response to chemical denaturants. Since hydrogen bonding remains the dominant interaction responsible for the mechanical unfolding barrier in the presence of GdmCl, the mechanical unfolding transition state does not change. This would account for why the unfolding distance of GB1 does not change in the presence of GdmCl.  90  3.3.3 The mechanical unfolding transition state of GB1 is more solvent exposed than the native state Since GdmCl interacts with proteins by non-specific direct binding, the binding energy will be proportional to the solvent accessible surface area (SASA) of the protein. ΔG = k denaturant ⋅ SASA ⋅ [denaturant ]  (3.2)  where ΔG is the free energy change upon binding to denaturant, kdenaturant is the prefactor, which is determined by the binding ability of denaturant to proteins and [denaturant] is the concentration of denaturant. Therefore, the change of free energy barrier is related to the change of SASA, ΔΔG = k denaturant ⋅ Δ SASA ⋅ [denaturant ]  (3.3)  The mechanical stability of GB1 is lowered by the presence of GdmCl, while the unfolding distance of GB1 is unaffected. This implies that GdmCl molecules bind to the transition state more strongly than to the native state. Since the transition state is lowered more than the native state, the mechanical unfolding barrier is decreased upon binding. The stronger binding of GdmCl to the mechanical unfolding transition state of GB1 indicates that the transition state is more unstructured that the native state and has larger SASA. However, the denatured state has the largest SASA and is stabilized more than the transition state. Thus, the folding barrier of GB1 is also lowered by the GdmCl, resulting in a lower folding constant, β0(GdmCl). Our single molecule AFM data can be well explained using the direct binding model. Combining the definition of m (m = ΔΔG/[denaturant]) (1) with eq. 3.3, one obtains m = ΔΔG /[denaturant ] = k denaturant ⋅ Δ SASA  (3.4)  91  Such that m is a direct measure of the change of SASA. If m is positive, the SASA increases and vice versa. Therefore, we can use mu from mechanical unfolding and chemical denaturation experiments to compare the SASA for the mechanical unfolding transition state and the chemical unfolding transition state. As shown in Fig. 3.5, the mechanical unfolding transition state of GB1 has similar SASA as the chemical unfolding transition state, as indicated by their similar mu values.  3.3.4 The effect of chemical denaturant on mechanical refolding The mechanical folding rate constant of GB1 is slowed down by GdmCl and shows a stronger logarithmic dependence on GdmCl concentration than that measured in stopped-flow experiments, revealing a difference between the chemical and mechanical folding pathways. In chemical folding studies, the folding reaction is initiated from a denatured state, which is not necessarily random and may contain residual structure (34, 35). In contrast, the folding reaction in mechanical folding studies is initiated from a well-defined fully stretched state (3, 36). The experimental setting in single molecule AFM folding experiments removes the complication of the residual structure that may still be present in the denatured protein. Moreover, it involves a collapsing phase into the kinetics, which corresponds to the collapsing of the fully extended protein chain to the random coil state when the force is relaxed to zero force (3, 36, 37). Therefore, carrying out folding experiments in the presence of chemical denaturant will have the potential to directly dissect the influence of the chemical denaturants on the hydrophobic collapse process, as well as the subsequent folding processes (38).  92  3.4 Conclusion We present the first single molecule AFM study on the effect of chemical denaturants on the mechanical folding and unfolding kinetics of the small protein GB1. Upon increasing the GdmCl concentration, we observed a systematic decreasing in the mechanical stability of GB1, demonstrating the softening effect of the chemical denaturant on the mechanical stability of proteins. This mechanical softening effect originates from the reduced free energy barrier between the folded and the transition state, which decreases linearly as a function of the denaturant concentration. Chemical denaturants, however, do not shift the mechanical unfolding pathway or alter the distance between the folded and transition state. We also found that the folding rate constant of GB1 is slowed down by GdmCl in the mechanical folding experiment. Combining the mechanical folding and unfolding kinetics of GB1 in GdmCl solution, we developed a “mechanical chevron plot” as a general tool to understand how chemical denaturants influence the mechanical folding and unfolding kinetics and free energy diagram in a quantitative fashion. This study demonstrates the great potential of combining chemical (or thermal) folding/unfolding techniques with single molecule AFM to reveal key features of the mechanical unfolding transition state. We anticipate that the combination of single molecule AFM, protein engineering, chemical denaturation and molecular dynamics simulations will provide invaluable information regarding the structure of the mechanical unfolding transition state and help to map the energy landscape in much greater detail. This in turn will contribute to a better understanding of the whole folding energy landscape of a protein.  93  3.5 Materials and methods 3.5.1 Protein engineering and expression The (GB1)8 polyprotein, containing 8 tandem repeats of GB1 domains, was engineered, expressed and purified as described in Chapter 2. Stopped-flow spectrofluorimetry experiments on (GB1)8 polyprotein indicated that the GB1 domains behave independently of each other (as shown in Fig. 3.5), in agreement with other polyprotein studies (39, 40).  3.5.2 Force spectroscopy of single proteins All the single-molecule force measurements were performed with a custom-built atomic force microscope as described in Chapter 2. The cantilevers were calibrated in PBS solution using the equipartition theorem with an average error of 10%. In order to minimize the errors from calibration, for all the experiments, around 50 curves of unfolding of (GB1)8 were obtained in PBS before switching the buffer to GdmCl solution. The average forces of these unfolding events served as an internal caliper. The unfolding forces of GB1 in GdmCl solution were corrected based on the difference of unfolding forces of GB1 in PBS of individual experiments and the average unfolding forces of all the experiments. In a typical unfolding experiment, 1μL of GB1 solution with a concentration of 741ng/μL was dropped onto a freshly cleaned glass coverslip containing ~50μL of PBS or GdmCl solution and stabilized for ~ 10 minutes before measurement. The pulling speed was 400 nm/s for all the unfolding experiments, except when reported otherwise. For the refolding experiments, only the cantilevers with minimum drift were used. During  94  the refolding experiments, we checked the starting position from time to time and adjusted the piezo position to balance the drift accordingly to ensure that the polyprotein was relaxed to zero extension.  3.5.3 Monte Carlo simulations The mechanical unfolding of GB1 domains was modeled as an all-or-none process with force-dependent rate constants, in which only the folded and unfolded state will be populated during the reaction (19, 20). The force-dependent unfolding rate constant can be described as: α(F,GdmCl)=α0(GdmCl)*exp(FΔxu/kBT), where, kB is the Boltzmann constant, T is the absolute temperature in Kelvin, α(F,GdmCl) is the unfolding rate constant at a stretching force of F in GdmCl solution, α0(GdmCl) is the unfolding rate constant at zero force in GdmCl solution, and Δxu is the distance between the folded state and the transition state along the direction of the force. Monte Carlo simulations were carried out according to published procedures (21, 41-43). The unfolding rate constants at different GdmCl concentrations were obtained by simulations with a fixed unfolding distance of 0.17nm.  95  3.6 References 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.  Fersht A (1999) Structure and mechanism in protein science : a guide to enzyme catalysis and protein folding (W.H. Freeman, New York). Li HB (2008) Advanced Functional Materials 18, 2643-2657. Fernandez JM & Li H (2004) Science 303, 1674-1678. Dietz H, Berkemeier F, Bertz M, & Rief M (2006) Proc Natl Acad Sci U S A 103, 12724-12728. Schlierf M & Rief M (2009) Angew Chem Int Ed Engl 48, 820-822. Schlierf M & Rief M (2005) J Mol Biol 354, 497-503. Law R, Liao G, Harper S, Yang G, Speicher DW, & Discher DE (2003) Biophys J 85, 3286-3293. Janovjak H, Kessler M, Oesterhelt D, Gaub H, & Muller DJ (2003) Embo J 22, 5220-5229. Taniguchi Y, Brockwell DJ, & Kawakami M (2008) Biophys J 95, 5296-5305. Sjobring U, Bjorck L, & Kastern W (1991) J Biol Chem 266, 399-405. Alexander P, Orban J, & Bryan P (1992) Biochemistry 31, 7243-7248. Alexander P, Fahnestock S, Lee T, Orban J, & Bryan P (1992) Biochemistry 31, 3597-3603. Park SH, Shastry MC, & Roder H (1999) Nat Struct Biol 6, 943-947. McCallister EL, Alm E, & Baker D (2000) Nat Struct Biol 7, 669-673. Krantz BA, Mayne L, Rumbley J, Englander SW, & Sosnick TR (2002) J Mol Biol 324, 359-371. Cao Y & Li H (2007) Nat Mater 6, 109-114. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Marko JF & Siggia ED (1995) Macromolecules 28, 8759-8770. Bell GI (1978) Science 200, 618-627. Evans E & Ritchie K (1997) Biophys J 72, 1541-1555. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Friedel M, Baumketner A, & Shea JE (2006) Proc Natl Acad Sci U S A 103, 8396-8401. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Best RB, Fowler SB, Herrera JL, Steward A, Paci E, & Clarke J (2003) J Mol Biol 330, 867-877. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Williams PM, Fowler SB, Best RB, Toca-Herrera JL, Scott KA, Steward A, & Clarke J (2003) Nature 422, 446-449. Robinson DR & Jencks WP (1965) Journal of the American Chemical Society 87, 2462-&. Mande SC & Sobhia ME (2000) Protein Eng 13, 133-141. O'Brien EP, Dima RI, Brooks B, & Thirumalai D (2007) J Am Chem Soc 129, 7346-7353. Nozaki Y & Tanford C (1963) J Biol Chem 238, 4074-4081.  96  31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.  Tanford C (1970) Adv Protein Chem 24, 1-95. Zou Q, Habermann-Rottinghaus SM, & Murphy KP (1998) Proteins 31, 107-115. Li PC, Huang L, & Makarov DE (2006) J Phys Chem B 110, 14469-14474. Alexandrescu AT, Abeygunawardana C, & Shortle D (1994) Biochemistry 33, 1063-1072. Shortle D & Ackerman MS (2001) Science 293, 487-489. Walther KA, Grater F, Dougan L, Badilla CL, Berne BJ, & Fernandez JM (2007) Proc Natl Acad Sci U S A 104, 7916-7921. Li MS, Hu CK, Klimov DK, & Thirumalai D (2006) Proc Natl Acad Sci U S A 103, 93-98. Dill KA (1990) Biochemistry 29, 7133-7155. Scott KA, Steward A, Fowler SB, & Clarke J (2002) J Mol Biol 315, 819-829. Rounsevell RW, Steward A, & Clarke J (2005) Biophys J 88, 2022-2029. Rief M, Fernandez JM, & Gaub HE (1998) Phys Rev Lett 81, 4764-4767. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. Rief M, Gautel M, Oesterhelt F, Fernandez JM, & Gaub HE (1997) Science 276, 1109-1112.  97  Chapter 4: Enhancing the mechanical stability of proteins by engineered metal chelation†  4.1 Introduction Elastomeric proteins function as molecular springs under biological settings and exhibit mechanical properties that underlie the elasticity of natural adhesives (1), cell adhesion proteins (2), and muscle proteins (3-5). They are also potential building blocks for the bottom-up construction of functional nanomechanical devices and biomaterials of superb mechanical properties (6, 7). To utilize elastomeric proteins as building blocks for various applications, it is essential to tailor the mechanical properties of elastomeric proteins at a molecular level. The development of single molecule atomic force microscopy (AFM) as well as computer modeling has made it possible to examine the mechanical properties of proteins at the single molecule level in vitro and in silico (8-12), allowing the elucidation of the relationships between the structure-mechanical properties of elastomeric proteins. Despite the tremendous progress in protein mechanics, generally it is not possible to enhance the mechanical stability of proteins in a rational fashion, with the exception of only a few isolated cases (13-17). Such a lack of knowledge has hindered the understanding of molecular design of naturally occurring elastomeric proteins and prevented rational design of novel protein-based materials. Mechanical stability is an intrinsic property of proteins and is commonly defined as the force required to mechanically unfold a given protein. Mechanical stability is determined by the mechanical unfolding energy barrier and the distance between the †  A version of this chapter has been published as: “Cao, Y.; Yoo, T.; Li, H., Single molecule force spectroscopy reveals engineered metal chelation is a general approach to enhance mechanical stability of proteins. Proc Natl Acad Sci U S A 2008, 105, (32), 11152-7.”.  98  native state and transition state (18, 19). It is thus different from thermodynamic stability of proteins (20). In contrast to the challenge in rationally enhancing the mechanical stability of proteins, many successful strategies to improve the thermodynamic stability of proteins (21-28) have been unveiled by extensive experimental and computational enzyme engineering work. Engineered metal chelation is one such general approach for protein stabilization (22, 29-31). In this method, a bi-histidine (bi-His) motif, which involves two histidines positioned to bind a bivalent metal ion (such as Ni2+), can be easily engineered on the surface of a protein. Due to preferential binding of the divalent metal ion to the native state over the denatured state, the protein can thus be stabilized. However, due to the differences between thermodynamic and mechanical stability, strategies to enhance the thermodynamic stability of proteins cannot be generally used to enhance mechanical stability. Here, we carry out a thermodynamic analysis to elucidate some general concepts on how to rationally improve the mechanical stability of proteins. As a proof of principle, we validate these concepts at the single molecule level in a small protein GB1 by using engineered metal chelation. Our results demonstrate that engineered metal chelation is a general and effective approach to rationally enhance the mechanical stability of proteins in a fully reversible fashion. To the best of our knowledge, this is the first general method in protein mechanics to rationally tune the mechanical stability of proteins. We anticipate that this approach will find a wide range of applications in engineering diverse elastomeric proteins.  99  4.2 Results 4.2.1 Rationale for enhancing the mechanical stability of proteins Thermodynamic stability is the free energy difference between the unfolded and folded states (ΔGU-N). In contrast, mechanical stability of proteins is determined by the mechanical unfolding energy barrier (ΔG‡-N) and the distance between the native state and the mechanical unfolding transition state Δxu (18, 19). To understand the general mechanism required to enhance the mechanical stability of proteins, we use the binding of metal ions as an example to carry out a simple thermodynamic cycle analysis for the mechanical unfolding reaction (Fig. 4.1A). Assuming that the unfolding distance does not change, the enhancement of the unfolding free energy barrier by metal ion binding (ΔΔG‡-N) is equal to the difference in binding energy of the metal ion to the native state and to the mechanical unfolding transition state (ΔGbind(‡)-ΔGbind(N)). Therefore, if the metal ions bind preferentially to the native state than to the transition state, the native state will be preferentially stabilized by the binding of metal ions and the unfolding energy barrier is thus increased. In this way, the enhancement of the mechanical stability of the protein can be achieved. If the metal ion stabilizes the transition state by the same amount as it does for the native state, the unfolding energy barrier will not change and there is no enhancement of mechanical stability, even though the thermodynamic stability of the protein will be enhanced. From this analysis, it becomes evident that enhancing the mechanical stability of a protein will be more demanding than enhancing the thermodynamic stability, as the former involves not only the native state, but also unfolding transition state, which is difficult to study. Therefore, to enhance the mechanical stability of proteins, one will need to preferentially stabilize the native state over the mechanical unfolding transition state. For metal chelation, we will need to 100  engineer a metal chelation site that will be somewhat disrupted in the mechanical unfolding transition state in order to achieve enhanced mechanical stability. As a proof of principle, here we use the well-characterized small protein GB1 as a model system to demonstrate the feasibility of realizing these general ideas. GB1 is a small α+β protein that is composed of a β sheet packed against an α helix (32) (Fig. 4.1B). Its mechanical unfolding has been well-characterized by single molecule AFM as described in chapter 2 (33, 34). Molecular dynamics simulation predicted that the main mechanical unfolding event corresponds to the rupture of the backbone hydrogen bonds between the force bearing β strands 1 and 4 (35, 36). In the mechanical unfolding transition state, the force-bearing strands 1 and 4 slide slightly against each other. Hence, if we engineer a metal chelation site across the two forcebearing β strands 1 and 4, the slight sliding of β strand 1 against 4 may distort the metal chelation site in the mechanical unfolding transition state and result in the preferential binding of metal ions to the native state over the transition state, thus heightening the unfolding energy barrier for GB1. Based on such reasoning, we have engineered bihistidine metal chelation sites into GB1, which are across the force-bearing β strands 1 and 4 and positioned to bind a bivalent metal ion.  101  Figure 4.1 Rationale for enhancing the mechanical stability of proteins. A) Thermodynamic cycle analysis describes enhancing the mechanical stability of a protein by preferential binding of a metal ion to the native state. The asterisks denote the protein in the metal ion bound state (both native state and unfolding transition state). ΔΔG‡-N is defined as the free energy difference of the mechanical unfolding barrier caused by metal chelation and equals to ΔG*‡-N-ΔG‡-N. ΔΔGbind is defined as ΔGbind(‡)-ΔGbind (N), where ΔGbind is the Gibbs free energy for the binding reaction. The thermodynamic cycle analysis shows that ΔΔG‡-N=ΔΔGbind. Hence, the difference in the mechanical unfolding free energy barrier upon chelation of metal ions is equal to the binding free energy difference between the mechanical unfolding transition state and the native state. B) Mechanical topology of GB1. The two force-bearing β strands of GB1 (colored in green) are the key region for the mechanical stability of GB1, as the rupture of the backbone hydrogen bonds (indicated by lines) connecting these two β strands are predicted to contribute most to the mechanical unfolding barrier of GB1. During the mechanical unfolding of GB1, there is slight sliding movement between the two force bearing β strands.  102  103  Figure 4.2 The thermodynamic stability of bi-His mutants of GB1 in the absence and presence of 14.3mM of NiCl2, measured by chemical denaturation. The equilibrium thermal stability was determined from GdmCl denaturation using tryptophan emission fluorescence (350nm) of GB1 mutants, which is excited at 280 nm. Protein concentrations were in the range of 1.5 to 33 μM in TrisHCl buffer (100mM, pH 7.4, containing 100mM NaCl) at 22±2oC. From A-E, the chemical denaturation curves of bi-His mutants of GB1 (G6-53, G4-55, G8-55, G32-36 and G4-6) in the absence of Ni2+ (filled cycles) and in the presence of 14.3mM of Ni2+ (open squares). The ΔGU-N of each mutants and ΔΔGU-N between metal chelated and metal free forms are summarized in Table 4.1.  Table 4.1 The free energy changes upon metal chelation of GB1 bi-His mutants.  Protein  ΔΔGU-N  in 14.3mM Ni2+  in H2O ΔGU-N  [GdmCl]0.5  ΔG*U-N  (ΔG*U-N -ΔGU-N)  (M)  (kcal/mol)  (M)  (kcal/mol)  (kcal/mol)  wt GB1  2.51  6.03  -  -  -  G6-53  1.34  3.26  2.01  4.85  1.59  G4-51  1.21  2.42  1.93  3.89  1.47  G8-55  2.21  3.96  2.41  4.32  0.36  G4-6  2.01  3.73  2.28  4.25  0.52  G32-36  2.08  3.74  2.63  4.76  1.02  [GdmCl]0.5  4.2.2 Binding of metal ions significantly enhances the mechanical stability of GB1 bi-His mutants. Using site-directed mutagenesis, we mutated residues 6 and 53 of GB1 to histidine to obtain bi-His mutant G6-53 (“G” represents GB1 and 6 and 53 indicate the positions of histidines in the GB1 mutants.) (Fig. 4.3A). Equilibrium chemical denaturation studies showed that the presence of Ni2+ increases the thermodynamic stability of G6-53, confirming the metal chelation capability of G6-53 (Fig. 4.2 and table 4.1). We then constructed polyprotein (G6-53)8, which is composed of eight identical tandem repeats of G6-53, and used single molecule AFM to examine its mechanical stability. Stretching polyprotein (G6-53)8 in the absence of metal ions results in forceextension curves of characteristic saw-tooth pattern appearance, where each individual sawtooth peak corresponds to the mechanical unfolding of the individual G6-53 domains in the polyprotein (Fig. 4.3B, black curve). The unfolding force peaks are equally spaced with a contour length increment (ΔLc) of ~18.0 nm, as measured by fitting the WormLike Chain (WLC) model of polymer elasticity(37) to consecutive unfolding force peaks. The average unfolding force of G6-53 in the absence of metal ions is 119±29 pN 104  (average±standard deviation, n=1927) at a pulling speed of 400 nm/s (Fig. 4.3C, black histogram), which is lower than the unfolding force for wt GB1 (184±41 pN)(34), indicating that the introduction of bi-His site into the force-bearing region of GB1 destabilizes GB1 mechanically.  Figure 4.3 The mechanical stability of GB1 bi-His mutants is enhanced by the binding of Ni2+. A, D and G) Engineered bi-His metal chelation sites in GB1. The binding of metal ions to the bi-His site will introduce a cross-strand bridge over the two force-bearing strands1 and 4. B, E and H) Typical force-extension curves of GB1 bi-His mutants G653, G4-51 and G8-55 in the absence of metal ions (black curves) and in the presence of 4mM Ni2+ (red curves). C, F and I) Unfolding force histograms of GB1 bi-His mutants G6-53, G4-51 and G8-55 in the absence (in black) and presence (in red) of 4mM Ni2+, respectively. It is evident that, upon binding of Ni2+, the mechanical unfolding force histograms of bi-His mutants shift towards higher forces, indicating that the mechanical stability of bi-His mutants are enhanced significantly by the binding of Ni2+. The average unfolding forces are 119 pN (n=1927), 120 pN (n=1345) and 160 pN (n=1637) for G653, G4-51 and G8-55 in the absence of Ni2+, respectively. In contrast, the average unfolding forces are 243 pN (n=2226), 198 pN (n=1609) and 219 pN (n=1098) for G653, G4-51 and G8-55 in the presence of 4mM Ni2+, respectively. 105  Stretching (G6-53)8 in the presence of 4mM Ni2+ results in sawtooth-like forceextension curves as the one shown in Fig. 4.3B (red curve). The unfolding force peaks of G6-53 in the presence of 4mM Ni2+ are equally spaced with ΔLc of ~18 nm, identical to that for G6-53 in the absence of Ni2+, suggesting that the unfolding force peaks indeed correspond to the mechanical unfolding of G6-53. However, the unfolding of G6-53 in the presence of 4mM Ni2+ occurs at much elevated forces. The average unfolding force of G6-53 in 4mM Ni2+ is 243±49 pN (n=2226) (Fig. 4.3C, red histogram), which is more than double of that for G6-53 in the absence of Ni2+. And this unfolding force (~243 pN) is also significantly higher than that for wt GB1 (~180 pN). These results clearly indicate that the binding of Ni2+ to the engineered bi-his metal chelation site in GB1 significantly enhances the mechanical stability of G6-53, just as we predicted.  4.2.3 The enhancement of mechanical stability by metal ion binding is fully reversible The enhancement of mechanical stability by the binding of metal ions is fully reversible. Upon addition of the Ni2+ competitive binding agent imidazole to the solution, the mechanical stability of bi-His mutant G6-53 can be fully reverted. Fig. 4.4 shows an example of such experiments. The binding of 4mM Ni2+ to G6-53 results in the increase of mechanical unfolding forces of G6-53 from 110 pN to ~240 pN, as evidenced by the shift of the unfolding force histogram shown in Fig. 4.4. Upon addition of 300 mM imidazole, imidazole will compete with the histidine residues in GB1 to bind Ni2+. This competitive binding will result in the dissociation of Ni2+ ions from G6-53. Accordingly, the unfolding forces of G6-53 were observed to drop back to ~110 pN. This process is  106  fully reversible and provides the possibility to tune the mechanical stability of G6-53 using environmental stimuli in a fully reversible fashion.  Figure 4.4 The mechanical stability of G6-53 can be regulated reversibly by the binding of Ni2+ ions as well as its competitive binding reagent imidazole. The unfolding force histogram of G6-53 is centered at ~ 110 pN in Tris buffer (10mM, pH7.4) (top panel, n=476). After adding 4mM Ni2+, the unfolding force histogram shifts towards higher force with an average unfolding force of ~ 250 pN (middle panel, n=567). Upon addition of 300mM imidazole to the solution, the unfolding force histogram shifts back to lower unfolding forces (bottom panel, n=429), which is indistinguishable from the unfolding force histogram in the absence of Ni2+. All these three unfolding force histograms were obtained from the same experiment.  107  4.2.4 The enhancement in mechanical stability by binding of metal ion is context dependent. Having demonstrated that the binding of a metal ion to an engineered metal chelation site can significantly enhance the mechanical stability of GB1, we investigate the influence of the location of the metal chelation site on the mechanical stabilization effect. For this purpose, we engineered bi-His mutants G4-51 (Fig. 4.3D) and G8-55 (Fig. 4.3G) and their corresponding polyproteins (G4-51)8 and (G8-55)8. The metal chelation properties of these bi-His mutants were also confirmed by the observed increase in their thermodynamic stability upon binding of Ni2+ in chemical denaturation studies (Fig. 4.2). It is noted that the thermodynamic stabilization effect of bi-His mutants upon metal chelation depends on the location of metal chelation site, as evidenced by the difference in the increase in thermodynamic stability among the different bi-His mutants (Fig. 4.2 and table 4.1). Using single molecule AFM, we measured the unfolding forces of G4-51 and G8-55 in the absence and presence of Ni2+. As shown in Fig. 4.3E and 4.3F, the introduction of a bi-His site in G4-51 results in a mechanical destabilization effect on GB1 and the mechanical unfolding of G4-51 occurs at ~110 pN in the absence of metal ions. In contrast, the mechanical destabilization effect by the introduction of a bi-His site in G8-55 is much milder than G4-51 and G6-53, and the average unfolding force of G855 is about 160 pN in the absence of metal ions, only ~20 pN lower than that for wt GB1 (Fig. 4.3H and I). Despite the mechanical destabilization effect, the binding of Ni2+ significantly enhances the mechanical stability of both G4-51 and G8-55 (Fig. 4.3E and H), and the mechanical unfolding of G4-51 and G8-55 in the presence of 4mM Ni2+ occurs at ~200 pN and ~210 pN (Fig. 4.3F and I), respectively.  108  It becomes evident that an engineered bi-His metal chelation site in the forcebearing region of GB1 offers an efficient approach to enhance the mechanical stability of GB1 by the binding of divalent metal ions. However, the magnitude of the enhancement of the mechanical stability is not the same but depends on the location where the metal chelation site is engineered. Our data suggest that the metal chelation site in the very center of the force-bearing region (site 6-53) has the strongest stabilization effect, while the sites at the periphery of the force-bearing region have a weaker effect. It is interesting to note that a similar trend was also observed on the relative increase in thermodynamic stability for the three bi-His mutants (table 4.2). Table 4.2 Summary of the unfolding force, unfolding distance (Δxu) and spontaneous unfolding rate constant α0 at zero force for bi-His mutants G4-51, G6-53 and G8-55 in the absence (-) and presence (+) of 4 mM Ni2+. G4-51 G6-53 G8-55 wt GB1 -Ni2+  +Ni2+  -Ni2+  +Ni2+  -Ni2+  +Ni2+  120±29  198±43  119±29  243±49  160±38  219±57  184±41  Δxu (nm)  0.20  0.17  0.20  0.17  0.20  0.17  0.17  α0 (s-1)  0.12  0.023  0.14  0.0071  0.029  0.014  0.039  Unfolding force (pN, ±S.D.)  ΔΔG‡-N (kCal/mol) (ΔΔRTlnα0)  0.99  1.79  0.46  -  4.2.5 The metal chelation sites engineered outside the force-bearing region do not affect the mechanical stability of GB1. As rationalized in the thermodynamic cycle analysis in section 4.2.1, preferential binding of metal ions to the metal chelation site in the native state over the transition state is the key to enhancing mechanical stability. If the metal chelation provides similar stabilization to the mechanical unfolding transition state as it does to the native state, the unfolding free energy barrier will not change and no net enhancement of mechanical 109  stability will be achieved. To further validate this rationale, we engineered two control biHis mutants G32-36 and G4-6, in which a metal chelation site was engineered in the αhelix and the first β strand of GB1(38), respectively. The metal chelation properties for G32-36 and G4-6 were confirmed by the increase of their thermodynamic stability in the presence of Ni2+ (Fig. 4.2).Since the α-helix is well within the core of GB1, it will not experience a mechanical stretching force until GB1 has completely unfolded. Therefore, the stretching force will not affect the binding affinity of the metal chelation to metal ions in the mechanical unfolding transition state of G32-36. Fig. 4.5A shows the unfolding force histograms for G32-36 in the absence and presence of 4mM Ni2+. As predicted, the binding of Ni2+ to G32-36 does not have any effect on the mechanical unfolding forces of G32-36, despite its clear effect in enhancing the thermodynamic stability of G32-36. Similarly, the binding of metal ions to bi-His mutant G4-6 does not have any effect on the mechanical stability. These results clearly demonstrate that distorting the metal chelation site in the mechanical unfolding transition state is the key to realizing the preferential binding of metal ions to the native state and the enhancement of mechanical stability.  110  Figure 4.5 Metal chelation site outside the force-bearing region does not enhance the mechanical stability of bi-His mutants G32-36 (A) and G4-6 (B), in which a metal chelation site was engineered into the α helix and the first β strand, respectively. The unfolding force histograms for the Ni2+-bound (in black) and unbound forms (in gray) of bi-His mutants are indistinguishable for both G32-36 and G4-6, respectively. It is of note that, in comparison with wt GB1, the mechanical stability of bi-His mutants G32-36 and G4-6 changes due to the double histidine mutations. The unfolding force is 149±37 pN (n=832) and 142±37 pN (n=341) for G32-36 and G4-6, respectively.  4.2.6 Enhancing the mechanical stability by increasing the free energy barrier To confirm the mechanism for enhancing mechanical stability of GB1, we need to quantify the change in the unfolding free energy barrier upon binding of Ni2+. Toward this goal, we carried out single molecule AFM stretching experiments on (G4-51)8, (G653)8 and (G8-55)8 at different pulling speeds. As shown in Fig. 4.6, the unfolding forces of bi-His mutants depend upon pulling speeds: the faster the pulling speed, the higher the required unfolding force. Using standard Monte Carlo simulation procedures (39, 40), we 111  can reproduce the force-extension curves of these three polyproteins. By fitting the unfolding force distributions and their dependence on pulling speeds (Fig. 4.6) simultaneously, we can then estimate the mechanical unfolding rate constant at zero force, α0, and the distance between the native state and transition state, Δxu. We found that a Δxu of 0.17 nm can describe the mechanical unfolding well for all Ni2+-bound biHis GB1 mutants, while Δxu of 0.20 nm is a good value for the Ni2+-free form of all biHis mutants. This result is consistent with the observation that the unfolding force distributions for Ni2+ bound bi-His mutants are somewhat broader than those for the Ni2+free bi-His mutants (Fig. 4.3C, F and I). The measured α0 and unfolding distance Δxu for bi-His mutants are tabulated in Table 4.2 together with those for wt GB1. The increase in the mechanical unfolding free energy barrier (ΔΔG‡-N) by metal chelation is equal to RTln(α0(Ni2+-bound)/α0(Ni2+-free)), where R is the gas constant and T is the temperature. Hence, it can be calculated that ΔΔG‡-N ranges from 0.4 kcal/mol to 1.8 kcal/mol for biHis mutants, indicating that the binding of Ni2+ stabilizes the native state more than the mechanical unfolding transition state, the very principle underlying using metal chelation to enhance mechanical stability. It is of note that, the binding of Ni2+ to the metal chelation site not only increases the mechanical unfolding energy barrier, but also reduces the unfolding distance Δxu. Both factors contribute to the enhancement of the mechanical stability. It seems that the introduction of a bi-His site into GB1 leads to the increase of Δxu from 0.17 nm for wt GB1 to 0.20 nm for bi-His mutants, but the binding of Ni2+ to the bi-His site brings Δxu back to 0.17 nm. The underlying detailed molecular mechanism remains unclear.  112  Figure 4.6 The pulling speed dependence of the mechanical unfolding forces of bi-His mutants in the absence and presence of 4mM Ni2+. The unfolding force of G6-53, G4-51 and G8-55 were measured at different pulling speeds in the absence (black) and in the presence (red) of 4mM of Ni2+. The solid lines correspond to Monte Carlo simulation fits to the experimental data using the parameters shown in Table 4.2. It is evident that the chelation of Ni2+ does not significantly change the slope of the speed-dependence of unfolding forces of bi-His mutants. In contrast, the spontaneous unfolding rate constant α0 deceased by ~2- 20 fold upon binding of Ni2+, indicating that the enhancement of mechanical stability by the binding of Ni2+ is largely the result of an increase of the mechanical unfolding free energy barrier. 113  4.3 Discussion Tuning the mechanical stability of elastomeric proteins, especially enhancing the mechanical stability, has been challenging in protein mechanics (11). Differing from thermodynamic stability (that is the free energy difference between the unfolded state and native state), mechanical stability is the “kinetic stability”, in that it is directly related to the free energy difference between the native state and the mechanical unfolding transition state (20). Because a mechanical unfolding transition state is difficult to characterize, it has not been possible to develop general and rational approaches toward enhancing the mechanical stability of proteins. Although there are many successful approaches in enzyme engineering to enhance thermodynamic stability of proteins, these approaches cannot be directly applied to enhancing the mechanical stability of proteins. Through a simple thermodynamic cycle analysis for the mechanical unfolding reaction, we discover that the key to enhance the mechanical stability of proteins is the preferential stabilization of the native state over the mechanical unfolding transition state. Hence, by engineering metal chelation bi-histidine sites across two force-bearing β strands, we successfully enhance the mechanical stability of GB1 in a fully reversible fashion. The net mechanical stabilization achieved by metal chelation in bi-His mutants of GB1 is substantial and ranges from 60 pN to 120 pN. Such an enhancement effect of mechanical stability is likely due to the distortion/disruption of the metal chelation site in the mechanical unfolding transition state by the stretching force, leading to the preferential stabilization of the native state over the transition state by the binding of divalent metal ions. The extreme case will be that the bonds between Ni2+ and bi-His, resulting from the metal chelation, are fully ruptured in the mechanical unfolding transition state. In this case, the thermodynamic stabilization of the native state by metal 114  binding can be fully converted into increasing the mechanical unfolding energy barrier, which will lead to a maximum enhancement of the mechanical stability of proteins. Therefore, the larger the thermodynamic stabilization effect is upon metal chelation, the larger the mechanical stabilization can be potentially achieved. However, it is important to note that a larger thermodynamic stabilization effect only provides the possibility for achieving a larger mechanical stabilization effect. The actual magnitude of mechanical stabilization depends on the degree of preferential stabilization of the native state over the transition state. In addition, it seems that the metal chelation in the bi-His site across the forcebearing strands also helps to consolidate the force-bearing region and limit the shearsliding movement of the two force-bearing strands. This effect is exemplified by the decrease in the mechanical unfolding distance Δxu of the bi-His mutants upon binding of metal ions. Previous single molecule AFM studies have revealed that ligand binding (14) and protein-protein interactions (15, 41) can enhance mechanical stability of some proteins. These methods are restricted to particular proteins that have unique ligand binding properties, and thus cannot be easily generalized to other protein systems. In contrast, biHis based metal chelation sites can be easily engineered into a wide range of proteins with little or no disruption of the native state and has been widely used in traditional enzyme engineering (38), as well as in protein folding studies (the so-called Ψ-value analysis) (42, 43). Therefore, the method of engineered metal chelation is not an approach unique to particular proteins. Instead, it can be used in a wide range of proteins and hence represents the first general approach in protein mechanics to rationally enhance the mechanical stability of elastomeric proteins. Moreover, different divalent metal ions, 115  such as Cu2+, Zn2+, Ni2+ and Co2+, exhibit different binding affinities to bi-His sites (44) and thus may provide additional control over the enhancement of mechanical stability. Our preliminary data has shown that different metal ions indeed can lead to different enhancements of the mechanical stability of GB1. The general ideas illustrated herein are not limited to metal chelation, but constitute rather general principles to enhance the mechanical stability of proteins. A wide variety of methodologies have been developed to improve thermodynamic stability of proteins (21-28), and these methods are also potential routes one can utilize to enhance mechanical stability of proteins, as long as one can find effective ways to selectively stabilize the native state over the transition state. The insights elucidated here will not only open up new avenues towards regulating the mechanical properties of elastomeric proteins in their biological settings, but also make it feasible to tailor the mechanical properties of elastomeric proteins for applications in bioengineering and material sciences. Furthermore, the method of engineered metal chelation demonstrated here also has important implications in elucidating the nature of the unfolding transition state along its mechanical unfolding pathways. Metal chelation has been used to probe the structure of the chemical folding/unfolding transition state in the so-called Ψ-value analysis(42, 43). Similarly, by engineering metal chelation sites in different regions of the protein of interest, it is now feasible to use single molecule AFM to probe their effects on the mechanical stability and deduce important information about the role of mutated sites in the formation of the mechanical unfolding transition state. These studies will not only make it possible to map the mechanical unfolding transition state of proteins, just as the Ψ-value analysis does in the classical protein folding/unfolding dynamics, but also 116  provide unprecedented opportunity to directly compare the mechanical and chemical unfolding pathways and understand the differences between them.  4.4 Materials and methods 4.4.1 Protein engineering Plasmids that encode wild type GB1 were generously provided by Prof. David Baker of the University of Washington. We mutated the solvent exposed residues on the surface of GB1 at desired positions to histidines to form the metal chelation sites, based on the crystal structure of GB1 (PDB code: 3GB1 or 1PGA). All the bi-His mutants were constructed using the mega primer method with a sense primer comprising one His mutation and an anti-sense primer comprising the other His mutation. The gene sequences of all bi-His mutants were confirmed by direct DNA sequencing. All the polyprotein genes were constructed as described in previous chapters and Appendix A (33, 34). The polyproteins were expressed in DH5α strain of E. Coli, purified by Co2+ affinity chromatography, and eluted in PBS buffer with 300 mM NaCl and 150 mM imidazole. 20 mM EDTA was added to the elution fractions to remove residual Co2+ that may exist in the elution fraction. The proteins were further dialyzed against Tris-HCl buffer (10mM, pH 7.4, containing 100 mM NaCl) to completely remove EDTA and imidazole. All bi-His mutants were folded to a similar structure at wild-type GB1, as measured by circular dichroism.  4.4.2 Single molecule AFM experiment Single molecule AFM experiments were carried out on a custom-built AFM as described in previous chapters (33, 34). All the force-extension measurements were 117  carried out either in Tris-HCl buffer (10mM, pH 7.4, containing 100 mM NaCl) or TrisHCl plus 4mM NiCl2. The spring constant of the AFM cantilevers (Si3N4 cantilevers from Veeco) was calibrated using the equipatition theorem before each experiment with a typical value of 60 pN nm-1. For Ni2+ binding studies, we first deposited the polyprotein and Ni2+ solution onto a glass coverslip containing 50uL of Tris-HCl buffer and mixed them in situ. The AFM experiments were carried out after allowing the mixture to equilibrate for ~30 mins. The pulling speed was 400 nms-1 for all experiments unless otherwise indicated.  118  4.5 References 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.  Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, & Hansma PK (1999) Nature 399, 761-763. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. Labeit S & Kolmerer B (1995) Science 270, 293-296. Rief M, Gautel M, Oesterhelt F, Fernandez JM, & Gaub HE (1997) Science 276, 1109-1112. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Becker N, Oroudjev E, Mutz S, Cleveland JP, Hansma PK, Hayashi CY, Makarov DE, & Hansma HG (2003) Nat Mater 2, 278-283. Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, & Savage K (2002) Philos Trans R Soc Lond B Biol Sci 357, 121-132. Carrion-Vazquez M, Oberhauser AF, Fisher TE, Marszalek PE, Li H, & Fernandez JM (2000) Prog Biophys Mol Biol 74, 63-91. Brockwell DJ (2007) Current Nanoscience 3, 3-15. Carrion-Vazquez M, Oberhauser A, Diez H, Hervas R, Oroz J, Fernandez J, & Martinez-Martin A (2006) in Advanced Techniques in Biophysics, eds. Arrondo J & Alonso A (Springer), pp. 163-236. Li H (2007) Organic & Biomolecular Chemistry 5, 3399-3406. Sotomayor M & Schulten K (2007) Science 316, 1144-1148. Sharma D, Perisic O, Peng Q, Cao Y, Lam C, Lu H, & Li H (2007) Proc Natl Acad Sci U S A 104, 9278-9283. Ainavarapu SR, Li L, Badilla CL, & Fernandez JM (2005) Biophys J 89, 33373344. Cao Y, Balamurali MM, Sharma D, & Li H (2007) Proc Natl Acad Sci U S A 104, 15677-15681. Dietz H, Berkemeier F, Bertz M, & Rief M (2006) Proc Natl Acad Sci U S A 103, 12724-12728. Ng SP, Billings KS, Ohashi T, Allen MD, Best RB, Randles LG, Erickson HP, & Clarke J (2007) Proc Natl Acad Sci U S A 104, 9633-9637. Evans E & Ritchie K (1997) Biophys J 72, 1541-1555. Evans E & Ritchie K (1999) Biophys J 76, 2439-2447. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Sanchez-Ruiz JM & Makhatadze GI (2001) Trends Biotechnol 19, 132-135. Kellis JT, Jr., Todd RJ, & Arnold FH (1991) Biotechnology (N Y) 9, 994-995. Malakauskas SM & Mayo SL (1998) Nat Struct Biol 5, 470-475. Nauli S, Kuhlman B, & Baker D (2001) Nat Struct Biol 8, 602-605. van den Burg B & Eijsink VG (2002) Curr Opin Biotechnol 13, 333-337. Sanchez-Ruiz JM (2007) Biophys Chem 126, 43-49. Perl D & Schmid FX (2002) Chembiochem 3, 39-44. Perl D, Mueller U, Heinemann U, & Schmid FX (2000) Nat Struct Biol 7, 380383. Schneider JP & Kelly JW (1995) Chemical Reviews 95, 2169-2187. Regan L (1995) Trends Biochem Sci 20, 280-285. 119  31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.  Lu Y, Berry SM, & Pfister TD (2001) Chemical Reviews 101, 3047-3080. Gronenborn AM, Filpula DR, Essig NZ, Achari A, Whitlow M, Wingfield PT, & Clore GM (1991) Science 253, 657-661. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Cao Y & Li H (2007) Nat Mater 6, 109-114. Li PC & Makarov DE (2004) Journal of Physical Chemistry B 108, 745-749. Li PC, Huang L, & Makarov DE (2006) J Phys Chem B 110, 14469-14474. Marko JF & Siggia ED (1995) Macromolecules 28, 8759-8770. Arnold FH & Haymore BL (1991) Science 252, 1796-1797. Rief M, Fernandez JM, & Gaub HE (1998) Phys Rev Lett 81, 4764-4767. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Cao Y, Yoo T, Zhuang S, & Li H (2008) J Mol Biol 378, 1132-1141. Krantz BA & Sosnick TR (2001) Nat Struct Biol 8, 1042-1047. Sosnick TR, Krantz BA, Dothager RS, & Baxa M (2006) Chemical Reviews 106, 1862-1876. Beyer MK & Clausen-Schaumann H (2005) Chemical Reviews 105, 2921-2948.  120  Chapter 5: Enhancing the mechanical stability of proteins by protein-protein interactions †  5.1 Introduction Elastomeric proteins function as molecular springs in a variety of biological machines and tissues as well as in bio-materials of superb mechanical properties to establish elastic connections and provide mechanical strength, elasticity and extensibility (1-9). The mechanical stability of constituting domains in elastomeric proteins plays an important role in determining the overall mechanical properties of elastomeric proteins. Tuning the mechanical properties of proteins is not only important for regulating various biological processes and activities, but is also critical to engineer artificial elastomeric proteins with defined mechanical properties for constructing smart materials and nanomechanical devices, such as force sensors and mechanical switches (10). As discussed in previous chapters, mechanical stability is an intrinsic property of proteins and is defined by the three dimensional structures of proteins (11-14). Single molecule force spectroscopy studies in combination with protein engineering techniques have revealed that site-directed mutagenesis (15-18) and recombination of protein fragments (19) are efficient approaches to tune the mechanical stability of a given protein. However, these are chemical methods resulting in mutant proteins that are permanently modified from wild type proteins. It remains challenging to efficiently regulate the mechanical  †  A version of this chapter has been published as: “Cao, Y.; Balamurali, M. M.; Sharma, D.; Li, H., A functional single-molecule binding assay via force spectroscopy. Proc Natl Acad Sci U S A 2007, 104, (40), 15677-81.” and “Cao, Y.; Yoo, T.; Zhuang, S.; Li, H., Protein-protein interaction regulates proteins' mechanical stability. J Mol Biol 2008, 378, (5), 1132-41.”.  121  stability of a protein using external factors in a fully reversible fashion. Recently, the effect of ligand binding on the mechanical stability of dihydrofolate reductase has been investigated using single molecule AFM (20-22). It was shown that ligand binding can mechanically stabilize dihydrofolate reductase from Chinese hamster (20), opening up a new perspective in protein mechanics. Compared with protein-ligand interactions, protein-protein interactions often have higher affinity and larger interaction surface (23). Considering the ubiquitous nature of protein-protein interactions in biology, we have started to explore the use of protein-protein interactions to tune the mechanical stability of proteins. Here, we use the binding of the Fc and Fab fragments of IgG to GB1 and its mutants as model systems to illustrate that protein-protein interactions represent an efficient and general appoach to regulate the mechanical stability of proteins. GB1, the B1 IgG binding domain of protein G from group G of Streptococcus, is a small α+β protein (Fig. 5.1) (24). The mechanical properties of GB1 have been extensively studied in chapter 2. Our single molecule atomic force microscopy (AFM) studies showed that an artificial polyprotein made of GB1 is an ideal elastomeric protein with mechanical properties comparable and even superior to those of naturally occurring elastomeric proteins (25, 26). It is well known that GB1 can bind Fc and Fab fragments of IgG with high affinity, providing the possibility to tune the mechanical stability of GB1 using IgG fragments (27-29). Fc interacts with GB1 in the region of the C-terminal part of the α-helix, the N-terminal part of the third β-strand and the loop between them (29). Contrary to Fc, Fab binds to GB1 on the second β-strand and the C-terminal end of the α-helix(28) (Fig. 5.1). Using single molecule AFM techniques, we will demonstrate that the binding of Fc and Fab fragments can significantly enhance the mechanical  122  stability of GB1. Thus protein-protein interactions can serve as an efficient and general approach to regulate the mechanical stability of GB1-like proteins. We anticipate that this new methodology will help to develop novel elastomeric proteins with tunable mechanical stability and compliance. 5.2 Results 5.2.1 Binding of IgG fragments can significantly enhance the mechanical stability of GB1. As we demonstrated before, polyprotein (GB1)8 has significant mechanical stability (25, 26). Stretching polyprotein (GB1)8 resulted in the sequential unraveling of the individual GB1 domains in the polyprotein, giving rise to the force-extension curves of characteristic saw-tooth pattern appearance (Fig.5.2A). The individual force peaks correspond to the mechanical unfolding of the individual GB1 domains in the polyprotein chain. The last peak corresponds to the stretching and the subsequent detachment of the completely unfolded polyprotein chain from either the cantilever tip or the glass substrate. In PBS, GB1 unfolds at an average force of 182±40 pN (n=2593) at a pulling speed of 400 nm/s.  123  Figure 5.1 hFc and hFab bind GB1 in the different regions of GB1. A) Sequence alignment of the GB1, C2 (GC2) and B3 (GB3) IgG binding domains of protein G. These three domains share high sequence identity. B) The three dimensional structures of GB1, GC2/hFc and GB3/hFab. Due to the high sequence identity between GB1, GC2 and GB3, the three dimensional structures of these three domains are very similar to each other. Thus, the three dimensional structures of GC2/hFc and GB3/hFab can be used as model structures for GB1/hFc and GB1/hFab.  124  Figure 5.2 The mechanical stability of protein GB1 is significantly enhanced by the binding of human IgG fragments hFc and hFab. A-C) Force-extension curves of GB1 polyprotein, GB1/hFc complex and GB1/hFab complex. The top panels show the schematics of stretching polyproteins of GB1, GB1 with hFc and GB1 with hFab between an AFM tip and glass substrate, respectively. Stretching polyprotein (GB1)8 results in force-extension curves of typical saw-tooth pattern that are characterized by unfolding forces of ~180 pN, as indicated by the line, and contour length increments ΔLc of ~18 nm (left, black). Each individual force peak corresponds to the mechanical unfolding of individual GB1 domains in the polyprotein. The mechanical stability of GB1 is enhanced by the binding of hFc (center, red) and hFab (right, blue). When pre-equilibrated with ~11-25 μM of hFc, the majority of GB1 domains unfold at much higher forces of ~270 pN, as indicated by the dashed line, which is ~90 pN higher than that for GB1 in the absence of hFc. The mechanical stability of GB1 can also be enhanced by the binding of hFab, despite the fact that hFab binds GB1 in a different region. Stretching (GB1)8 in the presence of 50−70μM of hFab results in the unfolding forces of ~270 pN, as indicated by the line (right panel, blue). D-E) Unfolding force histograms of GB1 (black), GB1/hFc complex (red) and GB1/Fab complex (blue). The binding of hFc and hFab to GB1 significantly enhances the mechanical stability of GB1.  125  In order to test whether protein-protein interactions can be used to effectively modulate the mechanical stability of proteins, we carried out single molecule AFM experiments in the presence of human Fc and Fab fragments (hFc and hFab) of IgG to investigate the effect of binding of IgG fragments on the mechanical stability of GB1. Fig. 5. 2B shows a typical force-extension curve of GB1 in the presence of 22 μM hFc, the concentration required for the complete binding of hFc to GB1 based on the binding affinity measured by surface plasmon resonance (SPR) spectroscopy (30). Stretching GB1 complexed with the hFc fragment results in force-extension curves of sawtooth-like appearance. The contour length increment (ΔLc) of GB1 in complex with hFc, as measured using the worm-like-chain (WLC) model of polymer elasticity (31) to fit consecutive unfolding force peaks, is ~18 nm, the same as that for GB1 in the absence of hFc. However, the unfolding of GB1/hFc complex occurs at much higher forces of ~270 pN, as compared with ~180 pN for the unfolding of GB1 alone. Fig. 5.2D shows the comparison of the unfolding force histogram for GB1 alone and in complex with hFc. It is evident that the binding of the hFc to GB1 significantly enhanced the mechanical stability of GB1. The average unfolding force of GB1 in complex with hFc is 263±62 pN (n=711), a ~80 pN increase in the mechanical stability for GB1. To further confirm that the high unfolding force events are indeed the result of the unfolding of GB1 in complex with hFc, we carried out single molecular AFM experiments using different concentrations of hFc. If the high unfolding force is indeed due to the unfolding of GB1 in complex with hFc, we should observe a bimodal distribution of the unfolding forces if the concentration of hFc is below the saturation concentration. Indeed,  126  at intermediate concentrations of hFc, we observed two populations of unfolding events of GB1, as shown in Fig. 5.3B-E: one has an average force of ~180 pN, which is due to the unfolding of GB1 free of hFc; while the other one has an average force of ~270 pN, which corresponds to the unfolding of GB1 in complex with hFc. It is worth noting that the population of high force events increases with increased hFc concentration, following the trend that more GB1 domains are bound with hFc at higher concentration of hFc. These results confirm that the binding of hFc indeed enhances the mechanical stability of GB1. It is also important to note that, from the bimodal distribution of the unfolding forces, we can directly determine the relative population of GB1 in the hFc-bound state and hFc-free state. As we demonstrated previously (32), this information can allow us to directly determine the dissociation constant of hFc/GB1 at the single molecule level using force-spectroscopy data.  127  Figure 5.3 The mechanical stability of GB1/hFc complex is independent on the concentration of hFc. A) Unfolding force histogram of GB1 in the absence of hFc. The average unfolding force is ~180 pN. Solid lines are Gaussian fits to the experimental data. B-E) Unfolding force histograms of GB1 show two clear populations in the presence of hFc: one is at 180 pN, which corresponds the unfolding of hFc-free GB1; and the other is at 270 pN, which corresponds the unfolding of GB1 in the complex with hFc. The initial concentration of hFc for each histogram is shown on the right. Each unfolding force histogram was fitted with two Gaussian functions (solid lines). It is clear that the increase of hFc concentration only shifts the population of bound and unbound form of GB1, and does not change the mechanical stability of GB1 in complex with hFc. Fab fragment of IgG binds GB1 on the second β-strand and the C-terminal end of the α-helix. This Fab binding site on GB1 is different from the binding site for the Fc fragment. Despite the difference in the binding region on GB1, Fab also exhibits a significant stabilization effect on the mechanical stability of GB1. Fig. 5.2C shows  128  representative force-extension curves of stretching GB1 in complex with hFab (50 μM). It is clear that the unfolding of GB1 in complex with hFab results in a similar ΔLc of 18 nm, but the unfolding occurs at much elevated forces. The average unfolding force for GB1 in complex with hFab measured from the unfolding force histogram is 262±76 pN (n=981) (Fig. 5.2E). The distribution of the unfolding forces for GB1/Fab is somewhat broader than that for GB1 alone and GB1/hFc. It is, however, important to note that the binding affinity of Fab to GB1 is much lower than that for hFc to GB1 (30). Hence, it is possible that the unfolding force histogram of GB1/Fab contains some unfolding events from free GB1, which broadens the unfolding force histogram and lowers the average unfolding force. These results demonstrate that protein-protein interactions can significantly stabilize the mechanical stability of protein GB1. The Fab and Fc fragments test here bind GB1 at different regions, but both have a stabilization effect on the mechanical stability of GB1. Although the stabilization effect is similar for the binding of the Fab and Fc fragments, we can envision that it is possible to engineer different GB1 mutants, in which one binding site is modified while the other remains intact, in order to tune the mechanical stability of proteins via different routes.  129  Figure 5.4 Enhancement of the mechanical stability by the binding of hFc is not an artifact due to the polymerized form of GB1. A) Force-extension curves of heteropolyprotein (GB1-TNfn3)4 in the absence (in black) and the presence (in red) of 11 μM hFc. The unfolding events of the third fribronectin type 3 domains of human tenascin C (TNfn3) are characterized by ΔLc of ~28 nm and occur at ~120 pN. The unfolding events of GB1 are characterized by ΔLc of ~18 nm. Since Tnfn3 does not bind to hFc, its unfolding force is unaffected in the presence of hFc. B) Unfolding force histogram for GB1 domains in the heteropolyprotein (GB1-TNfn3)4. The average unfolding force of GB1 in the absence of hFc is 173±39 pN (n=233), and the average force of GB1 increases to 263±57 pN (n=206) upon binding of hFc. Since the single molecule AFM experiments were carried out on polyproteins of (GB1)8, it is critical to ensure that the binding of hFc or hFab to GB1 is not affected by the polymerized form of GB1 and the mechanical stabilization effect we observed upon binding of GB1 to hFc or Fab is not an artifact due to the proximity of GB1 domains in 130  the polyprotein. To prove that the binding of hFc to GB1 is not affected by polymerization of GB1 domains into a polyprotein, we carried out surface plasmon resonance (SPR) measurements. We found that the binding constant measured for hFc to GB1 in homopolyprotein (GB1)8 is almost the same as that for hFc to monomeric GB1, suggesting that the binding properties of GB1 to hFc do not change in the polymerized form of GB1. To eliminate the possibility that the mechanical stabilization effect is not an artifact due to the proximity of GB1 domains in the polyprotein, we have constructed a control heteropolyprotein (GB1-TNfn3)4, in which GB1 is spaced by the third fibronectin type 3 domains of human tenascin C (TNfn3) domains. TNfn3 can not bind to hFc and unfolds at a force around 120 pN (33). When the TNfn3 domains are folded, the GB1 domains are spaced from each other by ~3.1 nm. Upon the unfolding of the TNfn3 domains, the GB1 domains will be separated by the unfolded TNfn3 domains by up to 30 nm. This construct ensures that the binding of Fc or Fab fragments to GB1 domains are not affected by neighboring GB1 domains, and thus GB1 domains in this heteropolyprotein can serve as a model system for the free GB1 domain. Fig. 5.4A shows typical force-extension curves for the heteropolyprotein (GB1-TNfn3)4 in the absence (top curve) and presence of 11 μM of hFc. The unfolding events of TNfn3 domains, which are characterized by ΔLc of ~28 nm, occur at ~120 pN and occur before the unfolding of the GB1 domains. We can then determine the unfolding forces of GB1 domains without any ambiguity. Fig. 5.4B shows the unfolding force histograms of GB1 in the absence (in black) and presence (in red) of hFc. It is evident that the binding of hFc to GB1 increases the mechanical stability of GB1 from ~180 pN to ~270 pN, which is in close agreement with results on polyprotein (GB1)8. These results indicate that the  131  enhancement of mechanical stability by the binding of hFc is not an artifact due the proximity of GB1 domains in the polyprotein, but rather that it is an intrinsic property of GB1.  5.2.2 The binding of the IgG fragments does not shift the mechanical unfolding transition state for GB1. To investigate how the binding of the IgG fragment affects the mechanical unfolding of GB1, we carried out force-extension measurements of GB1 in complex with hFc at different pulling speeds. As shown in Fig. 5.4, the unfolding force at which GB1/hFc in complex with hFc unfolds depends upon the pulling speed. The higher the pulling speed is, the higher the force is required to unfold GB1 in complex with hFc. The speed dependence of the unfolding force of GB1/hFc has a similar slope as that for GB1 alone. Since the slope of the speed dependence of unfolding forces is directly related to the distance between the folded state and the mechanical unfolding transition state (34), the similarity in slopes strongly suggest that the binding of the IgG fragment to GB1 does not change the location of the mechanical unfolding transition state. Therefore, the enhancement of the mechanical stability of GB1 is accomplished through an increase of the free energy barrier to unfolding, rather than the shifting of the location of the transition state. Using standard Monte Carlo simulation procedures, we estimated that the unfolding rate constant α0 at zero force for GB1 in complex with hFc is 2.9×10-4 s-1, while α0 for GB1 alone is 0.039 s-1. The change of α0 for GB1 upon binding of hFc corresponds to a 2.9 kcal mol-1 increase of the mechanical unfolding energy barrier (ΔΔRTlnα0).  132  Figure 5.5 The unfolding force increment of GB1 upon binding of hFc is independent of the pulling speed. The black squares show the average unfolding force of GB1 in the presence of 20 μM of hFc at pulling speeds ranging from 200 nm/s to 2000 nm/s. The grey squares show the speed dependence of the unfolding forces for wt GB1 in the absence of hFc. The slope of the speed-dependence of unfolding forces for the GB1/hFc complex is the same as that of GB1 in the absence of hFc. The solid lines correspond to the Monte Carlo simulation with the same unfolding distance of 0.17 nm and unfolding rate constant in the absence of force of 2.9×10-4 s-1 and 0.039 s-1, respectively. The error bars indicate the standard error of the measurements. The data for GB1 in the absence of hFc were taken from Fig. 2.2 and the error bars are not shown.  5.2.3 The amplitude of the mechanical stability enhancement does not correlate with the binding affinity. Fc interacts with GB1 in the region of the C-terminal part of the α-helix, the Nterminal part of the third β-strand and the loop between them (29). The six residues that are in direct contact with hFc are highlighted in Fig. 5.6A. These residues are directly located in the interface between GB1 and hFc, but are distant from the force-bearing β strand region (Fig. 5.6A), which is important to the mechanical stability of GB1 as shown  133  in Chapter 4 (35). In order to investigate whether the magnitude of the mechanical stability enhancement is correlated with the binding affinity, we focused on four out of these six residues and constructed four GB1 alanine mutants T25A, K28A, K31A and N35A, which have different binding affinity to hFc as determined by fluorescence titration (36). We did not construct mutants involving residues E27 and W43, as they were shown to be critical for the binding of hFc to GB1 and point mutations at these two positions almost completely abolish the binding of GB1 to hFc (36). The association constants of T25A, K28A, K31A and N35A to hFc are 1.5, 8.3, 354 and 54 times lower than that for wild type GB1 to hFc (36). Fig. 5.5B shows the representative forceextension curves of the polyproteins of the alanine mutants alone (left panel) as well as in the presence of 20-50 μM of hFc. It is clear that the alanine mutations, which are outside the key region (involving β strands 1 and 4) for the mechanical stability of GB1, do not alter the mechanical stability of GB1 significantly. In the presence of 20-50 μM of hFc, the force-extension curves of the alanine mutants exhibit two levels of unfolding forces: one at ~180 pN which is the same as the mutants alone, and the other at a much elevated level of forces (~300 pN for T25A, and ~260 pN for K28A, K31A and N35A), which corresponds to the mechanical stability of the alanine mutants in complex with hFc. The unfolding force histograms for these mutants in the presence and absence of hFc are shown in Fig. 5.6C. For T25A and K28A, we can clearly observe the stabilization effect upon the binding of hFc. However, for K31A and N35A, the chances of observing K31A (or N35A) in complex with hFc are significantly reduced due to the significantly reduced binding affinity. The unfolding force histograms for these two mutants are still dominated by the unfolding resulting from the hFc-free form of GB1. However, from the limited  134  trajectories that show the unfolding of hFc bound form of K31A and N35A, we did observe the enhancement in their mechanical stability, as exemplified by the two forceextension curves in Fig. 5.6B. These results clearly indicate that, although the binding affinity of GB1 and GB1 mutants to hFc varies by three orders of magnitude, the magnitude for the unfolding force enhancement falls in the narrow range of 70-120 pN, suggesting that there is no direct correlation between the binding affinity and the amplitude of the mechanical stability enhancement. Hence, upon binding, the GB1/hFc complex represents a unique state that is of well-defined mechanical stability. The binding affinity only affects the population of the GB1/hFc complex at a given concentration of hFc, but does not affect the intrinsic mechanical stability of the GB1/hFc complex. The finding of the near constant enhancement effect also raises interesting questions on what structural parameters determine the magnitude of the enhancement effect and how we can modulate this effect.  135  Figure 5.6 The mechanical stability of GB1/hFc complex does not correlate with the binding affinity. A) Mechanical topology of GB1 and the locations of point mutations. The two force-bearing β strands of GB1 (colored in green) are the key region for the mechanical stability of GB1, as the rupture of the backbone hydrogen bonds (indicated by lines) connecting these two β strands are predicted to be the mechanical unfolding barrier for GB1. Residues that are directly in contact with hFc are colored in yellow. It is evident that alanine mutations T25A, K28A, K31A and N35A are distant from the force-bearing region of GB1. B) Force-extension curves of polyproteins (GT25A)8, (GK28A)8, (GK31A)8, and (GN35A)8 in the presence (red) and absence (black) of hFc. The average  136  unfolding forces for GB1 alanine mutants remain ~180 pN, indicating that these alanine mutations do not change the mechanical stability of GB1. In the presence of hFc (20 μM for GT25A and GK28A and 50 uM for GK31A and GN35A), the force-extension curves clearly show two different levels of unfolding forces (indicated by solid lines): one is at ~180 pN, which corresponds to the unfolding of GB1 mutants that are free of hFc; the other one is at ~260 pN to ~300 pN, which corresponds to the unfolding of the GB1 mutants that are in complex with hFc. C) Unfolding force histograms of the alanine mutants of GB1 in the presence (red) and absence (black) of hFc. The unfolding force histograms for GT25A and GK28A show clear bimodal distributions, with one peak at ~180 pN and the other at ~300 pN for GT25A and ~260 pN for GK28A. Due to the low affinity of GK31A and GN35A to hFc, the unfolding events corresponding to the unfolding of hFc bound mutants are rather limited, and their unfolding force histograms do not show clear bimodal distribution. However, the enhanced mechanical stability of mutant GB1 in complex with hFc is evident in individual force-extension curves.  5.2.4 The enhancement of the mechanical stability of GB1 by IgG fragment is through long range coupling. Upon stretching, the stretching force is directly acting on the two force-bearing β strands 1 and 4 of GB1 (Fig. 5.6A). Molecular dynamics simulations on protein G, as well as its structural homologue protein L, revealed that the major unfolding event corresponds to the rupture of the hydrogen bonds connecting β strands 1 and 4 (35, 37). The hydrogen bonds within the two β hairpins remain largely intact at the transition state for mechanical unfolding. hFc binds GB1 in the region of the C-terminal part of the αhelix, the N-terminal part of the third β strand and the loop between the two structural elements. Hence, hFc does not directly interact with the key region (β strands 1 and 4) that is critical for the mechanical stability of GB1 (35). Although there is no direct interaction between hFc and the key region, the mechanical stability of GB1 is enhanced by the binding of hFc. This observation suggests that the modulation of the mechanical stability of GB1 by interacting with the hFc fragment is through long range coupling.  137  This effect is similar to the allosteric regulation of enzymatic activities found in a wide variety of enzymes (38, 39) and hence can serve as a novel mechanism for regulating the mechanical stability of proteins. In this sense, the Fc fragment functions as an “allosteric activator” for the mechanical stability of GB1. Understanding the molecular details of the allosteric modulation of the mechanical stability of proteins will require more detailed mechanistic studies, possibly involving detailed molecular dynamics simulations on the mechanical unfolding processes of GB1 in complex with hFc.  Figure 5.7 The long range coupling of mechanical stability is retained in GB1 mutants with large structural perturbation. A). Representative force-extension curves of NuG2 (upper panel) and Gc3b4 (bottom panel) in PBS. The contour length increments, ΔLc, of both mutants are ~18 nm, which is similar to wild-type GB1. However, the unfolding force of NuG2 and Gc3b4 are 105 pN and 210 pN respectively, which are distinct from wild-type GB1. B) Force-extension curves of NuG2 and Gc3b4 in the presence of 22 μM hFc. It is evident that upon binding of hFc, the mechanical stability of NuG2 and Gc3b4 is greatly enhanced. The average unfolding force is ~210 pN for the NuG2/hFc complex and ~300 pN for the Gc3b4/hFc complex. C) Unfolding force histograms of NuG2 (upper) and Gc3b4 (bottom) in the absence (in black) and presence (in red) of hFc. The increase of unfolding force for NuG2 upon binding to hFc is ~105 pN and is ~90 pN for Gc3b4.  138  5.2.5 The long range coupling regulation of the mechanical stability of GB1 is robust. Since the enhancement of mechanical stability by protein-protein interactions is through long range coupling instead of direct interactions, it will be interesting to examine whether such a coupling effect is robust if relatively large structural changes are introduced into GB1. Towards this goal, we used two GB1 mutants, NuG2 and Gc3b4. NuG2 is a fast folding mutant of GB1, which consists of a de novo designed version of GB1 in which 11 residues in the first β hairpin were mutated (40). Gc3b4 is a computationally designed hyperthermophilic variant of GB1, which contains 7 mutants in the hydrophobic core to optimize the core packing as well as other interactions of GB1 (41). Both NuG2 and Gc3b4 contain big perturbations to their primary sequence and have different kinetic and thermodynamic properties as compared to wild type GB1. The binding of NuG2 and Gc3b4 to hFc were confirmed by SPR experiments (41). In order to test the mechanical stabilization effect upon binding, we measured the mechanical stability of NuG2 and Gc3b4. As shown in Fig. 5.7A, stretching polyproteins made of (NuG2)8 and (Gc3b4)8 results in force-extension curves of saw-tooth pattern with average unfolding forces of 105 pN and 210 pN, respectively. These values clearly demonstrate the effect caused by the large perturbation of the primary sequence of GB1 on mechanical stability. However, the enhancement of mechanical stability of these two mutants by the binding of hFc is retained. Upon stretching in the presence of 22 μM hFc, the mechanical stability of both NuG2 and Gc3b4 increased significantly (Fig. 5. 6B). The unfolding force of NuG2 in complex with hFc is around 210 pN, whereas and the unfolding force of  139  Gc3b4 in complex with hFc is around 300 pN. The unfolding force histograms (Fig. 5.7C) clearly support the latter data. Evidently, the long range coupling between the hFc binding site and the mechanical “active” site is adequately maintained in NuG2 and Gc3b4, despite the large perturbation. 5.3 Discussion Protein-protein interactions are ubiquitous in biology and have been used extensively by nature to enhance the thermodynamic stability of proteins and proteincomplexes. However, since thermodynamic stability generally does not correlate with mechanical stability (42), it was unknown whether protein-protein interactions can also provide stabilization to the mechanical stability of proteins (43). Our studies provide the first example that protein-protein interactions can significantly enhance the mechanical stability of proteins. Considering the ubiquitous nature of protein-protein interactions, our results open up exciting possibilities towards using protein-protein interaction as an effective means to modulate the mechanical stability of proteins in a controlled and reversible fashion. We anticipate that such regulation mechanism for mechanical stability will be found in many other systems. An immediate possible candidate will be GB1’s structurally homologous proteins. For example, protein L is a structural homologue of GB1 and its mechanical stability has been investigated in detail (37). Protein L also binds to IgG fragments with high affinity. We anticipate that the methodology demonstrated here can be directly applied to protein L and an enhancement of the mechanical stability of protein L by the binding of IgG fragments can be readily achieved. It will be of interest to compare the magnitude of the mechanical enhancement between such structurally homologous proteins.  140  Furthermore, our results reveal that the binding of IgG fragments modulates the mechanical stability of GB1 via a long range coupling mechanism. Although not directly interacting with the key region that is important for mechanical stability of GB1, IgG fragments provide effective modulation of the mechanical stability of GB1. We also showed that the long range coupling between the force bearing region and the hFc binding region is robust to the overall structural alternation of GB1. These results demonstrate that a long range coupling mechanism is a novel and versatile mechanism for modulation of the mechanical stability of proteins. It is likely that such mechanisms are used in a range of biological processes, such as cell migration and adhesion, to finely regulate the mechanical properties and performance of the elastomeric proteins involved in these processes. Moreover, this mechanism also provides unprecedented possibilities to regulate the mechanical stability of proteins in a fully reversible fashion and help to develop smart artificial proteins with well-defined mechanical properties for a wide range of applications in material science and nanomechanics.  5.4 Materials and Methods 5.4.1 Protein Engineering Plasmids that encode wild type GB1 and NuG2 were generously provided by David Baker of the University of Washington. Plasmid that contains the Gc3b4 gene was a generous gift from Steven Mayo of Caltech. All the point mutants were constructed using standard PCR based site-directed mutagenesis techniques and confirmed by direct DNA sequencing. All the polyprotein genes were constructed using well-established procedures. The polyproteins were expressed in the DH5α strain of E. Coli, purified by  141  Co2+ affinity chromatography, and stored at 4oC in PBS buffer with 300 mM NaCl and 150 mM imidazole. hFc (Cat#16-16-090707-FC ) and hFab (Cat#16-16-090707-Fab) were purchased from Athens Research and Technology (Athens, GA).  5.4.2 Single molecule AFM experiment Single molecule AFM experiments were carried out on a custom-built atomic force microscope as described previously. All the force-extension measurements were carried out in PBS buffer (pH 7.4). For hFc binding studies, we first deposited polyprotein and hFc solution onto a glass coverslip and mixed them in situ. The AFM experiments were carried out after allowing the mixture to equilibrate for ~30mins. Two different types of AFM cantilevers (Si3N4 cantilevers from Veeco) were used in all the experiments with a spring constant of 60 pN nm-1 and 15 pN nm-1, respectively. All the cantilevers were calibrated in PBS buffer using the equipartition theorem before each experiment. The pulling speed used for all the pulling experiment was 400 nm s-1, unless otherwise indicated.  142  5.5 References 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.  Tatham AS & Shewry PR (2000) Trends Biochem Sci 25, 567-571. Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, & Savage K (2002) Philos Trans R Soc Lond B Biol Sci 357, 121-132. Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, & Hansma PK (1999) Nature 399, 761-763. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. Labeit S & Kolmerer B (1995) Science 270, 293-296. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Lee G, Abdi K, Jiang Y, Michaely P, Bennett V, & Marszalek PE (2006) Nature 440, 246-249. Bullard B, Garcia T, Benes V, Leake MC, Linke WA, & Oberhauser AF (2006) Proc Natl Acad Sci U S A 103, 4451-4456. Rief M, Clausen-Schaumann H, & Gaub HE (1999) Nat Struct Biol 6, 346-349. Dietz H & Rief M (2004) Proc Natl Acad Sci U S A 101, 16192-16197. Carrion-Vazquez M, Oberhauser AF, Fisher TE, Marszalek PE, Li H, & Fernandez JM (2000) Prog Biophys Mol Biol 74, 63-91. Lu H & Schulten K (2000) Biophys J 79, 51-65. Paci E & Karplus M (2000) Proc Natl Acad Sci U S A 97, 6521-6526. Klimov DK & Thirumalai D (2000) Proc Natl Acad Sci U S A 97, 7254-7259. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Brockwell DJ, Beddard GS, Clarkson J, Zinober RC, Blake AW, Trinick J, Olmsted PD, Smith DA, & Radford SE (2002) Biophys J 83, 458-472. Williams PM, Fowler SB, Best RB, Toca-Herrera JL, Scott KA, Steward A, & Clarke J (2003) Nature 422, 446-449. Li H (2007) Organic & Biomolecular Chemistry 5, 3399-3406. Sharma D, Cao Y, & Li H (2006) Angew Chem Int Ed Engl 45, 5633-5638. Ainavarapu SR, Li L, Badilla CL, & Fernandez JM (2005) Biophys J 89, 33373344. Junker JP, Hell K, Schlierf M, Neupert W, & Rief M (2005) Biophys J 89, L46-48. Wilcox AJ, Choy J, Bustamante C, & Matouschek A (2005) Proc Natl Acad Sci U S A 102, 15435-15440. Lo Conte L, Chothia C, & Janin J (1999) J Mol Biol 285, 2177-2198. Gronenborn AM, Filpula DR, Essig NZ, Achari A, Whitlow M, Wingfield PT, & Clore GM (1991) Science 253, 657-661. Cao Y & Li H (2007) Nat Mater 6, 109-114. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Akerstrom B, Brodin T, Reis K, & Bjorck L (1985) J Immunol 135, 2589-2592. Derrick JP & Wigley DB (1992) Nature 359, 752-754. Sauer-Eriksson AE, Kleywegt GJ, Uhlen M, & Jones TA (1995) Structure 3, 265278.  143  30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43.  Sagawa T, Oda M, Morii H, Takizawa H, Kozono H, & Azuma T (2005) Mol Immunol 42, 9-18. Marko JF & Siggia ED (1995) Macromolecules 28, 8759-8770. Cao Y, Balamurali MM, Sharma D, & Li H (2007) Proc Natl Acad Sci U S A 104, 15677-15681. Ng SP, Rounsevell RW, Steward A, Geierhaas CD, Williams PM, Paci E, & Clarke J (2005) J Mol Biol 350, 776-789. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Li PC & Makarov DE (2004) Journal of Physical Chemistry B 108, 745-749. Sloan DJ & Hellinga HW (1999) Protein Sci 8, 1643-1648. Brockwell DJ, Beddard GS, Paci E, West DK, Olmsted PD, Smith DA, & Radford SE (2005) Biophys J 89, 506-519. Swain JF & Gierasch LM (2006) Curr Opin Struct Biol 16, 102-108. Changeux JP & Edelstein SJ (2005) Science 308, 1424-1428. Nauli S, Kuhlman B, & Baker D (2001) Nat Struct Biol 8, 602-605. Malakauskas SM & Mayo SL (1998) Nat Struct Biol 5, 470-475. Li H, Oberhauser AF, Fowler SB, Clarke J, & Fernandez JM (2000) Proc Natl Acad Sci U S A 97, 6527-6531. Hann E, Kirkpatrick N, Kleanthous C, Smith DA, Radford SE, & Brockwell DJ (2007) Biophys J 92, L79-81.  144  Chapter 6: Engineered elastomeric proteins with dual elasticity can be controlled by a molecular regulator†  6.1 Introduction Elastomeric proteins are molecular springs found in many biological tissues and bio-materials of superb mechanical properties (1-5). Depending on their functions, elastomeric proteins can display distinct mechanical properties by functioning as entropic springs, whose elasticity is governed by configurational entropy (1, 2, 6, 7), or as shock absorbers (3-5, 8-10). Here we combine single-molecule atomic force microcopy (AFM) and protein engineering techniques to create the first elastomeric chameleon proteins having two distinct mechanical stabilities whose performances can change in response to a molecular regulator. These chameleon proteins are mechanically labile by design and behave as entropic springs. Upon binding of a molecular regulator, the chameleon proteins switch into a state of significant mechanical stability and serve as shockabsorbers. These engineered proteins effectively mimic and combine the two extreme elastic behaviors found in natural elastomeric proteins, and thus represent a new type of “smart” nanomaterial that will find potential applications in nanomechanics and material sciences. Mechanical stability is an intrinsic property of proteins and depends on the three dimensional structures of proteins (11-14). Mechanical stability can be measured quantitatively using single-molecule AFM. Tuning the mechanical properties of proteins  †  A version of this chapter has been published as “Cao, Y.; Li, H., Engineered elastomeric proteins with dual elasticity can be controlled by a molecular regulator. Nat Nanotechnol 2008, 3, (8), 512-6.”.  145  is not only important for regulating various biological processes and activities (15, 16), but is also critical to use artificial elastomeric proteins for constructing smart materials and nanomechanical devices (17, 18). Polyprotein of GB1, the B1 immunoglobulin G (IgG) binding domain of Streptococcal protein G (19), is an ideal artificial elastomeric protein (17). GB1 is a small α+β protein (Fig. 6.1A) and binds the constant Fc region of IgG with high affinity (20). In the previous chapter we demonstrated that protein-protein interactions can effectively regulate the mechanical stability of GB1 (21, 22): our singlemolecule AFM studies showed that GB1 domains unfold at ~180 pN with a contour length increment (ΔLc) of ~18 nm; upon binding of the Fc fragment of human IgG (hFc), the mechanical stability of the GB1 domains increases to ~260 pN. As discussed in previous chapters, it has been predicted by molecular dynamics simulations of mechanical unfolding process of GB1 and its structural homologue, the rupture of hydrogen bonds connecting β-strands 1 and 4 provides key resistance to mechanical unfolding of GB1 (23, 24). Thus β-strands 1 and 4 constitute the mechanoactive site for GB1 (Fig. 6.1A, in green). Although the Fc binding site of GB1, whose key residues are highlighted in Fig. 6.1A (dark blue) (20), is distant from the mechano-active site, the binding of Fc can significantly enhance the mechanical stability of GB1. Such unique properties suggest that it may be possible to perturb the mechano-active site of GB1 to significantly reduce its mechanical stability but not to dramatically alter its Fc binding property, so that Fc binding can still be used to enhance the mechanical stability of GB1. Based on such a hypothesis, here we endeavor to engineer GB1-based artificial polyprotein “chameleons” with distinctively different dual mechanical compliance (the inverse of stiffness) that can be controlled by the molecular regulator hFc.  146  Figure 6.1 Designing elastomeric chameleon proteins using proline mutagenesis A) The binding site for Fc (indicated in dark blue) on the GB1 protein is distant from the mechano-active site (green, with backbone hydrogen bonds show with black lines). B) The location where valine (Val54) and threonine (Thr18) are substituted with proline in GB1. C) Far-UV CD spectra shows that the V54P and T18P mutations disrupt the β sheet structure of GB1, but overall the mutant proteins GV54P and GT18P largely retained their α+β structure.  147  6.2 Results To design such elastomeric chameleon proteins, we employed proline mutagenesis. Proline mutations can block the formation of backbone hydrogen bonds and disrupt local β sheet structure (25). Thus, we constructed the mutant protein GV54P, in which valine on position 54 on β-strand 4 was substituted with a proline to disrupt the mechano-active site (Fig. 6.1B). The far-UV circular dichroism (CD) spectrum of GV54P (Fig. 6.1C) indicated that GV54P is folded with a typical α+β structure. However, compared with wild type (wt) GB1, the proline mutation V54P did cause some disruption in the β-sheet region, as evidenced by the decrease in intensity of the CD signal for βsheet at ~215 nm. To measure the mechanical stability of GV54P unambiguously using singlemolecule AFM, we constructed a heteropolyprotein (GV54P-I27)4, in which the wellcharacterized I27 domain from muscle protein titin (26) serve as fingerprints for discerning features of the mechanical unfolding of GV54P. The unfolding of I27 is characterized by ΔLc of ~28 nm and an unfolding force of ~200 pN (26, 27). Stretching (GV54P-I27)4 results in force-extension curves that are characterized by featureless long spacers followed by the characteristic unfolding force peaks of I27 (Fig. 6.2A). Because GV54P alternates with I27 in the heteropolyprotein, we are certain that the long featureless spacer preceding the I27 unfolding events must result from the stretching and unfolding of GV54P (28). The long featureless spacers indicate that most GV54P domains unfold at forces that are below the detection limit of our AFM (~20 pN). In a small percentage of force-extension curves, we also observed that a few GV54P domains unfold at low but detectable forces (bottom trace of Fig 6.2A). Fig. 6.2D (black curve)  148  shows the unfolding force histogram of GV54P. It is evident that GV54P does not show a well-defined peak and 60% of GV54P domains unfold at forces below 30 pN, indicating that the mutation V54P significantly disrupted the mechano-active site for GB1 and dramatically reduces its mechanical stability. Therefore, GV54P behaves largely like an entropic spring upon stretching. We then used surface plasmon resonance (SPR) spectroscopy to examine the binding properties of GV54P to hFc. Fig. 6.3A shows the binding and dissociation curves of GV54P to hFc measured using SPR. It is clear that GV54P can bind to and dissociate from hFc in a reversible fashion. The dissociation constant (Kd) of GV54P to hFc is 5.56×10-7 M, which is somewhat reduced as compared with that for wt GB1 (29). This indicates that, although the mechano-active site in GV54P is disrupted, GV54P does retain its binding affinity to hFc. Therefore, we could use hFc to enhance the mechanical stability of GV54P.  149  Figure 6.2 Elastomeric chameleon proteins show dual mechanical elasticity. Elastic behaviors of GV54P and GT18P can be switched from random coil-like behavior A), E) to sequential unfolding behavior characterized by unfolding events of ΔLc of ~18 nm upon binding of hFc B), F); and back to entropic spring behavior upon removing of hFc C), G). Top panels are schematic of the AFM experiments under given experimental conditions. Grey lines are fits to unfolding force peaks using the worm-like Chain (WLC) model of polymer elasticity. D), H) Unfolding force histograms of GV54P and GT18P in the presence (red, n=240 for GV54P and n=517 for GT18P) and absence of hFc (black, n=283), respectively. Stretching (I27-GV54P)4 in the presence of 11μM of hFc, which is close to the saturating concentration, results in force-extension curves of unfolding events showing mixed ΔLc of 18 nm and 28 nm (Fig. 6.2B). The unfolding events with ΔLc of ~28 nm (Fig. 6.2B, in black) correspond to the unfolding of the I27 domains, while the unfolding events with ΔLc of 18 nm (Fig. 6.2B, in red) correspond to the complete unfolding of GV54P in complex with hFc (GV54P/hFc) (21). The unfolding force of GV54P/hFc  150  show a broad distribution with the peak centered at much elevated force of ~200 pN (Fig. 6.2D, in red), which increases significantly as compared with that for GV54P alone.  Figure 6.3 SPR sensorgrams for the binding and dissociation of monomeric chameleon proteins GV54P (A) and GT18P (B) to hFc. Concentrations of analytes GV54P and GT18P are indicated on each individual curve. Using the Langmuir 1:1 association model, the dissociation constants for GV54P and GT18P to hFc were determined to be 5.56×10-7M and 4.73×10-9M, respectively. These results demonstrate that GV54P exhibits two distinct mechanical compliances depending on the presence of hFc: in the absence of hFc, GV54P is mechanically compliant and functions as an entropic spring; upon binding of hFc, GV54P  151  exhibits significant mechanical stability and unfolds sequentially upon stretching. Such distinct dual mechanical compliances are effectively incorporated into one protein and thus make GV54P a mechanical “chameleon”, which can change its mechanical properties in response to the molecular regulator hFc. It is of note that the regulation of the mechanical stability of GV54P by hFc is fully reversible (Fig. 6.2C): upon removing hFc from the solution by simple flushing with fresh PBS, the dissociation of hFc from GV54P converts GV54P back to a mechanically labile state and the force-extension relations of GV54P resume random coil-like elastic behaviors. We also explored the use of mutant GT18P to construct an elastomeric chameleon protein. Following a similar strategy, we constructed GT18P, in which threonine on position 18 was mutated to a proline in β-strand 2 to disrupt the mechano-active site and weaken its mechanical stability (Fig. 6.1B). Similar to GV54P, the CD spectrum of GT18P shows that GT18P retains an overall α+β structure but its β-sheet structure is somewhat disrupted (Fig. 6.1C). SPR experiments confirm that GT18P retains high binding affinity to hFc (Kd = 4.73×10-9 M) (Fig. 6.3B). Stretching the polyprotein (GT18P)8 results in force-extension curves characteristic of ideal entropic springs (Fig. 6.2E), which can be well-described by the worm-like-chain model (WLC) of polymer elasticity and are fully reversible upon relaxation (Fig. 6.4A). The lack of unfolding force peaks indicates that GT18P unfolds at forces below our detection limit and the polyprotein (GT18P)8 behaves like an ideal entropic spring. In comparison, the binding of hFc to GT18P changes the mechanical behavior of (GT18P)8 significantly. Stretching (GT18P)8 in the presence of 11 μM hFc results in sawtooth peaks with ΔLc of ~18 nm, corresponding to the unfolding of GT18P domains in complex with hFc (Fig. 6.2F). The  152  unfolding force histogram of GT18P in the presence of hFc (Fig. 6.2H) indicates that GT18P in complex with hFc has a well-defined mechanical stability of ~110 pN. Removing hFc from the solution will convert (GT18P)8 back to an entropic spring-like polymer (Fig. 6.2G).  Figure 6.4 Reversible elastic behaviors of polyprotein (GT18P)8. A), Representative stretching-relaxation cycles of the same (GT18P)8 molecule. The stretching and relaxation curves are superimposable with no hysteresis between the two, indicating the fully reversible and entropic nature of the elastic behavior of (GT18P)8. B), Stretchingrelaxation-stretching cycle of the same (GT18P)8 molecule in the presence of 11μM of hFc. The unfolded GT18P domains can fold back after relaxation and rebind with hFc to regain their mechanical stability, a typical feature of naturally occurring tandem modular elastomeric proteins.  153  Figure 6.5 Schematic illustration of the general concept of elastomeric chameleon proteins. A) Elastomeric chameleon proteins show dual elasticity. They can be reversibly switched between a compliant entropic spring and a mechanically stable shock absorber by a molecular regulator. The shaded areas under the force-extension curves show the energy required to stretch the protein to the given extension. B) Comparing the energy cost for stretching an entropic spring and a shock absorber protein to the same extension.  6.3 Discussion Compared with GV54P, (GT18P)8’s chameleon nature is even more striking: in response to the presence of hFc, (GT18P)8 exhibits completely different elastic behaviors that do not overlap with each other. Therefore, (GT18P)8 represents a perfect chameleon elastomeric protein whose idealized features are schematically depicted in Fig. 6.5. The distinct mechanical stability exhibited by the two forms of the chameleon elastomeric protein lead to very different mechanical performances. In the absence of the molecular regulator hFc, (GT18P)8 serves as an ideal elastic spring to provide elasticity and extensibility: stretching (GT18P)8 does not require significant amount of energy (shaded area under the force-extension curve) and stretching-relaxation is fully reversible. In contrast, the molecular regulator can convert the chameleon protein into a mechanically stable state that offers higher strength and toughness. For example, stretching (GT18P)8 in the presence of hFc require as much as 20 times the energy  154  required to stretch an entropic spring to the same extension (3) (Fig. 6.5B). Moreover, (GT18P)8 in the presence of hFc can fully refold to regain its mechanical stability after mechanical unfolding (Fig. 6.4B). These properties are similar to those of natural tandem modular proteins and make the chameleon protein in the presence of molecular regulator an ideal shock-absorber. Clearly, these two distinct mechanical behaviors, which closely mimic the two extreme behaviors observed in naturally occurring elastomeric proteins, are effectively incorporated into one protein. Therefore, the same chameleon protein can function as very different mechanical elements in response to a molecular regulator in a fully reversible fashion. Similar traits can also be found in other proteins, such as dihydropholate reductase (16). These materials represent a new type of engineered “smart” materials that possess unique mechanical features. Our study not only represents an important step forward in designing smart artificial proteins with well-defined mechanical properties, but will also pave the way to use engineered elastomeric proteins in a wide range of applications in nanomechanics as well as in material sciences.  6.4 Experimental section 6.4.1 Protein engineering Plasmids that encode wild type GB1 were generously provided by David Baker of the University of Washington. Mutants GV54P and GT18P were constructed using standard PCR based site-directed mutagenesis techniques and confirmed by direct DNA sequencing. Polyprotein genes of (GV54P-I27)4 and (GT18P)8 were constructed using well-established procedures (26). The polyproteins were expressed in the DH5α strain of E. coli, purified by Co2+ affinity chromatography, and stored at 4oC in PBS buffer with  155  300 mM NaCl and 150 mM imidazole. hFc (Cat#16-16-090707-FC ) was purchased from Athens Research and Technology (Athens, GA).  6.4.2 Single-molecule AFM experiment Single-molecule AFM experiments were carried out on a custom-built atomic force microscope as described in previous chapters (21). All the force-extension measurements were carried out in PBS buffer (pH 7.4). In a typical experiment, 1 μl of polyprotein sample was deposited onto a clean glass cover slip covered by 50 μl PBS buffer, and was allowed to adsorb for ~5 minutes before carrying out force-extension measurements. For hFc binding studies, we added hFc stock solution to the PBS buffer and mixed thoroughly to reach the desired hFc concentration. The concentration of hFc we used in our single molecule AFM experiments was kept constant in this study at ~11 μM in different experiments, which is close to the saturating concentration according to the dissociation constant measured from SPR experiments (29). At this concentration of hFc, at least 95% of the total amount of the polyproteins ((GV54P-I27)4 or (GT18P)8 are in the hFc bound form. The molar ratio of hFc to GB1 mutant domains in the polyprotein is around 10. The concentration of hFc was determined by the concentration of hFc stock solution and the dilution factor in our experiments. The AFM experiments were carried out after allowing the mixture to equilibrate for ~30 minutes. For hFc dissociation experiments, we replaced the PBS buffer containing ~11 μM hFc by flushing three times with pure PBS buffer to remove hFc. After equilibrating for ~30 minutes, the AFM experiments were carried out. Two different types of AFM cantilevers (Si3N4 cantilevers from Veeco) were used in all the experiments with a spring constant of 60 pN nm-1 and  156  15 pN nm-1, respectively. All the cantilevers were calibrated in PBS buffer using the equipartition theorem before each experiment. The pulling speed used for all the pulling experiment was 400 nm s-1.  6.4.3 Circular dichoism (CD) and Surface plasmon resonance (SPR) measurements CD spectra were recorded on a Jasco-J810 spectropolarimeter flushed with nitrogen gas. The spectra were recorded in a 0.2cm path length cuvette at a scan rate of 50nm min-1. For each protein sample, an average of three scans is reported. The protein samples were measured in 0.5×PBS at pH 7.4. Data have been corrected for buffer contributions. The SPR experiments were carried out on Biacore3000. hFc was immobilized onto CM5 chip (Biacore) until the SPR signal reaches ~1000 RU. Different concentrations of analytes, GT18P or GV54P monomeric proteins, were then passed through the CM5 chip to measure the binding of GT18P or GV54P to hFc. Dissociation of GT18P or GV54P from hFc was monitored by passing buffer alone through the CM5 chip.  157  6.5 References 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.  Tatham AS & Shewry PR (2000) Trends Biochem Sci 25, 567-571. Gosline J, Lillie M, Carrington E, Guerette P, Ortlepp C, & Savage K (2002) Philos Trans R Soc Lond B Biol Sci 357, 121-132. Smith BL, Schaffer TE, Viani M, Thompson JB, Frederick NA, Kindt J, Belcher A, Stucky GD, Morse DE, & Hansma PK (1999) Nature 399, 761-763. Oberhauser AF, Marszalek PE, Erickson HP, & Fernandez JM (1998) Nature 393, 181-185. Labeit S & Kolmerer B (1995) Science 270, 293-296. Elvin CM, Carr AG, Huson MG, Maxwell JM, Pearson RD, Vuocolo T, Liyou NE, Wong DC, Merritt DJ, & Dixon NE (2005) Nature 437, 999-1002. Urry DW, Hugel T, Seitz M, Gaub HE, Sheiba L, Dea J, Xu J, & Parker T (2002) Philos Trans R Soc Lond B Biol Sci 357, 169-184. Li H, Linke WA, Oberhauser AF, Carrion-Vazquez M, Kerkvliet JG, Lu H, Marszalek PE, & Fernandez JM (2002) Nature 418, 998-1002. Lee G, Abdi K, Jiang Y, Michaely P, Bennett V, & Marszalek PE (2006) Nature 440, 246-249. Bullard B, Garcia T, Benes V, Leake MC, Linke WA, & Oberhauser AF (2006) Proc Natl Acad Sci U S A 103, 4451-4456. Carrion-Vazquez M, Oberhauser AF, Fisher TE, Marszalek PE, Li H, & Fernandez JM (2000) Prog Biophys Mol Biol 74, 63-91. Lu H & Schulten K (2000) Biophys J 79, 51-65. Paci E & Karplus M (2000) Proc Natl Acad Sci U S A 97, 6521-6526. Klimov DK & Thirumalai D (2000) Proc Natl Acad Sci U S A 97, 7254-7259. Bustanji Y & Samori B (2002) Angewandte Chemie-International Edition 41, 1546-1548. Ainavarapu SR, Li L, Badilla CL, & Fernandez JM (2005) Biophys J 89, 33373344. Cao Y & Li H (2007) Nat Mater 6, 109-114. Dietz H & Rief M (2004) Proc Natl Acad Sci U S A 101, 16192-16197. Gronenborn AM, Filpula DR, Essig NZ, Achari A, Whitlow M, Wingfield PT, & Clore GM (1991) Science 253, 657-661. Sauer-Eriksson AE, Kleywegt GJ, Uhlen M, & Jones TA (1995) Structure 3, 265278. Cao Y, Balamurali MM, Sharma D, & Li H (2007) Proc Natl Acad Sci U S A 104, 15677-15681. Cao Y, Yoo T, Zhuang S, & Li H (2008) J Mol Biol 378, 1132-1141. Li PC & Makarov DE (2004) Journal of Physical Chemistry B 108, 745-749. Brockwell DJ, Beddard GS, Paci E, West DK, Olmsted PD, Smith DA, & Radford SE (2005) Biophys J 89, 506-519. Wood SJ, Wetzel R, Martin JD, & Hurle MR (1995) Biochemistry 34, 724-730. Carrion-Vazquez M, Oberhauser AF, Fowler SB, Marszalek PE, Broedel SE, Clarke J, & Fernandez JM (1999) Proc Natl Acad Sci U S A 96, 3694-3699. Li H, Oberhauser AF, Fowler SB, Clarke J, & Fernandez JM (2000) Proc Natl Acad Sci U S A 97, 6527-6531.  158  28. 29.  Li H, Oberhauser AF, Redick SD, Carrion-Vazquez M, Erickson HP, & Fernandez JM (2001) Proc Natl Acad Sci U S A 98, 10682-10686. Sagawa T, Oda M, Morii H, Takizawa H, Kozono H, & Azuma T (2005) Mol Immunol 42, 9-18.  159  Chapter 7: Engineering tandem modular protein-based reversible hydrogels†  7.1 Introduction Artificial protein hydrogels are of increasing interest because of their potential biomedical application as drug delivery carriers, synthetic extracellular matrices and tissue engineering materials (1-7). Pioneer work has demonstrated that leucine zipper domains are excellent building blocks to construct self-assembled protein hydrogels, opening up tremendous opportunities in using genetic engineering to tune the physical and functional properties of hydrogels (8-17) in a precisely controlled manner at the gene level. The use of a flexible random-coil like polypeptide or synthetic polymer as a center block has been the gold standard to construct leucine zipper based protein hydrogels (8, 11-17). Although recent progress has allowed for the splicing of single globular domain into a random coil like center block (15), the random coil nature of the center block remains the key design principle. However, due to their high flexibility, flexible random-coil like polypeptides are susceptible to forming intramolecular loops, leading to undesirable fast erosion rate of the hydrogel in solutions (13-15). Moreover, many extracellular matrix proteins are tandem modular elastomeric proteins, which are composed of individually folded domains (18). Such modular excellular matrix proteins not only possess important biological functions, but also display significant mechanical stability. Thus, incorporating tandem modular  †  A version of this chapter has been published as “Cao, Y.; Li, H., Engineering tandem modular protein based reversible hydrogels. Chem Commun (Camb) 2008, (35), 4144-6.”.  160  proteins into hydrogels will not only be important in creating synthetic extracellular matrices that closely mimic naturally occurring ones, but also may improve the physical properties of the resulting hydrogels. However, direct incorporation of such tandem modular proteins into hydrogels has been challenging. As a proof of principle, we demonstrate here the engineering of the first artificial tandem modular protein based reversible hydrogel and show that the engineered novel hydrogel exhibits unique properties combining much improved erosion properties, fast and reversible sol-gel transition and ability to bind IgG antibodies.  7.2 Results The model tandem modular protein we used here is an artificial polyprotein (GB1)8 (denoted as (G)8 hereafter), which is composed of eight GB1 domains arranged in tandem (Fig. 7.1A). Previous single molecule atomic force microscopy (AFM) experiments have demonstrated that polyprotein (G)8 exhibits excellent mechanical properties that are either comparable to or superior to those of naturally occurring elastomeric proteins (19, 20). To construct a (G)8 based hydrogel, we adopted the standard triblock protein design for a hydrogel: we use (G)8 as the center block to replace the commonly used random-coil like polypeptide, and use the well-characterized leucine zipper domains A (8) to flank the center block at its N- and C-termini (Fig. 7.1A). Leucine zipper A, designed by Tirrell and co-workers (8), can self-associate into oligomers and its use in constructing protein-based hydrogels has been studied in detail (8, 13, 15).  161  Figure 7.1 A) Schematic of the artificial protein A(G)8A. The green helices represent the leucine zipper domains “A” and the red globular proteins represent the GB1 domains. The amino acid sequences of leucine zipper domain A and globular GB1 domains are also shown. B) 12% denaturing SDS-PAGE gel of A(G)8A protein. The unit for the ladder on the right is kDa.  Using standard molecular biology techniques, we constructed the gene of A(G)8A and expressed the triblock protein in E. Coli. Fig. 7.1B shows the denaturing SDS-PAGE gel of the purified protein. The purified A(G)8A appears as a predominant band on SDSPAGE gel with an apparent molecular weight of ~57kDa, in close agreement with the theoretic molecular weight of A(G)8A of 63 kDa. Far UV circular dichroism (CD) spectroscopy results provide further supporting information for the structural integrity of A(G)8A (Fig. 7.2). The CD spectrum of polyprotein (G)8 is characterized by two broad negative minima at 208 nm and 222 nm, consistent with the α+β structure of the constituting GB1 domains (Fig. 7.2A) (21). In comparison, the CD spectrum of A(G)8A shows significant increase in the intensity of the bands characteristic of α-helix secondary structures, as would be expected upon addition of two leucine-zipper sequences to the N- and C- termini of (G)8. Furthermore, in dilute solution, A(G)8A  162  exhibits two distinct thermal unfolding transitions (Fig. 7.2B), as probed by the intensity of the mean residue ellipticity at 222 nm. The first transition occurs at a Tm (temperature of the transition midpoint) of 43 °C and corresponds to the thermal dissociation of the coiled-coil oligomers of A (8); the second transition occurs at a Tm of 75 °C and corresponds to the thermal denaturation of the folded GB1 domains (21). Moreover, the thermal melting behaviors of A(G)8A is fully reversible, providing the possibility of constructing reversible hydrogels. A)  -6  -1  [θ]MRE222 (deg•cm •dmol )  -1  -8  2  0  -10  3  x10  3  2  -5  -10  -15  200  210  220  230  240  Wavelength (nm)  250  x10  [θ]MRE (deg•cm •dmol )  B)  -12 -14 -16  20  40  60  80 0  Temperature ( C)  Figure 7.2 A) CD spectra of (G)8 (open circle) and A(G)8A (filled square) in 0.5×PBS, pH 7.4. B) Thermal melting of A(G)8A in 0.5×PBS pH 7.4. Two melting transitions occur at 43 °C and 75 °C, respectively.  The coiled-coil domains A can self-associate and dissociate depending on temperature, giving rise to the possibility for A(G)8A to form a thermal reversible hydrogel. Indeed, 7% (w/w) aqueous solution of A(G)8A in PBS buffer (pH 7.6) readily forms an opaque hydrogel (Fig. 7.3B). The gel can hang at the bottom of the vial without flowing down on the time scale of months, demonstrating its capability of retaining buffer in the protein matrix, the very character of typical hydrogels. For comparison, a 7% aqueous solution of (G)8 in PBS (pH 7.6) results in a clear transparent solution,  163  which flows readily (Fig. 7.3A). This result indicates that the gelation of A(G)8A is not an intrinsic property of polyprotein (G)8, instead, it is due to the self-aggregation of the coiled-coil A domains flanking the center block of tandem modular protein (G)8. The micro-structure of the hydrogel was investigated using scanning electron microscopy (SEM). As shown in Fig.7.3C, the 7% hydrogel shows an interconnected porous network structure, suggesting that the hydrogel is formed via physical cross-linking, mediated by the leucine zipper sequence.  Figure 7.3 A) Aqueous solution of 7% (G)8, B) Hydrogel of 7% A(G)8A in 100 mM phosphate buffer, pH 7.6, and C) SEM image of 7% A(G)8A hydrogel (The scale bar is 5μm).  Since the self-association and dissociation of coiled-coil domain A is temperature dependent, we expect that A(G)8A hydrogels should undergo sol-gel/gel-sol transitions at similar temperatures. Indeed, increasing temperature to 60 °C results in a gel-sol transition and the hydrogel turned into a viscous liquid. Cooling down the solution to room temperature results in the formation of the hydrogel again (Fig. 7.4). These results confirm that the gel formation is mediated by the formation of physical crosslinking between coiled-coil sequences. In addition, it is also of note that the sol-gel transition for  164  A(G)8A is very fast and can be accomplished typically within 30 seconds. Such a fast response time is very unique for a triblock protein of such high molecular weight. Its origin is not clear.  Figure 7.4 A) hydrogel of A(G)8A at room temperature (the right one; the left, 7% (G)8 solution for comparison ) B) and C) at 600C for 10 s.  Figure 7.5 Solutions of A(G)8A at different concentrations.  These results clearly demonstrate the feasibility of engineering tandem modular protein based hydrogels. Fig. 7.6 shows a schematic for the formation of an A(G)8A hydrogel network at the molecular level, in which the physical crosslinking is formed via self-association of leucine zipper domains. The use of rigid tandem modular proteins as the center block to construct leucine zipper based hydrogels is in contrast to the general understanding that flexible random coil-like polypeptides are a necessary requirement for  165  the center block, suggesting that flexible sequences may not be mandatory in constructing leucine zipper based hydrogels.  Figure 7.6 Schematic drawing of the A(G)8A hydrogel. Hydrogels are formed through the self-association of leucine zipper domains A (green helices).  A previous study of the AC10A hydrogel showed that leucine zipper domains in the triblock proteins cannot be fully utilized and remain dangling in the hydrogel, due to the formation of intramolecular loop between two terminal leucine zipper domains. This results in a high erosion rate and a low storage modulus of the hydrogel. This formation of intramolecular loops is mainly the result of the flexibility and the short end-to-end distance of the center flexible random coil-like polypeptide sequence (13, 15). The use of tandem modular proteins in constructing hydrogels is not only providing new building blocks for hydrogel construction, but also provides an efficient means to overcome this shortcoming and lead to improved properties. Since tandem modular proteins are more rigid than flexible random coil-like sequences, the end-to-end distance of the tandem modular proteins is much larger than that for a flexible random coil sequences. Consequently, the coiled-coil A domains at both ends of the (G)8 are unlikely to meet  166  with each other, thereby effectively preventing the formation of intramolecular loops and increasing the efficiency of intermolecular association. The improvement in the intermolecular association mediated by leucine zipper domains have led to some unique properties observed in our tandem modular protein based hydrogel: low gelling point and low erosion rate. 1.0  Erosion fraction  0.8  0.6  0.4  0.2  0.0 0  500  1000  1500  2000  2500  Time (min) Figure 7.7 Erosion profile of 100mg of 7% A(G)8A hydrogel with a surface area of 0.86 cm2 at room temperature. A linear regression (solid line) measures an erosion rate of 3.23×10-3 mg cm-2min-1.  We estimated the gelling point of A(G)8A by preparing A(G)8A aqueous solutions at different concentrations (pH 7.6). The photographs of these solutions are shown in Fig. 7.5. It is remarkable that A(G)8A can form a gel at a concentration as low as 3.1%. Such a gelation concentration is lower than that for the well-characterized protein AC10A, in which the center block is made of a random coiled sequence (8). Considering that the molecular weight of the center block (G)8 is four times more than  167  that of the C10 sequence, the actual content of coiled-coil domains A in A(G)8A is even lower than that in AC10A. Similarly, the erosion properties of hydrogel A(G)8A also show significant improvements over the random coil sequence based hydrogel AC10A. The erosion rate is a measure of the stability of physically crosslinked hydrogel in open solution. The higher the erosion rate is, the less stable the hydrogel is. Improving erosion properties is key for the use of hydrogels in many biomedical applications. As shown in Fig. 7.7, the erosion profile of A(G)8A in open solution shows a linear mass loss versus time, indicating that the erosion is occurring at the surface. The measured erosion rate is 3.23 ×10-3 mg cm-2 min-1, which is one order of magnitude slower than that for AC10A measured under similar conditions (13, 15). This result is consistent with a previous study on hydrogel AC10-GFP-A, in which a GFP domain is inserted at the end of the random hydrophilic sequence in AC10A (15). Such a low erosion rate makes A(G)8A a desirable hydrogel for biomedical applications, in that it can retain its shape for sufficient long time periods. It is important to point out that the low gelling point and slower erosion rate are not likely the result of the low solubility of (G)8, as (G)8 has excellent water solubility and can dissolve in PBS to a concentration of as high as 200 mg/ml. Furthermore, GB1 domains are well-known for their ability to bind IgG antibodies (22). Therefore, our engineered A(G)8A hydrogel also carries IgG binding properties. Detailed studies on the binding behavior of IgG antibody to A(G)8A hydrogel are currently under way.  168  7.3 Conclusion In summary, we demonstrated here the engineering of the first tandem modular protein based thermo-reversible hydrogel. This novel type of genetically engineered protein-based hydrogel incorporates an artificial tandem modular protein into the hydrogel matrix, thus paving the way to engineer intact extracellular matrix protein-based hydrogels. Moreover, the use of tandem modular proteins makes it possible to engineer hydrogels with much improved physical properties and can be used as artificial extracellular matrix and tissue engineering materials.  7.4 Experimental section 7.4.1 Protein engineering and hydrogel formation The gene encoding protein GB1 was a generous gift from David Baker of the University of Washington. The gene that encodes G8 was constructed as previously reported in Chapter 1. The DNA sequence of leucine zipper based coiled coil domain A, flanked with a 5’ BamHI restriction site and 3’ BglII and KpnI restriction sites, was synthesized by PCR (polymerase chain reaction) based oligonucleotides assembly. The expression vectors of pQE80L-A(G)8A was constructed by iterative cloning A, (G)8 and A genes into the empty pQE80L vector, on the basis of the identity of the sticky ends generated by BamHI and BglII restriction enzymes. The resulting sequence of the entire A(G)8A protein is shown in Fig. 7.8. The expression vector was transformed into Escherichia coli strain DH5α. Cultures were grown at 37 oC in 2.5% LB containing 100mg/L ampicillin, and induced with 0.8mM isopropyl-1-β-D-thiogalactoside (IPTG) when its optical density reached ~1. Protein expression continued for 5 hours after  169  induction. The cells were harvested by centrifugation at 15,000g for 15min and lysed using French press. The soluble fraction was purified using Co2+ affinity chromatography. 10mM dithiothreitol (DTT) was added to both washing and elution buffers to avoid the oxidation of terminal cysteines in A(G)8A. The yield of the protein A(G)8A was in the range of 40mg to 80mg per liter of culture. The purity of the purified A(G)8A is around 90%, as estimated from the SDS-PAGE using AlphaEaseFC software (Version 4.0.0, Alpha Innotech Coporation, San Leandro, CA 94577) . The 10% “impurity” was likely to be truncated fragments of A(G)8A, which were frequently observed in the expression of polyproteins, such as (G)8. The purified protein was then dialyzed against deionized water for 3 days to remove all the salt from the elution buffer. During dialysis, the water was changed every 12 hours. The protein was then lyophilized after dialysis. The hydrogel was made by redissolving the protein sample into phosphate buffer (100mM, pH 7.6). Vigorous mixing helps in the dissolution of the protein. The trapped air bubbles can be removed by fast spinning.  Figure 7.8 Protein sequence of A(G)8A. The sequence MRGSHHHHHH is from the expression vector pQE80L; residues RS between leucine zipper A and GB1 domains in the protein sequence are from either BglII site or the fusion site of BamHI and BglII.  170  7.4.2 Circular dichroism measurement (CD) CD spectra were recorded on Jasco-J810 spectropolarimeter flushed with nitrogen gas. The spectra were recorded in a 0.2cm path length cuvette at a scan rate of 50nm/min. The protein samples were measured in 0.5×PBS at pH 7.4. The data have been corrected for buffer contributions. For each protein sample, an average of three scans is reported and the CD signal was converted into mean residue ellipticity (MRE) using following equation: θMRE = (100·θobs)/[dC(n-1)], where θobs is the observed ellipticity (in deg), d is path length (in cm), C is concentration of protein samples (in M), and n is total number of amino acids in the protein. For thermal melting measurements, the temperature was increased from 15oC to 90oC with a rate of 3oC/min.  7.4.3 Scanning electron microscopy (SEM) SEM imaging of the hydrogel was performed using a Hitachi S4700 Scanning electron microscope. The 7% hydrogel sample was prepared in Eppendorf tubes and left at room temperature for 24 hours to allow the complete formation of hydrogels. Then the samples were shock-frozen in liquid nitrogen, quickly transferred to a freeze drier and lyophilized for 12 hours. The dry samples were then carefully fractured in liquid nitrogen and fixed on aluminum stubs. The surface of the sample was coated by 1nm of gold before SEM observation.  7.4.4 Erosion measurement The erosion rate of the 7% hydrogel A(G)8A was measured using a similar method as reported by Shen et al. 100mg of hydrogel was transferred into a cylindrical  171  glass tube with a flat bottom (1.05cm diameter). The glass tube with the hydrogel was then centrifuged at 1700g for 10 minutes to completely push the hydrogel sample to the bottom and smooth the surface of the hydrogel. The thin gel film together with the glass tube was then soaked in 5mL of 100mM phosphate buffer, pH 7.6, in a scintillation vial. The whole setup was placed on a compact rocker (FINEPCR) tilting at 50 rpm with amplitude of ± 9o, at room temperature. The erosion profiles were determined by measuring the protein absorbance at 280 nm of the supernatant at successive time points using a Nano-drop ultraviolet-visible spectrophotometer. Two different samples were measured and the average value was reported.  172  7.5 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.  Cushing MC & Anseth KS (2007) Science 316, 1133-1134. Fedorovich NE, Alblas J, de Wijn JR, Hennink WE, Verbout AJ, & Dhert WJ (2007) Tissue Eng 13, 1905-1925. Kopecek J (2007) Biomaterials 28, 5185-5192. Lee KY & Mooney DJ (2001) Chem Rev 101, 1869-1879. Lutolf MP & Hubbell JA (2005) Nat Biotechnol 23, 47-55. Peppas NA, Hilt JZ, Khademhosseini A, & Langer R (2006) Advanced Materials 18, 1345-1360. Ulijn RV, Bibi N, Jayawarna V, Thornton PD, Todd SJ, Mart RJ, Smith AM, & Gough JE (2007) Materials Today 10, 40-48. Petka WA, Harden JL, McGrath KP, Wirtz D, & Tirrell DA (1998) Science 281, 389-392. Wang C, Stewart RJ, & Kopecek J (1999) Nature 397, 417-420. Wang C, Kopecek J, & Stewart RJ (2001) Biomacromolecules 2, 912-920. Shen W, Lammertink RGH, Sakata JK, Kornfield JA, & Tirrell DA (2005) Macromolecules 38, 3909-3916. Xu C, Breedveld V, & Kopecek J (2005) Biomacromolecules 6, 1739-1749. Shen W, Zhang K, Kornfield JA, & Tirrell DA (2006) Nat Mater 5, 153-158. Shen W, Kornfield JA, & Tirrell DA (2007) Macromolecules 40, 689-692. Wheeldon IR, Barton SC, & Banta S (2007) Biomacromolecules 8, 2990-2994. Mi L, Fischer S, Chung B, Sundelacruz S, & Harden JL (2006) Biomacromolecules 7, 38-47. Fischer SE, Liu X, Mao HQ, & Harden JL (2007) Biomaterials 28, 3325-3337. Vakonakis I & Campbell ID (2007) Curr Opin Cell Biol 19, 578-583. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Cao Y & Li H (2007) Nat Mater 6, 109-114. Alexander P, Fahnestock S, Lee T, Orban J, & Bryan P (1992) Biochemistry 31, 3597-3603. Akerstrom B, Brodin T, Reis K, & Bjorck L (1985) J Immunol 135, 2589-2592.  173  Chapter 8: Summary and future directions  8.1 Summary “Protein mechanics” is a burgeoning field and great progress has been made in recent years (1-9). Single molecule AFM studies on proteins offer the opportunity to understand the mechanical properties at a single molecule level. To discover and design proteins with novel mechanical properties will be a direct test of our knowledge on the mechanical design of proteins and a challenging task for protein engineers. We have discovered that non-mechanical protein GB1, the B1 immunoglobulin (IgG) binding domain of protein G from group G Streptococcus, has remarkable mechanical stability comparable to many naturally occurring elastomeric proteins (10). Our results highlighted the importance of topology to the mechanical stability of proteins. We proposed a simple way to identify mechanically stable proteins using shear topology as a search criterion, which significantly narrows down the searching for mechanically stable proteins. In addition to its high mechanical stability, we have shown that GB1 polyprotein exhibits a unique combination of mechanical features, including the fastest folding kinetics, high folding fidelity, and low mechanical fatigue during repeated stretching-relaxation cycles and ability to fold against residual forces (11). These superior mechanical properties make GB1 polyprotein a very valuable artificial protein-based molecular spring that could function in challenging working conditions requiring repeated stretching-relaxation. Our study represents an important step towards engineering artificial tandem modular protein based molecular springs with tailored nanomechanical properties for bottom-up construction of new devices and materials.  174  We have developed three methods to modulate the mechanical stability of GB1. These three methods are based on non-covalent interactions of proteins with either small molecules, such as denaturants and metal ions, or large protein molecules. First, we studied the effect of a chemical denaturant, guanidium hydrochloride (GdmCl), on the mechanical stability of GB1 (12). Upon increasing the GdmCl concentration, we observed a systematic decreasing in the mechanical stability of GB1, indicating the softening effect of the chemical denaturant on the mechanical stability of proteins. This mechanical softening effect originates from the reduced free energy barrier between the native and the unfolding transition state, which decreases linearly as a function of the denaturant concentration. Chemical denaturants, however, do not shift the mechanical unfolding pathway or alter the distance between the native and the transition state. We also found that the folding rate constant of GB1 is slowed down by GdmCl in a mechanical folding experiment. Combining the mechanical folding and unfolding kinetics of GB1 in GdmCl solution, we developed a “mechanical chevron plot” as a general tool to understand how chemical denaturants influence the mechanical folding and unfolding kinetics and free energy landscape in a quantitative fashion. Secondly, we demonstrated that engineered bi-histidine metal chelation can enhance the mechanical stability of proteins significantly and reversibly (13). Based on a simple thermodynamic cycle analysis, we engineered bi-histidine metal chelation site into various locations of GB1 to achieve preferential stabilization of the native state over the mechanical unfolding transition state of GB1 by the binding of metal ions. The metal chelation can enhance the mechanical stability of GB1 by as much as 100 pN. Since bi-histidine metal chelation sites can be easily implemented, engineered metal chelation provides a general  175  methodology to enhance mechanical stability of a wide variety of proteins. To the best of our knowledge, this is the first general approach in protein mechanics enabling the rational tuning of the mechanical stability of proteins. This new approach will not only open new avenues towards engineering proteins of tailored nanomechanical properties, but also provide new approaches to systematically map the mechanical unfolding pathway of proteins. Thirdly, we demonstrated that protein–protein interactions can potentially serve as an effective means to regulate the mechanical properties of elastomeric proteins (14, 15). We show that the binding of Fc and Fab fragments of IgG antibody to GB1 (16) can significantly enhance the mechanical stability of GB1. The regulation of the mechanical stability of GB1 by IgG fragments is not through direct modification of the interactions in the mechanically key region of GB1; instead, it is accomplished via the long-range coupling between the IgG binding site and the mechanically key region of GB1 (14). Although Fc and Fab bind GB1 at different regions of GB1, their binding to GB1 can significantly increase the mechanical stability of GB1. Using alanine point mutants of GB1, we showed that the amplitude of mechanical stability enhancement of GB1 by Fc does not correlate with the binding affinity, suggesting that the binding affinity only affects the population of GB1/human Fc (hFc) complex at a given concentration of hFc, but does not affect the intrinsic mechanical stability of the GB1/hFc complex. Furthermore, our results indicate that the mechanical stability enhancement by IgG fragments is robust and can tolerate sequence/structural perturbation to GB1. Our results also demonstrate that the use of protein–protein interactions is an efficient approach to regulate the mechanical stability of GB1-like proteins.  176  Figure 8.1 Tuning the mechanical stability of proteins by modulating the free energy difference between the native state and the transition state (ΔΔGN-T). A) The free energy diagram of the mechanical unfolding N, T and U represent the native, transition and unfolded states, respectively. B)-D) Three different scenarios of the change of ΔΔGN-T in different conditions. B) Because the chemical denaturant stabilizes the transition state more than the native state, the mechanical stability of a protein is decreased. (Chapter 3) C) For the use of metal chelation and protein-ligand interactions, if the native state is preferentially stabilized over the transition state, the mechanical stability of proteins in enhanced. (Chapter 4 and 5) D) If the metal chelation sites and the ligand binding regions of proteins are undisturbed at the mechanical unfolding transition state, the net change of ΔΔGN-T is zero. The mechanical stability of a protein is unchanged in this case. (Chapter 4)  From these three examples, we proposed a general mechanism to tune the mechanical stability of proteins (Fig. 8.1): to change the relative free energy between the native state and the mechanical unfolding transition state (ΔΔGN-T). Since the transition state is more unstructured than the native state, the binding of chemical denaturant to the  177  transition state will be stronger than to the native state. Therefore, ΔΔGN-T becomes smaller in the presence of chemical denaturant. For the metal chelation method, the native state is preferentially stabilized against the unfolding transition state. Therefore, the mechanical stability of proteins can be enhanced upon binding of metal ions. For the protein-ligand interaction, if the ligand can bind to the native state but not the transition state, the native state is preferentially stabilized and the mechanical stability is enhanced. If the ligand can still bind to the transition state, the mechanical stability is unchanged. We also showed the power of applying protein-protein interactions to engineer an elastomeric protein with dual mechanical stability. We found that the Fc fragment of IgG modulates the mechanical stability of GB1 via a long-range coupling mechanism, and the Fc binding site of GB1 is distant from the key region of GB1 that is responsible for its mechanical stability (14). Such unique properties make it possible to perturb the mechano-active site of GB1 to significantly reduce its mechanical stability but not to dramatically alter its Fc binding affinity, so that Fc binding can still be used to enhance the mechanical stability of GB1. Based on this unique property, we engineered GB1based artificial polyprotein “chameleons” with distinctively different dual mechanical compliance that can be controlled by a molecular regulator hFc (17). In the absence of hFc, the chameleon proteins are mechanically compliant and function as an entropic spring; upon binding of hFc, they exhibit significant mechanical stability and unfold sequentially upon stretching. We engineered the first tandem modular protein based hydrogel (18). We adopted the standard triblock protein design for hydrogels and used (GB1)8 as the center block and the well-characterized leucine zipper domain A to flank the center block at its N- and  178  C-termini. Leucine zipper A can self-associate into oligomers and trigger the gelation of the protein (19). We found that the hydrogel exhibits unique properties combining low erosion rate, fast and reversible sol–gel transition and antibody binding ability Although the overall mechanical property of the tandem modular elastomeric protein based hydrogel has not been tested yet, we anticipate that incorporating the elastomeric protein (GB1)8 into hydrogels will significantly improve the strength and elasticity of the hydrogel.  8.2 Future directions 8.2.1  Can non-mechanical proteins with other folds have significant  mechanical stability? Although GB1 is a non-mechanical protein and shows high mechanical stability, it is a β-grasp protein, similar to many mechanical proteins. It is natural to ask whether non-mechanical proteins with other folds are mechanically stable. It is possible that not all the mechanically stable folds have been sampled by nature for mechanical functions. To identify mechanically stable protein folds is therefore not only important for the application of elastomeric proteins as biomaterials, but is also of fundamental importance. In our group, we have extended the single molecule AFM studies to other nonmechanical proteins and even computer designed proteins with different folds. We have evaluated the important roles of different structural motifs to the mechanical stability of a protein. We have found that shear topology is a common feature among mechanically stable proteins (10). Due to the shear topology, the interactions, including backbone hydrogen bonding and hydrophobic interactions, between two force-bearing strands can  179  work collectively against the stretching force and unravel more-or-less simultaneously during the unfolding process. Therefore, this very region serves as a mechanical clamp to resist mechanical unfolding and imparts mechanical stability to the proteins. However, we further found that the direct connection between the N- and C- terminal β strands by hydrogen bonding is not necessary (20). The computer designed protein Top7 is a good example of a protein without direct interactions between N- and C- terminal β strands (21). Nonetheless, there are still many other aspects which have not been explored. For example, it is still unknown whether the length of the N- and C- terminal β strands has any effect on the mechanical stability; why a protein with the same fold as another can have diverse mechanical stability; what the contribution of the structure besides the mechanical clamp region to the mechanical stability of a protein is. More extensive single molecule AFM studies on non-mechanical proteins are required to answer these questions and to advance our understanding on the molecular determinants of mechanical stability.  8.2.2  Rational tuning the mechanical stability of proteins  Rational tuning the mechanical stability of proteins is still a challenging task facing all protein engineers. All methods reported so far are along two lines: a) change the unfolding pathway to shift the position of the unfolding transition state. b) change the relative energy between the native state and transition state. A summary of current methods are shown in Fig. 8.2 (12-14, 20, 22-27). Among them, the rational methods to enhance the mechanical stability of proteins have only been reported in limited cases (13, 14, 20, 22, 26, 27). We have demonstrated that engineered bi-histidine metal chelation can be a general method to enhance the mechanical stability of GB1 (13). The key to  180  enhance the mechanical stability of proteins by metal chelation is preferentially stabilizing the native state over the mechanical unfolding transition state. However, the presence of metal ions is indispensable for the application of this method, which may not be satisfied in many conditions. It will be important to develop an even more general method to enhance the mechanical stability of a protein by directly optimizing its amino acid sequence. Following the same line as the engineered metal chelation site method, we may be able to preferentially stabilize the native state over the transition state by optimizing the residues in the mechanical clamp region of proteins. If the protein with a new sequence results in enhanced thermodynamic stability, the native state of the protein will be stabilized. However, at the transition state, the interactions among these residues will be disrupted and there is no stabilization effect for the transition state. The optimization of the sequences of the mechanical clamp region of proteins can be achieved either by computer design or directed evolution.  181  182  Figure 8.2 Summary of currently available methods to tune the mechanical stability of proteins  8.2.3  Beyond single protein domains  So far the majority of single molecule AFM experiments are focused on individually folded protein domains. From a materials perspective, these domains are not isolated but interacting with other molecules to give the overall mechanical properties of the materials. These interactions can be non-specific bindings and specific interactions. It is still not known how strong these interactions are and whether these interactions affect the mechanical properties of proteins. On the other hand, nature has designed many elastomeric proteins whose mechanical properties are not only determined by the tertiary structure of individually folded protein domains but also by the quaternary structure of the whole elastomeric proteins. For example, ankyrin repeats show a linear spring behaviour before the breaking of its spiral structure at a force much higher than the unfolding of protein domains (28). The titin Z1Z2-telethonin complex is able to bear strong forces generated in muscle contraction and stretching. The hydrogen bonding between Titin Z2 and telethonin is critical to the mechanical stability of the complex. Without telethonin, apo-titin Z1Z2 exhibits much reduced mechanical resistance. In addition, the N-terminal segment of telethonin is not properly folded without titin Z1Z2 (29). These experiments highlight the importance of interdomain interactions to the overall mechanical property of elastomeric proteins and elastomeric protein based materials. The study of these interactions will be the focus of our future endeavours.  183  8.3 References 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.  Li HB (2008) Advanced Functional Materials 18, 2643-2657. Li H (2007) Organic & Biomolecular Chemistry 5, 3399-3406. Carrion-Vazquez M, Oberhauser A, Diez H, Hervas R, Oroz J, Fernandez J, & Martinez-Martin A (2006) in Advanced Techniques in Biophysics, eds. Arrondo J & Alonso A (Springer), pp. 163-236. Fisher TE, Oberhauser AF, Carrion-Vazquez M, Marszalek PE, & Fernandez JM (1999) Trends Biochem Sci 24, 379-384. Oberhauser AF & Carrion-Vazquez M (2008) J Biol Chem 283, 6617-6621. Borgia A, Williams PM, & Clarke J (2008) Annu Rev Biochem 77, 101-125. Brockwell DJ (2007) Biochem Soc Trans 35, 1564-1568. Brockwell DJ (2007) Current Nanoscience 3, 3-15. Carrion-Vazquez M, Oberhauser AF, Fisher TE, Marszalek PE, Li H, & Fernandez JM (2000) Prog Biophys Mol Biol 74, 63-91. Cao Y, Lam C, Wang M, & Li H (2006) Angew Chem Int Ed Engl 45, 642-645. Cao Y & Li H (2007) Nat Mater 6, 109-114. Cao Y & Li H (2008) J Mol Biol 375, 316-324. Cao Y, Yoo T, & Li H (2008) Proc Natl Acad Sci U S A 105, 11152-11157. Cao Y, Yoo T, Zhuang S, & Li H (2008) J Mol Biol 378, 1132-1141. Cao Y, Balamurali MM, Sharma D, & Li H (2007) Proc Natl Acad Sci U S A 104, 15677-15681. Gronenborn AM, Filpula DR, Essig NZ, Achari A, Whitlow M, Wingfield PT, & Clore GM (1991) Science 253, 657-661. Cao Y & Li H (2008) Nat Nanotechnol 3, 512-516. Cao Y & Li H (2008) Chem Commun (Camb), 4144-4146. Petka WA, Harden JL, McGrath KP, Wirtz D, & Tirrell DA (1998) Science 281, 389-392. Sharma D, Perisic O, Peng Q, Cao Y, Lam C, Lu H, & Li H (2007) Proc Natl Acad Sci U S A 104, 9278-9283. Kuhlman B, Dantas G, Ireton GC, Varani G, Stoddard BL, & Baker D (2003) Science 302, 1364-1368. Li H, Wang HC, Cao Y, Sharma D, & Wang M (2008) J Mol Biol 379, 871-880. Dougan L, Feng G, Lu H, & Fernandez JM (2008) Proc Natl Acad Sci U S A 105, 3185-3190. Dietz H, Berkemeier F, Bertz M, & Rief M (2006) Proc Natl Acad Sci U S A 103, 12724-12728. Li H, Carrion-Vazquez M, Oberhauser AF, Marszalek PE, & Fernandez JM (2000) Nat Struct Biol 7, 1117-1120. Borgia A, Steward A, & Clarke J (2008) Angew Chem Int Ed Engl 47, 6900-6903. Ng SP, Billings KS, Ohashi T, Allen MD, Best RB, Randles LG, Erickson HP, & Clarke J (2007) Proc Natl Acad Sci U S A 104, 9633-9637. Lee G, Abdi K, Jiang Y, Michaely P, Bennett V, & Marszalek PE (2006) Nature 440, 246-249. Lee EH, Gao M, Pinotsis N, Wilmanns M, & Schulten K (2006) Structure 14, 497-509.  184  Appendix A: Polyprotein engineering A1. Engineering of plasmids containing octameric genes encoding proteins of interest. We mainly used multiple-cloning strategy to engineer the plasmid of octamer genes. First, we introduce the restriction sites to the gene of protein of interest by polymerase chain reaction (PCR). The typical set of restriction sites is BamHI, BlgII and KpnI. The recognition sequences of them are list in Table A1.  Table A1. The restriction sites typically used for protein engineering.  Cut site Overhang  BamHI  BglII  KpnI  5’ GGATCC 3’ 3’ CCTAGG 5’  5’ AGATCT 3’ 3’ TCTAGA 5’  5’ GGTACC 3’ 3’ CCATGG 5’  5’-GATC  5’-GATC  CATG-3’  It is of note that the plasmid digested by BamHI and BglII have the same overhang and can be ligated together. Moreover, after ligation, the new position (AGATCC) can not be digested by either BamHI or BglII. As shown in Fig. A1, through multiple clone strategy, the gene of polyproteins by multiple cloning steps. Each step includes i. digestion of the plasmid by BamHI and KpnI to prepare the insert; ii. digestion of the plasmid by BglII and KpnI to prepare the vector; and iii. ligation of insert into vector; iv. transform the ligation mixture into E coli. (XL1/Blue strain); and v. screening of the clones containing the correct inserts by a double digestion using restriction enzyme BamHI and KpnI followed by agarose gel electrophoresis. The representative gel pictures  185  are shown in Fig. A2. We use pUC19 (New England Biolabs) for typical multiple cloning steps, as pUC19 is a high copy vector in Escherichia coli (E. coli). However, since the inserts for tetramer and octamer are typically more than 1kb, which are comparable with the size of pUC19 empty vector (~ 2.7 kb), it is difficult for the tetramer-to-octamer cloning step to be done in pUC19 vector. Therefore, we transfer the tetramer insert into the expression vector pQE80L (Qiagen, Valencia, CA) (~4.8 kb) for the tetramer-to-octamer cloning step. pQE80L vector has a T5 promoter and encodes a 6histidine tag at the N-terminal of the polyprotein for affinity column purification.  186  187  Figure A1. General procedure for polyprotein gene engineering.  Figure A2. The representative agarose gel DNA electrophoresis pictures for plasmids containing monomer, dimer, tetramer and octamer of the gene encoding the protein of interest. The gel is stained by ethidium bromide and imaged by AlphaImager from Alpha Innotech (San Leandro, CA) under 365 nm UV light. The vector for monomer, dimer and tetramer is pUC19. The vector for tetramer* and octamer is pQE80L. The right lane of each picture was the 2-log DNA ladder (New England Biolabs) and its size was shown on the right. The plasmids containing genes in different stages (monomer, dimer, tetramer or octamer) were digested by BamHI and BglII restriction enzymes (New England Biolabs). The arrows indicate the correct size of the genes in different stages. It is of noting that above the band of insert (indicated by arrows), there are multiple bands for the vectors due to incomplete enzymatic digestion. The top band is the linear form digested by one enzyme, the middle band is the completely digested vector and the lowest band is the super coiled undigested plasmid.  A2. Polyprotein expression We use E. coli expression system for our protein expression. First, the plasmid was transformed to DH5α competent cell, plated on an agar plate containing 2.5% LuriaBertani broth (LB), 2% agar and 10mg/cm3 ampicillin), and grew overnight at 37 oC. Then the biggest colony was chosen for expression and grew in 10mL 2.5% LB containing 10 mg/L ampicillin over night at 37 oC and 225 rpm. The overnight culture was transferred to 1L of LB medium (1/100 dilution) containing 10 mg/L ampicillin and grew at 37 oC and 225 rpm till the optical density (OD) of the culture reached 0.6-1 (log  188  phase). This typically took around 2.5 hours. Then 1mM of isopropyl-1-β-Dthiogalactoside (IPTG) was added to the medium to induce the protein expression. Protein expression proceeded for 3-4 hours at 37oC and 225 rpm. The cells were harvested by centrifugation at 15,000g for 15min and lysised by incubation with 1mg/mL lysozyme for 30 min. The soluble fraction was passed through Co2+ affinity chromatography, washed by 10mM phosphate buffer containing 300mM NaCl and 7mM imidazole and eluted in phosphate buffer saline (PBS) containing 250mM imidazole. 5mM dithiothreitol (DTT) was added to both washing and elution buffers to avoid the oxidation of terminal cysteins at the C-terminus of the polyprotein. The yield of polyproteins ranged from 10mg to 80mg per liter of culture, depends on the proteins. The purity of the purified proteins is around 90%, as estimated from the SDS-PAGE using AlphaEaseFC software (Version 4.0.0, Alpha Innotech, San Leandro, CA 94577). Fig. A3 shows a representative SDS-PAGE gel for a purified protein (GB1)8.  Protein  Figure A3. 12% denaturing SDS-PAGE picture of octamer polyprotein of (GB1)8. The electrophoresis was run using Tris-Glycine buffer at 200V for 1hour. The gel was stained using Brilliant Blue G-250 (Fisher biotech) and destained using water containing 10% acetic acid and 30% methanol. The protein marker is prestained broad range marker (7175kDa) from New England Biolabs.  189  A3. Sequences of proteins and the encoding cDNAs A3.1 wild type GB1 Protein: MDTYKLILNGKTLKGETTTEAVDAATAEKVFKQYANDNG VDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.2 G6-53 Protein: MDTYKLHLNGKTLKGETTTEAVDAATAEKVFKQYANDN GVDGEWTYDDATKTFHVTE cDNA: atggacacctacaaactgcatctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttccatgttacc gaa  A3.3 G4-51 Protein: MDTYHLILNGKTLKGETTTEAVDAATAEKVFKQYANDNG VDGEWTYDDATKHFTVTE cDNA: atggacacctaccatctgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcagaa aaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaacatttcacggttaccg aa  A3.4 G8-55 Protein: MDTYKLILHGKTLKGETTTEAVDAATAEKVFKQYANDNG VDGEWTYDDATKTFTVHE cDNA:  190  atggacacctacaaactgatcctgcatggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcagaa aaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttcatg aa  A3.5 G4-6 Protein: MDTYHLHLNGKTLKGETTTEAVDAATAEKVFKQYANDN GVDGEWTYDDATKTFTVTE cDNA: atggacacctaccacctgcatctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.6 G32-36 Protein: MDTYKLILNGKTLKGETTTEAVDAATAEKVFKHYANHNG VDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacattacgctaaccacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttacc gaa A3.7 GT25A Protein: MDTYKLILNGKTLKGETTTEAVDAAAAEKVFKQYANDN GVDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctgctgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.8 GK28A Protein: MDTYKLILNGKTLKGETTTEAVDAATAEAVFKQYANDNG VDGEWTYDDATKTFTVTE cDNA:  191  atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga agcagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.9  GK31A  Protein: MDTYKLILNGKTLKGETTTEAVDAATAEKVFAQYANDNG VDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcgcacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.10 GN35A Protein: MDTYKLILNGKTLKGETTTEAVDAATAEKVFKQYAADNG VDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctgccgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa A3.11 NuG2 Protein: MDTYKLVIVLNGTTFTYTTEAVDAATAEKVFKQYANDNG VDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactggttattgttcttaacggaaccacctttacctataccaccgaagctgtagacgctgctactgcagaaaaa gttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttaccgaa A3.12 Gc3b4 Protein: MTTFKLIINGKTLKGEITIEAVDAAEAEKIFKQYANDNGI DGEWTYDDATKTFTVTE cDNA: atgactactttcaaattaatcattaatggtaaaacattgaaaggcgaaatcactatcgaagctgttgatgctgctgaagcagaaaaa atcttcaaacaatacgctaacgacaacggtattgacggtgaatggacttacgacgatgcgactaagacctttacagttactgaa  192  A3.13 GT18P Protein: Met D T Y K L I L N G K T L K G E T T P E A V D A A T A E K V F K Q Y A N D N GVDGEWTYDDATKTFTVTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccccggaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacggttac cgaa  A3.14 GV54P Protein: Met D T Y K L I L N G K T L K G E T T T E A V D A A T A E K V F K Q Y A N D N GVDGEWTYDDATKTFTPTE cDNA: atggacacctacaaactgatcctgaacggtaaaaccctgaaaggtgaaaccaccaccgaagctgtagacgctgctactgcaga aaaagttttcaaacagtacgctaacgacaacggtgtcgacggtgaatggacctacgacgacgctaccaaaaccttcacgccgac cgaa A3.15 Leucine zipper A Protein: GSMRGDDSGDLENEVAQLEREVRSLEDEAAELEQKVSRL K N E I E D L K A E G D H V A P R S C C Stop G T cDNA: ggatccatgcgtggagatgatagcggtgatctggaaaacgaagtggcccagctggaaagggaagttagatcgctggaagatg aagcggctgaactggaacaaaaagtctctcgcctgaaaaacgaaatcgaagacctgaaagccgagggtgatcatgtggcgcct agatcttgttgctaaggtacc Note: The cDNA was synthesized using overlap polymerase chain reaction (PCR) method. The primer sequences are as following: SENSE PRIMERS: AFOR-1 CGAGGATCCATGCGTGGAGATGATAGCGGTGATCTG AFOR-2 GAAAACGAAGTGGCCCAGCTGGAAAGGGAAGTTAGATCGCTG AFOR-3 GAAGATGAAGCGGCTGAACTGGAACAAAAAGTCTCTCGCCTGAAAAAC AFOR-4 GAAATCGAAGACCTGAAAGCCGAGGGTGATCATGTGGCGCCTAG  193  ANTI-SENSE PRIMERS: AREV-1 CTGGGCCACTTCGTTTTCCAGATCACCGCTATCATCTC AREV-2 CAGTTCAGCCGCTTCATCTTCCAGCGATCTAACTTCCCTTTCCAG AREV-3 GCTTTCAGGTCTTCGATTTCGTTTTTCAGGCGAGAGACTTTTTGTTC AREV-4 CTCGGTACCTTAGCAACAAGATCTAGGCGCCACATGATCACCCTCG  A3.16 Leucine zipper P Protein: GSMRGSGDLAPQMLRELQETNAALQDVRELLRQQVKEIT F L K N T V M E S D A S G K L N D R S C C Stop G T cDNA: ggatccatgagaggatcgggtgatctggcgccgcagatgctgcgtgaactgcaggaaaccaacgcggcgctgcaggacgttc gtgaactgctgcgtcagcaggttaaagaaatcaccttcctgaaaaacaccgttatggaatctgacgcgtctggtaaactgaacgat agatcttgttgctaaggtacc Note: The cDNA was synthesized using overlap polymerase chain reaction (PCR) method. The primer sequences are as following: SENSE PRIMERS: PFOR-1 CGAGGATCCATGAGAGGATCGGGTGATCTGGCGCCGCAGATG PFOR-2 CTGCGTGAACTGCAGGAAACCAACGCGGCGCTGCAGGAC PFOR-3 GTTCGTGAACTGCTGCGTCAGCAGGTTAAAGAAATCACCTTC PFOR-4 CTGAAAAACACCGTTATGGAATCTGACGCGTCTGGTAAACTGAACG ANTI-SENSE PRIMERS: PREV-1 GTTTCCTGCAGTTCACGCAGCATCTGCGGCGCCAGATCAC PREV-2 CTGACGCAGCAGTTCACGAACGTCCTGCAGCGCCGCGTTG PREV-3 GTCAGATTCCATAACGGTGTTTTTCAGGAAGGTGATTTCTTTAACCTG PREV-4 CTCGGTACCTTAGCAACAAGATCTATCGTTCAGTTTACCAGACGC  194  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0061613/manifest

Comment

Related Items