MIAMI UNIVERSITY The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

of

Whitney Richert Craig

Candidate for the Degree

DOCTOR OF PHILOSOPHY

______Dr. David Tierney, Director

______Dr. Stacey Lowery Bretz, Reader

______Dr. Michael Crowder, Reader

______Dr. A. Carole Dabney-Smith, Reader

______Dr. Jiangjiang Zhu, Graduate School Representative

ABSTRACT

EVALUATION OF ZINC BINDING GROUPS (ZBGs) AS INHIBITOR BUILDING BLOCKS USING CARBONIC ANHYDRASE AND THE CATALYTIC DOMAIN OF 12 (cdMMP-12)

by

Whitney Richert Craig

Carbonic anhydrase (CA) was used as a model for matrix metalloproteinases (MMPs). CA is a more ubiquitous protein that binds zinc in the same manner as MMPs, His3(OH). Current MMP inhibitors target both unregulated and normal functioning MMPs at the catalytic Zn(II) site, but show minimal specificity and high toxicity. A new class of inhibitor building blocks are being evaluated as potential MMP inhibitor scaffolds, based on their interactions with Zn- and Co- containing CA.

EVALUATION OF ZINC BINDING GROUPS (ZBGs) AS INHIBITOR BUILDING BLOCKS USING CARBONIC ANHYDRASE AND THE CATALYTIC DOMAIN OF MATRIX METALLOPROTEINASE 12 (cdMMP-12)

A DISSERTATION

Presented to the Faculty of

Miami University in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Chemistry and Biochemistry

by

Whitney Richert Craig

The Graduate School Miami University Oxford, Ohio

2017

Dissertation Director: Dr. David L. Tierney

©

Whitney Richert Craig

2017 TABLE OF CONTENTS Chapter 1: Introduction ...... 1 1.1 Metalloenzyme overview...... 2 1.2 Matrix metalloproteinases...... 2 1.2.1 Structure of matrix metalloproteinases...... 3 1.2.2 Catalytic mechanism of matrix metalloproteinases...... 4 1.3 Matrix metalloproteinase 12...... 6 1.4 Matrix metalloproteinase inhibitors...... 6 1.5 Modeling matrix metalloproteinases...... 7 1.6 Carbonic anhydrase as a model...... 8 1.7 Dissertation goals and specific aims...... 9 1.8 References...... 9 1.9 Figures...... 20

Chapter 2: Evaluation of zinc binding groups (ZBGs) thiomaltol, thiopyromechonic acid, and allothiomaltol using carbonic anhydrase as a model ...... 25 2.1 Abstract...... 27 2.2 Introduction...... 27 2.3 Experimental procedures...... 29 2.3.1 Materials...... 29 2.3.2 Methods...... 29 2.4 Results and Discussion...... 31 2.4.1 X-Ray Absorption Spectroscopy...... 31 2.4.2 UV-visible spectroscopy...... 32 2.4.3 Nuclear magnetic resonance spectroscopy...... 34 2.4.4 Electron paramagnetic resonance spectroscopy...... 34 2.5 Conclusion...... 35 2.6 Associated content for supporting information...... 35 2.7 References...... 35 2.8 Figures...... 38

iii Chapter 3 : Evaluation of zinc binding groups (ZBGs) 5-chloro-8-quinolinol (B12), 8-hydroxy-2- quinolinecarboxylic acid (B2) 3-hydroxypyridine-2(1H)-thione (D5), and 2-hydroxy-2,4,6- cycloheptatrienone (G11) using model enzyme carbonic anhydrase ...... 50 3.1 Abstract...... 52 3.2 Introduction...... 52 3.3 Experimental procedure...... 53 3.3.1 Materials...... 53 3.3.2 Methods...... 54 3.4 Results and discussion...... 57 3.4.1 Dissociation of chelator fragment protons via UV-visible spectroscopy...... 58 3.4.2 5-Chloro-8-quinolinol (B12) coordination assessment via UV-visible spectroscopy. 58 3.4.3 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via UV-visible spectroscopy...... 59 3.4.4 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via nuclear magnetic resonance spectroscopy...... 59 3.4.5 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via electron paramagnetic resonance spectroscopy...... 60 3.4.6 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via UV-visible spectroscopy...... 60 3.4.7 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via nuclear magnetic resonance spectroscopy...... 61 3.4.8 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via electron paramagnetic resonance spectroscopy...... 61 3.4.9 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via UV-visible spectroscopy...... 62 3.4.10 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via nuclear magnetic resonance spectroscopy...... 62 3.4.11 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via electron paramagnetic resonance spectroscopy...... 63 3.5 Conclusions...... 63 3.7 References...... 64 3.8 Figures...... 67

iv Chapter 4: Evaluation of curcumin and functional group analysis as a zinc binding groups (ZBGs) using model enzyme carbonic anhydrase ...... 84 4.1 Abstract...... 86 4.2 Introduction...... 86 4.3 Experimental procedure...... 88 4.3.1 Materials...... 88 4.3.2 Methods...... 88 4.4 Results and discussion...... 90 4.4.1 UV-visible spectroscopy...... 91 4.4.2 Nuclear magnetic resonance spectroscopy...... 92 4.4.3 Electron paramagnetic resonance spectroscopy...... 94 4.5 Conclusion...... 94 4.7 References...... 95 4.8 Figures...... 99

Chapter 5: Over-expression, purification, and characterization of the catalytic domain of MMP- 12 (cdMMP-12) ...... 114 5.1 Abstract...... 116 5.2 Introduction...... 116 5.3 Experimental procedure...... 118 5.3.1 Materials...... 118 5.3.2 Methods...... 118 5.4 Results and discussion...... 120 5.5 Conclusions...... 122 5.6 Acknowledgements...... 123 5.7 References...... 123 5.8 Figures...... 126

Chapter 6: Concluding Remarks ...... 130 6.1 Conclusions ...... 132 6.2 References...... 134

v LIST OF TABLES

Table 2-1. Best fits of CoCA+MBG EXAFS……………………………………………………40 Table 2-2. Binding constants from optical titrations…………………………………………….43 Table 2- S1. Detailed EXAFS curve fitting results for CoCA …………………………………..46 Table 2-S2. Detailed EXAFS curve fitting results for CoCA+TM ……………………………..47 Table 2- S3. Detailed EXAFS curve fitting results for CoCA+TPMA …………………………48 Table 2- S4. Detailed EXAFS curve fitting results for CoCA+ATM …………………………..49 Table 3-1: Approximate calculated pKa values for each ZBG- ZBG – B12, B2, D5, and G11 ……………………………………………………………………………………………………69 Table 3-2. Expected g-values for each perpendicular EPR spectra - ZBG – B12, B2, D5, and G11 .……………………………………………………………………………………………...82 Table 4-1. Expected g-values for each perpendicular EPR spectra- curcumin, guaiacol, and acac ……………………………………………………………………………..……………………112 Table 5-1. Metal content of cdMMP-12 samples ……………………………………………...129

vi LIST OF FIGURES

Figure 1-1: Domain structures of MMP classes…………………………………………………20 Figure 1-2: Catalytic Site of MMP-12…………………………………………………………...21 Figure 1-3: Structures of MMP-12 domains……………………………………………………..21 Figure 1-4: 3-Dimensional structure of MMP-8…………………………………………………22 Figure 1-5: Proposed catalytic mechanisms for MMP…………………………………………..23 Figure 1-6: 3-Dimensional comparison of MMP to CA…………………………………………24 Figure 2.1. Structures of metal binding groups (ZBGs) under examination (top) and ZnCA binding modes observed previously.……………………………………………………………..38 Figure 2.2. X-ray absorption spectroscopy of CoCA, and CoCA with added TM, TPMA, and ATM……………………………………………………………………………………………...39 Figure 2-3. Optical spectra of zinc binding groups as a function of pH…………………………41 Figure 2-4: Optical titration of 38 M CoCA with increasing amounts of (A) thiomaltol, (B) thiopyromeconic acid, and (C) allothiomaltol…………………………………………………...42 Figure 2-5: Optical spectra of CoCA with increasing ATM additions at low concentrations…...43 Figure 2-6: 200 MHz 1H NMR spectra of CoCA with increasing additions of ZBG…………....44 Figure 2-7: X-band EPR spectra of CoCA, and CoCA in complex……………………………...45 Figure 2-S1. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA………………….46 Figure 2-S2. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+TM ……………47 Figure 2-S3. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+TPMA…………48 Figure 2-S4. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+ATM ………….49 Figure 3-1: Structure of fragments, name, and abbreviations……………………………………67 Figure 3-2: Optical spectra of zinc binding groups at varying pH………………………………68 Figure 3-3: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 5-chloro-8-quinolinol (B12)………………………………………………………………..69 Figure 3-4: The change in absorbance for the substrate specific bands for titrated amounts of 5- chloro-8-quinolinol (B12) added to 25 µM CoCA………………………………………………70 Figure 3-5: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 8-hydroxy-2-quinolinecarboxylic acid (B2)………………………………………………..71 Figure 3-6: The change in absorbance for the substrate specific bands for titrated amounts of 8- hydroxy-2-quinolinecarboxylic acid (B2) added to 25 µM CoCA………………………………72

vii Figure 3-7: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 8-hydroxy-2-quinolinecarboxylic acid (B2)…………………………………….73 Figure 3-8. X-band perpendicular mode EPR of CoCA and CoCA + B2……………………….74 Figure 3-9: UV-visible spectra of the titration of 215 µM cobalt(II) substituted carbonic anhydrase with 3-hydroxypyridine-2(1H)-thione (D5)………………………………………….75 Figure 3-10: The change in absorbance for the substrate specific bands for titrated amounts of 3- hydroxypyridine-2(1H)-thione (D5) added to 215 µM CoCA…………………………………..76 Figure 3-11: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 3-hydroxypyridine-2(1H)-thione (D5)………………………………………….77 Figure 3-12. X-band perpendicular mode EPR of CoCA and CoCA + D5……………………..78 Figure 3-13: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 2-hydroxy-2,4,6-cycloheptatrienone (G11)……………………………………..79 Figure 3-14: The change in absorbance for the substrate specific bands for titrated amounts of 2- hydroxy-2,4,6-cycloheptatrienone (G11) added to 25 µM CoCA……………………………….80 Figure 3-15: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 2-hydroxy-2,4,6-cycloheptatrienone (G11)…………………………………...... 81 Figure 3-16. X-band perpendicular mode EPR of CoCA and CoCA + G11…………………….82 Figure 3-17: Proposed coordination of each ZBG fragments……………………………………83 Figure 4-1. Structure of curcumin and moieties…………………………………………………99 Figure 4-2. UV-visible spectra of the titration of 760 µM cobalt(II) substituted carbonic anhydrase with curcumin……………………………………………………………………….100 Figure 4-3. The change in absorbance for the d-d bands for titrated amounts of curcumin added to 760 µM CoCA……………………………………………………………………………….101 Figure 4-4. The change in absorbance for the substrate specific bands for titrated amounts of curcumin added to 760 µM CoCA..…………………………………………………………….102 Figure 4-5. UV-visible spectra of the titration of 760 µM cobalt(II) substituted carbonic anhydrase with acetylacetonate (acac). ..………………………………………………………103 Figure 4-6. The change in absorbance for the d-d bands for titrated amounts of acac added to 760 µM CoCA....…………………………………………………………………………………….104 Figure 4-7: UV-visible spectra of the titration of 760 µM cobalt(II) substituted carbonic anhydrase with guaiacol....……………………………………………………………………...105 Figure 4-8: The change in absorbance for the d-d bands for titrated amounts of guaiacol added to 760 µM CoCA..…………………………………………………………………………………106

viii Figure 4-9: The change in absorbance for the substrate specific bands for titrated amounts of guaiacol added to 760 µM CoCA....……………………………………………………………107 Figure 4-10: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with curcumin....…………………………………………………………………….108 Figure 4-11: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with acac....………………………………………………………………………….109 Figure 4-12: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with guaiacol....……………………………………………………………………...110 Figure 4-13. X-band EPR of perpendicular mode of CoCA, CoCA + 2 molar equivalence each inhibitor at specified attenuation………………………………………………………………..111 Figure 4-14: Proposed coordination of curcumin, guaiacol, and acac in comparison to resting enzyme binding...………………………………………………………………………………112

Figure 5-1. Crystal structure of Zn2-cdMMP-12 using the coordinates (PDB 3NX7)…………126 Figure 5-2. Overexpression and refolding of cdMMP-12 using SDS-PAGE…………………..127 Figure 5-3. Q-Sepharose ion exchange chromatographic purification of cdMMP-12 using SDS- PAGE...…………………………………………………………………………………………128 Figure 5-4. MALDI-TOF mass spectrum of trypsin digested cdMMP-12……………………..129

ix DEDICATION

To Mia and Joseph, may you find the niche where you excel and surround yourself with those who help you surpass your goals.

In Memory of Wilhelm “Bill” Richert, 1935 – 2014 and Bobbie “Alice” Howard, 1934 – 1998.

x ACKNOWLEDGEMENTS

I would like to take the opportunity to thank several people who have contributed to the completion of my dissertation. First, I must sincerely thank my advisor, Dr. David Tierney, for his guidance and benevolence that helped me progress forward with research. Thank you for encouraging me to grow as a scientist and to become an independent thinker. I am grateful for the experience and know it has pushed me to grow. I want to thank the members of my dissertation committee, Drs. Michael Crowder, Carole Dabney-Smith, Stacey Lowery Bretz, and Luis Actis, for their direction and assistance throughout my graduate career. Also, Stacey, it has been a privilege to work with you in order to develop my teaching methods. Your thoughtfulness and feedback have been invaluable in shaping my teaching ideals. I also need to thank members of the Tierney and Crowder lab, past and current, graduate and undergraduate. To the undergraduates who worked with me, thank you for your willingness to learn and persistence to achieve despite research difficulties. The perspective each of you brought was refreshing and kept me grounded with your humor – from overflowing carboys to oranges, thank you! There are many within the department who have enriched my years at Miami. Amanda Storm, you are a blessing beyond measure and a beacon of faith! I am delighted our paths crossed. Zahilyn Allred, you and Zach are such a gift to Andrew and me. Jess Bases, Melissa Dougherty, Ashley Richardson your friendship made this journey better; Mahesh Aitha, Robb Baum, Hao Yang, and Colin Jack, thank you for passing on your expertise and insights. A deep thanks to my church family and friends. Don and Jane Dawley – thank you for opening your home and heart to us graduate students. You have shaped my spirit and fueled my spiritual growth. It has meant more than can be expressed. Krysti Glasscock and Kate Godwin, thank you for unending friendship that has endured time, distance, and my quirks. Most importantly, I express my gratitude to my family. To my mom and dad, thank you for encouraging me to do my best in all aspects of life, especially in times where it was difficult to press onward. I am here because of your unwavering faith in my abilities from day one. Though my dreams have caused me to spread my wings, I am always grateful that I can fly away home. I thank my husband, Andrew, for your support throughout this journey. Thank you for your encouragement through adversity, pushing me to persevere when I did not see how. Thank you for riding the roller coaster of graduate school with me as a companion and friend. I can’t wait to see where our journey takes us!

xi

Chapter 1: Introduction

1 1.1 Metalloenzyme overview. Metalloenzymes are widespread in their bioavailability and applications as they play key roles physiologically. Approximately, one third of structurally characterized proteins and half of all proteins contain either a metal with catalytic function or a metal with a structural purpose 1. It is not a surprise that metalloproteins and metalloenzymes have a wide array of applications including cosmetics2, 3, pesticides4, 5, and preservatives3. Metalloenzymes have been implicated in pathological processes ranging from cancers6-11, cardiovascular diseases12, ocular diseases13-15, infectious diseases16-18, inflammatory diseases19-21, and neurological diseases14, 22. The pharmacological need to probe and analyze these arises from the fact that an estimated one quarter to one third of proteins require a metal in order to support the protein’s structure or to carry out the protein’s functions23.

Recent literature in the past 40 years has focused on matrix metalloproteinases, a family of Ca(II)- and Zn(II)-dependent enzymes that aid in the maintenance of the extracellular matrix components24-28. Like many metalloenzymes, the calcium and zinc metals are essential for structure and function. These metals are coordinated by electron donating side chains of amino acids. These amino acids include histidine, cysteine, aspartate, and glutamate residues29, 30.

1.2 Matrix metalloproteinases. Matrix Metalloproteinases (MMPs) are a family of Zn(II)-dependent peptidases that play intracellular and extracellular functions in many physiological and pathological processes31. Since the first account of MMPs in the 1960s32, twenty-six isoforms of MMPs have been discovered and grouped into subclasses based on the substrate degraded. The subgroups consist of (MMP-1, -8, - 13, -18), (MMP-2, -9), matrilysins (MMP-7, -26), membrane-type MMPs (MMP-14, -15, -16, -17, -24, -25), metalloelastase (MMP-12) stromelysins (MMP-3, -10), stromelysin-like (MMP-11), and other uncategorized MMPs12, 33. Currently, there have been 23 MMPs identified in human tissues34, eleven of which are localized to the same chromosomal location 11q2335. The main role of MMPs biologically is connective tissue degradation in the extracellular matrix36. Over-expression and misregulation of MMPs can lead to many illnesses such as arthritis, cancer, and heart disease37. Due to the implications in diseases, MMPs have been a desired target for small molecule inhibitors. The role of MMPs in

2 humans has inspired the engineering of potential inhibitors, which can reduce and regulate these diseases by binding to the zinc(II) metal in the of the enzyme37.

1.2.1 Structure of matrix metalloproteinases. MMPs range in size from 28 to 92 kDa38. All MMPs consist of a pro-peptide domain of approximately 80 amino acids, a catalytic domain of approximately 180 – 200 amino acids, and most also have additional domains important in substrate recognition and inhibitor binding39. The catalytic region is followed by a short hinge region that is highly variable. Additional domains are a C-terminal glycosylphosphatidylinositol (GPI) anchor, fibronectin type II-related modules, hemopexin-like domains located on the C-termini of the enzymes40, 41, and a transmembrane domain42. Most MMPs possess a four-bladed, fan shaped structure consisting of approximately 210 amino acids and are linked to the catalytic domain C-terminal through a linker region rich in proline and vary in length43. It is hypothesized that the catalytic domain and the hemopexin-like domain act in conjunction to destabilize and unwind substrate for hydrolysis. Also, the hemopexin-like domain can bind tissue inhibitors of metalloproteinases (TIMPs)44. These binding interactions of the hemopexin-like domain give evidence to possible regulation of MMP activity due to this region12, 43. Figure 1-1 illustrates the structure of each MMP subclass.

Across the MMP family of enzymes, conservation exists in the active site topology where a zinc(II) binding motif HEXXHXXGXXH residues (Figure 1-2), in which the three histidine side chains coordinate the metal45. A fourth coordination to the Zn(II) active site is from a cysteine residue in the propeptide domain. This coordination acts as a blocker, preventing the binding of substrate to the Zn(II) and effectively yielding a catalytically-inactive enzyme 41, 46. Once the prodomain is cleaved, the fourth coordination shifts to water molecule until a substrate is present, resulting in a tetrahedral geometry as is observed in several other metalloezymes 46. This cleavage yields a smaller amino acid construct and a catalytically-active enzyme (Figure 1-3) with a spherically shaped active site 40 Å in diameter34. From this resting state information, it is postulated that when inhibition occurs, the coordinating water is replaced by a zinc binding group (ZBG) in either a monodentate or bidentate fashion47 within the shallow active site.

3 Another Zn(II) is coordinated in a tetrahedral manner to three histidines (His147, His162, and His175) and a monodentate aspartate (Asp149)48. This zinc ion is imperative to enzyme activity, and the binding sequence is highly conserved through all subclasses. Inactivity is due to folding issues, yielding an active site region with a misaligned tris-histidine motif49. Also necessary to structural stability are two calcium ions. The first is coordinated in an octahedral fashion by three amino acids and three backbone carbonyls50. A mutant of this site also resulted in a catalytically- inactive enzyme, unable to cleave a peptide substrate51. The second calcium ion is coordinated by two solvent water molecules in lieu of backbone carbonyls.

Surrounding the active site are six pockets, S1’ – S3’ and S1 – S3, that play a key role in substrate selectivity52 (Figure 1-4). These pockets are moderately hydrophobic as well as relatively similar in shape throughout the MMPs53. The S1’ pocket has a major role in binding selectivity due to its variable size, ranging from shallow (MMP-1, -7, and 17), to moderate (MMP-2, -9), to deep (MMP-3,-8,-12, -13)43 in relative depth. This difference in accessory site shapes between different MMPs allows for the opportunity of subclass or even isozyme specificity in inhibitor structure design. Coupling these differences with a strong zinc binding group would provide a means by which to optimize the interactions of an inhibitor with a specific MMP.

1.2.2 Catalytic mechanism of matrix metalloproteinases. To date, limited information has been reported on the reaction mechanism of MMPs. Four mechanisms have been proposed (Figure 1-5), all varying in the methods used to inform the model. Computational, biophysical, and mimetic methods have been utilized, giving rise to an assortment of active site coordination numbers, mechanisms, and justifications for the model proposed. The first mechanism, proposed by Pelmenschikov and Siegbhan, uses a two-layer ONIOM computational approach54. This model draws from previously studies of the mechanisms of and carboxypeptidase to propose a 4-coordinate Zn(II)- resting enzyme. Active site coordination of three histidines and a water mimic the resting active site of both thermolysin and carboxypeptidase. As catalysis occurs, a hydrogen bonded glutamate activates a water molecule. The carbonyl of a scissile bond in the peptide back bone coordinates

4 the active site Zn(II) as a 5-coordinate compound throughout the reaction. As product is released, a 4-coordinate active site is regenerated.

A second proposed mechanism by the Bertini group uses “snapshots” from crystal images of MMP-12 during catalysis55. A 6-coordinate active site metal is proposed. The coordination sphere consists of three histidine amino acids and three water molecules, one of which is hydrogen bonded to an active site glutamate. As substrate is introduced, two water molecule coordination sites are replaced and a 5-coordinate metal in a gem-diol bound state is generated. One water molecule is consumed during this proposed mechanism, and a resting 6-coordinate active site is regenerated.

A third mechanism relies heavily on data from a matrilysn (MMP-7) – inhibitor complex crystal structure 50, 56. This mechanism also proposes an active site glutamate acting through acid/base catalysis to nucleophilically attack the scissile bond of the substrate. A hydroxamic acid is hydrolyzed in catalysis, returning the active site to resting state. This mechanism does not detail active site coordination during catalysis but illustrates the structures determined through crystal structure analysis56.

The fourth mechanism proposed is based largely on MMP-3 modeling and simulations57. A unique active site is suggested, containing no water bound to the Zn(II) metal during resting state. It is postulated that this 3-coordinate Zn(II) is stabilized by the glutamate and hydrophobic conditions in the active site. Upon substrate addition, glutamate directly attacks the scissile bond to generate a 5-coordinate intermediate that displaces a histidine. This displacement remains present as hydrolysis of the substrate continues and a water nucleophilically attacks the scissile bond. A 4-coordinate Zn(II) is generated and returns to a 3-coordinate metal site upon the release of product.

There are several unique differences among these models, both in coordination number of the metal active site and in the steps of the mechanism sequence. Further exploration is essential to determine a mechanism for MMPs or for a subgroup of MMPs, if there are differences throughout catalysis. Focus on an MMP catalytic event in solution in lieu of a computational or

5 crystallographic method may provide deeper insight, as the current proposed mechanisms do not fully mimic the native catalytic environment.

1.3 Matrix metalloproteinase 12. Matrix Metalloproteinase-12 (MMP-12, Fig 1.) is a Zn(II)-dependent metalloelastase that was first identified in inflammatory macrophages. This enzyme can hydrolyze portions of the extracellular matrix (ECM)58. MMP-12 plays a crucial role in various physiological processes including ovulation and the regulation of local blood vessels58-60. When unregulated, MMP-12 can cause inflammation, cancer progression, and degradation in the lung alveolar wall59 due to the degradation of soluble and insoluble elastin and basement membrane material58, 61-63. Cleavage of elastin bonds favor an isoleucine or leucine in the position prior to hydrolysis if not proceeded by proline64. Because of the unregulated effects, MMP-12 is an appealing therapeutic target. The main therapeutic strategy to treat this disregulation is through the structure based design of drugs that target the MMP catalytic domain. Many efforts have been made to design an inhibitor specifically for MMP; however, these inhibitors tend to target over-expressed MMPs and MMPs functioning normally in tissue degradation, signaling, and other physiological processes. A majority of drug inhibitors designed to date have failed in clinical trials due to toxicity and lack of binding specificity to solely MMP-1265-68. By investigating MMP-12 and model complex binding mechanisms via the catalytic Zn(II) center and Co(II)-substituted analogs, I hope to screen zinc binding group (ZBG) inhibitor building blocks that have improved binding specificity and increased binding affinity by studying the catalytic Zn(II) of MMP-12.

1.4 Matrix metalloproteinase inhibitors. MMPs are considered a desirable drug target because of their multiple disease implications. Clinical development of an inhibitor has been largely unsuccessful to date apart from doxycycline used to treat periodontal disease69. The mechanism by which doxycycline inhibits MMP has not been fully elucidated70, and numerous interactions with other metalloenzymes has been reported with doxycycline71-73. Many issues in clinical trials have been attributed to the nonselective nature of inhibitors. Current inhibitors tend to be indiscriminate in binding and can bind to either an individual a subgroup of MMP or any member of the family of enzymes.74 MMPs remain an attractive drug target, however, an inhibitor with increased selectivity, and

6 predictive animal models of a developed discriminatory inhibitor are necessary for superior selectivity75. The utilization of metal ion cofactors as a target is widely popular as MMP inhibitors share structural features of a metal binding group to bind the catalytic zinc, a sidechain to bind MMP specific regions, and a mimic of a peptide backbone19, 43. To date, four major types of zinc binding group have been proposed for a matrix metalloproteinase inhibitor (MMPi) base. These structures include carboxylates, hydroxamic acids, cyclic and aromatic amines, and phosphate-based zinc binding groups67. A fragment-based lead design was used as an approach to identify new chelating groups that would provide a backbone framework as a basis for the synthesis and characterization of an inhibitor76, 77. The use of low molecular weight compounds as fragments for drug target design provides a method of improving specificity by limiting the steric constrains and hydrogen bonding constraints that larger inhibitor complexes experience in an active site. To optimize metal binding, a multitude of metal chelators have been highlighted as candidates for screening against MMPs in order to suppress hydrolytic activity78-82. This dissertation focuses on screening zinc binding group (ZBG) inhibitors that have been shown to have improved binding potential with MMPs, specifically MMP-12,47, 76, 83 as well as promising compound analogs 84, 85. By screening these molecules, insight into the orientation and the extent of binding of each ZBG can be elucidated, providing much needed insight to the mechanism by which these therapeutic agents inhibit MMPs.

1.5 Modeling matrix metalloproteinases. The binding motif and coordination geometry of the aforementioned inhibitors have been investigated using computational models and scaffold models such as tris(pyrazolyl)borate complexes84, 86. Though using these scaffolds can yield beneficial insight into protein binding, they do not mimic the interactions, dynamics, and nuances of a metalloenzyme active site. These studies do not provide complete insight to binding as they do not model hydrophobic binding pockets many metalloenzymes possess79, 80. Moving to a model enzyme is a logical step, but active site metal coordination must be addressed. Zinc(II) is a d10 diamagnetic metal that does not have a free electron required for most spectroscopic methods. Due to the lack of spectroscopic techniques available to probe zinc ions, both as solvated aqua complex and when coordinated to ligand or proteins, understanding the method by which an inhibitor binds to an active site is a challenging task22. Isomorphic substitution has been proposed as a method by

7 which to replace spectroscopically silent metal ions with another metal ion47. Through substitution, spectroscopically silent, d10 native Zn(II) can be substituted with spectroscopically active, d7 Co(II)87. This technique allows for the zinc(II) in MMPs to be replaced with spectroscopically active cobalt(II)37. Due to the relative similarity in binding bond lengths, similar binding geometries, and activity of cobalt(II) and zinc(II), cobalt(II) complexes have been widely utilized to study zinc(II) proteins 88-92. Therefore, utilizing a cobalt(II) model enzyme can give valuable information that would be otherwise inaccessible with the use of native zinc(II) containing enzymes.

Previous investigation of MMPs 1,3,7, and 16 have been conducted by cohorts85, 93, 94. Within these studies, the catalytic domains of the MMPs were over-expressed in E. coli and purified in a manner specific to each MMP. The resulting purified protein was characterized through circular dichroism spectroscopy, fluorescence studies, mass spectroscopy, metal analysis, and steady- state kinetics. Zinc binding groups provided by collaborators82, 95 for screening. Thiomaltol and allothiomaltol, two zinc binding groups investigated in Chapter 2, have been probed in docking studies to determine IC50 measurements. These findings further provide a basis to understand the binding orientation and strength of a ZBG. This supports data from Chapter 2 of this dissertation and establishes information of specificity for the MMP family as well as off-target binding affinity of the building block.

1.6 Carbonic anhydrase as a model. Carbonic anhydrases (CAs, EC 4.2.1.1) are ubiquitous zinc enzymes, present in multiple isoforms throughout all kingdoms13. Carbonic anhydrase II (CA- II) was the first zinc metalloenzyme discovered in 1940, 96 and the use of CA-II as a bridge from model complex to protein target studies is extensive in the literature, 97-135. CA possesses a well-studied mononuclear Zn(II) active site, and the Co(II)-substituted active site has been well characterized88-90. Another benefit CA holds for modeling is its highly surface accessible active site. The active site of Zn(II) CA has a solvent accessible binding pocket and has a binding motif similar in shape with the catalytic Zn(II) of MMPs. Unlike the shallow, bowl-shaped active site of MMPs, the CA active site is more broad in dimension, allowing for direct assessment of binding orinetations when an inhibitor is used. The hydrophobic nature of the CA active site

8 models that of MMP but with less steric strain. The characterization of preliminary ZBG inhibitors was previously conducted using carbonic anhydrase (CA)47 as a model. Such ZBGs preliminarily investigated include acetohydroxamic acid (AHA), which is used as a benchmark for kinetic inhibition studies in MMP related studies.

1.7 Dissertation goals and specific aims. The ultimate goal of this dissertation is to provide support to drug development by providing insight to ZBG binding in order to find an effective molecule as a candidate for optimization. As described above, there is a need to investigate the coordination of zinc binding groups as a means by which to rationally design an inhibitor that is both enzyme or subclass specific and possesses a decreased toxicity to healthy cells. Due to the difficulty in over-expressing matrix metalloproteinases, the majority of binding information herein is addressed using a model enzyme. This enzyme modeling serves two purposes: first and foremost, the use of CA provides a means by which to screen inhibitors in solution. Using an enzyme to model more closely mimics the native environment compared to model complexes and computational investigations. Secondly, the use of CA provides a basis for non-specific binding. The analysis of ZBG coordination provides general understanding of off-target binding that can be used as a comparison as functional groups are added to the ZBG in endeavors to increase specificity.

Chapter two and three describe the analysis of fragments from a ZBG library27, 47, 76, 83, 84. A proposed coordination of each ZBG is depicted, supporting previous modeling. Chapter four demonstrates a continuance of inhibition, focusing on plant based molecules136 and structural functional groups to probe the binding mechanisms deeper. Chapter five describes the protocol for isolation of over-expressed, soluble MMP-12 in low yields and the use of isolated protein for metal content and catalytic analysis. Expression of MMPs in E. coli systems posed a number of challenges as discussed in Chapter 5 of this dissertation.

1.8 References. [1] Thomson, A. J., and Gray, H. B. (1998) Bio-inorganic Chemistry, Current Opinion in Chemical Biology 2, 155-158. [2] Smit, N., Vicanova, J., and Pavel, S. (2009) The hunt for natural skin whitening agents, International Journal of Molecular Sciences 10, 5326-5349.

9 [3] Parvez, S., Kang, M., Chung, H. S., and Bae, H. (2007) Naturally occurring tyrosinase inhibitors: mechanism and applications in skin health, cosmetics and agriculture industries, Phytotherapy Research : PTR 21, 805-816. [4] Biou, V., Dumas, R., Cohen-Addad, C., Douce, R., Job, D., and Pebay-Peyroula, E. (1997) The crystal structure of plant acetohydroxy acid isomeroreductase complexed with NADPH, two magnesium ions and a herbicidal transition state analog determined at 1.65 A resolution, The EMBO Journal 16, 3405-3415. [5] Cai-Xia Hou, L. M. A. D., Jack P. Goodman and Mark A. Williams. (2006) Metabolism of the peptide deformylase inhibitor actinonin in tobacco, Weed Science 54, 246-254. [6] Paris, M., Porcelloni, M., Binaschi, M., and Fattori, D. (2008) Histone deacetylase inhibitors: from bench to clinic, J Med Chem 51, 1505-1529. [7] Chan, M. M., Huang, H. I., Fenton, M. R., and Fong, D. (1998) In vivo inhibition of nitric oxide synthase gene expression by curcumin, a cancer preventive natural product with anti-inflammatory properties, Biochem Pharmacol 55, 1955-1962. [8] Decock, J., Thirkettle, S., Wagstaff, L., and Edwards, D. R. (2011) Matrix metalloproteinases: protective roles in cancer, J Cell Mol Med 15, 1254-1265. [9] Deryugina, E. I., and Quigley, J. P. (2006) Matrix metalloproteinases and tumor metastasis, Cancer Metastasis Rev 25, 9-34. [10] Gialeli, C., Theocharis, A. D., and Karamanos, N. K. (2011) Roles of matrix metalloproteinases in cancer progression and their pharmacological targeting, FEBS J 278, 16-27. [11] Sousa, S. F., Fernandes, P. A., and Ramos, M. J. (2008) Farnesyltransferase inhibitors: a detailed chemical view on an elusive biological problem, Curr Med Chem 15, 1478-1492. [12] Nagase, H., Visse, R., and Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs, Cardiovasc Res 69, 562-573. [13] Supuran, C. T., Scozzafava, A., and Casini, A. (2003) Carbonic anhydrase inhibitors, Medicinal Research Reviews 23, 146-189. [14] Supuran, C. T. (2008) Carbonic anhydrases: novel therapeutic applications for inhibitors and activators, Nature Reviews Drug Discovery 7, 168-181. [15] Shirasaki, Y. (2008) Molecular Design for Enhancement of Ocular Penetration, Journal of Pharmaceutical Sciences 97, 2462-2496. [16] Fernandez-Montero, J. V., Vispo, E., and Soriano, V. (2014) Emerging antiretroviral drugs, Expert Opin Pharmacother 15, 211-219. [17] Hernick, M., and Fierke, C. A. (2005) Zinc : the mechanisms of zinc-dependent deacetylases, Arch Biochem Biophys 433, 71-84. [18] Dubey, S., Satyanarayana, Y. D., and Lavania, H. (2007) Development of integrase inhibitors for treatment of AIDS: an overview, Eur J Med Chem 42, 1159-1168.

10 [19] Mark Whittaker, Christopher D. Floyd, Peter Brown, a., and Gearing, A. J. H. (1999) Design and therapeutic application of matrix metalloproteinase inhibitors, Chemical Reviews 99, 41. [20] Cristalli, G., Costanzi, S., Lambertucci, C., Lupidi, G., Vittori, S., Volpini, R., and Camaioni, E. (2001) Adenosine deaminase: functional implications and different classes of inhibitors, Med Res Rev 21, 105-128. [21] Moss, M. L., Sklair-Tavron, L., and Nudelman, R. (2008) Drug insight: tumor necrosis factor-converting enzyme as a pharmaceutical target for rheumatoid arthritis, Nat Clin Pract Rheumatol 4, 300-309. [22] McRae, R., Bagchi, P., Sumalekshmy, S., and Fahrni, C. J. (2009) In situ imaging of metals in cells and tissues, Chem Rev 109, 4780-4827. [23] Waldron, K. J., and Robinson, N. J. (2009) How do bacterial cells ensure that metalloproteins get the correct metal?, Nature Reviews Microbiology 7, 25-35. [24] Bertini, I., Calderone, V., Fragai, M., Giachetti, A., Loconte, M., Luchinat, C., Maletta, M., Nativi, C., and Yeo, K. J. (2007) Exploring the subtleties of drug-receptor interactions: The case of matrix metalloproteinases, Journal of the American Chemical Society 129, 2466-2475. [25] Alcaraz, L. A., Banci, L., Bertini, I., Cantini, F., Donaire, A., and Gonnelli, L. (2007) Matrix metalloproteinase-inhibitor interaction: the solution structure of the catalytic domain of human matrix metalloproteinase-3 with different inhibitors, J Biol Inorg Chem 12, 1197-1206. [26] Lauer-Fields, J., Brew, K., Whitehead, J. K., Li, S., Hammer, R. P., and Fields, G. B. (2007) Triple-helical transition state analogues: a new class of selective matrix metalloproteinase inhibitors, J Am Chem Soc 129, 10408-10417. [27] Jacobsen, F. E., Lewis, J. A., and Cohen, S. M. (2007) The design of inhibitors for medicinally relevant metalloproteins, ChemMedChem 2, 152-171. [28] Rosenblum, G., Meroueh, S., Toth, M., Fisher, J. F., Fridman, R., Mobashery, S., and Sagi, I. (2007) Molecular structures and dynamics of the stepwise activation mechanism of a matrix metalloproteinase zymogen: challenging the cysteine switch dogma, J Am Chem Soc 129, 13566-13574. [29] Holm, R. H., Kennepohl, P., and Solomon, E. I. (1996) Structural and Functional Aspects of Metal Sites in Biology, Chemical Reviews 96, 2239-2314. [30] Ulmer, D. D., and Vallee, B. L. (1971) Structure and Function of Metalloenzymes, In Bioinorganic Chemistry, pp 187-218, American Chemical Society. [31] Kumar, D., Kumar, M., Saravanan, C., and Singh, S. K. (2012) Curcumin: a potential candidate for matrix metalloproteinase inhibitors, Expert Opinion on Therapeutic Targets 16, 959-972.

11 [32] Gross, J., and Lapiere, C. M. (1962) Collagenolytic activity in amphibian tissues: a tissue culture assay, Proc Natl Acad Sci U S A 48, 1014-1022. [33] Sekhon, B. S. (2010) Matrix metalloproteinases - an overview, Research and Reports in Biology 2010, 1-20. [34] Tallant, C., Marrero, A., and Gomis-Ruth, F. X. (2010) Matrix metalloproteinases: fold and function of their catalytic domains, Biochim. et Biophy. Atca 1803, 20-28. [35] Yadav, L., Puri, N., Rastogi, V., Satpute, P., Ahmad, R., and Kaur, G. (2014) Matrix metalloproteinases and cancer - roles in threat and therapy, Asian Pac J Cancer Prev 15, 1085-1091. [36] Yang, H., Makaroff, K., Paz, N., Aitha, M., Crowder, M. W., and Tierney, D. L. (2015) Metal Ion Dependence of the Matrix Metalloproteinase-1 Mechanism, Biochemistry 54, 3631-3639. [37] Jacobsen, F. E., Breece, R. M., Myers, W. K., Tierney, D. L., and Cohen, S. M. (2006) Model complexes of cobalt-substituted matrix metalloproteinases: Tools for inhibitor design, Inorganic Chemistry 45, 7306-7315. [38] Woessner, J. F. (1991) Matrix metalloproteinases and their inhibitors in connective-tissue remodeling, Faseb Journal 5, 2145-2154. [39] Yiotakis, A., and Dive, V. (2008) Synthetic active site-directed inhibitors of metzincins: achievement and perspectives, Mol Aspects Med 29, 329-338. [40] Thomas, G. T., Lewis, M. P., and Speight, P. M. (1999) Matrix metalloproteinases and oral cancer, Oral oncology 35, 227-233. [41] Visse, R., and Nagase, H. (2003) Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry, Circ Res 92, 827-839. [42] Gomis-Ruth, F. X. (2009) Catalytic domain architecture of metzincin metalloproteases, J Biol Chem 284, 15353-15357. [43] Dorman, G., Cseh, S., Hajdu, I., Barna, L., Konya, D., Kupai, K., Kovacs, L., and Ferdinandy, P. (2010) Matrix metalloproteinase inhibitors: a critical appraisal of design principles and proposed therapeutic utility, Drugs 70, 949-964. [44] Nagase, H., and Brew, K. (2003) Designing TIMP (tissue inhibitor of metalloproteinases) variants that are selective metalloproteinase inhibitors, Biochem Soc Symp, 201-212. [45] Nagase, H., Visse, R., and Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs, Cardiovascular Research 69, 562-573. [46] Van Wart, H. E., and Birkedal-Hansen, H. (1990) The cysteine switch: a principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metalloproteinase gene family, Proc Natl Acad Sci U S A 87, 5578-5582. [47] Jacobsen, F. E., Breece, R. M., Myers, W. K., Tierney, D. L., and Cohen, S. M. (2006) Model complexes of cobalt-substituted matrix metalloproteinases: Tools for inhibitor design, Inorganic Chemistry 45, 7306-7315.

12 [48] Maskos, K. (2005) Crystal structures of MMPs in complex with physiological and pharmacological inhibitors, Biochimie 87, 249-263. [49] Windsor, L. J., Bodden, M. K., Birkedal-Hansen, B., Engler, J. A., and Birkedal-Hansen, H. (1994) Mutational analysis of residues in and around the active site of human fibroblast- type , J Biol Chem 269, 26201-26207. [50] Bode, W., Fernandez-Catalan, C., Tschesche, H., Grams, F., Nagase, H., and Maskos, K. (1999) Structural properties of matrix metalloproteinases, Cell Mol Life Sci 55, 639-652. [51] Williams, D. H., and Murray, E. J. (1994) Specific amino acid substitutions in human collagenase cause decreased autoproteolysis and reveal a requirement for a second zinc atom for catalytic activity, FEBS Lett 354, 267-270. [52] Pirard, B. (2007) Insight into the structural determinants for selective inhibition of matrix metalloproteinases, Drug Discov Today 12, 640-646. [53] Lukacova, V., Zhang, Y., Mackov, M., Baricic, P., Raha, S., Calvo, J. A., and Balaz, S. (2004) Similarity of binding sites of human matrix metalloproteinases, J Biol Chem 279, 14194-14200. [54] Pelmenschikov, V., and Siegbahn, P. E. (2002) Catalytic mechanism of matrix metalloproteinases: two-layered ONIOM study, Inorg Chem 41, 5659-5666. [55] Bertini, I., Calderone, V., Fragai, M., Luchinat, C., Maletta, M., and Yeo, K. J. (2006) Snapshots of the reaction mechanism of matrix metalloproteinases, Angewandte Chemie 45, 7952-7955. [56] Browner, M. F., Smith, W. W., and Castelhano, A. L. (1995) Matrilysin-inhibitor complexes: common themes among metalloproteases, Biochemistry 34, 6602-6610. [57] Manzetti, S., McCulloch, D. R., Herington, A. C., and van der Spoel, D. (2003) Modeling of enzyme-substrate complexes for the metalloproteases MMP-3, ADAM-9 and ADAM-10, Journal of computer-aided molecular design 17, 551-565. [58] Shapiro, S. D., Kobayashi, D. K., and Ley, T. J. (1993) Cloning and characterization of a unique elastolytic metalloproteinase produced by human alveolar macrophages, J Biol Chem 268, 23824-23829. [59] Kessenbrock, K., Plaks, V., and Werb, Z. (2010) Matrix metalloproteinases: regulators of the tumor microenvironment, Cell 141, 52-67. [60] Gross, J., and Lapiere, C. M. (1962) Collagenlytic activity in amphibian tissue: a tissue culture assay, Proc Natl Acad Sci U S A 48, 1014-1022. [61] Pellicoro, A., Aucott, R. L., Ramachandran, P., Robson, A. J., Fallowfield, J. A., Snowdon, V. K., Hartland, S. N., Vernon, M., Duffield, J. S., Benyon, R. C., Forbes, S. J., and Iredale, J. P. (2012) Elastin accumulation is regulated at the level of degradation by macrophage metalloelastase (MMP-12) during experimental liver fibrosis, Hepatology 55, 1965-1975.

13 [62] Stracke, J. O., Hutton, M., Stewart, M., Pendas, A. M., Smith, B., Lopez-Otin, C., Murphy, G., and Knauper, V. (2000) Biochemical characterization of the catalytic domain of human matrix metalloproteinase 19. Evidence for a role as a potent basement membrane degrading enzyme, J Biol Chem 275, 14809-14816. [63] Shapiro, S. D., Griffin, G. L., Gilbert, D. J., Jenkins, N. A., Copeland, N. G., Welgus, H. G., Senior, R. M., and Ley, T. J. (1992) Molecular cloning, chromosomal localization, and bacterial expression of a murine macrophage metalloelastase, J Biol Chem 267, 4664- 4671. [64] Mecham, R. P., Broekelmann, T. J., Fliszar, C. J., Shapiro, S. D., Welgus, H. G., and Senior, R. M. (1997) Elastin degradation by matrix metalloproteinases. Cleavage site specificity and mechanisms of elastolysis, J Biol Chem 272, 18071-18076. [65] Fulcher, Y. G., and Van Doren, S. R. (2011) Remote Exosites of the Catalytic Domain of Matrix Metalloproteinase-12 Enhanse Elastin Degradation, Biochemistry 50, 9488–9499. [66] Parkar, A. A., Stow, M. D., Smith, K., Panicker, A. K., Guilloteau, J. P., Jupp, R., and Crowe, S. J. (2000) Large-scale expression, refolding, and purification of the catalytic domain of human macrophage metalloelastase (MMP-12) in Escherichia coli, Protein Expr Purif 20, 152-161. [67] Jacobsen, J. A., Major Jourden, J. L., Miller, M. T., and Cohen, S. M. (2010) To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition, Biochim. Biophys. Acta 1803, 72-94. [68] Yang, W., Arii, S., Gorrin-Rivas, M. J., Mori, A., Onodera, H., and Imamura, M. (2001) Human macrophage metalloelastase gene expression in colorectal carcinoma and its clinicopathologic significance, Cancer 91, 1277-1283. [69] Peterson, J. T., Li, H., Dillon, L., and Bryant, J. W. (2000) Evolution of matrix metalloprotease and tissue inhibitor expression during heart failure progression in the infarcted rat, Cardiovascular research 46, 307-315. [70] Garcia, R. A., Pantazatos, D. P., Gessner, C. R., Go, K. V., Woods, V. L., Jr., and Villarreal, F. J. (2005) Molecular interactions between matrilysin and the matrix metalloproteinase inhibitor doxycycline investigated by deuterium exchange mass spectrometry, Mol Pharmacol 67, 1128-1136. [71] Brown, S., Meroueh, S. O., Fridman, R., and Mobashery, S. (2004) Quest for selectivity in inhibition of matrix metalloproteinases, Curr Top Med Chem 4, 1227-1238. [72] Martinot, M. (2007) [Microbiological and pharmacological data useful for the treatment of Lyme disease. Treatment and follow up of early Lyme disease (erythema migrans)], Med Mal Infect 37, 394-409. [73] Sapadin, A. N., and Fleischmajer, R. (2006) Tetracyclines: nonantibiotic properties and their clinical implications, J Am Acad Dermatol 54, 258-265.

14 [74] Johnson, A. R., Pavlovsky, A. G., Ortwine, D. F., Prior, F., Man, C. F., Bornemeier, D. A., Banotai, C. A., Mueller, W. T., McConnell, P., Yan, C., Baragi, V., Lesch, C., Roark, W. H., Wilson, M., Datta, K., Guzman, R., Han, H. K., and Dyer, R. D. (2007) Discovery and characterization of a novel inhibitor of matrix metalloprotease-13 that reduces cartilage damage in vivo without joint fibroplasia side effects, J Biol Chem 282, 27781- 27791. [75] Dorman, G., Kocsis-Szommer, K., Spadoni, C., and Ferdinandy, P. (2007) MMP inhibitors in cardiac diseases: an update, Recent Pat Cardiovasc Drug Discov 2, 186-194. [76] Jacobsen, J. A., Fullagar, J. L., Miller, M. T., and Cohen, S. M. (2011) Identifying chelators for metalloprotein inhibitors using a fragment-based approach. - Abstract - Europe PMC, Journal of Medicinal Chemistry 54, 591-602. [77] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., de Oliveira, C. A. F., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) 'Unconcentional' coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-5406. [78] Jacobsen, J. A., Fullagar, J. L., Miller, M. T., and Cohen, S. M. (2011) Identifying chelators for metalloprotein inhibitors using a fragment-based approach, J Med Chem 54, 591-602. [79] Puerta, D. T., and Cohen, S. M. (2003) Examination of Novel Zinc-Binding Groups for Use in Matrix Metalloproteinase Inhibitors, Inorganic Chemistry 42, 3423-3430. [80] Puerta, D. T., Lewis, J. A., and Cohen, S. M. (2004) New Beginnings for Matrix Metalloproteinase Inhibitors: Identification of High-Affinity Zinc-Binding Groups, Journal of the American Chemical Society 126, 8388–8389. [81] He, H., Puerta, D. T., Cohen, S. M., and Rodgers, K. R. (2005) Structural and Spectroscopic Study of Reactions between Chelating Zinc-Binding Groups and Mimics of the Matrix Metalloproteinase and Disintegrin Metalloprotease Catalytic Sites: The Coordination Chemistry of Metalloprotease Inhibition, Inorganic Chemistry 44. [82] Puerta, D. T., Griffin, M. O., Lewis, J. A., Romero-Perez, D., Garcia, R., Villarreal, F. J., and Cohen, S. M. (2006) Heterocyclic zinc-binding groups for use in next-generation matrix metalloproteinase inhibitors: potency, toxicity, and reactivity, Journal of Biological Inorganic Chemistry 11, 131-138. [83] Dick, B. L., Patel, A., McCammon, J. A., and Cohen, S. M. (2017) Effect of donor atom identity on metal-binding pharmacophore coordination, J Biol Inorg Chem. [84] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., Oliveira, C. A. F. d., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) ‘Unconventional’ coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-54056. [85] Meng, F., Yang, H., Jack, C., Zhang, H., Moller, A., Spivey, D., Page, R. C., Tierney, D. L., and Crowder, M. W. (2016) Biochemical characterization and zinc binding group (ZBGs)

15 inhibition studies on the catalytic domain of MMP7 (cdMMP7), J Inorg Biochem 165, 7- 17. [86] Parkin, G. (2004) Synthetic Analogues Relevant to the Structure and Function of Zinc Enzymes, Chemical Reviews 104, 699−768. [87] Maret, W., and Vallee, B. L. (1993) Cobalt as probe and label of proteins, Methods Enzymol 226, 52-71. [88] Bencini, A., Bertini, I., Canti, G., Gatteschi, D., and Luchinat, C. (1981) The epr spectra of the inhibitor derivatives of cobalt carbonic anhydrase, J Inorg Biochem 14, 81-93. [89] Bertini, I., Jonsson, B. H., Luchinat, C., Pierattelli, R., and Vila, A. J. (1994) Strategies of signal assignments in paramagnetic metalloproteins. An NMR investigation of the thiocyanate adduct of the cobalt (II)-substituted human carbonic anhydrase II, J Magn Reson B 104, 230-239. [90] Bertini, I., Canti, G., Luchinat, C., and Scozzafava, A. (2002) Characterization of cobalt(II) bovine carbonic anhydrase and of its derivatives, Journal of the American Chemical Society 100, 4873-4877. [91] Banci, L., Bertini, I., Cramaro, F., Del Conte, R., and Viezzoli, M. S. (2003) Solution structure of Apo Cu,Zn superoxide dismutase: role of metal ions in protein folding, Biochemistry 42, 9543-9553. [92] Kremer-Aach, A., Klaui, W., Bell, R., Strerath, A., Wunderlich, H., and Mootz, D. (1997) Cobalt as a probe for zinc in metalloenzyme model compounds? A comparison of spectroscopic features and coordination geometry of four- and five-coordinate complexes. Crystal and molecular structures of [Co(eta(3)-Tp(Ph))(eta(2)-Tp(Ph))], [(eta(3)-Tp(Ph))Zn(anthranilate)], and [(eta(3)-Tp(Ph))M(eta(2)-acac)] (Tp(Ph) = Hydrotris(3-phenylpyrazol-1-yl)borate, acac = Pentane-2,4-dionate, and M = Zn, Co), Inorg Chem 36, 1552-1563. [93] Meng, F., Yang, H., Aitha, M., George, S., Tierney, D., and Crowder, M. (2016) Biochemical and spectroscopic characterization of the catalytic domain of MMP16 (cdMMP16) | SpringerLink, journal of biological inorganic chemistry 21, 523-535. [94] Yang, H., Makaroff, K., Paz, N., Aitha, M., Crowder, M. W., and Tierney, D. L. (2015) Metal ion dependence of the matrix metalloproteinase-1 mechanism, Biochemistry 54, 3631-3639. [95] Agrawal, A., Johnson, S. L., Jacobsen, J. A., Miller, M. T., Chen, L. H., Pellecchia, M., and Cohen, S. M. (2010) Chelator fragment libraries for targeting metalloproteinases, ChemMedChem 5, 195-199. [96] Keilin, D., and Mann, T. (1940) Carbonic anhydrase. Purification and nature of the enzyme, Biochem J 34, 1163-1176. [97] Armstrong, J. M., Myers, D. V., Verpoorte, J. A., and Edsall, J. T. (1966) Purification and properties of human erthrocyte carbonic anhydrase, J. Biol. Chem. 241, 5137-5149.

16 [98] Bertini, I., Canti, G., Luchinat, C., and Scozzafava, A. (1978) Characterization of cobalt(II) bovine carbonic anhydrase and of its derivatives, Journal of the American Chemical Society 100, 4873-4877. [99] Bertini, I., Luchinat, C., and Scozzafava, A. (1980) The acid base equilibria of carbonic anhydrase, Inorg. Chim. Acta 46, 85-89. [100] Bracey, M. H., Christiansen, J., Tovar, P., Cramer, S. P., and Bartlett, S. G. (1994) Spinach carbonic anhydrase: investigation of the zinc-binding ligands by site-directed mutagenesis, Elemental Analysis and EXAFS, Biochemistry 33, 13126-13131. [101] Carlsson, U., Henderson, L. E., and Linkskog, S. (1973) Denaturation and reactivation of human carbonic anhydrases in guandine hydrochloride and urea, Biochimica et Biophysica Acta 310, 376-387. [102] Cheng, X., Chen, R., Bruce, J. E., Schwartz, B. L., Anderson, G. A., Hofstadler, S. A., Gale, D. C., and Smith, R. D. (1995) Using electrospray ionization FTICR mass spectrometry to study competitive binding of inhibitors to carbonic anhydrase, Journal of the American Chemical Society 117, 8859-8860. [103] Denisov, V. P., Jonsson, B. H., and Halle, B. (1999) Dynamics of functional water in the active site of native carbonic anhydrase from 17O magnetic relaxation dispersion, Journal of the American Chemical Society 121, 2327-2328. [104] Garmer, D. R., and Krauss, M. (1993) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, Journal of the American Chemical Society 115, 10247-10257. [105] Hakansson, K., and Liljas, A. (1994) The structure of a complex between carbonic anhydrase II and a new inhibitor, trifluoromethane sulphonamide, FEBS Letters 350, 319- 322. [106] Hunt, J. A., Ahmed, M., and Fierke, C. A. (1999) Metal binding specificity in carbonic anhydrase is influenced by conserved hydrophobic core residues, Biochemistry 38, 9054- 9062. [107] Keilin, D., and Mann, T. (1940) Carbonic anhydrase. Purification and nature of the enzyme, Biochemical Journal 34, 1163-1176. [108] Kiefer, L. L., and Fierke, C. A. (1994) Functional characterization of human carbonic anhydrase II variants with altered zinc binding sites, Biochemistry 33, 15233-15240. [109] Kiefer, L. L., Paterno, S. A., and Fierke, C. A. (1995) Hydrogen bond network in the metal of carbonic anhydrase enhances zinc affinity and catalytic efficiency, Journal of the American Chemical Society 117, 6831-6837. [110] Kim, G., Selengut, J., and Levine, R. L. (2000) Carbonic anhydrase III: the phosphatase activity is extrinsic, Arch. Biochem. Biophys. 377, 334-340.

17 [111] Lesburg, C. A., Huang, C. C., Christianson, D. W., and Fierke, C. A. (1997) Histidine- carboxamide ligand substitutions in the zinc binding Site of carbonic anhydrase II alter metal coordination geometry but retain catalytic activity, Biochemistry 36, 15780-15791. [112] Lindskog, S. (1963) Effects of pH and inhibitors on some properties related to metal Binding in Bovine Carbonic Anhydrase, Journal of Biological Chemistry 238, 945-951. [113] Lindskog, S., and Malmstrom, B. G. (1962) Metal binding and catalytic activity in bovine carbonic anhydrase, Journal of Biological Chemistry 237, 1129-1137. [114] Lipton, A. S., Heck, R. W., and Ellis, P. D. (2004) Zinc solid-state NMR spectroscopy of human carbonic anhydrase: Implications for the enzymatic mechanism, Journal of the American Chemical Society 126, 4735-4739. [115] Qian, M., Earnhardt, J. N., Wadhwa, N. R., Tu, C., Laipis, P. J., and Silverman, D. N. (1999) Proton transfer to residues of basic pKa during catalysis by carbonic anhydrase, Biochimica et Biophysica Acta 1434, 1-5. [116] Qian, M., Tu, C., Earnhardt, J. N., Laipis, P. J., and Silverman, D. N. (1997) Glutamate and aspartate as proton shuttles in mutants of carbonic anhydrase, Biochemistry 36, 15758-15764. [117] Rowlett, R. S., Chance, M. R., Wirt, M. D., Sidelinger, D. E., Royal, J. R., Woodroffe, M., Wang, Y. A., Saha, R. P., and Lam, M. G. (1994) Kinetic and structural characterization of spinach carbonic anhydrase, Biochemistry 33, 13967-13976. [118] Rowlett, R. S., Gargiulo, N. J., Santoli, F. A., Jackson, J. M., and Corbett, A. H. (1991) Activation and Inhibition of Bovine Carbonic Anhydrase III by Dianions, Journal of Biological Chemistry 266, 933-941. [119] Roy, B. C., Banerjee, B. C., Swanson, M., Jia, X. G., Haldar, M. K., Mallik, S., and Srivastava, D. K. (2004) Two-prong inhibitors for human carbonic anhydrase II, J. Am. Chem. Soc. 126, 13206-13207. [120] Scozzafava, A., Menabuoni, L., Mincione, F., and Supuran, C. T. (2002) Carbonic anhydrase inhibitors. A general approach for the preparation of water-soluble sulfonamides incorporating polyamino-polycarboxylate tails and of their metal complexes possessing long-lasting, topical intraocular pressure-lowering properties, Journal of Medicinal Chemistry 45, 1466-1476. [121] Strasdeit, H. (2001) The first cadmium specific enzyme, Angew. Chem. 40, 707-709. [122] Xue, Y., Jonsson, B. H., Liljas, A., and Lindskog, S. (1994) Modification of a metal ligand in carbonic anhydrase: Crystal structure of His94-Glu suman isozyme II, FEBS Letters 352, 137-140. [123] Banci, L., Dugad, L. B., La Mar, G. N., Keating, K. A., Luchinat, C., and Pierattelli, R. (1992) 1-H nuclear magnetic resonance investigation of cobalt(II) substituted carbonic anhydrase, Biophys. J. 63, 530-543.

18 [124] Bencini, A., Bertini, I., Canti, G., Gatteschi, D., and Luchinat, C. (1981) The epr spectra of the inhibitor derivatives of cobalt carbonic anhydrase, Journal of Inorganic Biochemistry 14, 81-93. [125] Bertini, I., Jonsson, B. H., Luchinat, C., Pierattelli, R., and Vila, A. J. (1994) Strategies of signal assignments in paramagnetic metalloproteins. An NMR investigation of the thiocyanate adduct of the cobalt (II)-substituted human carbonic anhydrase II, J Magn Reson B 104, 230-239. [126] Alberti, G., Bertini, I., Luchinat, C., and Scozzafava, A. (1981) A new class of inhibitors capable of binding both the acidic and alkaline forms of carbonic anhydrase, Biochimica Et Biophysica Acta 668, 16-26. [127] Bertini, I., Canti, G., Luchinat, C., and Borghi, E. (1983) Investigation of the system copper(II) carbonic anhydrase and HCO3-/CO2, Journal of Inorganic Biochemistry 18, 221-229. [128] Bertini, I., Luchinat, C., Pierattelli, R., and Vila, A. J. (1992) The interaction of acetate and formate with cobalt carbonic anhydrase. An NMR study, European journal of biochemistry / FEBS 208, 607-615. [129] Boriack, P. A., Christianson, D. W., Kingery-Wood, J., and Whitesides, G. M. (1995) Secondary interactions significantly removed from the sulfonamide binding pocket of carbonic anhydrase II influence inhibitor binding constants, Journal of Medicinal Chemistry 38, 2286-2291. [130] Cockle, S. A., Lindskog, S., and Grell, E. (1974) Electron-paramagnetic-resonance studies on cobalt(II) carbonic anhydrase-sulphonamide complexes, The Biochemical journal 143, 703-715. [131] Cockle, S. A. (1974) Electron-paramagnetic-resonance studies on cobalt(II) carbonic anhydrase. Low-spin cyanide complexes, The Biochemical journal 137, 587-596. [132] Coleman, J. E., and Coleman, R. V. (1972) Magnetic circular dichroism of Co(II) carbonic anhydrase, The Journal of biological chemistry 247, 4718-4728. [133] Grell, E., and Bray, R. C. (1971) Electron paramagnetic resonance spectroscopy of bovine cobalt carbonic anhydrase B, Biochimica Et Biophysica Acta 236, 503-506. [134] Jacob, G. S., Brown, R. D., 3rd, and Koenig, S. H. (1978) Relaxation of solvent protons by cobalt bovine carbonic anhydrase, Biochem. Biophys. Res. Commun. 82, 203-209. [135] Moratal, J. M., Donaire, A., Salgado, J., and Martinez-Ferrer, M. J. (1990) Interaction of sulphate and chloride with cobalt(II)-carbonic anhydrase, Journal of Inorganic Biochemistry 40, 245-253. [136] Kumar, D., Kumar, M., Saravanan, C., and Singh, S. K. (2012) Curcumin: a potential candidate for matrix metalloproteinase inhibitors, Expert opinion on therapeutic targets 16, 959-972.

19 1.9 Figures.

Figure 1-1: Domain structures of MMP classes. Green block – signal peptide; Blue block – propeptide domain, cleaved to yield active protein; Red line – cysteine switch sequence of PRCG(V/N)PD, cysteine residue coordinates to the catalytic zinc (gray sphere) when propeptide domain is present; Black stripe block – furin-recognition sequence consisting of 10-14 amino acid repeats of RX(K/R)R; Purple block – catalytic domain containing two metal binding sites, one structural (not shown) and one with catalytic function (gray sphere); F-II – fibronectin type II domain consisting of three repeats of 58 amino acids spliced between the catalytic domain, MMP-9 also possesses a collagen domain at the end of the catalytic section; Yellow block – linker region; Pink block – hemopexin domain; Gray block - glycosylphosphatidylinositol membrane-anchoring domain and trans-membrane domain.

20

Figure 1-2: Catalytic Site of MMP-1254. Ribbon represents the protein backbone surrounding the active site. Coordinating ligands are labeled and numbered using the full sequence.

Figure 1-3: Structures of MMP-12 domains. (Left) X-ray Structure of MMP-12 Full Length Enzyme (PDB code 3AB0)130. (Right) X-ray structure of MMP-12 catalytic domain (PDB code 1OS2)131. Acetohydroxamic acid is coordinated in the active site of both structures.

21

Figure 1-4: 3-Dimensional structure of MMP843. S1′–S3′ and S1–S3 pockets are labeled in relation to the active site. The active site is indicated with Zn2+ ion.

22 A B

D C

Figure 1-5: Proposed catalytic mechanisms for MMPs. A: Pelmenschikov and Siegbahn model. This model utilizes thermolysin and carboxypeptidase mechanism information42, 132to propose an active site Zn(II) is 4-coordinate in the resting form and 5-coordinate in all catalytic intermediates. B: Bertini model. Crystalline “snapshots” of MMP-12 during catalysis were used to propose a 6-coordinate Zn(II) in the resting form. Coordination of three water molecules are lost during catalysis to yield a 5-coordinate active site Zn(II) intermediate and 4-coordinate Zn(II) before the product is released.55 C: Browner. This model utilizes crystal structures of matrilysin-inhibitor complexes to propose an acid/ base-catalysis mechanism involving a glutamate to undergo a nucleophilic attack of the scissile bond. Details of the active site coordination is not investigated though a 5-coordinate Zn(II) is employed.56 D: Manzetti model. MMP-3 modeling studies were used to propose a closed-off, water displaced active site and the formation of a 4- and 5- coordinate Zn(II) during catalysis57. Glutamate acts as a schiff base to displace a histidine during the reaction.

23

Figure 1-6: 3-Dimensional comparison of MMP to CA. Left – MMP-3, orange sphere indicates catalytic Zn(II) ion. Histidine side chains that coordinate the metal are present in green and blue. Bowl-shaped active site is shown in a front-on manner. Right – carbonic anhydrase, orange sphere indicates Zn(II) ion present in the active site. Histidine side chains that coordinate the metal are present in green and blue. Active site is largely surface accessible.

24 Chapter 2: Evaluation of zinc binding groups (ZBGs) thiomaltol, thiopyromechonic acid, and allothiomaltol using carbonic anhydrase as a model enzyme

25 Unconventional Coordination Chemistry of Metal Binding Fragments and Donor Group pKa Whitney R. Craig,† Amy R. Marts,† Tessa W. Baker,† David P. Martin,‡ Garrett C. Reed,† Anthony A. Forchonie,† Daniel T. DeGenova,† Rithvik Venna,† Rahil H. Patil,† Ania Plonski,† Robert McCarrick,† Michael W. Crowder,† Seth M. Cohen,‡* and David L. Tierney†*

† Department of Chemistry and Biochemistry, Miami University, 651 East High Street, Oxford, Ohio 45056, United States ‡ Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093, United States Corresponding Author. *E-mail: [email protected]. Phone:513- 529-8234.

Funding Source Statement. This work was supported by the National Institutes of Health (GM093987 to MWC and DLT, and P30-EB-009998 to the Center for Synchrotron Biosciences from the National Institute of Biomedical Imaging and Bioengineering, which supports beamline X3B at the National Synchrotron Light Source) and the National Science Foundation (CHE- 1151658 AMD CHE-1509285 to MWC and DLT, as well as CHE-1152755 to DLT).

Author contributions: WRC and DLT contributed writing of manuscript; WRC, ARM, and DLT contributed to the data analysis. WRC, TWB, AAF, DTD, RP, RV, AP, TB, NM, MKG, WM and JC contributed to the preparation of protein. WRC and TWB contributed to the UV-visible collection. WRC, ARM, and TWB conducted all nuclear magnetic resonance spectra. DLT and ARM contributed the preparation and acquisition of all electron paramagnetic resonance.

Notes. The authors declare no competing financial interest.

Abbreviations. CA, carbonic anhydrase; ZBG, zinc binding group; TM, thiomaltol; TPHA, thiopyromechonic acid; ATM, allothiomaltol; NMR, nuclear magnetic resonance; EPR, electron paramagnetic resonance; EXAFS, extended X-ray absorption fine structure.

26 2.1 Abstract. The combination of XAS, UV-vis, NMR, and EPR was used to examine the binding of a series of a-hydroxythiones to CoCA. All three appear to bind preferentially in their neutral, protonated forms. Two of the three clearly bind in a monodentate fashion, through the thione sulfur alone. Thiomaltol (TM) appears to show some orientational preference, based on the NMR, while it appears thiopyromeconic acid (TPMA) retains rotational freedom. In contrast, allothiomaltol (ATM) forms a five-coordinate complex via bidentate coordination. Based on optical titrations, we speculate that this may be due to the lower initial pKa of ATM (8.3) relative to TM (9.0) and

TPMA (9.5). Binding through the thione is shown to reduce the hydroxyl pKa by ~0.7 pH units on metal binding, bringing only ATM’s pKa close to the pH of the experiment, facilitating deprotonation and subsequent coordination of the hydroxyl. The data predict the presence of a solvent-exchangeable proton on TM and TPMA, and Q-band 2-pulse ESEEM experiments on CoCA+TM suggest the proton is present. ESE-detected EPR also showed a surprising frequency dependence, giving only a subset of the expected resonances at X-band.

2.2 Introduction. Metalloproteins comprise a diverse class of proteins, embodying an estimated fifty percent of proteins1. Due to both the widespread biological availability and pharmacological relevance, metalloproteins remain an attractive therapeutic target. Multiple efforts have been directed to inhibit many metalloproteins; however, many inhibitors exhibit poor selectivity and toxicity2, 3. These zinc-binding groups (ZBGs) are small molecules that bind the metal ion active site of the enzyme2, 4. ZBG design has little variety, consisting of thiols, carboxylic acids, phosphates, and hydroxamic acids5. This lack of diversity leads to a desire for new ZBGs with a different backbone moiety.

Structure-aided rational design has been utilized for next generation ZBGs to decrease toxicity, improve potency, and improve selectivity of metalloprotein-targeted inhibitors to minimize off- target binding5, 6. The documented instability and off-target binding of hydroxamic acids deepen the need to explore other metal binding motifs that increase stability. Small molecule models that mimic metalloprotein active sites have been employed to aid in the understanding of ZBG binding motifs and used to predict ligand binding in the metalloprotein active site7-9. These

27 models have provided screening information but cannot mimic the restrictions of a metalloprotein active site that arise from large amino acid side chains 6, 10, 11. This steric side chain interference gives rise to changes in the metal-ligand coordination and relatively weak metal coordination, further complicating computational efforts to predict the binding mode and binding strength of ZBGs5, 12, 13. Further, models cannot mimic the hydrophobic binding pockets that affect pKa values of potential metal binding amino acid side chains. In addition to these limitations, literature describing the orientation and denticity of ZBGs can change based on minor changes in the backbone structure of the inhibitor for which modeling does not fully account5. Due to the limitations of many structural models, coupled with the difficulty of purifying large biological enzymes, a utilization of smaller, more readily accessible and more commercially available proteins has been used to mimic biological proteins and probe off-target binding.

Carbonic anhydrase (CA) is utilized as an intermediate step between model scaffold mimicking systems and larger metalloproteins to elucidate the binding mode and potency of ZBGs. CA is widely used as a model system due to its ubiquitous nature, stability, solubility, rigid structure, 2+ 14-16 ease of crystallization, and His3-Zn binding motif that is relatively open to solvent. Small molecule inhibitors have been developed and approved for clinical investigation that possess binding of Zn2+-dependent enzymes, such as CA, through similar functional group motifs.17 These motifs predominately include carboxylates and hydroxamic acids; however. recent efforts have been made to investigate alternate motifs. 18, 19 Heterocyclic chelators based on a hydroxythiopyrone scaffold have been investigated with human carbonic anhydrase II as well as inorganic model complexes in crystallographic studies and computational model studies.6, 15

In this study, a new generation of ZBGs based on hydroxythiopyrones were investigated. These specific hydroxythiopyrones are 5-hydroxy-2-methyl-4H-pyran-4-thione (allothiomaltol, ATM), 3-hydroxy-2-methyl-4H-pyran-4-thione (thiomaltol, THM), and 3-hydroxy-4H-pyran-4-thione (thiopyromeconic acid, TPMA), differing only by the presence and/or position of a methyl group. These three ZBGs have been investigated in the model complex tris(pyrazolyl)borate (Tp) and found to bind in the same manner.11 However, these ZBGs coordinate with the Zn2+ ion in CA differently than predicted.6 These differences contribute to the lack of implementation of

28 motif diversity in metalloprotein drug design. The results presented within demonstrate the unexpected binding modes of these hydroxythiopyrones and the need to adapt current drug design methods.

2.3 Experimental procedures. 2.3.1 Materials. Commercially available, lyophilized CA-II from bovine erythrocytes (Sigma) was used for all of the studies presented here, as purchased. All reagents were used, as purchased, without further purification. All buffer solutions were made with 18.1 MΩ water from a Barnstead NANOpure system. Metal binding groups (MBGs) were prepared according to published procedures20. All MBG complexes for spectroscopy were prepared with 3 molar equivalents of the MBG, at the stated concentrations.

2.3.2 Methods. 2.3.2.1 Preparation of Co(II)-substituted CA-II. CA-II was dissolved at ~ 10 mM in Chelex- treated (Sigma Chelex® 100 sodium form) 50 mM phosphate, pH 7.5, (referred to as “phosphate buffer” from here out). The concentrated protein was dialyzed for 48 hours total against 4 x 2 L of phosphate buffer containing a 10-fold excess of 2,6-pyridinedicarboxylic acid. The resulting apo-CA-II contained less than 0.05 molar equivalents of Zn(II), as determined from ICP-MS.

Five molar eq of Co(II) (10 mM CoCl2·6H2O; 99.999 %, Strem Chemicals) was added, and the mixture was incubated at 4 °C for 30 minutes. To remove excess metal, the protein mixture was treated with 30% (v/v) of Chelex-treated phosphate buffer for 15 minutes. Co(II)-substituted CA (CoCA) was then isolated from the Chelex using a small volume gravity flow column.

2.3.2.2 X-ray absorption spectroscopy. Samples for EXAFS (~ 1.5 mM) were prepared with 20% (v/v) glycerol as a glassing agent. Samples were loaded into Lucite cuvettes with 6 µm polypropylene windows, and frozen in liquid nitrogen. Data was collected at the National Synchrotron Light Source (NSLS), beamline X3B, using a Si(111) double-crystal monochromator. Harmonic rejection was accomplished using a Ni focusing mirror. Fluorescence excitation spectra were measured with a 31-element solid-state Ge detector array; samples were held at approximately 15 K using a Displex cryostat. EXAFS data collection and reduction was performed according to published procedures21. Duplicate data sets were collected for each

29 sample, eight scans per sample. As the individual data sets gave similar results they were averaged; it is these average data that are presented here. Conversion from energy to k-space used E0 = 7735 eV for the Co K-edge. EXAFS data were fitted using the nonlinear least-squares engine of IFEFFIT (distributed with SixPack, available free of charge from http://www- ssrl.slac.stanford.edu/~swebb/sixpack.html). Theoretical amplitude and phase functions were calculated with FEFF v. 80022.

2.3.2.3 UV-visible spectroscopy. Optical spectra were obtained using a Perkin-Elmer Lambda 750 spectrophotometer. Quartz cuvettes (Perkin-Elmer; 1 cm path length) were used; the spectrometer was blanked against Chelex-treated 50 mM phosphate buffer, pH 7.5. MBG pKas were determined by monitoring the optical spectrum of a 78 µM solution as a function of pH. Binding assays were conducted by the addition 0.1-3.0 molar equivalents of selected the MBG.

2.3.2.4 Nuclear magnetic resonance spectroscopy. NMR samples were prepared using 50 mM phosphate buffer at pD 7.5, made with 99.9% D2O (Cambridge Isotopes) and concentrated by centrifugation (Millipore Amicon® Ultra 15). Spectra were collected on ~1.5 mM samples with a Bruker ASX 200 NMR spectrometer (νH = 200.13 MHz) at approximately 290K. Water suppression was accomplished using a long, low-power pulse (100-150 ms, ~1 W) on the water signal, before moving the transmitter frequency to the region of interest for data collection23. Spectra were acquired using a 3 µs pulse at full power and referenced to the residual water resonance at 4.7 ppm. Titrations were conducted by the addition 0.5, 1, 2, and 3 molar equivalents of the selected MBG. Each spectrum consists of 4000 scans of 16k data points over a 75 kHz (375 ppm) window. All FIDs were apodized using a simple exponential that incorporated an additional 5 Hz line width.

2.3.2.5 Electron paramagnetic resonance spectroscopy. Low temperature X-band CW EPR spectra were collected using a Bruker EMX EPR spectrometer, equipped with an ER4116DM dual-mode cavity and an Oxford Instruments ESR900 liquid helium flow cryostat. Samples (~ 1.5 mM CoCA) contained ~ 20% (v/v) glycerol as a glassing agent. The spectra presented here were recorded at 9.61 GHz using the following parameters: 20 µW microwave power, 10 G

30 magnetic field modulation (100 kHz); time constant/conversion time = 82 ms; receiver gain = 1 x 5 10 , 16 scans.

2.4 Results and Discussion. In previous work on this set of MBGs interacting with native Zn(II)-containing CA, we used a combination of X-ray diffraction and X-ray absorption spectroscopy (XAS) to show that all three MBGs bind to the active site metal ion through the thione sulfur atom, and only allothiomaltol binds in a bidentate fashion (see Figure 2-1)20. This binding mode is in contrast to model studies, which showed bidentate coordination of all to a metal-trispyrazylborate scaffold and was attributed to steric interactions with the 2-methyl group of thiomaltol. Our interests here were to examine this coordination more closely, using the Co(II)-substituted enzyme. We begin with XAS of CoCA and its complexes with TM, TPMA, and ATM, as the XAS provides the only direct comparison between the Zn(II) and Co(II) enzymes.

2.4.1 X-Ray Absorption Spectroscopy. The XANES region of the spectra is shown in Figure 2-2A. The MBG complexes show a clear shift to lower energy (~ 0.5 eV), consistent with the incorporation of a soft donor in the metal ion’s coordination sphere, and greatest for the ATM complex. This shift is also reflected in the 1s63d transitions shown in Figure 2-2B, where the center of gravity of the transition shifts nearly 1 eV to lower energy. Integration of the 1s63d transitions shows a clear shift to lower intensity on addition of an MBG, from 20.3 for CoCA to 12.1-16.1 for the MBG complexes. There is a much greater loss of 1s63d intensity to higher energy in the MBG complexes relative to the resting enzyme, suggesting much of the change can be attributed to the loss of an available higher-energy bound-state MO.

The EXAFS data also support the binding modes indicated in Fig. 2-1. A comparison of the EXAFS FTs is presented in Figure 2-2C, and the best fit results are summarized in Table 2-1. Detailed fitting results are presented in Supporting Information, Figures 2-S1 - 2-S4 and Tables 2-S1 - 2-S4. The CoCA data are well fited with the expected model of 4 low-Z (N/O) donors, including 3 histidine ligands (Fig. 2-S1 and Table 2-S1). Addition of TM leads to a broadening of the first shell along with the appearance of a shoulder to high-R on the first-shell

31 peak (consistent with addition of a sulfur ligand), and substantial diminution of the outer shell scattering. The curve fitting results for the TM complex show a nearly 50% reduction in fit residual on inclusion of a Co-S interaction in the first shell, although the overall coordination number is poorly determined between (N/O)3S vs. (N/O)4S (Fig. 2-S2 and Table 2-S2). Fits to the outer shell scattering were less satisfactory, most likely due to destructive interference from Co-TM multiple scattering. This is in contrast to the EXAFS of TPMA complex, which shows the same first shell effects, but nearly identical outer shell scattering. The curve fits support sulfur coordination, with an accompanying 58 % reduction in fit residual, while the coordination number is again poorly determined between (N/O)3S vs. (N/O)4S (Fig. 2-S3 and Table 2-S3).

The ATM complex, which showed the smallest 1s63d transition (Table 2-1), also showed the most dramatically altered first shell peak, with addition of a shoulder to high R, and broadening to lower R, suggestive of a greater variation in metal-N/O bond lengths. The high-R component is consistent with S coordination, based on a near 50 % reduction in fit residual (compare fits S4- 2 and S4-5), but attempts to include distinct Co-N and Co-O shells did not significantly improve the fits. However, the ATM complex presents the only data set that shows a clear minimum at 5- coordination (compare fits S4-1 and S4-2 to S4-3 and S4-4 to S4-5). Multiple scattering fits to these data were again unsatisfying, likely due to competing Co-ATM multiple-scattering interactions, which should be more significant in a bidentate MBG complex.

2.4.2 UV-visible spectroscopy. Optical titrations were performed to examine MBG binding to CoCA. The conjugated thione leads to strong p-p* tranistions in the visible, between 350 and 400 nm (see Figure 2-3). All three MBG spectra are strongly pH dependent, allowing for determination of the hydroxyl pKa for each molecule. For TM and ATM, bands for the protonated (low pH, red spectra in Fig. 2-3), and unprotonated (high pH, blue spectra) forms are well resolved, whereas they overlap for TPMA. TPMA shows lower energy transitions at 415 (OH) and 476 (O-) nm that proved difficult to analyze. However, based on the spectra in Fig. 2-3, the hydroxyl pKas rank from TPMA highest at ~ 9.5, to TM at ~ 9.0, and ATM lowest at ~ 8.3. All three pKas are above physiological pH, indicating the protonated forms should predominate for all, though a significant amount of the unprotonated form should be present for ATM.

32

After establishing the MBG pKas and the wavelengths characteristic to their protonation state, CoCA was titrated with each MBG in increasing increments up to 3 molar equivalents. Figure 2- 4 shows the full titrations, which show that both protonated and unprotonated MBGs accumulate in the solution. However, a cursory examination of the ligand-field bands (insets to Fig. 2-4), shows that the optical titrations are consistent with buildup of the species identified by XAS. That is, both TM and TPMA, as they accumulate in solution, begin to shift the overall spectrum upwards, but they do not affect the general shape of the ligand field bands, particularly for TPMA. This trend suggests that binding of TM and TPMA must preserve the original 4- coordinate structure of resting CoCA. Meanwhile, addition of a single molar eq of ATM very quickly abolishes the ligand field bands (only the 515 nm band remains discernible), consistent with an increase in the Co(II) ion’s coordination number.

To explore the equilibria involved, we examined the buildup of the various species. Plots of change in of absorbance (DA) vs. concentration of the MBG are shown in Figure 2-5. The simplest curve is seen for TM (Fig. 2-5A). The TM data are well fited with the Michaelis-

Menten (M-M) equation, across the span of the data. Fits to DA356 (protonated) and DA386

(unprotonated) give similar binding constants of 350 and 480 µM, respectively, compared to a Ki of 1.4 mM for TM with ZnCA20. In contrast, the full TPMA titration can be fited, within error, with straight lines (dashed lines in Fig. 2-5B). However, closer inspection of the early points in the titration, at less than 1 molar eq of MBG, show the expected influence of metal binding, and the data can be fit at 341 nm (which is present for both protonated and unprotonated TPMA and would therefore be an overall binding constant, at 198 µM) and 476 nm (unprotonated, 320 µM). The fits to the early titration suggest potentially tighter binding of TPMA than TM.

Surprisingly, CoCA appears to show a preference for the protonated form of ATM. The fit lines in Fig. 2-5C are both simple M-M fits, but the fit to DA354 (protonated form) includes an x-offset to allow for a lag in its appearance. The lag of ~ 9 µM ATM means the protonated form does not appear in solution up to that concentration. Again, we examined the early points in the titration (< 1 eq MBG) (Fig. 2-5D). The M-M fits (to the first four data points) shown as solid lines indicate fairly tight binding of ATM, compared to the fits from part C that are superimposed as

33 dashed lines. Comparison of the two red lines makes the lag phase discussed above more apparent, supporting the suggestion that CoCA shows a preference for the protonated form of ATM.

2.4.3 Nuclear magnetic resonance spectroscopy. Proton NMR titrations are shown in Figure 2-6. The spectrum of resting CoCA is characterized by a pair of intense 2H-exchangeable resonances at 62 and 52 ppm, corresponding to the NH protons of the three coordinated histidines, along with broad resonances at 40 and 93 ppm, and a number of sharp lines between 20 and -10 ppm from secondary interactions within the active site. At 0.5 eq of TM (Fig. 2-6A), the His resonance at 52 ppm nearly disappears and new resonances at 68, 79, and 102 ppm appear. At 1 eq, the line at 52 ppm is fully attenuated, and further additions do not lead to further changes, showing that a tight 1:1 complex is formed at the 1.5 mM concentrations used for NMR. The breadth of the lines is consistent with retention of four-coordination, while the dramatic change in chemical shift pattern suggests there some orientational ordering of bound TM, consistent with the crystallography of the Zn enzyme20. By comparison, addition of 0.5 eq of TPMA leads to only a small upfield shift of the 52 ppm resonance that continues to its final position at 1 eq, again showing that a tight 1:1 complex is formed. The lack of significant rearrangement in the TPMA complex suggests that it retains largely free rotation when bound to the enzyme.

These are in stark contrast to the addition of ATM to CoCA, which leads to the appearance of a number of new resonances and some rearrangement of the His resonances. Three of these can be attributed to the constitutive protons of ATM, although it is not possible to assign the spectra at present. The appearance of a number of resonances between 35 and -20 ppm are indicative of strong second sphere interactions in the active site, while the sharpness of the lines is clearly suggestive of five-coordination.

2.4.4 Electron paramagnetic resonance spectroscopy. To probe the symmetry of the coordination environment, EPR spectra were collected for CoCA with two molar equivalents of each MBG (Figure 2-7). The data show each MBG alters the

Co(II) EPR spectrum, sharpening the low field (g1) feature, while broadening the signal at high

34 field. That is, the g2 region (~ 2000 G) shifts to higher field and features at g3 (~ 3500 G) appear to broaden to high field. We found it difficult to interpret these data quantitatively, beyond the level of addition of an MBG leads to a consistent shift in the spectra, with the greatest effect induced by binding of ATM, consistent with the other techniques described above.

2.5 Conclusion. The combination of XAS, UV-vis, NMR and EPR spectroscopies was used to examine the binding of a series of a-hydroxythiones to CA. Two of the three bind clearly in a monodentate fashion, through the thione sulfur. Thiomaltol (TM) appears to show some orientational preference, based on the NMR, while it appears thiopyromeconic acid (TPMA) retains rotational freedom. In contrast, allothiomaltol (ATM) forms a five-coordinate complex via bidentate coordination. Optical titrations indicate the TM binds with equal affinity whether it is protonated or not, and similar arguments can be made for TPMA. ATM, meanwhile, shows a clear preference for binding in its protonated form. We speculate that this may be due to the lower pKa of ATM, which is expected to be reduced by ~ 0.5 pH units on metal binding, bringing the hydroxyl pKa close to the pH of the experiment, facilitating proton transfer to the solvent for dissociation.

2.6 Associated content for supporting information. Supporting Information. Detailed EXAFS fitting results are presented (Figures 2-S1 – 2-S4, Tables 2-S1 - 2-S4).

2.7 References. [1] Thomson, A. J., and Gray, H. B. (1998) Bio-inorganic Chemistry, Current Opinion in Chemical Biology 2, 155-158. [2] Jacobsen, J. A., Major Jourden, J. L., Miller, M. T., and Cohen, S. M. (2010) To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition, Biochim. Biophys. Acta 1803, 72-94. [3] Tallant, C., Marrero, A., and Gomis-Ruth, F. X. (2010) Matrix metalloproteinases: Fold and function of their catalytic domains, Biochim. et Biophy. Atca 1803, 20-28. [4] Fulcher, Y. G., and Van Doren, S. R. (2011) Remote exosites of the catalytic domain of matrix metalloproteinase-12 enhanse elastin degradation, Biochemistry 50, 9488–9499.

35 [5] Martin, D. P., Blachly, P. G., McCammon, J. A., and Cohen, S. M. (2014) Exploring the Influence of the Protein Environment on Metal-Binding Pharmacophores, J. Med. Chem. 57, 7216-7135. [6] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., Oliveira, C. A. F. d., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) ‘Unconventional’ coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-54056. [7] Parkin, G. (2004) Synthetic analogues relevant to the structure and function of zinc enzymes, Chemical Reviews 104, 699−768. [8] He, H., Puerta, D. T., Cohen, S. M., and Rodgers, K. R. (2005) Structural and spectroscopic study of reactions between chelating zinc-binding groups and mimics of the matrix metalloproteinase and disintegrin metalloprotease catalytic sites: The Coordination Chemistry of Metalloprotease Inhibition, Inorganic Chemistry 44. [9] Jacobsen, F. E., Breece, R. M., Myers, W. K., Tierney, D. L., and Cohen, S. M. (2006) Model Complexes of Cobalt-Substituted Matrix Metalloproteinases: Tools for Inhibitor Design, Inorganic Chemistry 45, 7306-7315. [10] Puerta, D. T., and Cohen, S. M. (2003) Examination of novel zinc-binding groups for use in matrix metalloproteinase inhibitors, Inorganic Chemistry 42, 3423-3430. [11] Puerta, D. T., Lewis, J. A., and Cohen, S. M. (2004) New beginnings for matrix metalloproteinase inhibitors: Identification of high-affinity zinc-binding groups, Journal of the American Chemical Society 126, 8388–8389. [12] Pottel, J., Therrien, E., Gleason, J. L., and Moitessier, N. (2014) Docking ligands into flexible and solvated macromolecules. 6. Development and application to the docking of HDACs and other zinc metalloenzymes inhibitors -, Journal of Chemical Information and Modeling 54, 254-656. [13] Seebeck, B., Reulecke, I., Kämper, A., and Rarey, M. (2014) Modeling of metal interaction geometries for protein–ligand docking, Proteins: Structure, Function, and Bioinformatics 71, 1237-1254. [14] Krishnamurthy, V. M., Kaufman, G. K., Urbach, A. R., Gitlin, I., Gudiksen, K. L., and Whiteside, G. M. (2008) Carbonic anhydrase as a model for biophysical and physical- organic studies of proteins and protein-ligand binding, Chemical Reviews 108, 946–1051. [15] Martin, D. P., Hann, Z. S., and Cohen, S. M. (2013) Metalloprotein-inhibitor binding: Human carbonic anhydrase II as a model for probing metal-ligand interactions in a metalloprotein active site, Inorganic Chemistry 52, 12207−12215. [16] Breiten, B., Lockett, M. R., Sherman, W., Fujita, S., Al-Sayah, M. H., Lange, H., Bowers, C. M., Heroux, A., Krilov, G., and Whitesides, G. M. (2013) Water networks contribute to enthalpy/entropy compensation in protein-ligand binding., Journal of the American Chemical Society 135, 15579-15584. [17] Supuran, C. T., and Winum, J. Y. (2009) Introduction to zinc enzymes as drug targets, In Drug Design of Zinc-Enzyme Inhibitors: Functional, Structural, and Disease Applications (Supuran, C. T. a. W., J.-Y., Ed.), pp 1-12, John Wiley & Sons, Inc., Hoboken, NJ.

36 [18] Kawai, K., and Nagata, N. (2012) Metal-ligand interactions: an analysis of zinc binding groups using the Protein Data Bank. - Abstract - Europe PMC, European Journal of Medicinal Chemistry 51, 271-276. [19] Jacobsen, J. A., Fullagar, J. L., Miller, M. T., and Cohen, S. M. (2011) Identifying chelators for metalloprotein inhibitors using a fragment-based approach. - Abstract - Europe PMC, Journal of Medicinal Chemistry 54, 591-602. [20] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., de Oliveira, C. A. F., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) 'Unconcentional' coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-5406. [21] Tierney, D. L., and Schenk, G. (2014) X-ray absorption spectroscopy of dinuclear metallohydrolases, Biophysical Journal 107, 1263-1272. [22] Ankudinov, A. L., Ravel, B., Rehr, J. J., and Conradson, S. D. (1998) Real-space multiple- scattering calculation and interpretation of x-ray-absorption near-edge structure, Physical Review B 58, 7565-7576. [23] Riley, E. A., Petros, A. K., Smith, K. A., Gibney, B. R., and Tierney,, and L., D. (2006) Frequency-switching inversion-recovery for severely hyperfine shifted NMR: Evidence of asymmetric electron relaxation in high-spin Co(II). Inorganic Chemistry 45, 10016- 10018.

37

2.8 Figures.

Figure 2-1. Structures of metal binding groups (MBGs) under examination (top) and ZnCA binding modes observed previously.

38

Figure 2-2. X-ray Absorption spectroscopy of CoCA, and CoCA with added TM, TPMA and ATM. (A) XANES region. Each of the MBG-added spectra are overlaid with the CoCA spectrum (thin lines), for comparison. (B) 1s63d transitions. CoCA is shown in black, while ZBG-added data are shown in gray. (C) EXAFS Fourier transforms. Each of the MBG-added FTs are overlaid with the CoCA FT (thin lines) for comparison. Fitting results are shown and summarized in Figures 2-S1 - 2-S4 and Tables 2-S1 - 2-S4.

39 Table 2-1. Best fits to CoCA+MBG EXAFS. Sample Model Fit 1s63d area a CoCA 4 N/O (2.02 Å) S1-4 20.3 CoCA+TM 3 N/O (2.04 Å) + 1 S (2.29 Å) S2-4 13.7 CoCA+TPMA 3 N/O (2.04 Å) + 1 S (2.25 Å) S3-4 16.1 CoCA+ATM 4 N/O (2.09 Å) + 1 S (2.28 Å) S4-5 12.1 a Areas in units 10-2 eV.

40

Figure 2-3. Optical spectra of metal binding groups as a function of pH. (A) TM at pH 7.5 (red), 9.2 (gray) and 11.0 (blue). (B) TPMA at pH 5.8 (red), 7.6 (gray), 9.2 (black), and 11.0 (blue). (C) ATM at pH 5.9 (red), 7.5 (gray), 9 (black), and 10 (blue). The red (OH) and blue (O-) labels indicate, respectively, bands that are characteristic to the protonated and deprotonated forms of the MBG. Y-axis is corresponds to absorbance.

41

TM

Figure 2-4: Optical titration of 38 µM CoCA with increasing amounts of (A) thiomaltol, (B) thiopyromeconic acid and (C) allothiomaltol. Bands characteristic to the protonated (OH, red) and deprotonated (O-, blue) state of the MBG are as indicated (see Figure 2-2). Insets: Expansion of the ligand-field region of the spectra, with CoCA shown in black and successive additions of 1, 2, and 3 molar equivalents of the MBG in gray.

42

Figure 2-5: Titrations of CoCA with increasing amounts of (A) TM, (B) TPMA and (C) ATM. Red symbols indicate absorption bands associated with protonated MBG, blue symbols represent predominately deprotonated MBG bands and green symbols are used for bands that are overlapped for the two species. (D) Expanded view of the ATM titration from 0 to ~ 1.2 eq. Fit lines are as described in the text.

Table 2-2. Binding constants from optical titrations.a

a All values in mM. b From Martin, et al [20]. c From 341 nm fit, with contributions from both OH and O- forms.

43

1 Figure 2-6: 200 MHz H NMR of CoCA with increasing additions of MBG in 90 % H2O.

44

Figure 2-7: X-band EPR spectra of CoCA, and CoCA in complex with 2 molar eq of the indicated MBG.

45

Figure 2-S1. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA. Detailed fitting results are presented in Table 2-S1.

Table S1. Detailed EXAFS curve fitting results for CoCA.a

b c c Fit Model Co-N/O Co-His Rf Ru 1 4 N/O 2.02 (2.8) 37 361 2 5 N/O 2.02 (4.3) 48 374 3 6 N/O 2.02 (5.6) 74 400 2.96 (2.3) 3.18 (8.3) 4 4 N/O (3 His) 2.03 (2.3) 62 201 4.22 (6.1) 4.55 (13) a Distances (Å) and disorder parameters (in parentheses, s2 (10-3 Å2)) derive from fits to filtered EXAFS data [k = 1.5-12.0 Å-1; R = 0.8-2.1 Å (fits 1-3) or 0.1-4.2 Å (fit 4)], fixing coordination numbers as indicated. b Multiple scattering paths represent combined paths, as described previously (see Materials and Methods). c Goodness of fit (Rf for fits to filtered data; Ru for fits to unfiltered data) defined as 1000* N 2 2 2 2 å {[Re(c i,obs ) + Im(c i,obs ) ]- [Re(c i,calc ) + Im(c i,calc ) ]} i =1 N 2 2 å {[Re(c i,obs ) + Im(c i,obs ) ]} i =1 , where N is the number of data points.

46

Figure 2-S2. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+TM. Detailed fitting results are presented in Table 2-S2.

Table 2-S2. Detailed EXAFS curve fitting results for CoCA+TM.a

b c c Fit Model Co-N/O Co-S Co-His Rf Ru 1 4 N/O 2.06 (4.9) 40 164 2 5 N/O 2.06 (6.7) 53 174 3 6 N/O 2.06 (8.3) 80 199

4 3 N/O + 1 S 2.04 (3.0) 2.29 (4.2) 23 154 5 4 N/O + 1 S 2.04 (4.7) 2.34 (10) 25 157 2.94 (7.3) 3.19 (16) 6 4 N/O (3 His) 2.06 (4.8) 145 106 3.83 (15) 4.13 (24) 3 N/O (3 His) + 2.93 (7.0) 3.18 (16) 7 2.04 (3.3) 2.29 (3.9) 107 82 1 S 3.82 (14) 4.11 (21) 4 N/O (3 His) + 2.90 (6.6) 3.12 (17) 8 2.05 (5.2) 2.32 (10) 126 94 1 S 3.81 (17) 4.15 (22) a Distances (Å) and disorder parameters (in parentheses, s2 (10-3 Å2)) derive from fits to filtered EXAFS data [k = 1.5-12.0 Å-1; R = 0.7-2.2 Å (fits 1-5) or 0.1-4.2 Å (fits 6-8)], fixing coordination numbers as indicated. b Multiple scattering paths represent combined paths, as described previously (see Materials and Methods). c Goodness of fit (Rf for fits to filtered data; Ru for fits to unfiltered data) defined as 1000* N 2 2 2 2 , where N is the number of data points. å{[Re(ci,obs ) + Im(ci,obs ) ]- [Re(ci,calc ) + Im(ci,calc ) ]} i=1 N 2 2 å{[Re(ci,obs ) + Im(ci,obs ) ]} i=1

47

Figure 2-S3. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+TPMA. Detailed fitting results are presented in Table 2-S3.

Table 2-S3. Detailed EXAFS curve fitting results for CoCA+TPMA.a

b c c Fit Model Co-N/O Co-S Co-His Rf Ru 1 4 N/O 2.04 (4.9) 64 383 2 5 N/O 2.04 (6.2) 67 389 3 6 N/O 2.04 (7.6) 87 409 4 3 N/O + 1 S 2.04 (5.3) 2.25 (8.5) 26 371 5 4 N/O + 1 S 2.03 (11) 2.24 (5.9) 28 368 2.94 (6.6) 3.26 (17) 6 4 N/O (3 His) 2.04 (5.0) 228 297 4.06 (6.7) 4.26 (8.3) 3 N/O (3 His) 2.94 (6.2) 3.25 (16) 7 2.02 (7.2) 2.25 (6.6) 179 277 + 1 S 4.06 (6.7) 4.26 (11) a Distances (Å) and disorder parameters (in parentheses, s2 (10-3 Å2)) derive from fits to filtered EXAFS data [k = 1.5-12.0 Å-1; R = 0.7-2.2 Å (fits 1-5) or 0.1-4.2 Å (fits 6-7)], fixing coordination numbers as indicated. b Multiple scattering paths represent combined paths, as described previously (see Materials and Methods). c Goodness of fit (Rf for fits to filtered data; Ru for fits to unfiltered data) defined as 1000* N 2 2 2 2 , where N is the number of data points. å{[Re(ci,obs ) + Im(ci,obs ) ]- [Re(ci,calc ) + Im(ci,calc ) ]} i=1 N 2 2 å{[Re(ci,obs ) + Im(ci,obs ) ]} i=1

48

Figure 2-S4. Fourier transforms (A) of k3-weighted EXAFS (B) from CoCA+ATM. Detailed fitting results are presented in Table 2-S4.

Table 2-S4. Detailed EXAFS curve fitting results for CoCA+ATM.a

b c c Fit Model Co-N/O Co-S Co-His Rf Ru 1 4 N/O 2.06 (6.2) 74 214 2 5 N/O 2.06 (7.3) 62 212 (7.2) 3 6 N/O 2.06 (8.9) 86 227 (8.8) 4 3 N/O + 1 S 2.07 (6.4) 2.29 (9.4) 47 197 5 4 N/O + 1 S 2.09 (9.9) 2.28 (6.4) 33 184 4 N/O (3 His) 2.99 (9.0) 3.28 (14) 6 2.09 (13) 2.29 (5.2) 99 97 + 1 S + 1 O 3.97 (22) 4.22 (17) a Distances (Å) and disorder parameters (in parentheses, s2 (10-3 Å2)) derive from fits to filtered EXAFS data [k = 1.5-12.2 Å-1; R = 0.5-2.2 Å (fits 1-6) or 0.1-4.2 Å (fits 7-9)], fixing coordination numbers as indicated. b Multiple scattering paths represent combined paths, as described previously (see Materials and Methods). c Goodness of fit (Rf for fits to filtered data; Ru for fits to unfiltered data) defined as 1000*

N 2 2 2 2 , where N is the number of data point å{[Re(ci,obs ) + Im(ci,obs ) ]- [Re(ci,calc ) + Im(ci,calc ) ]} i=1 N 2 2 å{[Re(ci,obs ) + Im(ci,obs ) ]} i=1

49

Chapter 3 : Evaluation of zinc binding groups (ZBGs) 5-chloro-8-quinolinol (B12), 8- hydroxy-2-quinolinecarboxylic acid (B2) 3-hydroxypyridine-2(1H)-thione (D5), and 2- hydroxy-2,4,6-cycloheptatrienone (G11) using model enzyme carbonic anhydrase

50 Evaluation of zinc binding groups (ZBGs) 5-chloro-8-quinolinol (B12), 8-hydroxy-2- quinolinecarboxylic acid (B2) 3-hydroxypyridine-2(1H)-thione (D5), and 2-hydroxy-2,4,6- cycloheptatrienone (G11) using model enzyme carbonic anhydrase Whitney R. Craig,† Micah E. Morris, † Anthony A. Forchonie,† Daniel T. DeGenova,† Rahil Patel,† Rithvik Venna,† Ania Plonski,† Seth M. Cohen,‡ Robert McCarrick,† M. Sameer Al-Abdul- Wahid, † Michael W. Crowder,† and David L. Tierney†* † Department of Chemistry and Biochemistry, Miami University, 651 East High Street, Oxford, Ohio 45056, United States ‡ Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093, United States

Corresponding Author. *E-mail: [email protected]. Phone:513- 529-8234.

Funding Source Statement. This work was supported by the National Institutes of Health (GM093987 to M.W.C. and D.L.T.) and the National Science Foundation (CHE-1151658 to M.W.C. and D.L.T.).

Author contributions: WRC, MEM, and DLT contributed writing of manuscript; WRC and DLT contributed to the data analysis. WRC, AAF, DTD, RP, RV, AP, and GCR contributed to pH UV-visible collection. WRC conducted all nuclear magnetic resonance spectra. SAAW contributed to the acquisition of nuclear magnetic resonance conditions. DLT contributed the preparation and acquisition of all electron paramagnetic resonance.

Abbreviations. CA, carbonic anhydrase; MBG, metal binding group; UV-vis, UV-visible spectroscopy; NMR, nuclear magnetic resonance; EPR, electron paramagnetic resonance; MES, 2-(N-morpholino) ethanesulfonic acid hydrate; B2, 8-hydroxy-2-quinolinecarboxylic acid; B12, 5-chloro-8-quinolinol; D5, 3-hydroxypyridine-2(1H)-thione; Gll, 2-hydroxy-2,4,6- cycloheptatrienone; hCAII, Human Carbonic Anhydrase II; EDTA, ethylene-diamine-tetraacetic acid; DTT, Dithiothreitol.

51 3.1 Abstract. The binding of four metal binding compounds, 8-hydroxy-2-quinolinecarboxylic acid (B2), 5- chloro-8-quinolinol (B12), 3-hydroxypyridine-2(1H)-thione (D5), and 2-hydroxy-2,4,6- cycloheptatrienone (G11) to human carbonic anhydrase II were investigated. The metal binding groups (MBGs) B2 and D5 displayed monodentate coordination of the active site metal ion, Co(II). G11 displayed bidentate coordination of the active site metal ion as characterized by binding assays using UV-visible spectroscopy (UV-vis), nuclear magnetic resonance spectroscopy (NMR), and electron paramagnetic resonance spectroscopy (EPR). 5-Chloro-8- quinolinol (B12) exhibited poor solubility and was determined unsuitable to pursue without structural modification. These results suggest that though all MBGs investigated coordinate to the metal active site. G11, in particular, has a binding mode that may aid in the design and development of potent metalloenzymatic inhibitors.

Keywords. Human Carbonic Anhydrase, CA Over-expression; CA Purification; CA Refolding; Metal Binding Group Evaluation.

3.2 Introduction. Matrix metalloproteinases (MMPs) are a class of zinc hydrolases that are calcium- and zinc- dependent28 and found throughout multiple tissue and organ systems29. MMPs primarily degrade connective tissue and intercellular connections, playing an important role in growth, wound healing, and reproduction28-30. Misregulation of MMPs can cause uncontrolled degradation of intercellular connections, progressing to issues such as inflammation, arthritis, heart disease, glaucoma, and cancer31-33. As a result of MMPs being implicated in these diseases, the pursuit of an inhibitor that would bind the metal active site strongly, causing hydrolytic activity of the MMP to cease is desired. The active site of MMPs contains a tris-histidine metal binding motif and a utilizes a glutamate residue that acts as a general acid/base during catalysis3, 34. A methionine residue within the active site pocket allows for hydrophobic contacts to be made with ligands3. Difficulties in over-expression, purification, and stability make MMPs challenging to study with MBGs35. These challenges leads to a need for an active site model to characterize these interactions.

52 Model complexes have been used to examine the environment of an MMP-like active site10, 11, 36. These models are limited by their lack of a secondary coordination sphere that would be present in an active site. Further, electrostatic, intermolecular forces, and steric strains are not present in most model complexes that are imperative to probe the orientation and selectivity of MBG binding to MMPs.37 Human Carbonic Anhydrase II (hCAII) is an attractive active site model for MMPs due to its conservation of active site tris-histidine motif, shared active site geometry, and active site surface accessability15. hCAII has been used as a model for Zn(II) metalloproteins previously6, 14, 15, 38. Further, hCAII is a highly stable protein, commercially-available, water soluble, and easy to crystallize, making it an ideal protein to assess MBGs through spectroscopy and crystallography.

In this present work, Co(II)-substituted hCAII (CoCA) is used to assess inhibitor coordination and binding affinity using UV-visible spectroscopy and paramagnetic NMR spectroscopy. Potential inhibitors, 8-hydroxy-2-quinolinecarboxylic acid (B2) , 5-chloro-8-quinolinol (B12), 3- hydroxypyridine-2(1H)-thione (D5), and 2-hydroxy-2,4,6-cycloheptatrienone (G11), originate from a fragment library of metalloprotein inhibitors that have been screened against several MMPs for potential binding at high concentrationsand further assessed by density functional theory DFT studies19. Lower concentration levels must be screened in order determine subtle differences and distinctions between binding motifs of each MBG at levels more reflective of medicinal dosages. Herein, the aforementioned inhibitors are probed using UV-visible spectroscopy (UV-vis), nuclear magnetic resonance spectroscopy (NMR), and electron paramagnetic resonance spectroscopy (EPR) to elucidate binding information. The results reveal that the low molar equivalents are sufficient to determine the coordination geometry for each complex.

3.3 Experimental procedure. 3.3.1 Materials. The gene for hCAII was received from professor Carole Fierke’s laboratory in a storage cell line and was subcloned into a pET26b vector. The pET26b vector was transformed into E.coli BL21(DE3) cells (Invitrogen) and used to over-express human carbonic anhydrase II. Lysogeny broth (LB) medium was obtained from Invitrogen (Carlsbad, CA) and prepared to standard

53 concentrations (25 g/L), pH 7.5. Isopropyl-β-D-thiogalactoside (IPTG) (Bold Biotechnologies) and 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride (AEBSF) (GoldBio) were used for over-expression. All other over-expression materials were purchased from Fisher Scientific. DEAE Sephacel (GE Healthcare) weak anion exchange and SP-Sepharose (GE Healthcare) cation exchange columns were used to purify recombinant hCA-II. Purified protein solutions were pooled and concentrated with an Amicon ultrafiltration cell using a YM-10 DIAFLO membrane from Amicon, Inc. (Beverly, MA). Lyophilized carbonic anhydrase-II (CA-II) from bovine erythrocytes (Sigma) was reconstituted in 50 mM 2-(N-morpholino) ethanesulfonic acid hydrate (MES, Fisher) buffer at pH 7.5. Barnstead NANOpure water (18.1 MΩ) was used to prepare all buffer solutions.

3.3.2 Methods. 3.3.2.1 Overexpression of hCA-II in LB. Plasmid pET26b-hCA-II was transformed into BL21(DE3) E. coli cells, and the cell mixture was plated on LB-agar plates containing 25 µg/mL ampicillin and incubated overnight at 37 °C. A single colony was transferred into 50 mL of LB (25 g/L, pH 7.5) containing 25 µg/mL ampicillin, and the culture was allowed to shake at 37 °C and 230 rpm until cultures reach an

OD600 = 0.8-1. Cultures were induced with ZnSO4 and IPTG to a final concentration of 0.45 mM and 0.25 mM, respectively. Cultures were allowed to grow at 30 °C for 3 hours. 4-(2- Aminoethyl)-benzenesulfonylfluoride hydrochloride (AEBSF) (8µg/L total volume) and Nα-p- tosyl-L-arginine methyl ester (TAME) (1µg/L total volume) were added, and the culture remained in shaker for 3 hours longer. Cells were harvested by centrifugation at 4 °C at approximately 5,000 rpm for 15 minutes. Cell pellets were resuspended in approximately 50 mL/

L grown lysis buffer (50 mM Tris-SO4, 50 mM NaCl, 10 mM ethylenediaminetetraacetic acid

(EDTA), 200 µM ZnSO4, 1 mM dithiothreitol (DTT), 10 µg/ml AEBSF, 1 µg/ml TAME, and 1 mM benzamidine, pH 8) on ice and flash frozen until needed. Cell pellets were stored at -80 °C.

3.3.2.2 Purification of hCA-II in LB. Frozen cell pellets were thawed at 4 °C and endonuclease (Sigma) was added, stirring for 20 minutes at 25 °C with slight agitation. A French press was used to lyse cells and the resulting solution was centrifuged at 14K rpm and 4 °C for 45 minutes. The resulting solution was

54 dialyzed against 10 mM Tris-SO4, pH 8.0, containing 200 µM ZnSO4 and 1 mM DTT using Spectra/Por dialysis tubing (Spectrum) overnight at 4 °C. A DEAE Sephacel resin column (25 mL/L culture) was equilibrated with 10 mM Tris-SO4, pH 8.0, containing 100 µM ZnSO4 and 1 mM DTT and incubated with the dialyzed solution for 30 minutes. Resin mixture was filtered and washed with one resin volume. The resulting sample was dialyzed against 10 mM MES, pH

7.0, 100 µM ZnSO4 and 1 mM tris(2-carboxyethyl)phosphine (TCEP) overnight at 4 °C. S- Sepharose resin (30 mL/L culture) was equilibrated with 10 mM MES, pH 7.0, containing 100

µM ZnSO4 and 1mM TCEP (MES equilibrium buffer). The sample was loaded onto the column and eluted at a flow of 3 mL/min at 4 °C. The column was washed with MES equilibration buffer until the A280 drops below 0.1. Carbonic anhydrase was eluted with buffers 10 mM MES, pH 7.0, containing 0.1 mM ZnSO4, 1 mM TCEP (low salt), and 10 mM MES, pH 7.0, containing

0.1 mM ZnSO4, 0.5 M (NH4)2SO4, 1 mM TCEP (high salt), set at a linear gradient. Fractions were collected at 4 °C, and the A280 was measured to determine which fractions to pool. Samples were screened using SDS-PAGE. Over-expressed CA was used in the study of B12, B2, and G11.

3.3.2.3 Preparation of Co(II)-substituted CA-II. Commercially-available CA (400 mg) was solubilized in 10 mM MES buffer, pH 7.5, and used for screening D5. Both lyophilized and over-expressed CA was dialyzed against chelex-filtered (Sigma Chelex® 100 sodium form) buffer containing 50 mM 2,6-pyradinedicarboxylic acid (PDA) (Sigma) and 50 mM MES, pH 7.5, for 48 hours, being replenished three times daily at 4 °C. Protein was further dialyzed against 50 mM of MES, pH 7.5, for 24 hours at 4 °C, being replaced with fresh buffer twice daily. Five molar equivalents of a CoCl2 (cobalt (II) chloride hexahydrate, 99.999% crystal ACS reagent, Spectrum) dissolved in 10 mM MES buffer, pH 7.5, as described above was added and was incubated with CA at 4 °C for 30 minutes. The resulting solution was incubated with Chelex® 100 sodium form (Sigma) to final amount of one-third solution volume for 15 minutes. The Chelex® 100 sodium form containing protein solution was passed through an empty gravity flow column with a porous 30 µm polyethylene bed support (econo-pack Bio-Rad) at 22 °C, and the resulting CA-II solution was pink and remained pink for several months at 4 °C. Chelex® 100 sodium form was regenerated by passing several column volumes of 50 mM of MES, pH 7.5, containing 50 mM PDA until the pour through was clear

55 and Chelex® 100 sodium form was white. The Chelex® 100 sodium form was stored at room temperature for later use.

3.3.2.4 Metal analysis. The metal content of the over-expressed and lyophilized CA-II samples was determined using a Perkin-Elmer Optima 7300V Inductively Coupled Plasma with Optical Emission Spectroscopy (ICP-OES) spectrometer. Protein samples were diluted to 1-3 µM with 1% acidified triply distilled water. Calibration curves were generated using several dilutions of metal standards (Inorganic Ventures - Zn, Co, Fe, Ni, and Cu) ranging from 0.008 to 1 µM (5 ppb to 500 ppb). Emission lines at 202.548, 228.616, 238.196, and 327.394 nm were chosen to ensure the lowest possible detection limits of zinc, cobalt, iron, nickel, and copper, respectively. Metal analysis of apo-CA-II by ICP showed less than 0.05 molar equivalents of Zn(II) remained after dialysis. Metal analysis of Co(II) CA-II by ICP-OES showed approximately 0.98 molar equivalents of cobalt(II) per protein in solution (98% efficiency).

3.3.2.5 UV-visible spectroscopy. Optical spectra were obtained using a PerkinElmer Lambda 750 spectrophotometer. CoCA binding assays were conducted by the addition 0.1-5.0 molar equivalents of a selected inhibitor in increasing increments to CoCA in 50 mM MES pH 7.5. Ki (inhibitor binding affinity) values were determined for all compounds by monitoring the inhibitor protonation and deprotonation peaks of each ZBG. Metal coordination numbers were assessed by monitoring the metal d-d transition bands (550 nm and 615 nm). Quartz cuvettes (Starna Cells Inc.) were used for all spectra, and the spectra of 50 mM MES, pH 7.5, was subtracted for a solution background. Graphical representations presented herein were generated using Kaleidagraph v4.5 (software available for purchase at http://synergy.com)

3.3.2.6 Nuclear magnetic resonance spectroscopy.

All samples for NMR were prepared using 99.9% D2O (Cambridge Isotopes) to solubilize all buffer salts. A Bruker ASX 300 NMR spectrometer (νH=300.1 MHz) was used to collect spectra on samples containing 1-2 mM CoCA at approximately 290K. Samples were prepared using 50 22 1 mM MES in D2O, pH 7.1, using water suppression parameters . Paramagnetically-shifted H

56 resonances arose due to proximity to the cobalt(II) metal center. Resonances were monitored with water resonance reference at 4.7 ppm. Inhibition binding assays were performed by the addition of molar equivalents of the selected molecule. Selected inhibitors were also monitored at an addition of 3 molar equivalents of inhibitor to protein concentrations. The method for water suppression and the number of scans were held constant throughout each inhibition experiment. The pre-saturation pulse was typically 100 -150 ms (approximately 1 W), centered at 4.7 ppm (the water frequency), while the acquisition pulse was 3 µs at full power, centered at 4.18 ppm. NUTS (NMR data processing available from http://acornnmr.com/nuts.htm) was used to Fourier- transform all spectra, and Kaleidagraph (software available for purchase at http://synergy.com) was used to generate all graphical representations.

3.3.2.7 Electron paramagnetic resonance spectroscopy. Low temperature X-band CW EPR spectra were obtained using a liquid helium flow cryostat in conjunction with a Bruker EMX EPR spectrometer equipped with an Oxford Instruments ESR900 liquid helium flow cryostat. The temperature was set to approximately 4.5K for each experiment. Samples (1-2 mM) with 20% (v/v) glycerol as a glassing agent were analyzed.

Spectra were collected at 9.64 GHz (B0^ B0) with parameters: magnetic field modulation = 10G (100kHz); time constant/conversion time = 82 ms; receiver gain = 1x104; number of scans = 4.

3.4 Results and discussion. The goal of this study was to determine if potential inhibitors of MMPs bind to the Co(II) metal center of the active site in CoCA. Carbonic anhydrase was used as a model in this study due to the greater surface accessibility to the metal in CA38-43 and the prototype for cobalt(II)- substitution being well-documented44. This groundwork for direct assessment of interactions of fragments 5-chloro-8-quinolinol (B12), quinolinecarboxylic acid (B2), 3-hydroxypyridine- 2(1H)-thione (D5), and 2-hydroxy-2,4,6-cycloheptatrienone (G11) in the metal binding pocket of CA. Fragments, present in Figure 3-1, were chosen based on screening of a chelator fragment library19 and due to their structural similarities to successful fragments.

57 3.4.1 Dissociation of chelator fragment protons via UV-visible spectroscopy. Optical spectra of CoCA with each ZBG were conducted. The optical pH dependence of each

ZBG allowed for pKa determination for each molecule (found in Figures 3-2). The hydroxyl proton pKa of B2, B12, D5, and G11, as well as the carboxy proton pKa of B2 are presented in

Table 3-1. Two pKa values were below or at physiological pH, particularly that of G11. Figure 3-2 illustrates the ZBG addition to CoCA during the optical titration. The spectrum spans across a broad wavelength scan to observe all charge transfers upon additions. Acidic species are featured in red, whereas basic species are featured in blue. pKa values were determined as follows. From the acidic spectra, an extinction coefficient was calculated for the acidic species and its absorbance at a maximum. The same was conducted for the basic spectra. For all subsequent pH sample spectra, the concentration of the protonated and deprotonated species was calculated using Beer’s law. Employing the Henderson-Hasslebalch equation, the log of concentrations of the protonated and deprotonated species can be substituted to determine the point at which pH is equal to the pKa of a specific species.

3.4.2 5-Chloro-8-quinolinol (B12) coordination assessment via UV-visible spectroscopy.

After the establishment of pKas, each ZBG was added in increasing increments to monitor binding to CoCA. The binding environment of CoCA was investigated by adding increasing molar equivalents of each inhibitor fragment to CoCA at 25 µM. Figure 3-3 illustrates ZBG addition over a broad wavelength scan to observe all changes upon addition of each ZBG. Attention was paid to the transition intensity at 550 nm and 615 nm, corresponding to ligand field transitions of high-spin Co(II)23, 45. For 5-chloro-8-quinolinol (B12), the d-d band transitions are not altered from those of the resting enzyme spectrum at 615/650 nm. The fanning pattern does affect the 515/550nm d-d bands, however, general d-d features are retained. The peak corresponding to the deprotonation event (369 nm) was fitted for pseudo Michaelis-Menten binding kinetics (Figure-4). Plotting the change in absorbance exhibits a break in the binding affinity for B12 coordinating to CoCA. A linear trend in absorbance beyond one molar equivalent suggests the binding mode remains consistent throughout all additions. Due to poor solubility issues, further binding information of 5-chloro-8-quinolinol (B12) could not be obtained.

58 3.4.3 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via UV-visible spectroscopy. As seen in the B12 peak spectra (Figure 3-5), B2 also retains the same line shape for d-d transitions as that of the resting CoCA as additions of B2 are added. This line shape retention suggests the coordination of the metal site is not altered from resting state upon incorporation of B2, and a 4-coordinate system is maintained. Peak intensities in the 550 nm and 615 nm region decrease however. This reduction is hypothesized to be due to the removal of Co(II) from the enzyme via chelation. Metal binding up to the point that stripping began was further investigated to understand how binding strength is affected by B2. Increases in overall absorbance suggests B2 does not remove the metal initially. A shift in binding affinity can also be seen after one molar equivalent addition of MBG is added, suggesting a change in the active site possibly due to its π-acceptor properties46, 47 of deprotonated B2. This alteration of the active site could lead to a shift in the extinction coefficient of the compound (Figure 3-6). This shift suggests a transition from measuring binding affinity to measuring the extinction coefficient of B2 itself, resulting in the determination that protein saturation was reached at one molar equivalent of B2.

3.4.4 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via nuclear magnetic resonance spectroscopy.

The NMR spectra, Figure 3-7, represents a sample of CoCA exchanged into 90% D2O. Though

D2O is utilized, the residual H2O and protons on CA that are not shifted due to proximity to the metal center yields a large peak that spans approximately -5 ppm to 15 ppm48. Broad and sharp signals span -10 to 100 ppm due to paramagnetic shifts of proton signals that arise from the hyperfine coupling of the unpaired Co2+ electrons with protons in the active site of CoCA35, 48.

The lack of D2O exchangeable protons at approximately 61 and 75 ppm in the CoCA spectra are due to the NH proton of the three γ- or ε-NH His in active site9. These signals can be used to evaluate changes at the cobalt(II) active site upon ZBG addition.

Figure 3-7 displays the 1H NMR spectra for CoCA with B2. The top spectrum was obtained with one equivalent of cobalt(II) and 5 equivalents of B2 in the absence of protein. This spectra represents a B2-cobalt(II) complex, and comparisons between spectra features can be made to visualize the resulting proton peak pattern when associated with a high spin Co(II) metal center.

59 It has peaks at approximately -5 ppm to 15 ppm (water), 20 ppm, 28 ppm, and 88 ppm that are unique to the coordination of B2 to cobalt(II). These indicator peaks provide insight to the affinity of binding during the titration of ZBG to CoCA, allowing for direct monitoring of the formation of a CoCA-B2 complex. The bottom spectrum shows resting CoCA, peaks at 50 ppm (δ-N3), and 60 ppm (δ-NH and ε-NH protons). Peaks near 20 ppm and 38 ppm begin to arise with the addition of 0.25 molar equivalent B2, suggesting B2 is coordinated to the cobalt metal center in CoCA and does so with high affinity. The peak intensities at 50 ppm (δ-N3), and 60 ppm (δ-NH and ε-NH protons) weaken as more potential inhibitor is added. We determine that B2 is binding with a high affinity to the enzyme, ultimately resulting in the metal being removed from the center.

3.4.5 8-Hydroxy-2-quinolinecarboxylic acid (B2) coordination assessment via electron paramagnetic resonance spectroscopy. The EPR spectra of CoCA and CoCA titrated with B2 (Figure 3-8) exhibit a broad, axial signal in the perpendicular spectra, B0^B1. Effective-g values can be found in Table 3-2. A peak at approximately g’ = 5.5 (1250 G) suggests a high-spin Co(II) metal center. Upon the addition B2, effective g-values are not shifted, suggesting the coordination of the metal does not change in the active site of CoCA. This spectra shape similarity is consistent with the 4-coordinate model from UV-visible d-d band data. Though there is a slight shift in the high-spin Co(II) peak shape, the decreased signal to noise in the B2 sprectra precluded further investigation.

3.4.6 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via UV-visible spectroscopy. Similar to B12 and B2 peak spectra, the spectra of D5-CoCA also retains the same line shape for d-d transitions as that of the resting CoCA with the incorporation of molar equivalent additions, suggesting the coordination of the metal site is 4-coordinate. Peak intensities in the 550 nm and 615 nm region fan out upon additions (Figure 3-9); however, the d-d band transitions are not altered from those of the resting enzyme spectrum, especially in the 615/650 nm region. The binding affinity does not exhibit a break in the fit for D5 as seen in B2 until approximately two molar equivalent. This suggests a somewhat weak binding affinity. The peak corresponding to the maximum absorbance in pKa determination (351 nm) was also investigated and fitted for

60 pseudo Michaelis-Menten binding kinetics Figure 3-10. This shift suggests a transition from measuring binding affinity to measuring the extinction coefficient of D5 itself, resulting in the determination that protein saturation was reached. The binding affinity is well past the concentration of substrate and protein, suggesting weak binding that does not result in metal chelation.

3.4.7 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via nuclear magnetic resonance spectroscopy. Figure 3-11 illustrates the 1H NMR spectra for CoCA with D5. The top spectrum was obtained from one equivalent of cobalt(II) and 5 equivalents of D5 in the absence of protein. This spectrum is utilized to determine spectra features of binding and metal removal by the ZBG. Peaks at approximately -5 ppm to 15 ppm are obscured by the water and diamagnetic proton resonances. A peak at 53 ppm and a broad shoulder near the water peak are unique to the coordination of D5 to cobalt(II). These indicator peaks provide insight to the extent of binding during the titration of ZBG to CoCA, allowing for direct monitoring of the formation of a CoCA- D5 complex due to metal removal. The bottom spectrum shows resting CoCA, peaks at 66 ppm (δ-N3) and 78 ppm (δ-NH and ε-NH protons). The addition of 0.25 molar equivalent of D5 gives rise to a unique peak at approximately 53 ppm, suggesting D5 is coordinated to the cobalt metal center in CoCA. The peak intensities at 66 ppm (δ-N3), and 78 ppm (δ-NH and ε-NH protons) split in a different line pattern as more potential inhibitor is added, indicating a change in the active site. This shift is greater evident at the two molar equivalent addition. A possible consideration to the change in line shape of Figure 3-10 and 3-11 is the possibility of coordination through the thione37. However, the NMR spectra presented here does not encompass the range to evaluate this fully.

3.4.8 3-Hydroxypyridine-2(1H)-thione (D5) coordination assessment via electron paramagnetic resonance spectroscopy. The EPR spectra of CoCA and CoCA titrated with D5 (Figure 3-12) exhibit a broad signal in the perpendicular spectra, B0^B1. Upon the addition D5, effective g-values are not shifted, suggesting the coordination of the metal does not change in the active site of CoCA. This line shape retention is consistent with the 4-coordinate model from UV-visible d-d band data. Though

61 there is no shift in the high- spin Co(II) peak shape and agrees with the line shape of resting CoCA. Effective-g values can be found in Table 3-2.

3.4.9 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via UV-visible spectroscopy. A fanning pattern occurred upon additions of G11 (Figure 3-13), however the transitions are distinctly different as there is little line feature at 615 nm as seen in the spectrum of them resting CoCA spectrum. Analysis of MBG specific protonation and deprotonation wavelength obtained from pKa analysis (Figure 3-14) is largely linear, suggesting weak binding to the metal center. Pseudo Michalis-Menten binding affinity of G11 yields a binding affinity well exceeding the concentration of the substrate used as well. Another difference between G11 and the other fragments herein is a pKa slightly below physiological pH. Other investigations of the IC50 of 19 G11 suggests effective binding of G11 to a metal center , possibly due to the pKa.

3.4.10 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via nuclear magnetic resonance spectroscopy. The top spectrum, which was obtained with one equivalent of cobalt(II) and 5 equivalents of G11 in the absence of protein and gives rise to the comparison between spectra features of binding (Figure 3-15). Peaks at approximately -5 ppm to 10 ppm indicate the presence of water. The peak at -15 ppm and a broad peak (15 ppm) near the water peak are unique to the coordination of G11 to cobalt(II). These indicator peaks provide insight to the affinity of binding during the titration of ZBG to CoCA, allowing for direct monitoring of the formation of a CoCA-G11 complex. The bottom spectrum shows resting CoCA, peaks at 51 ppm (δ-N3) and 62 ppm (δ-NH and ε-NH protons). The addition of 0.5 molar equivalent of G11 drastically changes the region between 15 and 30 ppm. Further, a decrease in peak intensity at 51 ppm further suggests a change in active site geometry, as suggested by the d-d transition bands in UV-visible spectroscopy. G11 is coordinated to the cobalt metal center in CoCA; however, not in a manner that suggests chelation, as the peak at 62 ppm (δ-NH and ε-NH protons) does not decrease in intensity as well. The increase in intensity at -15 ppm further suggests that G11 does bind to the metal. Due to the change in the histidine proton peaks, specifically the peal at 51 ppm in the

62 NMR spectra, a shift from 4-coordinate geometry upon the addition of G11 is evident. Perturbation of the active site geometry must arise for this change to occur.

3.4.11 2-Hydroxy-2,4,6-cycloheptatrienone (G11) coordination assessment via electron paramagnetic resonance spectroscopy. The EPR spectra of CoCA and CoCA titrated with D5 (Figure 3-16) exhibit a broad signal in the perpendicular spectra, B0^B1. Upon the addition G11, effective g-values are shifted, suggesting the coordination of the metal changes in the active site of CoCA. This shift is consistent with from UV-visible d-d band data, suggesting a shift from 4-coordinate geometry. A distinct shift in the maximum peak shape compared to the resting CoCA was present in the spectra. Upon closer investigation of this maximum, a line splitting pattern indicative of high- spin Co(II) is present30. Similar to DFT calculations and the study of G11 with model complexes37, G11 appears to bind in a weak 5-coordinate fashion in the CoCA active site. This bidentate coordination makes it an attractive fragment to proceed forward pursuing as an MMP inhibitor despite weak binding. Effective-g values can be found in Table 3-2.

3.5 Conclusions. The investigation of metal binding groups as a manner to achieve more effective therapeutic targets to MMP largely depends on the binding affinity and binding mode of a fragment. Herein, fragments 5-chloro-8-quinolinol (B12), quinolinecarboxylic acid (B2), 3-hydroxypyridine- 2(1H)-thione (D5), and 2-hydroxy-2,4,6-cycloheptatrienone (G11) (Figure 3-1) were investigated for their effectiveness in binding the Co(II) metal center of CAII. Proposed coordination orientations are given in Figure 3-17. Solubility issues were experienced with 5- chloro-8-quinolinol (B12), and it was concluded that it would not be an effective therapeutic without alteration to the structure of the molecule. Perhaps the addition of a more hydrophilic functional group would increase the solubility of the molecule and thus increase the likelihood of yielding a more effective binding fragment. Quinolinecarboxylic acid (B2) was determined to strip CA of the Co(II) in the active site, as seen through the decrease in d-d bands in UV-visible spectra and decrease of histidine proton peak intensities in the NMR spectra. This tight binding to the metal center is attractive when designing a therapeutic target; however, the metal stripping issue must be addressed. The thione moiety present in 3-hydroxypyridine-2(1H)-thione (D5)

63 holds promise if a binding environment or substitution addition for specificity could allow D5 to 6, 9 bind similar to other thione compounds with lower pKa values . The change in binding at higher concentrations supports the possibility for further investigation into the binding mode and is therefore a molecule that could be considered for the basis of an inhibitor design. G11 holds the highest promise as a therapeutic fragment to be used in inhibitor design. The shift in d-d bands as well as the perturbation of the active site seen in NMR suggests a 5 coordinate metal, three γ- or ε-NH His coordinated and a bidentate ZBG coordinated. Upon the addition of G11 to CoCA, the EPR line spitting pattern centered at an effective g’» 6.5 (1035G) suggests a shift in metal coordination from resting CoCA. For these reasons, G11 stands out among the MBGs examined herein and should be considered for screening against MMPs.

3.7 References. [1] Massova, I., Kotra, L. P., Fridman, R., and Mobashery, S. (1998) Matrix metalloproteinases: structures, evolution, and diversification, FASEB J 12, 1075-1095. [2] Nagase, H., Visse, R., and Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs, Cardiovasc Res 69, 562-573. [3] Massova, I., Kotra, L. P., and Mobashery, S. (1998) Structural insight into the binding motifs for the calcium ion and the non-catalytic zinc in matrix metalloproteases, Bioorg Med Chem Lett 8, 853-858. [4] Gialeli, C., Theocharis, A. D., and Karamanos, N. K. (2011) Roles of matrix metalloproteinases in cancer progression and their pharmacological targeting, FEBS J 278, 16-27. [5] Kessenbrock, K., Plaks, V., and Werb, Z. (2010) Matrix metalloproteinases: regulators of the tumor microenvironment, Cell 141, 52-67. [6] Supuran, C. T., Scozzafava, A., and Casini, A. (2003) Carbonic anhydrase inhibitors, Medicinal Research Reviews 23, 146-189. [7] Bode, W., Fernandez-Catalan, C., Tschesche, H., Grams, F., Nagase, H., and Maskos, K. (1999) Structural properties of matrix metalloproteinases, Cell Mol Life Sci 55, 639-652. [8] Tallant, C., Marrero, A., and Gomis-Ruth, F. X. (2010) Matrix metalloproteinases: fold and function of their catalytic domains, Biochim. et Biophy. Atca 1803, 20-28. [9] Meng, F., Yang, H., Aitha, M., George, S., Tierney, D., and Crowder, M. (2016) Biochemical and spectroscopic characterization of the catalytic domain of MMP16 (cdMMP16) | SpringerLink, journal of biological inorganic chemistry 21, 523-535. [10] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., de Oliveira, C. A. F., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) 'Unconcentional' coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-5406.

64 [11] Puerta, D. T., and Cohen, S. M. (2003) Examination of novel zinc-binding groups for use in matrix metalloproteinase inhibitors, Inorganic Chemistry 42, 3423-3430. [12] Puerta, D. T., Lewis, J. A., and Cohen, S. M. (2004) New beginnings for matrix metalloproteinase inhibitors: Identification of high-affinity zinc-binding groups, Journal of the American Chemical Society 126, 8388–8389. [13] Dick, B. L., Patel, A., McCammon, J. A., and Cohen, S. M. (2017) Effect of donor atom identity on metal-binding pharmacophore coordination, J Biol Inorg Chem. [14] Martin, D. P., Hann, Z. S., and Cohen, S. M. (2013) Metalloprotein-inhibitor binding: Human carbonic anhydrase II as a model for probing metal-ligand interactions in a metalloprotein active site, Inorganic Chemistry 52, 12207−12215. [15] Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., Oliveira, C. A. F. d., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) ‘Unconventional’ coordination chemistry by metal chelating fragments in a metalloprotein active site, Journal of the American Chemical Society 136, 5400-54056. [16] Scozzafava, A., and Supuran, C. T. (2000) Carbonic anhydrase and matrix metalloproteinase inhibitors: Sulfonylated amino acid hydroxamates with MMP inhibitory properties act as efficient inhibitors of CA isozymes I, II, and IV, and N- hydroxysulfonamides inhibit both these zinc enzymes, Journal of Medicinal Chemistry 43, 10. [17] Krishnamurthy, V. M., Kaufman, G. K., Urbach, A. R., Gitlin, I., Gudiksen, K. L., and Whiteside, G. M. (2008) Carbonic anhydrase as a model for biophysical and physical- organic studies of proteins and protein-ligand binding, Chemical Reviews 108, 946–1051. [18] Jacobsen, J. A., Fullagar, J. L., Miller, M. T., and Cohen, S. M. (2011) Identifying chelators for metalloprotein inhibitors using a fragment-based approach. - Abstract - Europe PMC, Journal of Medicinal Chemistry 54, 591-602. [19] Riley, E. A., Petros, A. K., Smith, K. A., Gibney, B. R., and Tierney,, and L., D. (2006) Frequency-switching inversion-recovery for severely hyperfine shifted NMR: Evidence of asymmetric electron relaxation in high-spin Co(II). Inorganic Chemistry 45, 10016- 10018. [20] Bertini, I., Canti, G., Luchinat, C., and Scozzafava, A. (2002) Characterization of cobalt(II) bovine carbonic anhydrase and of its derivatives, Journal of the American Chemical Society 100, 4873-4877. [21] Garmer, D. R., and Krauss, M. (2002) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, Journal of the American Chemical society 115, 10247-10257. [22] Rowlett, R. S., Chance, M. R., Wirt, M. D., Sidelinger, D. E., Royal, J. R., Woodroffe, M., Wang, Y. F., Saha, R. P., and Lam, M. G. (1994) Kinetic and structural characterization of spinach carbonic anhydrase, Biochemistry 33, 13967-13976. [23] Rowlett, R. S. (2010) Structure and catalytic mechanism of the β-carbonic anhydrases, 1804, 362–373. [24] Bertini, I., Jonsson, B. H., Luchinat, C., Pierattelli, R., and Vila, A. J. (1994) Strategies of signal assignments in paramagnetic metalloproteins. An NMR investigation of the thiocyanate adduct of the cobalt (II)-substituted human carbonic anhydrase II, J Magn Reson B 104, 230-239.

65 [25] Supuran, C. T. (2008) Carbonic anhydrases: novel therapeutic applications for inhibitors and activators, Nature Reviews Drug Discovery 7, 168-181. [26] Garmer, D. R., and Krauss, M. (2002) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, J. Am. Chem. Soc. 115, 10247. [27] Llarrull, L. I., Tioni, M. F., Kowalski, J., Bennett, B., and Vila, A. J. (2007) Evidence for a dinuclear active site in the metallo-beta-lactamase BcII with substoichiometric Co(II). A new model for metal uptake, J Biol Chem 282, 30586-30595. [28] Hawk, M. J., Breece, R. M., Hajdin, C. E., Bender, K. M., Hu, Z., Costello, A. L., Bennett, B., Tierney, D. L., and Crowder, M. W. (2009) Differential binding of Co(II) and Zn(II) to metallo-β-lactamase Bla2 from Bacillus anthracis, Journal of the American Chemical Society 131, 10753-10762. [29] Simpson, R. T., and Vallee, B. L. (1968) Two differentiable classes of metal atoms in alkaline phosphatase of Escherichia coli, Biochemistry 7, 4343-4350. [30] Yang, H., Makaroff, K., Paz, N., Aitha, M., Crowder, M. W., and Tierney, D. L. (2015) Metal ion dependence of the matrix metalloproteinase-1 Mechanism, Biochemistry 54, 3631-3639. [31] Jacobsen, F. E., Breece, R. M., Myers, W. K., Tierney, D. L., and Cohen, S. M. (2006) Model complexes of cobalt-substituted matrix metalloproteinases: Tools for inhibitor design, Inorganic Chemistry 45, 7306-7315.

66

3.8 Figures

Figure 3-1: Structures of fragments, name, and abbreviations. Red letters indicate the protons associated with the corresponding measured pKa value.

67 8-hydroxy-2-quinolinecarboxylic acid (B2) 5-chloro-8-quinolinol (B12) Cl

OH N OH O N OH

3-hydroxypyridine-2(1H)-thione (D5) 2-hydroxy-2,4,6-cycloheptatrienone (G11) OH HN

S OH O

Figure 3-2: Optical spectra of zinc binding groups at varying pH. Structures of the ZBGs and the name of the ZBG are shown at the top right. The pKa values for compounds B2, B12, D5, and G11 as well as specific wavelengths to determine binding information were derived from the pH spectra.

68 Table 3-1: Approximate calculated pKa values for each ZBG – B12, B2, D5, and G11.

ZBG pKa 5-chloro-8-quinolinol (B12) 9.1 3.9 (carboxy) quinolinecarboxylic acid (B2) 9.0 (alcohol) 3-hydroxypyridine-2(1H)- 11.1 thione (D5) 2-hydroxy-2,4,6- 6.9 cycloheptatrienone (G11)

Figure 3-3: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 5-chloro-8-quinolinol (B12). The pink line indicates CoCA with no equivalents of B12 added. The black line represents CoCA titrated with 1.0 equivalents of B12. The green line represents CoCA titrated with 2.0 equivalents of B12. The blue line represents CoCA titrated with 3.0 equivalents of B12. The gray lines indicate the titration of CoCA with B12 equivalents ranging from 0.1 to 1.9 and 2.5. Inlay at top right is a zoom of larger graph, displaying d-d transitions of cobalt(II) substituted carbonic anhydrase throughout the titration.

69

Figure 3-4: The change in absorbance for the substrate specific bands for titrated amounts of 5-chloro-8-quinolinol (B12) added to 25 µM CoCA. The vertical line indicates the concentration at which CoCA (25 µM) is equal to the concentration of curcumin titrated into solution. Purple circles represent data obtained at 369 nm, the maximum light absorption wavelength of B12. Data sets are fitted with pseudo Michaelis–Menten binding to determine binding affinity (purple dashed line) to vertical line. After the vertical line, data sets are fitted with a line representing the absorbance due to unbound B12 in solution.

70

Figure 3-5: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 8-hydroxy-2-quinolinecarboxylic acid (B2). The pink line indicates CoCA with no equivalents of B2 added. The black line represents CoCA titrated with 1.0 equivalents of B2. The green line represents CoCA titrated with 2.0 equivalents of B2. The blue line represents CoCA titrated with 3.0 equivalents of B2. The gray lines indicate the titration of CoCA with B2 equivalents ranging from 0.1 to 1.9 and 2.5. Inlay at top right is a zoom of larger graph, displaying d-d transitions of cobalt(II) substituted carbonic anhydrase throughout the titration.

71

Figure 3-6: The change in absorbance for the substrate specific bands for titrated amounts of 8-hydroxy-2-quinolinecarboxylic acid (B2) added to 25 µM CoCA. The vertical line indicates the concentration at which CoCA (25 µM) is equal to the concentration of B2 titrated into solution. Purple circles represent data obtained at 340 nm, the maximum light absorption wavelength of B2. Data sets are fitted with pseudo Michaelis–Menten binding to determine binding affinity (purple dashed line).

72 * * * * *

Figure 3-7: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 8-hydroxy-2-quinolinecarboxylic acid (B2). A cobalt(II)/ B2 saturated standard is present as an internal reference. Asterisks indicate subtle changes in the equivalent addition that they first appear.

73

Figure 3-8. X-band perpendicular mode EPR of CoCA and CoCA + B2. CoCA + 3 molar equivalents 8-hydroxy-2-quinolinecarboxylic acid (B2) – black line. Gray line is CoCA with no zinc binding group added.

74

Figure 3-9: UV-visible spectra of the titration of 215 µM cobalt(II)-substituted carbonic anhydrase with 3-hydroxypyridine-2(1H)-thione (D5). The pink line indicates CoCA with no equivalents of D5 added. The black line represents CoCA titrated with 1.0 equivalents of D5. The green line represents CoCA titrated with 3.0 equivalents of D5. The blue line represents CoCA titrated with 5.0 equivalents of D5. The gray lines indicate the titration of CoCA with D5 equivalents ranging from 0.1 to 1.9 and 2.5. Inlay at top right is a zoom of larger graph, displaying d-d transitions of cobalt(II) substituted carbonic anhydrase throughout the titration.

75

Figure 3-10: The change in absorbance for the substrate specific bands for titrated amounts of 3-hydroxypyridine-2(1H)-thione (D5) added to 215 µM CoCA. The vertical line indicates the concentration at which CoCA (215 µM) is equal to the concentration of 3- hydroxypyridine-2(1H)-thione (D5) titrated into solution. Purple circles represent data obtained at 351 nm, the maximum light absorption wavelength of 3-hydroxypyridine-2(1H)-thione (D5). Data sets are fitted with pseudo Michaelis–Menten binding to determine binding affinity (dashed line).

76

Figure 3-11: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 3-hydroxypyridine-2(1H)-thione (D5). A cobalt(II)/ D5 saturated standard is present as an internal reference. Asterisks indicate subtle changes in the equivalent addition that they first appear.

77

Figure 3-12. X-band perpendicular mode EPR of CoCA and CoCA + D5. CoCA + 3 molar equivalents 3-hydroxypyridine-2(1H)-thione (D5)– black line. Gray line is CoCA with no zinc binding group added. Left- Full spectra, Right- Zoom on first effective g region.

78

Figure 3-13: UV-visible spectra of the titration of 25 µM cobalt(II) substituted carbonic anhydrase with 2-hydroxy-2,4,6-cycloheptatrienone (G11). The pink line indicates CoCA with no equivalents of G11 added. The black line represents CoCA titrated with 1.0 equivalents of B2. The green line represents CoCA titrated with 2.0 equivalents of G11. The blue line represents CoCA titrated with 3.0 equivalents of B2. The gray lines indicate the titration of CoCA with G11 equivalents ranging from 0.1 to 1.9 and 2.5. Inlay at top right is a zoom of larger graph, displaying d-d transitions of cobalt(II) substituted carbonic anhydrase throughout the titration.

79

Figure 3-14: The change in absorbance for the substrate specific bands for titrated amounts of 2-hydroxy-2,4,6-cycloheptatrienone (G11) added to 25 µM CoCA. The vertical line indicates the concentration at which CoCA (25 µM) is equal to the concentration of curcumin titrated into solution. Purple circles represent data obtained at 369 nm, the maximum light absorption wavelength of G11. Data sets are fitted with pseudo Michaelis–Menten binding to determine binding affinity (purple dashed line) to vertical line. After the vertical line, data sets are fitted with a line representing the absorbance due to unbound G11 in solution.

80

Figure 3-15: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with 2-hydroxy-2,4,6-cycloheptatrienone (G11). A cobalt(II)/ G11 standard is present as an internal reference. Asterisks indicate subtle changes in the equivalent addition that they first appear.

81

Figure 3-16. X-band perpendicular mode EPR of CoCA and CoCA + G11. CoCA + 3 molar equivalents 2-hydroxy-2,4,6-cycloheptatrienone (G11) – black line. Gray line is CoCA with no zinc binding group added. Left- Full spectra, Right- Zoom on first effective g region.

Table 3-2. G-values for each perpendicular EPR spectra – B12, B2, D5, and G11.

B2 D5 G11 CoCA Maximum Peak 6.0 5.4 5.9 5.5 Minimum 3.0 3.3 3.0 3.3 Zero Crossing 1.4 1.4 1.2 1.4

82 HN HO N O- S

O O

Co2+ Co2+ N N N N N N

N N N N

N N

OH2 O O Co2+ N Co2+ N N N N N N N N N N N

Figure 3-17: Proposed coordination of each ZBG fragments. Fragments quinolinecarboxylic acid (B2), 3-hydroxypyridine-2(1H)-thione (D5), 2-hydroxy-2,4,6-cycloheptatrienone (G11), and resting CoCA.

83 Chapter 4: Evaluation of curcumin and functional group analysis as a zinc binding groups (ZBGs) using model enzyme carbonic anhydrase

84 Evaluation of Curcumin and Functional Group Analysis as a Zinc Binding Groups (ZBGs) Using Model Enzyme Carbonic Anhydrase

Whitney R. Craig, Daniel T. DeGenova, Garrett C. Reed, Robert McCarrick, M. Sameer Al- Abdul-Wahid, and David L. Tierney*

Department of Chemistry and Biochemistry, Miami University, 651 East High Street, Oxford, Ohio 45056, United States

Corresponding Author. *E-mail: [email protected]. Phone: (513) 529-8234.

Funding Source Statement. This work was supported by the National Institutes of Health (GM093987 to M.W.C. and D.L.T. and the National Science Foundation (CHE-1151658 to M.W.C. and D.L.T.).

Author contributions: WRC, DTD, and DLT contributed writing of manuscript; WRC and DLT contributed to the data analysis. WRC, DTD, and GCR contributed to UV-visible collection.

WRC conducted all nuclear magnetic resonance spectra. SAAW contributed to the acquisition of nuclear magnetic resonance conditions. WRC and DLT contributed the preparation and acquisition of all electron paramagnetic resonance.

Abbreviations. CA, carbonic anhydrase; ZBG, zinc binding group; Acac, pentane-2,4-dione;

NMR, nuclear magnetic resonance; EPR, electron paramagnetic resonance

85 4.1 Abstract. Matrix metalloproteinases (MMPs) are a group of peptidases that play roles in disease-related intracellular and extracellular processes in the body, such as arthritis and cancer. Inhibiting MMPs may regulate processes important to these diseases. Metal chelators may be effective inhibitors of MMPs due to the zinc(II) in the active site. Curcumin, a potential MMP inhibitor, has been investigated in vivo at length, without literature discussing binding on a functional group level. Guaiacol (2-Methoxyphenol) and acac (Pentane-2,4-dione), are two components of curcumin that can aid in the investigation of curcumin as a potential inhibitor. These compounds were screened against carbonic anhydrase (CA), a model for the MMPs, to study their affinity to the metal active site. Spectroscopic data suggests guaiacol is a weak metal coordinating ligand, whereas acac is a known metal chelator at high concentrations. Curcumin displays binding affinity like acac with minimal metal removal, partially from the guaiacol influence.

4.2 Introduction. Curcumin, (1E,6E)-1,7-bis(4-hydroxy-3-methoxyphenyl)hepta-1,6-diene-3,5-dione, (Figure 4-1) is a natural yellow pigment primarily derived from Curcuma longa Linn. This compound comprises 2-9% of turmeric1, 2 and has been used in Eastern medicine as far back as the eleventh century. It is also used as in cuisine for coloration, preservation, and flavor enhancement3. Curcumin has been studied for its possible pharmacological properties in conditions, such as type 2 diabetes mellitus4, depression5, amyloid beta (metal chelation)6, arthritis7, 8, and cancer remediation9-11 with varying levels of success.

One issue many in vivo studies have found is a decrease in the bioavailability of curcumin due to rapid metabolism by the liver and fecal elimination12. Rodent studies estimate approximately 75% of a 1g / kg dose administered orally is eliminated in fecal matter prior to the documentation of therapeutic effects13, 14. Intravenous and intraperitoneal administration of curcumin showed an increased uptake when hepatic secretion was measured, compared to oral administration15. Clinical trials measuring curcumin concentration after various doses have been measured16. Doses of 4 to 8 g of curcumin peaked at serum levels ranging from 0.41 µM to 1.75 µM depending on dosage17. Another clinical trial investigating plasma concentrations of

86 curcumin and curcumin metabolites one hour after an oral dose of 3.6 g determined the concentration range to be 0.01 µM18.

In colorectal cancer patients given a 3.6 g dosage, accumulation of curcumin was measured in both cancerous and non-cancerous tissue19. Studies of orally administered curcumin and curcumin metabolites have found gaps between the administration of curcumin in vitro and in vivo20, suggesting a need for a deeper investigation of curcumin prior to oral administration. Because of bioavailability concerns, largely due to low aqueous solubility, possibilities for inhibitor frameworks must be investigated in order to tune curcumin to be competitive with conventional drugs. Furthermore, evidence of high in vivo excretion gives rise to the need for non-specific binding investigations.

Multiple types of chemopreventative effects have been reported concerning curcumin, specifically in hepatic and mammary tissues21. Numerous cell adhesion molecules involved in tumor growth and metastasis have demonstrated a reduction in binding after curcumin addition as well. Further, curcumin and curcumin analogs have also exhibited inhibitory properties with matrix metalloproteinases (MMPs), a group of zinc containing endopeptidases that have been implicated in cancer metastasis and progression. Curcumin has been reported to suppress the invasion of tumors and the migration of lung cancer cells. Further, studies conclude curcumin reduces local inflammation through inhibition of select MMPs as well as down-regulation in the expression of MMPs12. MMP-2 and MMP-9 expression levels have been found to be reduced, and decreased levels of angiogenesis have been reported, as has the inhibition of MMP transcription factors16, 22. Although curcumin has displayed an ability to interact with MMPs, curcumin is relatively insoluble in water, displays poor biological uptake and may exhibit non- specific binding23.

The efficacy and safety of curcumin for medical use have been explored. Low bioavailability, lack of specificity in vivo, and structural characteristics suggest metal binding possibilities that are not yet fully investigated. The electron withdrawing groups in curcumin have been hypothesized to be the predominant structural feature that allows curcumin the ability to access and bind to zinc ions of multiple proteins, specifically due to the stability of the enolate formed

87 at higher pH12. When screened against several MMP isozymes including MMP-2, -8, -9, and -13, the IC50 (half maximal inhibitory concentration) values ranged from 14 µM to 110 µM. Upon removal of the aryl moieties of curcumin, dose response is not observed. From this range in binding response, there is a growing need to investigate curcumin as a metal binding inhibitor, as well as screen its off-target binding potential. To better comprehend the binding of curcumin, lower concentrations representative of bioavailable quantities of the curcumin must be investigated.

To better understand the interaction between potential zinc binding groups (ZBGs) and MMPs, carbonic anhydrase (CA) has been used as a model for preliminary studies24, 25. The choice of CA provides an intermediate step between small molecule structural mimics and larger, less well behaved proteins. CA is a ubiquitous metalloprotein and binds an active site zinc(II) with the same tris-histidine ligand set26, making CA an ideal model to study inhibitor binding to zinc(II) hydrolases, especially MMPs. Carbonic anhydrase is one of the most well-studied mononuclear Zn(II) enzymes, and the prototype for cobalt(II)-substitution. There is greater surface accessibility to the metal in CA26-31, allowing for a direct assessment of the inhibitor interactions in the metal binding pocket32. Due to these key features and the relative instability of MMPs, CA has been used as a model for MMPs. We examine here the interaction of curcumin with the metal binding site of CA. Acetylacetonate and guaiacol were also investigated as metal binding moieties of curcumin to examine the ability of each part of curcumin to interact with the metal ion.

4.3 Experimental procedure. 4.3.1 Materials. Lyophilized carbonic anhydrase-II (CA-II) from bovine erythrocytes (Sigma) was reconstituted in 50 mM 2-(N-morpholino) ethanesulfonic acid hydrate (MES, Fisher) buffer at pH 7.5. Barnstead NANOpure water (18.1 MΩ) was used to prepare all buffer solutions.

4.3.2 Methods. 4.3.2.1 Preparation of Co(II)-substituted rehydrated CA-II. Native CA (400 mg) was dialyzed against Chelex-filtered (Sigma Chelex® 100 sodium form) buffer containing 50 mM

88 2,6-pyradinedicarboxylic acid (PDA) (Sigma) and 50 mM MES, pH 7.5, for 48 hours, being replenished three times daily. Protein was further dialyzed against 50 mM of MES, pH 7.5, for

24 hours, being replenished with new buffer twice daily. Five molar equivalents of a CoCl2 (cobalt (II) chloride hexahydrate, 99.999% crystal ACS reagent, Spectrum) were dissolved in MES, pH 7.5, as described above. The was incubated with CA at 4 °C for 30 minutes. The resulting solution was incubated with one-third-solution volume of Chelex® 100 sodium form (Sigma) for 15 minutes. A PerkinElmer model 150 inductively coupled plasma atomic emission spectrometer (ICP-AES) was used to measure the metal content of all CA samples. For the ICP- AES measurements, the protein concentrations were 0.2 µM, and ICP-AES standards containing Zn(II) and Co(II) at concentrations from 0.008 to 1 µM were prepared by serial dilution of standard solutions (Inorganic Ventures). Metal analysis of apo-CA-II by ICP showed less than 0.05 molar equivalents of Zn(II) remained after dialysis. The Chelex® 100 sodium form containing protein solution was passed through a gravity flow column (Bio-Rad) at 22 °C, and the resulting CA-II solution was pink and remained pink for several months at 4 °C. Metal analysis of Co(II)CA by ICP-OES showed approximately 0.98 molar equivalents of cobalt(II) per protein in solution. Chelex® 100 sodium form was regenerated by passing several column volumes of 50 mM of MES, pH 7.5, with 50 mM PDA until the pour through was clear and Chelex® 100 sodium form was white. The Chelex® 100 sodium form was stored at room temperature for later use.

4.3.2.2 UV-visible spectroscopy. Optical spectra were obtained using a PerkinElmer Lambda 750 Diode array spectrophotometer. 76 µM of each ZBG was analyzed to determine the spectra of protonated and deprotonated inhibitors. Spectroscopic peaks were analyzed and the wavelengths corresponding to the protonated and deprotonated species were determined. Inhibition assays were conducted by the addition 0.1-5.0 molar equivalents of a selected inhibitor in increasing increments to Co(II)CA in 50 mM MES, pH 7.5. Ki (inhibitor binding affinity) values were determined for acac, guaiacol, and curcumin by monitoring the inhibitor protonation and deprotonation peaks of each ZBG. Metal coordination number was assessed by monitoring the metal d-d transition bands (550 nm and 615 nm). Quartz cuvettes (Starna Cells Inc.) were used for all spectra, and the spectrum of 50 mM MES, pH 7.5, was subtracted for a solution

89 background. Graphical representations presented herein were generated using Kaleidagraph v4.5 (software available for purchase at http://synergy.com)

4.3.2.3 Nuclear magnetic resonance spectroscopy. All samples for NMR were prepared using

99.9% D2O (Cambridge Isotopes) to solubilize all buffer salts. A Bruker ASX 300 NMR spectrometer (νH=300.1 MHz) was used to collect spectra on samples containing 1-4 mM

Co(II)CA at approximately 290K. Samples were prepared using 50 mM MES in D2O, pH 7.1 using water suppression parameters33. Paramagnetically-shifted 1H resonances arose due to proximity to the cobalt(II) metal center. Resonances were monitored with water resonance reference at 4.7 ppm. Inhibition binding assays were conducted by the addition of varying molar equivalents of the selected molecule. Selected inhibitors were also monitored at an addition of 5 molar equivalents of inhibitor to protein concentrations. The method for water suppression and the number of scans were held constant throughout each inhibition experiment. The pre- saturation pulse was typically 100 -150 ms (approximately 1 W), centered at 4.7 ppm (the water frequency), while the acquisition pulse was 3 µs at full power, typically centered between 4.18 ppm. NUTS (NMR data processing available from http://acornnmr.com/nuts.htm) was used to fourier-transform all spectra and Kaleidagraph (software available for purchase at http://synergy.com) was used to generate all graphical representations.

4.3.2.4 Electron paramagnetic resonance spectroscopy. Low temperature X-band CW EPR spectra were obtained using a liquid helium flow cryostat in conjunction with a Bruker Elexsys EMX EPR spectrometer equipped with an Oxford Instruments ESR900 liquid helium flow cryostat. The temperature was set to approximately 4.5K for each experiment. Samples ranging from 1-4 mM with 20% (v/v) glycerol as a glassing agent were analyzed. Spectra were collected at 9.38 (B0 || B0) or 9.64 GHz (B0^ B0) with parameters: magnetic field modulation = 10G (100kHz); time constant/conversion time = 82 ms; receiver gain = 1x104; number of scans = 4.

4.4 Results and discussion. To date, there have been no biophysical studies of curcumin binding to metallonzymes. The goal of this study was to examine the potential metal-binding interactions of curcumin, along with acetylacetonate (acac) and guaiacol representing its component parts (Figure 4-1). Acac models

90 the central beta-diketonate, while guaiacol models the pendant methoxyphenol functional groups of curcumin.

4.4.1 UV-visible spectroscopy. The binding environment of CoCA was investigated by adding increasing molar equivalents of each inhibitor fragment to CoCA at 76 µM. Figures 4-2, 4-5, and 4-7 illustrate this ZBG addition over a broad wavelength scan to observe all changes upon additions of each ZBG. Particular attention was paid to the transition intensity at 550 nm and 615 nm, corresponding to ligand field transitions of high-spin Co(II)34, 35. For curcumin, a fanning pattern can be observed upon additions; however, the d-d band transitions are not altered from the resting enzyme spectrum at 615/650 nm (Figure 4-3). The fanning pattern does affect the 515/550nm d-d bands; however, general d-d features are retained. Plotting the change in absorbance exhibits a break in the binding affinity for curcumin coordinating to CoCA. The change in line shape at one molar equivalent suggests a change in binding mode when higher concentrations of curcumin are added. The peak corresponding to the greatest absorption in during pKa determination was also investigated and fitted for pseudo Michalis-Menten binding kinetics (Figure 4-4). As seen in the guaiacol signature peak spectra, curcumin also has a shift in binding affinity after one molar equivalents addition of ZBG. This shift suggests a transition from measuring binding affinity to measuring the extinction coefficient of curcumin itself, resulting in the determination that protein saturation was reached.

In the acac spectra, Figure 4-5, the line shape and peak pattern in this region is not altered, suggesting the coordination of the metal site is not altered from resting state upon incorporation of acac. Acac seems to coordinate the metal center in a monodentate fashion, replacing the water molecule in the active site. Upon the addition of greater than two molar equivalents, however, the measured absorbance value begin to decrease. Acac is a known metal chelator36 coordinating in a multidentate fashion. Therefore, it is hypothesized that these transition bands observed when the intensity decreases 550 nm and 615 nm is due to the removal of Co(II) from the enzyme via metal stripping. Metal binding up to the point stripping begins, two molar equivalents, are further investigated to understand how binding strength is affected by acac, Figure 4-6. Increases in overall absorbance suggests acac does not remove the metal initially but does shift the ligand-to-

91 metal charge transfer peaks of high-spin Co(II) upon the coordination of acac due to its π- acceptor properties37, 38. Because of the strength of metal coordination, details of curcumin orientation can be considered.

For guaiacol, Figure 4-7, line shape retention corresponding to d-d band transition of high-spin Co(II) peaks is observed. Line changes are not observed at 615/650 nm bands, however there is an increase in intensity of bands at 515/550 nm. Further investigation of metal binding strength, illustrated in Figure 4-8, denotes weaker coordination to the metal center in CoCA, as there is not a leveling event observed. The peak from pKa studies guaiacol, 344 nm, was probed to determine the strength of protonated guaiacol binding to CA (Figure 4-9). Monitoring this peak provided an opportunity to focus analysis on the interaction between the ZBG and CoCA. The change in absorbance was fitted with pseudo Michalis-Menten binding kinetics and calculated to have a binding affinity of ~0.177 mM at one molar equivalent of guaiacol to CoCA. However, here is an observed break in the fit at one molar equivalent, exhibiting a binding shift similar to that of curcumin and similar conclusions were drawn regarding the shift from Ki to an extinction coefficient.

4.4.2 Nuclear magnetic resonance spectroscopy. The NMR spectra are shown in Figure 4-10, 4-11, and 4-12. Samples of CoCA were exchanged into 90% D2O. Though D2O is utilized, the residual H2O and diamagnetic enzyme protons yields a large peak that spans approximately -5 ppm to 15 ppm39 Broad and sharp signals spanning -10 to 100 ppm due to paramagnetic shifts of proton signals that arise from the unpaired electron of Co2+ in the active site of CoCA39, 40. The lack of exchangeable protons at approximately 61 and 75 ppm in CoCA spectra represent the NH proton of the three γ- or ε-NH His in active site41 and can be used to evaluate changes at the cobalt(II) active site upon ZBG addition.

The 1H NMR spectra in Figure 4-10 exhibits CoCA titration additions with curcumin. The top spectrum, which contains one equivalent of elemental cobalt(II) and 5 equivalents of curcumin, represents curcumin coordinated to cobalt(II). All spectra in this figure have peaks at approximately -5 ppm to 15 ppm (water) and a broad peak at approximately 35 ppm that is unique to the coordination of curcumin to cobalt(II). This broad peaks provide insight to the

92 extent to which curcumin binds to CoCA. There is change upon a 0.25 equivalent addition of curcumin, with peaks at approximate 20 ppm and -20 ppm becoming more defined, indicated with asterisks. These peaks persist with an additional 0.25 equivalent addition (CoCA + 0.5 eq) but the features at 20 ppm are lost at one equivalent addition and beyond. A more significant change occurs at two equivalent additions as a peak at approximate 75 ppm, corresponding to the NH proton of the three γ- or ε-NH His in active site41, is no longer present, indicated by asterisks. This change suggests a change in the active site with a shift from three γ- or ε-NH His coordinating the metal center to two γ- or ε-NH His. Though there is no signature peak of curcumin fully chelating the Co2+ active site, the removal of a His-coordination suggests curcumin does interact to some extent.

Figure 4-11 displays the 1H NMR spectra for CoCA with acac. The top spectrum was obtained with one equivalent of cobalt(II) and 5 equivalents of acac in the absence of protein. This spectrum represents acac coordinating to cobalt(II), and comparisons between spectra features can be made. It has peaks at approximately -5 ppm to 15 ppm (water), 20 ppm and 35 ppm that are unique to the coordination of acac to cobalt(II). These indicator peaks provide insight to the affinity of binding during the titration of ZBG to CoCA, allowing for direct monitoring of the formation of a CoCA-acac complex. The bottom spectrum shows resting CoCA, peaks at 50 ppm (δ-N3), and 75 ppm (δ-NH and ε-NH protons). Peaks near 20 ppm and 35 ppm begin to arise with the first addition, allowing for determination of acac coordinated to the cobalt metal center in CoCA.

The CoCA-acac peaks in the range of 50 to 100 ppm decline in intensity as more inhibitor is added. The CoCA peaks in the 1:5 CoCA:acac have a much lower intensity than the peaks in the 1:1 CoCA:acac spectrum. We suggest that acac is binding tightly to the enzyme metal center, and ultimately strips the metal from the active site of the enzyme.

Figure 4-12 shows the 1H NMR spectra for CoCA titrated with guaiacol. The top spectrum, which contains one equivalent of elemental cobalt(II) and 5 equivalents of guaiacol, represents guaiacol coordinated to cobalt(II). This spectrum contains the large water peaks found around -5 ppm to 15 ppm. This spectrum contains a broad peak around -75 ppm, which is a unique peak

93 that was used as a marker for the CoCA-guaiacol complex. Each of the other spectra represent CoCA with various amounts of guaiacol. The spectra include peaks around -5 ppm to 15 ppm (water), 50 ppm (δ-N3), and 75 ppm (δ-NH and ε-NH protons2). The middle three spectra, which contain CoCA with 0.5, 1, or 2 equivalents of guaiacol, each lack the peak around -75 ppm. This peak represents the marker for metal stripping by guaiacol. Shoulder peaks alongside the water peak, indicated by the asterisks at 0.5 equivalent spectra, these shoulders become more prominent as greater amounts of guaiacol are added. From this analysis, we conclude that guaiacol does not significantly bind to the metal center in CoCA at higher concentrations. This moiety could cause issues of biological retention upon ingestion and change binding affinity of the overall curcumin complex.

4.4.3 Electron paramagnetic resonance spectroscopy. The EPR spectra of CoCA and CoCA titrated with each ZBG (Figure 4-13) exhibit a broad, axial signal in the perpendicular spectra, B0^B1. Effective-g values can be found in Table 4-1. A peak at approximately g’ = 5.5 (1250 G) suggests a high-spin Co(II) metal center42. Upon the addition curcumin, effective g-values are not shifted, suggesting the coordination of the metal does not change in the metal center. This is consistent with the 4-coordinate model from UV- visible d-d band data. A decrease of peak intensities in this area upon the acac additions further suggest small changes in the effective values, however for those measured, are comparable to CoCA. The most evident change in line shape appears at the dip of the first intense peak. Relative consistency of the effective-g values of each ZBG with resting CoCA suggested a 4- coordinate binding with Co(II) metal centers

4.5 Conclusion. Two potential moieties for investigation are guaiacol and acac as a model for curcumin, a potential MMP inhibitor (Figure 4-1). Guaiacol (2-methoxyphenol) is a plant-derived molecule, as an analog of curcumin aryl ring structures. The phenol ring present in guaiacol possesses hydroxyl and methoxyl substitutions reflective of native curcumin. Acac (pentane-2,4- dione) is a tautomer and known metal chelator due to its ability to make hydrogen bonds43. This analog is reflective of the ketone moiety of curcumin between the two phenyl rings. These complexes

94 were studied using UV-visible, NMR, and EPR spectroscopies. Herein, we determined acac and guaiacol are 4-coordinate whereas curcumin is intermediate 4/5 coordinate. At higher concentrations, NMR elucidates acac and curcumin as binding tightly, while guaiacol is a weak binding moiety. Minimal changes above 1:1 stoichiometry is exhibited for guaiacol and acac, yet observed in curcumin, suggesting a weaker binding for curcumin or a shift of metal coordination at higher concentrations. From this binding evaluation, insight to the bioavailability can be understood due to the concentration dependence of curcumin coordination to CoCA.

4.7 References. [1] Tayyem, R. F., Heath, D. D., Al-Delaimy, W. K., and Rock, C. L. (2009) Curcumin content of turmeric and curry powders, Nutrition and cancer 55, 126-131. [2] Lechtenberg, M., Institute of Pharmaceutical Biology and Phytochemistry, H., D‐48149 Münster, Germany, Quandt, B., Institute of Pharmaceutical Biology and Phytochemistry, H., D‐48149 Münster, Germany, Nahrstedt, A., Institute of Pharmaceutical Biology and Phytochemistry, H., D‐48149 Münster, Germany, and Institute of Pharmaceutical Biology and Phytochemistry, H., D‐48149 Münster, Germany. (2016) Quantitative determination of curcuminoids in Curcuma rhizomes and rapid differentiation of Curcuma domestica Val. and Curcuma xanthorrhiza Roxb. by capillary electrophoresis, Phytochemical Analysis 15, 152-158. [3] Sharma, R. A., Gescher, A. J., and Steward, A. J. (2005) Curcumin: The story so far, European Journal of Cancer 41, 1955–1968. [4] Chuengsamarn, S., Rattanamongkolgul, S., Luechapudiporn, R., Phisalaphong, C., and Jirawatnotai, S. (2012) Curcumin extract for prevention of type 2 diabetes, Diabetes Care 35, 2121-2127. [5] Bhutani, M. K., Bishnoi, M., and Kulkarni, S. K. (2009) Anti-depressant like effect of curcumin and its combination with piperine in unpredictable chronic stress-induced behavioral, biochemical and neurochemical changes, Pharmacol Biochem Behav 92, 39- 43. [6] Shakeri, A., and Sahebkar, A. (2016) Optimized curcumin formulations for the treatment of Alzheimer's disease: A patent evaluation, Journal of Neuroscience Research 94, 111-113. [7] Moona, D.-O., Kima, M.-O., Choib, Y. H., Parkc, Y.-M., and Kima, G.-Y. (2010) Curcumin attenuates inflammatory response in IL-1β-induced human synovial fibroblasts and collagen-induced arthritis in mouse model, International Immunopharmacology 10, 605– 610. [8] Shakibaei, M., John, T., Schulze-Tanzil, G., Lehmann, I., and Mobasheri, A. (2007) Suppression of NF-κB activation by curcumin leads to inhibition of expression of cyclo- oxygenase-2 and matrix metalloproteinase-9 in human articular chondrocytes: Implications for the treatment of osteoarthritis, Biochemical Pharmacology 73, 1434– 1445.

95 [9] Singh, S., and Khar, A. (2006) Biological effects of curcumin and its role in cancer chemoprevention and therapy, Anti-Cancer Agents in Medicinal Chemistry 6, 259-270. [10] Chang, L. C., and Yu, Y. L. (2016) Dietary components as epigenetic-regulating agents against cancer, In Biomedicine (Taipei). [11] Duvoix, A., Blasius, R., Delhalle, S., Schnekenburger, M., Morceau, F., Henry, E., Dicato, M., and Diederich, M. (2005) Chemopreventive and therapeutic effects of curcumin, Cancer Letters 223, 181–190. [12] Kumar, D., Kumar, M., Saravanan, C., and Singh, S. K. (2012) Curcumin: a potential candidate for matrix metalloproteinase inhibitors, Expert opinion on therapeutic targets 16, 959-972. [13] Wahlstorm, B., and Blennow, G. (1978) A study on the fate of curcumin in the rat, Acta Pharmacologica et Toxicol ogica 43, 86-92. [14] Ravindranath, V., and Chandrasekhara, N. (1981) Metabolism of curcumin--studies with [3H]curcumin, Toxicology 22, 337-344. [15] Ravindranath, V., and Chandrasekhara, N. (1981) In vitro studies on the intestinal absorption of curcumin in rats, Toxicology 20, 251-257. [16] Yodkeeree, S., Chaiwangyen, W., Garbisa, S., and Limtrakul, P. (2009) Curcumin, demethoxycurcumin and bisdemethoxycurcumin differentially inhibit cancer cell invasion through the down-regulation of MMPs and uPA, J Nutr Biochem 20, 87-95. [17] Cheng, A. L., Hsu, C. H., Lin, J. K., Hsu, M. M., Ho, Y. F., Shen, T. S., Ko, J. Y., Lin, J. T., Lin, B. R., Ming-Shiang, W., Yu, H. S., Jee, S. H., Chen, G. S., Chen, T. M., Chen, C. A., Lai, M. K., Pu, Y. S., Pan, M. H., Wang, Y. J., T sai, C. C., and Hsieh, C. Y. (2001) Phase I clinical trial of curcumin, a chemopreventive agent, in patients with high-risk or pre-malignant lesions, Anticancer Research 21, 895-900. [18] Sharma, R. A., Euden, S. A., Platton, S. L., Cooke, D. N., Shafayat, A., Hewitt, H. R., Marczylo, T. H., Morgan, B., Hemingway, D., Plummer, S. M., Pirmohamed, M., Gescher, A. J., and Steward, W. P. (2004) Phase I clinical trial of oral curcumin: biomarkers of systemic activity and compliance, Clin Cancer Res 10, 6847-6854. [19] Garcea, G., Jones, D. J., Singh, R., Dennison, A. R., Farmer, P. B., Sharma, R. A., Steward, W. P., Gescher, A. J., and Berry, D. P. (2004) Detection of curcumin and its metabolites in hepatic tissue and portal blood of patients following oral administration, Br J Cancer 90, 1011-1015. [20] Shoji, M., Nakagawa, K., Watanabe, A., Tsuduki, T., Yamada, T., Kuwahara, S., Kimura, F., and Miyazawa, T. (2014) Comparison of the effects of curcumin and curcumin glucuronide in human hepatocellular carcinoma HepG2 cells, Food Chem 151, 126-132. [21] Chan, M. M., Huang, H. I., Fenton, M. R., and Fong, D. (1998) In vivo inhibition of nitric oxide synthase gene expression by curcumin, a cancer preventive natural product with anti-inflammatory properties, Biochem Pharmacol 55, 1955-1962. [22] Mohan, R., Sivak, J., Ashton, P., Russo, L. A., Pham, B. Q., Kasahara, N., Raizman, M. B., and Fini, M. E. (2000) Curcuminoids inhibit the angiogenic response stimulated by fibroblast growth factor-2, including expression of matrix metalloproteinase B, J Biol Chem 275, 10405-10412.

96 [23] Carvalho, D. d. M., Takeuchi, K. P., Geraldine, R. M., Moura, C. J. d., Torres, M. C. L., Carvalho, D. d. M., Takeuchi, K. P., Geraldine, R. M., Moura, C. J. d., and Torres, M. C. L. (2015) Production, solubility and antioxidant activity of curcumin nanosuspension, Food Sci. Technol (Campinas) 35, 115-119. [24] Krishnamurthy, V. M., Kaufman, G. K., Urbach, A. R., Gitlin, I., Gudiksen, K. L., and Whiteside, G. M. (2008) Carbonic anhydrase as a model for biophysical and physical- organic studies of proteins and protein-ligand binding, Chemical Reviews 108, 946–1051. [25] Martin, D. P., Hann, Z. S., and Cohen, S. M. (2013) Metalloprotein-inhibitor Binding: Human Carbonic Anhydrase II as a Model for Probing Metal-ligand Interactions in a Metalloprotein Active Site, Inorganic Chemistry 52, 12207−12215. [26] Scozzafava, A., and Supuran, C. T. (2000) Carbonic anhydrase and matrix metalloproteinase inhibitors: sulfonylated amino acid hydroxamates with MMP inhibitory properties act as efficient inhibitors of CA isozymes I, II, and IV, and N- hydroxysulfonamides inhibit both these zinc enzymes, Journal of Medicinal Chemistry 43, 10. [27] Bertini, I., Canti, G., Luchinat, C., and Scozzafava, A. (2002) Characterization of cobalt(II) bovine carbonic anhydrase and of its derivatives, Journal of the American Chemical Society 100, 4873-4877. [28] Garmer, D. R., and Krauss, M. (2002) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, Journal of the American Chemical society 115, 10247-10257. [29] Rowlett, R. S., Chance, M. R., Wirt, M. D., Sidelinger, D. E., Royal, J. R., Woodroffe, M., Wang, Y. F., Saha, R. P., and Lam, M. G. (1994) Kinetic and structural characterization of spinach carbonic anhydrase, Biochemistry 33, 13967-13976. [30] Rowlett, R. S. (2010) Structure and catalytic mechanism of the β-carbonic anhydrases, 1804, 362–373. [31] Bertini, I., Jonsson, B. H., Luchinat, C., Pierattelli, R., and Vila, A. J. (1994) Strategies of signal assignments in paramagnetic metalloproteins. An NMR investigation of the thiocyanate adduct of the cobalt (II)-substituted human carbonic anhydrase II, J Magn Reson B 104, 230-239. [32] Supuran, C. T. (2008) Carbonic anhydrases: novel therapeutic applications for inhibitors and activators, Nature Reviews Drug Discovery 7, 168-181. [33] Riley, E. A., Petros, A. K., Smith, K. A., Gibney, B. R., and Tierney,, and L., D. (2006) requency-Switching Inversion-Recovery for Severely Hyperfine Shifted NMR: Evidence of Asymmetric Electron Relaxation in High-Spin Co(II). Inorganic Chemistry 45, 10016-10018. [34] Garmer, D. R., and Krauss, M. (2002) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, J. Am. Chem. Soc. 115, 10247. [35] Llarrull, L. I., Tioni, M. F., Kowalski, J., Bennett, B., and Vila, A. J. (2007) Evidence for a dinuclear active site in the metallo-beta-lactamase BcII with substoichiometric Co(II). A new model for metal uptake, J Biol Chem 282, 30586-30595. [36] Miessler, G. L., Fischer, P. J., and A., T. (2013) Inorganic Chemistry 5ed., Pearson.

97 [37] Hawk, M. J., Breece, R. M., Hajdin, C. E., Bender, K. M., Hu, Z., Costello, A. L., Bennett, B., Tierney, D. L., and Crowder, M. W. (2009) Differential binding of Co(II) and Zn(II) to metallo-β-lactamase Bla2 from Bacillus anthracis, Journal of the American Chemical Society 131, 10753-10762. [38] Simpson, R. T., and Vallee, B. L. (1968) Two differentiable classes of metal atoms in alkaline phosphatase of Escherichia coli, Biochemistry 7, 4343-4350. [39] Yang, H., Makaroff, K., Paz, N., Aitha, M., Crowder, M. W., and Tierney, D. L. (2015) Metal ion dependence of the matrix metalloproteinase-1 mechanism, Biochemistry 54, 3631-3639. [40] Meng, F., Yang, H., Aitha, M., George, S., Tierney, D., and Crowder, M. (2016) Biochemical and spectroscopic characterization of the catalytic domain of MMP16 (cdMMP16) | SpringerLink, journal of biological inorganic chemistry 21, 523-535. [41] Jacobsen, F. E., Breece, R. M., Myers, W. K., Tierney, D. L., and Cohen, S. M. (2006) Model complexes of cobalt-substituted matrix metalloproteinases: Tools for Inhibitor Design, Inorganic Chemistry 45, 7306-7315. [42] Yang, H., Aitha, M., Marts, A. R., Hetrick, A., Bennett, B., Crowder, M. W., and Tierney, D. L. (2014) Spectroscopic and mechanistic studies of heterodimetallic forms of metallo- β-lactamase NDM-1., Journal of the American Chemical Society. 5/21/2014 136, 7273. [43] Camerman, A., Mastropaolo, D., and Camerman, N. (1983) Molecular structure of acetylacetone. A crystallographic determination, Journal of the American Chemistry society 105, 1584-1586.

98 4.8 Figures. A (a) OH O O O

Curcumin HO pK (a) ~ 7.8, (b) ~8.5, and (c) 9.0 OH (c) a (b) B C

O O OH O OH

O Acetylacetonate Guaiacol (acac) pKa ~ 9.0 pKa ~ 10.0

Figure 4-1. Structure of curcumin and moieties. A. Curcumin structure as tautomer and pKa 1 2 3 values . B. Acetylacetonate tautomer and pKa value . C. Guaiacol structure and pKa value .

99

Figure 4-2. UV-visible spectra of the titration of 760 µM cobalt(II) substituted carbonic anhydrase with curcumin. The pink line indicates CoCA with no equivalents of curcumin. The green line represents CoCA titrated with 2.0 equivalents of curcumin. The blue line represents CoCA titrated with 5.0 equivalents of curcumin. The gray lines indicate the titration of CoCA with curcumin equivalents ranging from 0.1 to 1.9 and 2.5. Inlay at top right is a zoom of larger graph, displaying d-d transitions of cobalt(II)-substituted carbonic anhydrase throughout the titration.

100

Figure 4-3. The change in absorbance for the d-d bands for titrated amounts of curcumin added to 760 µM CoCA. The vertical line indicates the concentration at which CoCA (760 µM) is equal to the concentration of curcumin titrated into solution. Red circles represent data obtained at 550 nm and the blue squares indicate 615 data obtained at nm. Data sets were fitted with pseudo Michaelis–Menten binding in order to determine the inhibitor binding affinity.

101

Figure 4-4. The change in absorbance for the substrate specific bands for titrated amounts of curcumin added to 760 µM CoCA. The vertical line indicates the concentration at which CoCA (760 µM) is equal to the concentration of curcumin titrated into solution. Red circles represent data obtained at 420 nm, the maximum light absorption wavelength of curcumin51. Data sets were fitted with pseudo Michaelis–Menten binding in order to determine the inhibitor binding affinity.

102

Figure 4-5. UV-visible spectra of the titration of 760 µM cobalt(II)-substituted carbonic anhydrase with acetylacetonate (acac). The pink line indicates CoCA with no equivalents of Acac. The green line represents CoCA titrated with 2.0 equivalents of acac. The blue line represents CoCA titrated with 5.0 equivalents of acac. The gray lines indicate the titration of CoCA with acac equivalents ranging from 0.1 to 1.9 and 2.5.

103

Figure 4-6. The change in absorbance for the d-d bands for titrated amounts of acac added to 760 µM CoCA. The vertical line indicates the concentration at which CoCA (760 µM) is equal to the concentration of acac titrated into solution. Red circles represent data obtained at 550 nm and the blue squares indicate 615 data obtained at nm. Data sets were fitted with pseudo Michaelis–Menten binding in order to determine the inhibitor binding affinity. (red and blue lines).

104

Figure 4-7: UV-visible spectra of the titration of 760 µM cobalt(II)-substituted carbonic anhydrase with guaiacol. The pink line indicates CoCA with no equivalents of guaiacol. The green line represents CoCA titrated with 2.0 equivalents of guaiacol. The blue line represents CoCA titrated with 5.0 equivalents of guaiacol. The gray lines indicate the titration of CoCA with guaiacol equivalents ranging from 0.1 to 1.9 and 2.5.

105

Figure 4-8: The change in absorbance for the d-d bands for titrated amounts of guaiacol added to 760 µM CoCA. The vertical line indicates the concentration at which CoCA (760 µM) is equal to the concentration of guaiacol titrated into solution. Red circles represent data obtained at 550 nm and the blue squares indicate 615 data obtained at nm. Data sets were fitted with pseudo Michaelis–Menten binding in order to determine the inhibitor binding affinity (red and blue lines).

106

Figure 4-9: The change in absorbance for the substrate specific bands for titrated amounts of guaiacol added to 760 µM CoCA. The vertical line indicates the concentration at which CoCA (760 µM) is equal to the concentration of guaiacol titrated into solution. Purple circles represent data obtained at 344 nm, the maximum light absorption wavelength of guaiacol. Data sets were fitted with pseudo Michaelis–Menten binding in order to determine the inhibitor binding affinity (blue dashed line).

107

Figure 4-10: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with curcumin. A cobalt(II)/ curcumin saturated standard is present as an internal reference.

108

Figure 4-11: 200 MHz 1H NMR spectra of the titration of cobalt(II) substituted carbonic anhydrase with acac. A cobalt(II)/ acac saturated standard is present as an internal reference. Asterisks indicate subtle changes in the equivalent addition that they first appear.

109

Figure 4-12: 200 MHz 1H NMR spectra of the titration of cobalt(II)-substituted carbonic anhydrase with guaiacol. A cobalt(II)/ guaiacol saturated standard is present as an internal reference. Asterisks indicate subtle changes in the equivalent addition that they first appear.

110

Figure 4-13. X-band EPR of perpendicular mode of CoCA, CoCA + 2 molar equivalence each inhibitor at specified attenuation. Black line indicates specified CoCA + zinc binding group. Gray line is CoCA with no zinc binding group added. Left- Full spectra, Right- Zoom on first effective g region. Peaks in the curcumin and acac spectra at approximately 3200 gauss arise from copper contaminants in the inhibitor prior to addition to CoCA

111

Table 4-1. Expected g-values for each perpendicular EPR spectra.

Curcumin Acac Guaiacol CoCA Maximum Peak 5.5 5.4 5.6 5.5 Minimum 3.3 3.2 3.2 3.3 Zero Crossing 1.3 1.5 1.3 1.4

O O O O

HO OH HO OH

O OH O O

Co2+ Co2+ N NH N N N N

N N N N

N N

A

B C D

OH2 O O OH Co2+ O N Co2+ N N N Co2+ N N N N N N N N N N N N N N

Figure 4-14: Proposed coordination of curcumin, guaiacol, and acac in comparison to resting enzyme binding. A) Binding modes of curcumin with CoCA. Left motif is proposed for equivalents below 2. Right movif reflects the His-coordination loss observed in 2 equivalent addition NMR of curcumin. B) Binding motif of guaiacol with CoCA. C) Binding motif of acac with CoCA. D) Binding motif of CoCA resting active site.

112 [1] Tonnesen, H. H., Masson, M., and Loftsson, T. (2002) Studies of curcumin and curcuminoids. XXVII. Cyclodextrin complexation: solubility, chemical and photochemical stability, International journal of pharmaceutics 244, 127-135. [2] Stary, J., and Liljenzin, J. O. (1982) Critical evaluation of equilibrium constants involving acetylacetone and its metal chelates, Pure and Applied Chemistry 54, 2557–2592. [3] Pearce, P. J., and Simkins, R. J. J. (1968) Acid strengths of some substituted picric acids, Canadian Journal of Chemistry 46, 241-248. [4] Sharma, R. A., Gescher, A. J., and Steward, A. J. (2005) Curcumin: The story so far, European Journal of Cancer 41, 1955–1968.

113 Chapter 5: Over-expression, purification, and characterization of the catalytic domain of MMP-12 (cdMMP-12)

114 Over-expression, purification, and characterization of the catalytic domain of MMP-12 (cdMMP-12)

Whitney Richert Craig, Sam George, Aaron Abraham, Michael W. Crowder and David L. Tierney

Department of Chemistry and Biochemistry, Miami University, 650 East High Street, Oxford, Ohio 45056, United States

Corresponding Author: Corresponding Authors *E-mail: [email protected]. Phone: (513) 529-8234.

Funding Source Statement. This work was supported by the National Institutes of Health (GM093987 and CHE-1509285) to M.W.C. and D.L.T.

Author contributions: WRC, SG, and AA contributed writing of manuscript; WRC and MWC contributed to the data analysis. WRC, AA, and SG contributed to the overexpression and purification. SG and AA contributed to the acquisition of MALDI-TOF.

Abbreviations. MMP-12, matrix metalloproteinase; cdMMP-12, catalytic domain of matrix metalloproteinase 12; DTT, 1,4-dithiothreitol; Hepes, 4-(2-hydroxyethyl)-1-piperazineethane- sulfonic acid; IPTG, isopropyl-β-D-thiogalactoside; MMP, matrix metalloproteinases; SDS- PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; Tris, Tris(hydroxymethyl)aminomethane; EDTA-ethylene-diamine-tetraacetic acid; MCA, (7- methoxycoumarin-4-yl) acetic acid.

115 5.1 Abstract. Matrix metalloproteinases 12 (MMP-12), macrophage elastase, is an attractive drug target due to its role in multiple tissue remodeling activities. When these are misregulated, MMPs are implicated in disease processes such as emphysema, cancer, and inflammation. Herein, we report the over-expression of the catalytic domain of MMP-12 (cdMMP-12) in Escherichia coli, as well as purification and refolding. Structural and kinetic characterization of MMP-12 was investigated, and ICP-AES was conducted in order to assess metal content in cdMMP-12 samples. MALDI-TOF mass spectrometry was conducted to determine the molecular weight of cdMMP-12 and to verify the described over-expression and purification protocol.

Keywords. Matrix metalloproteinases; Over-expression; Purification; Refolding

5.2 Introduction. Matrix metalloproteinases, or MMP’s, are a family of zinc-dependent endopeptidases1, 2 that are most known for the maintenance and degradation of the extracellular matrix (ECM) and its various components 3, 4. In addition, MMP’s play a vital role in all stages of life, encompassing functions ranging from angiogenesis, mammary gland ductal branching, innate immune defense, and wound healing and repair5. Due to the variety of functions and large possible amounts of substrates, MMP transcription and translation is a tightly regulated process, as demonstrated in the evolutionarily conserved “cysteine switch” mechanism that maintains enzyme latency6. Numerous studies have shown that unregulated over-expression of MMP’s has been implicated in the accelerated breakdown of connective tissue associated with cancer metastasis and tumorigenesis4, 7, 8. This biological implication makes MMP’s a potential anti-cancer target and calls forth the need for specific inhibitors. MMP inhibitors, therefore, must be specific to target only one MMP in order to minimize side effects. Unfortunately, many proposed MMP inhibitors have failed during clinical trials9-11 due to the lack of specificity. Detailed information regarding MMP’s from each class is necessary in order to achieve the task of finding effective inhibitors.1

Structurally, MMP’s have the same basic structure consisting of a N-terminal propeptide domain, a catalytic domain, and a C-terminal hemopexin domain12. The propeptide domain helps protect the zinc coordination sphere via a cysteine in the pro-domain. Upon activation, the coordination

116 between the cysteine and Zn(II) catalytic site is cleaved6, 13. Specifically in looking at the catalytic domain, the polypeptide chain folds are nearly identical across all MMP’s. In looking at the active site of MMP’s, the catalytic zinc is coordinated by three histidine residues followed by a conserved methionine residue and an axial water molecule used for substrate hydrolysis. In the zymogen form of MMP’s, the catalytic zinc also coordinates with a cysteine. This polypeptide chain consists of a 5-stranded β-sheet, three α-helices, and connecting random loops. Furthermore, the active site binds 2 equivalents of zinc, and typically 2-3 equivalents of calcium14. The catalytic zinc is adjacent to three pockets on both the left (S1, S2, S3) and right (S1', S2', S3'). The three pockets on the right are of primary interest for zinc binding group inhibition and targeting due to variability in the S1' pocket among the various MMP’s 6. The hemopexin domain structure of 4 repeating loops is conserved among all MMP’s and is thought to be involved in protein-protein interactions, substrate specificity, and full cell migratory function in some MMP’s, such as MMP-915, 16. MMP-12 has a unique proclivity to cleave the hemopexin domain in order to result in a mature and active protein consisting of the catalytic domain12, 17, 18.

MMP-12, which belongs to the metalloelastase family, localizes on macrophages (especially in the alveoli) and chondrocytes19, 20. MMP-12 has been shown to be involved in antitumorigenic, anti-inflammatory, and antibacterial activities, as well as innate immunity21-24. MMP-12 is also noted for its ability to degrade a broad spectrum of ECM components, predominately soluble and insoluble in lung connective tissue25. In addition, MMP-12 is critical for macrophage entry during anti-inflammatory responses. Up-regulated MMP-12 activity has been found in colon cancer, epithelial cancer, lung cancer, and breast cancer20, 26-28. Almost paradoxically, however, MMP-12 has been shown to lead to a decrease in tumor growth via cleaving plasminogen to angiostatin to stop angiogenesis29.

The crystal structure of cdMMP-12 with both Zn and Ca equivalents has been reported, as shown in Figure 5-1. In regards to the substrate binding pockets of cdMMP-12, the S1' pocket is characterized as large and open, and closely resembles that of MMP-8. MMP-12 is unique in that the S1' pocket contains a polar Thr215 residue (which allows for the binding of polar substrates), compared to the typical hydrophobic Val or Ala residue present in other MMP’s6. In this paper,

117 we described a detailed study encompassing the over-expression and purification of cdMMP-12, along with metal content analysis of the enzyme.

5.3 Experimental procedure. 5.3.1 Materials. Luria-Broth medium (LB), Isopropyl-β-D-thiogalactoside (IPTG), Tris (hydroxymethyl) aminomethane (Tris), 4-(2-hydroxyethyl)-1-pipera-zineethanesulfonic acid (Hepes), urea, NaCl, guanidine HCl, dithiothreitol (DTT), Triton X-100, and CaCl2 were purchased from Fisher Scientific (Hampton, New Hampshire). Roche cOmplete™ Inhibitor Cocktail EASYpacks were purchased from Sigma Aldrich. E. coli BL21(DE3) Tuner cells were purchased from Novagen (Madison, WI, USA). Barnstead NANOpure water (18.1 MΩ) was used to prepare all buffer solutions. Protein samples were concentrated using centrifugation units purchased from EMD Millipore (Billerica, MA, USA). Q-Sepharose resin was used to purify cdMMP-12 from GE Healthcare Biosciences (Marlborough, MA, USA).

5.3.2 Methods. 5.3.2.1 Cloning, over-expression, and inclusion body preparation of MMP-12. The over-expression plasmid, which codes for the catalytic domain of MMP-12 (amino acid residues 106-267), was purchased from GenScript (Piscataway, NJ) with restriction sites for NdeI and BamHI at the 5' and 3' ends, respectively. The cdMMP-12 sequence contained mutations previously found to increase solubility30. The DNA sequence was verified by sequencing. BL21(DE3) Tuner E. coli cells were transformed with pET26b/cdMMP-12. Cells were then plated onto Luria-Bertani (LB) agar plates containing 25 µg/mL of kanamycin (kan). A single colony was used to inoculate a 50 mL starter culture of LB-Kan. The starter culture was shaken overnight at 37 ºC and 100 rpm. Approximately 10 mL of the culture was transferred to 1 L of LB-Kan; the culture was allowed to shake at 37 ºC and 100 rpm until the culture reached an

OD600nm of 0.8, approximately 3 hours. One milliliter of 1 M IPTG was added to a final concertation of 1 mM, and protein production was allowed to occur at 37 ºC for 3-4 hr. The resulting E. coli cells were collected via centrifugation at 6,000 rpm for 20 minutes. Unless otherwise specified, all steps to solubilize inclusion bodies were performed at 4 ºC. Cells were brought to 80 mL total volume with 100 mM Tris, pH 8.0. Two protease inhibitor tablets were

118 added, and the mixture was incubated on ice for 10 minutes. DNAse (final concentration of 5

µg/mL) and buffer containing 10 mM Tris pH 8.0 and MgCl2 (final concentration of 10 µM) were added, and the mixture was homogenized and centrifuged at 15,000 rpm for 30 minutes. The cell pellet was resuspended in 80 mL of 0.1% Triton X-100 (v/v) and centrifuged at 15,000 rpm for 20 minutes. The resulting cell pellet was resuspended in 80 mL of triply-distilled H2O and centrifuged at 15,000 rpm for 20 minutes. The supernatant was discarded. The resulting inclusion bodies were resuspended in 10 mL of triply-distilled H2O and 5 mL of butan-1-ol and centrifuged at 12,000 rpm for 20 minutes. Inclusion bodies were rapidly frozen using liquid nitrogen and stored at -80 ºC unless purification occurred. SDS-PAGE was used to evaluate the purity of the washing fractions as well as over-expression.

5.3.2.2 Purification and refolding of cdMMP-12. The inclusion bodies were solubilized in 5 mL of 8 M guanidine-HCl, 50 mM Tris pH 8.0, and 30 mM DTT. The mixture was incubated at 37 ºC for 60-90 minutes. Insolubilized matter was removed by centrifugation at 12,000 rpm for 20 minute,s and the solution then filtered through a 0.2 micron filter. The solubilized filtrate mixture was dialyzed and diluted in a two-step process; a first dialysis against refolding buffer A comprised of 3 M urea, 50 mM Hepes, 10 mM CaCl2, 30 mM NaCl, 0.1 mM zinc acetate, pH 7.2, followed by a second dialysis against cation exchange buffer 1 comprised of 3 M urea, 10 mM Hepes, 2 mM CaCl2, pH 7.0. All dialysis was conducted using eight volumes of buffer to filtrate and performed at 4 ºC for 24 hours, with buffer changes three times daily. The resulting dialyzed protein was incubated with rotation in 6 mL bead volume of Q-Sepharose resin (GE Healthcare) for 60 minutes at 4 ºC. The bound protein-resin mixture was collected using centrifugation at 6,000 rpm for 15 minutes. The resin was washed once with cation exchange buffer 2 comprised of 2 M urea, 10 mM Hepes, and 2 mM CaCl2, pH 7.0, once with cation exchange buffer 3 comprised of 1 M urea, 10 mM Hepes, and 2 mM CaCl2, pH 7.0, and twice with cation exchange buffer 4 comprised of 10 mM Hepes, and 2 mM CaCl2, pH 7.0. Washes were conducted with 10 bed volumes of buffer and incubated at 4 ºC with rotation for 10 minutes. cdMMP-12 protein was eluted with elution buffer comprised of 10 mM Hepes, 2 mM CaCl2, and 500 mM NaCl, pH 7.0. Five elutions of 8 mL were collected and pooled. cdMMP-12 was concentrated to approximately 40 mL using an Amicon fitted with a YM10 membrane and stored at -20 ºC until characterized.

119

5.3.2.3 Characterization of purified cdMMP-12 using SDS-PAGE and MALDI-TOF MS. Purity of the refolded protein was determined via MALDI-TOF mass spectrometry using a Bruker AutoFlex MALDI-TOF mass spectrometer. cdMMP-12 was mixed with 5 mg/mL 2- cyano-4-hydroxycinnamic acid (HCCA) in volume ratio of 1:5, spotted, and analyzed using a Bruker AutoFlex MALDI-TOF mass spectrometer.

5.3.2.4 Metal analyses of cdMMP-12. The metal content of cdMMP-12 samples was determined using a Perkin Elmer Optima 7300DV Inductively-Couple Plasma with Atomic Emission Spectrometer. Protein samples were diluted to a final concentration 1-3 µM with 1% acidified triply-distilled water (H2SO4). Calibration curves were generated using several dilutions of metal standards (Inorganic Ventures – Zn and Co) ranging from 0.008 to 1 µM (5 ppb to 500 ppb). The emission wavelengths were set to 213.8 nm and 238.9 nm for Zn and Co, respectively.

5.4 Results and discussion. The gene for cdMMP-12 (amino acid span 106-26430) was ligated into pET26b. Native construct MMP12 was reported to be safely concentrated to approximately 250 µM. The construct used herein contained two point mutations and a shortening of the amino acid span for the catalytic domain that were reported to increase the expression yield and solubility of MMP12. These mutations consisted of two Met residues at 104 and 105 that were introduced prior to Gly106, thus removing portion of the linker to the prodomain reducing autoproteolysis30. Phe171 is a residue in a largely hydrophobic region contributing significantly to the lack of solubility. Through computer simulations, the mutation to Asp171 resulted in comparable fluorogenic substrate turnover and a four-fold increase in solubility30. The mutation to Asp171 was included in the construct utilized herein. The removal of residues 264-267 eliminated a portion of the linker domain that was largely disordered and was therefore removed to increase the solubility of MMP12.

This construct resulted in a plasmid called cdMMP-12/pET26B, which was transformed into BL21(DE3) Tuner E. coli cells. Cells were cultured at 37 ºC, and protein production was induced

120 via IPTG addition as described in “Methods”. Previously, Parkar et al. described a procedure to over-express and purify cdMMP-1231. This method involved over-expression of the enzyme in a BL21(DE3) E. coli system using a pet21M over-expression plasmid. Using this method, cdMMP-12 in a BL21(DE3) Tuner E. coli system using a pet26b vector was processed into inclusion bodies, the enzyme was solubilized using a multi-step process, purified through gel filtration column chromatography, refolded via dialysis, and further purified through cation exchange chromatography via batch binding and centrifugation. This published method was used, however, resulted in low yield of cdMMP-12. Optimizations were made to the procedure using the following steps: (1) refolding the enzyme via dialysis following solubilization of the inclusion bodies, thus eliminating size exclusion chromatography where large quantities of cdMMP-12 from over-expression were lost due to degradation, (2) Cation exchange chromatography modified to use gravity filtration rather than the centrifugation method as previously used. Further, during cation exchange chromatography, (3) the NaCl concentration in the elution buffer was increased from 250 mM to 500 mM, and (4) flow rate during gel filtration column chromatography was decreased from 1 mL/minutes to 0.1 mL/minutes. Following these modifications, cdMMP-12 was processed into inclusion bodies, which were solubilized using a multi-step buffer exchange procedure. Soluble cdMMP-12 was obtained by multiple dialysis steps, as described in “Materials” (Figure 5-2). This yield was improved compared to previous over-expression with unmodified cdMMP-12 (amino acid span 106-264) but not to the levels described by Banci et al.

The enzyme was purified using a Q-Sepharose column and analyzed via SDS-PAGE. Protein bands, including bands with the molecular weight corresponding to cdMMP-12, were present (Figure 5-3). Purification of refolded cdMMP-12 was attempted on a 20 mL Q-sepharose column, which was pre-equilibrated with 10 mM MES, pH 7.0. Protein was eluted from the column using a linear gradient of cation exchange buffer 4 (10 mM Hepes, 2 mM CaCl2, 500 mM NaCl, pH 7.0). After analysis with UV-visible spectrophotometry (A280 nm) and SDS- PAGE (Figure 5-3), it was determined that while cdMMP-12 was eluted, other contaminants were also eluted. Further optimization of the protocol is imperative in order to improve the quantity and purity of cdMMP-12 produced, as current methods do not provide the means by which to pursue spectroscopic studies. Possible avenues include changing the salt gradient of the

121 Q-Sepharose column, shifting to an alternate ion exchange column and altering the pH of the exchange buffers accordingly, or using affinity-based chromatographic methods. This purification procedure described herein yielded approximately 0.5 -1 mg soluble cdMMP-12 per liter growth medium, whereas the unmodified method did not yield measurable quantities of purified cdMMP-12. Expression of MMPs in E.coli systems poses a number of challenges as discussed in Chapter 5 of this dissertation. Namely, their propensity to be cytotoxic and largely insoluble, thus placed into inclusion bodies. Further, MMPs are susceptible to autoproteolysis. Issues with purification prevented full optimization of large-scale over-expression of MMP-12.

The mass spectrum of the suspected cdMMP-12 band confirmed the presence of multiple proteins being shown on the gel. This MALDI-TOF MS showed a dominant peak at 4370 m/z. Also observed were numerous peaks below 1060 m/z, suggesting an impure product (Figure 5- 4). Prominent theoretical peaks of the trypsin digestion of cdMMP-12, as determined via Expasy’s Peptide Cutter tool, were not identified above 2:1 signal to noise baseline levels, presenting a further need to adjust the purification methods used herein. Despite modifications made to the DNA sequence utilized for increased solubility and ativity30 and high overexpression levels, purification is not effective to produce quantities of cdMMP-12 as desired.

ICP-AES revealed that cdMMP-12 contains 1.82 ± 0.01 equivalents of Zn(II). The content of Co(II) was determined to be less than 0.1 ± 0.01 equivalents. Attempts to cobalt-substitute native 28, 29 Zn2 – cdMMP-12 through exhaustive dialysis procedures were largely unsuccessful, as the apo-cdMMP-12 created in both procedures was unable to fold around the excess Co (Table 5-1).

5.5 Conclusions. It is apparent that MMPs represent important potential drug targets. Unfortunately, MMPs tend to be difficult enzymes to work with, as noted by instability and being over-expressed in insoluble inclusion bodies. This work describes efforts to attempt to overcome these challenges and prepare a stable analog of cdMMP-12. In order to produce a stable enzyme that can be used in inhibition studies, an alternate purification method is necessary. In addition, Co(II)-substituted analogs (both Co2 and ZnCo) will have to be prepared to spectroscopically probe the active site of cdMMP-12 and allow for a deeper understanding the binding kinetics and the mechanism of

122 the enzyme. It has been shown previously that a Co-substituted analog of cdMMP-1632 and could be used to gain insight into procedures for Co substitution of MMP’s did not yield successful results. We continue to make strides that yield quantities of purified cdMMP-12 can be used to investigate steady-state and stopped-flow kinetics, apo-protein metal stability, circular dichroism spectroscopy, fluorescence spectra, UV-visible spectroscopy of metal binding groups, NMR spectroscopy of Zn(II)Co(II) and Co(II)2 substituted protein, EPR of Zn(II)Co(II) and Co(II)2 substituted protein, IC50 assays of metal binding groups, and ITC binding studies. We believe that with this detailed information, comprehension of metal binding can be used to inform the creation of a cdMMP-12 specific inhibitor.

5.6 Acknowledgements. The work was supported with funds from the National Institute of Health, Miami University’s Undergraduate Summer Scholars Program, as well as Miami University’s Department of Chemistry and Biochemistry Hughes Intern Summer Program.

5.7 References. [1] Nagase, H., Visse, R., and Murphy, G. (2006) Structure and function of matrix metalloproteinases and TIMPs, Cardiovasc Res 69, 562-573. [2] Gross, J., and Lapiere, C. M. (1962) Collagenolytic activity in amphibian tissues: a tissue culture assay, Proc Natl Acad Sci U S A 48, 1014-1022. [3] Egeblad, M., and Werb, Z. (2002) New functions for the matrix metalloproteinases in cancer progression, Nat Rev Cancer 2, 161-174. [4] Hadler-Olsen, E., Winberg, J. O., and Uhlin-Hansen, L. (2013) Matrix metalloproteinases in cancer: their value as diagnostic and prognostic markers and therapeutic targets, Tumour Biol 34, 2041-2051. [5] Rodriguez, D., Morrison, C. J., and Overall, C. M. (2010) Matrix metalloproteinases: what do they not do? New substrates and biological roles identified by murine models and proteomics, Biochim Biophys Acta 1803, 39-54. [6] Jacobsen, J. A., Major Jourden, J. L., Miller, M. T., and Cohen, S. M. (2010) To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition, Biochim. Biophys. Acta 1803, 72-94. [7] Kessenbrock, K., Plaks, V., and Werb, Z. (2010) Matrix metalloproteinases: regulators of the tumor microenvironment, Cell 141, 52-67. [8] Deryugina, E. I., and Quigley, J. P. (2006) Matrix metalloproteinases and tumor metastasis, Cancer Metastasis Rev 25, 9-34. [9] Cathcart, J. M., and Cao, J. (2015) MMP Inhibitors: Past, present and future, Front Biosci (Landmark Ed) 20, 1164-1178.

123 [10] Mak, I. W., Evaniew, N., and Ghert, M. (2014) Lost in translation: animal models and clinical trials in cancer treatment, Am J Transl Res 6, 114-118. [11] Zucker, S., and Cao, J. (2009) Selective matrix metalloproteinase (MMP) inhibitors in cancer therapy: ready for prime time?, Cancer Biol Ther 8, 2371-2373. [12] Shapiro, S. D., Kobayashi, D. K., and Ley, T. J. (1993) Cloning and characterization of a unique elastolytic metalloproteinase produced by human alveolar macrophages, J Biol Chem 268, 23824-23829. [13] Vandenbroucke, R. E., and Libert, C. (2014) Is there new hope for therapeutic matrix metalloproteinase inhibition?, Nat Rev Drug Discov 13, 904-927. [14] Visse, R., and Nagase, H. (2003) Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function, and biochemistry, Circ Res 92, 827-839. [15] Glasheen, B. M., Kabra, A. T., and Page-McCaw, A. (2009) Distinct functions for the catalytic and hemopexin domains of a Drosophila matrix metalloproteinase, Proc Natl Acad Sci U S A 106, 2659-2664. [16] Dufour, A., Sampson, N. S., Zucker, S., and Cao, J. (2008) Role of the hemopexin domain of matrix metalloproteinases in cell migration, J Cell Physiol 217, 643-651. [17] Belaaouaj, A., Shipley, J. M., Kobayashi, D. K., Zimonjic, D. B., Popescu, N., Silverman, G. A., and Shapiro, S. D. (1995) Human macrophage metalloelastase. Genomic organization, chromosomal location, gene linkage, and tissue-specific expression, J Biol Chem 270, 14568-14575. [18] Shapiro, S. D., Griffin, G. L., Gilbert, D. J., Jenkins, N. A., Copeland, N. G., Welgus, H. G., Senior, R. M., and Ley, T. J. (1992) Molecular cloning, chromosomal localization, and bacterial expression of a murine macrophage metalloelastase, J Biol Chem 267, 4664- 4671. [19] Babusyte, A., Stravinskaite, K., Jeroch, J., Lotvall, J., Sakalauskas, R., and Sitkauskiene, B. (2007) Patterns of airway inflammation and MMP-12 expression in smokers and ex- smokers with COPD, Respir Res 8, 81. [20] Kerkela, E., Bohling, T., Herva, R., Uria, J. A., and Saarialho-Kere, U. (2001) Human macrophage metalloelastase (MMP-12) expression is induced in chondrocytes during fetal development and malignant transformation, Bone 29, 487-493. [21] Houghton, A. M., Grisolano, J. L., Baumann, M. L., Kobayashi, D. K., Hautamaki, R. D., Nehring, L. C., Cornelius, L. A., and Shapiro, S. D. (2006) Macrophage elastase (matrix metalloproteinase-12) suppresses growth of lung metastases, Cancer Res 66, 6149-6155. [22] Houghton, A. M., Hartzell, W. O., Robbins, C. S., Gomis-Ruth, F. X., and Shapiro, S. D. (2009) Macrophage elastase kills bacteria within murine macrophages, Nature 460, 637- 641. [23] Parks, W. C., Wilson, C. L., and Lopez-Boado, Y. S. (2004) Matrix metalloproteinases as modulators of inflammation and innate immunity, Nat Rev Immunol 4, 617-629. [24] Yamada, S., Wang, K. Y., Tanimoto, A., Fan, J., Shimajiri, S., Kitajima, S., Morimoto, M., Tsutsui, M., Watanabe, T., Yasumoto, K., and Sasaguri, Y. (2008) Matrix metalloproteinase 12 accelerates the initiation of atherosclerosis and stimulates the progression of fatty streaks to fibrous plaques in transgenic rabbits, Am J Pathol 172, 1419-1429.

124 [25] Vaalamo, M., Karjalainen-Lindsberg, M. L., Puolakkainen, P., Kere, J., and Saarialho-Kere, U. (1998) Distinct expression profiles of stromelysin-2 (MMP-10), collagenase-3 (MMP- 13), macrophage metalloelastase (MMP-12), and tissue inhibitor of metalloproteinases-3 (TIMP-3) in intestinal ulcerations, Am J Pathol 152, 1005-1014. [26] Klupp, F., Neumann, L., Kahlert, C., Diers, J., Halama, N., Franz, C., Schmidt, T., Koch, M., Weitz, J., Schneider, M., and Ulrich, A. (2016) Serum MMP7, MMP10 and MMP12 level as negative prognostic markers in colon cancer patients, BMC Cancer 16, 494. [27] Kerkela, E., Ala-Aho, R., Jeskanen, L., Rechardt, O., Grenman, R., Shapiro, S. D., Kahari, V. M., and Saarialho-Kere, U. (2000) Expression of human macrophage metalloelastase (MMP-12) by tumor cells in skin cancer, J Invest Dermatol 114, 1113-1119. [28] Decock, J., Thirkettle, S., Wagstaff, L., and Edwards, D. R. (2011) Matrix metalloproteinases: protective roles in cancer, J Cell Mol Med 15, 1254-1265. [29] Impola, U., Uitto, V. J., Hietanen, J., Hakkinen, L., Zhang, L., Larjava, H., Isaka, K., and Saarialho-Kere, U. (2004) Differential expression of matrilysin-1 (MMP-7), 92 kD gelatinase (MMP-9), and metalloelastase (MMP-12) in oral verrucous and squamous cell cancer, J Pathol 202, 14-22. [30] Banci, L., Bertini, I., Ciulli, Alessio, Fragai, M., and Ternia, B. (2003) Expression and high yield production of the catalytic domain of matrix metalloproteinase 12 and of an active mutant with increased solubility, Journal of Molecular Catalysis A: Chemical Volumes 204–205, 401–408. [31] Parkar, A. A., Stow, M. D., Smith, K., Panicker, A. K., Guilloteau, J. P., Jupp, R., and Crowe, S. J. (2000) Large-scale expression, refolding, and purification of the catalytic domain of human macrophage metalloelastase (MMP-12) in Escherichia coli, Protein Expr Purif 20, 152-161. [32] Meng, F., Yang, H., Aitha, M., George, S., Tierney, D., and Crowder, M. (2016) Biochemical and spectroscopic characterization of the catalytic domain of MMP16 (cdMMP16) | SpringerLink, journal of biological inorganic chemistry 21, 523-535.

125 5.8 Figures

106 Figure 5-1. Crystal structure of Zn2-cdMMP-12 using the coordinates (PDB 3NX7) . The central grey atom represents the catalytic zinc, whereas the grey atom on the right represents the structural zinc. The black atoms represent the two structural calcium atoms.

126 1 2 3 4 5 6 7 8 9 10

118 85 47 36 26

18

Figure 5-2. Overexpression and refolding of cdMMP-12 using SDS-PAGE. Lane 1. Perfect protein molecular weight markers, with weights in kilodaltons. Lane 2. Pre-induction sample. Lanes 3-6. Post-induction samples taken 1, 2, 3, 4 hours after induction. Lanes 7-9. Supernatants (S1, S2, S3) taken after inclusion body solubilization buffer washing. Lane 10. Solubilized inclusion bodies. The presence of a band at higher molecular weight suggests dimerization of cdMMP-12.

127 1 2 3 4

M W

250 150

100 75

50 25

Figure 5-3. Q-Sepharose ion exchange chromatographic purification of cdMMP-12 using SDS-PAGE. Lane 1. Perfect protein molecular weight marker, with weights in kilodaltons. Lane 2. Fraction 1 from protein elution. Lane 3. Concentrated Fractions 38+39 from protein elution. Lane 4. Concentrated Fractions 40+41 from protein elution. Lane 5. Fraction 26 from protein elution. The red box in Lanes 3 and 4 represent eluted cdMMP-12.

128

Figure 5-4. MALDI-TOF mass spectrum of trypsin digested cdMMP-12. Bands were cut from fractions from Q-Sepharose purification ran on the SDS-PAGE gel shown in Figure 5-3.

Table 5-1. Metal content of cdMMP-12 samples.

Protein/Metal equiv. Zn Protein/Metal equiv. Co

Zn2 – cdMMP-12 1.82 <0.1

Co2 – cdMMP-12 1.80 <0.1

129 Chapter 6: Concluding Remarks

130 Conclusion remarks

Whitney Richert Craig

Department of Chemistry and Biochemistry, Miami University, 650 East High Street, Oxford, Ohio 45056, United States

Abbreviations cdMMP-12, catalytic domain of matrix metalloproteinase 12; MMP, matrix metalloproteinases; MMP-12, matrix metalloproteinase; ZBG, zinc binding group; TPMA, thiopyromechonic acid; TM, thiomaltol; ATM, allothiomaltol; B12, 5-chloro-8-quinolinol; B2, quinolinecarboxylic acid; D5, 3-hydroxypyridine-2(1H)-thione; G11, 2-hydroxy-2,4,6- cycolheptatrienone; acac, pentane-2,4-dione.

131 6.1 Conclusions Matrix metalloproteinases (MMPs) are a class of calcium and zinc-containing that have been associated with multiple pathological conditions such as arthritis, inflammation, connective tissue degradation, and cancer progression8. Due to the disease implications, MMPs are an attractive target for inhibitor studies9. The main therapeutic strategy to treat this misregulation is through the design of drugs that target the MMP catalytic domain. Many efforts have been made to design an inhibitor; however, they target over-expressed MMPs and MMPs functioning normally in tissue degradation, signaling, and other physiological processes. The majority of drug inhibitors designed to date have failed in clinical trials due to toxicity and lack of binding specificity to a single MMP subclass10-12. By investigating MMP-12 and model complex binding mechanisms via the catalytic Zn(II) center and Co(II)-substituted analogs, a detailed understanding of how the coordination of a zinc binding group (ZBG) inhibitor influences or alters the active site can be gained.

The first set of ZBGs are thiol derivatives (Chapter 2) thiopyromechonic acid (TPMA) base with two methyl substitution additions (thiomaltol (TM) and allothiomaltol (ATM)). These additions are utilized to probe binding orientation and understand the electrostatic interactions that govern binding modes within the active site. The resulting zinc binding affinity ranged from micromolar to millimolar for human carbonic anhydrase (CA-II), with ATM exhibiting the tightest association with the metal center. The presence of a methyl group on the same face of the molecule ring structure as the hydroxide is hypothesized to aid in the tighter affinity. Removal of methyl substitution, seen in TPMA, lead to a decrease in CA inhibitory properties. The monodentate binding mode for TM and TPMA, as well as bidentate binding mode for ATM to the cobalt ion in the CA-II active site is proposed herein. This coordination occurs predominately through the sulfur atom on the molecule. These molecules have been screened against multiple MMPs13-15. Going forward, further substitutions must be added to these molecules, particularly ATM, to probe specificity to MMPs.

Cohen Library Fragments screened against CA-II (Chapter 3) include quinoloinol, carboxylic acid, pyridine, and tropolone derivatives. Each fragment showed promise when screened against MMP-12 using DFT calculations16. The poor solubility of 5-chloro-8-quinolinol (B12) requires

132 the need of further investigation, whether with alterations to the structure of the molecule in order to improve solubility or non-biological conditions, in order to be considered as a candidate for MMP-12. Quinolinecarboxylic acid (B2) was determined to be an ineffective fragment as well due to its metal removal properties. Structural changes could be explored to maintain tight binding to the metal center, which is attractive when designing a therapeutic target and decrease the chelation properties. Removal or substitution of a hydroxyl group could greatly change the ability to chelate. The thione moiety present in 3-hydroxypyridine-2(1H)-thione (D5) could be successful for condition where D5 could bind similar to other thione compounds with lower pKa values13, 17. The monodentate binding mode encourages further investigation and therefore a molecule that could be considered for the basis of an inhibitor design. Drastic spectral changes for compound G11, 2-hydroxy-2,4,6-cycolheptatrienone, suggests the highest promise as a therapeutic fragment within this group. The 5-coordinate binding system is highly advantageous to utilize as binding strength is an issue for specificity of molecules screened against the MMP family. Steady-state kinetics and IC50 measurements against each MMP subgroup would provide fruitful insight to affinity differences to inform selectivity efforts. Further, X-ray crystallography and EXAFS would probe binding further and should be pursued.

A shift to investigate a plant derived inhibitor and two potential substituent moieties fragments, guaiacol and pentane-2,4-dione (acac), were used are as a model for curcumin, a potential MMP inhibitor (Chapter 4). By investigating molecules that are already ingestible, toxicity to the system is already decreased. From this binding evaluation, active site coordination chemistry was elucidated. It was determined that acac and guaiacol are 4-coordinate whereas curcumin is intermediate 4/5 coordinate. At higher concentrations, acac and curcumin are determined to bind tightly to the cobalt center whereas guaiacol is a weak binding moiety. Monitoring of binding also suggests curcumin binds the metal center with a low binding affinity when compared to acac. Future investigation of curcumin is necessary in order to understand the binding affinity with MMP. Though much speculation has been made for curcumin or a curcumin derivative to be an effective inhibitor for MMPs18, screening must be conducted against MMP-12 and other MMPs to determine the extent of binding. The active site pocket of MMPs are considerably more shallow and rigid than CA and binding affinity to MMPs must be investigated first with methods similar to those in Chapter 4. Further, off-target binding would need to be remedied due to the

133 information presented here with CA-II. Specificity would need to be addressed to fully optimize inhibition of a specific MMP.

Purification attempts of MMP-12 (Chapter 5) have been conducted with challenges to prepare a stable analog of cdMMP-12. MMPs tend to be difficult enzymes to work with issues including instability, autoproteolysis, and loss of protein from insoluble inclusion body purification. Many efforts herein have been made to overcome these issues with marginal success. Mutations for increased expression and solubility19 were utilized in this construct; however, the purification performed by Banci et al.19, Parkar et al.11, and adaptations did not produce stable MMP-12 that can be used in inhibition studies. Due to the low yield of purified cdMMP-12, an alternate purification method is necessary. Changes in ion exchange column/pH combination should be explored during purification as should the incorporation of acetohydroxamic acid, a mild inhibitor of MMPs with dissociation constant around 20 mM, to reduce autoproteolysis19. We continue to make strides that yield quantities of purified cdMMP-12. In addition, Co(II)- substituted analogs (both Co2 and ZnCo) must be prepared in order to study zinc binding group coordinaton to the active site of cdMMP-12. Comprehension of metal binding is imperative to inform the creation of a cdMMP-12 specific inhibitor.

Although the ZBGs investigated herein are not optimized for binding, we advocate the examination of the interaction of the aforementioned ZBGs as possible building blocks for inhibition, particularly TM, G11, and curcumin. Greater specificity will need to be incorporated to these ZBGs to minimize off-target binding, as this investigation focuses on a model protein, where significant binding is observed. Screening against cdMMP-12 is necessary, giving rise to the need to continue efforts to purify the protein.

6.2 References. [1] Bertini, I., Canti, G., Luchinat, C., and Scozzafava, A. (2002) Characterization of cobalt(II) bovine carbonic anhydrase and of its derivatives, Journal of the American Chemical Society 100, 4873-4877. [2] Garmer, D. R., and Krauss, M. (2002) Ab initio quantum chemical study of the cobalt d-d spectroscopy of several substituted zinc enzymes, Journal of the American Chemical Society 115, 10247-10257.

134 [3] Rowlett, R. S., Chance, M. R., Wirt, M. D., Sidelinger, D. E., Royal, J. R., Woodroffe, M., Wang, Y. F., Saha, R. P., and Lam, M. G. (1994) Kinetic and structural characterization of spinach carbonic anhydrase, Biochemistry 33, 13967-13976. [4] Rowlett, R. S. (2010) Structure and catalytic mechanism of the β-carbonic anhydrases, 1804, 362–373. [5] Bertini, I., Jonsson, B. H., Luchinat, C., Pierattelli, R., and Vila, A. J. (1994) Strategies of signal assignments in paramagnetic metalloproteins. An NMR investigation of the thiocyanate adduct of the cobalt (II)-substituted human carbonic anhydrase II, J Magn Reson B 104, 230-239. [6] Scozzafava, A., and Supuran, C. T. (2000) Carbonic anhydrase and matrix metalloproteinase inhibitors: sulfonylated amino acid hydroxamates with MMP inhibitory properties act as efficient inhibitors of CA isozymes I, II, and IV, and N-hydroxysulfonamides inhibit both these zinc enzymes, Journal of Medicinal Chemistry 43, 10. [7] Supuran, C. T. (2008) Carbonic anhydrases: novel therapeutic applications for inhibitors and activators, Nature Reviews Drug Discovery 7, 168-181. [8] Tallant, C., Marrero, A., and Gomis-Ruth, F. X. (2010) Matrix Metalloproteinases: Fold and function of their catalytic domains, Biochim. et Biophy. Atca 1803, 20-28. [9] Jacobsen, J. A., Major Jourden, J. L., Miller, M. T., and Cohen, S. M. (2010) To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition, Biochim. Biophys. Acta 1803, 72-94. [10] Fulcher, Y. G., and Van Doren, S. R. (2011) Remote exosites of the catalytic domain of matrix metalloproteinase-12 enhanse elastin degradation, Biochemistry 50, 9488–9499. [11] Parkar, A. A., Stow, M. D., Smith, K., Panicker, A. K., Guilloteau, J. P., Jupp, R., and Crowe, S. J. (2000) Large-scale expression, refolding, and purification of the catalytic domain of human macrophage metalloelastase (MMP-12) in Escherichia coli, Protein Expr Purif 20, 152-161. [12] Yang, W., Arii, S., Gorrin-Rivas, M. J., Mori, A., Onodera, H., and Imamura, M. (2001) Human macrophage metalloelastase gene expression in colorectal carcinoma and its clinicopathologic significance, Cancer 91, 1277-1283. [13] Meng, F., Yang, H., Aitha, M., George, S., Tierney, D., and Crowder, M. (2016) Biochemical and spectroscopic characterization of the catalytic domain of MMP16 (cdMMP16) | SpringerLink, journal of biological inorganic chemistry 21, 523-535. [14] Meng, F., Yang, H., Jack, C., Zhang, H., Moller, A., Spivey, D., Page, R. C., Tierney, D. L., and Crowder, M. W. (2016) Biochemical characterization and zinc binding group (ZBGs) inhibition studies on the catalytic domain of MMP7 (cdMMP7), J Inorg Biochem 165, 7- 17.

135 [15] Yang, H., Makaroff, K., Paz, N., Aitha, M., Crowder, M. W., and Tierney, D. L. (2015) Metal ion dependence of the matrix metalloproteinase-1 mechanism, Biochemistry 54, 3631-3639. [16] Jacobsen, J. A., Fullagar, J. L., Miller, M. T., and Cohen, S. M. (2011) Identifying chelators for metalloprotein inhibitors using a fragment-based approach. - Abstract - Europe PMC, Journal of Medicinal Chemistry 54, 591-602. [17] Supuran, C. T., Scozzafava, A., and Casini, A. (2003) Carbonic anhydrase inhibitors, Medicinal Research Reviews 23, 146-189. [18] Kumar, D., Kumar, M., Saravanan, C., and Singh, S. K. (2012) Curcumin: a potential candidate for matrix metalloproteinase inhibitors, Expert opinion on therapeutic targets 16, 959-972. [19] Banci, L., Bertini, I., Ciulli, Alessio, Fragai, M., and Ternia, B. (2003) Expression and high yield production of the catalytic domain of matrix metalloproteinase 12 and of an active mutant with increased solubility , Journal of Molecular Catalysis A: Chemical Volumes 204–205, 401–408.

136