This item is only available as the following downloads:
TACN AND JIBING TOWARD SYNTHETIC MODELS OF OXALATE DEGRADING METALLOENZYMES BY ERINN BRIGHAM A Thesis Submitted to the Division of Natural Sciences New College of Florida In partial fulfillment of the requirements for the degree of Bachelor of Arts Under the sponsorship of Dr. Suzanne E. Sherman Sarasota, Florida May, 2011
ii Acknowledgements I would like to acknowledge the New College Foundat ion and the Council of Academic Affairs for generously funding my thesis r esearch. Drs. Scudder, Sherman, Shipman, and Walstrom, and Alexandria Liang have taught me all the chemistry I know and I am very gr ateful. Dr. Scudders genius for teaching made the once dreaded organic chemistry ex citing and intellectually fulfilling. I will always be grateful that he lured me into the f ield. Ali Liang was the teaching assistant when I took or ganic chemistrybetween her and Dr. Scudder, I began to learn how to think chem ically. A year later I started working with Ali in Dr. Shermans lab. I was intimidated, b ut she reassured me and she taught me how to be a scientist and how to respond to obstacl es. Even in her absence she has been a constant source of support and inspiration througho ut this year. I would also like to thank Dr. Steven Shipman for his constant encouragement and understanding. He taught me how to solder, helped m e develop my computational method, and allowed me to use his computers for cal culations. I am very grateful to Dr. Sherman for her advice a nd supportand seemingly infinite patience. She has always been very underst anding and her guidance and encouragement has made a world of difference in my experience at New College and life afterward. Of course, I must thank my friends and family who have stuck with me through many highs and lows. I would particularly like to t hank my mother, my father, and Jesse ODell for their constant encouragement, and the ma ny late nights and lunches in lab. Finally, I am grateful to Leah Burgerwithout who m I would be entirely lost.
iii Contents Chapter I Introduction .................................................. ................................................ 1 Oxalate I.A ....................................... ................................................... ............................. 1 Enzymes of oxalate degradation I.B ................ ................................................... ............. 5 Oxalate decarboxylase I.B.I .................................................. ...................................... 6 Oxalate oxidase I.B.II ................................................... ............................................. 17 Research on small molecule models I.C. ............ ................................................... ........ 30 Berreau and Dunbar I.C.I .................................................. ......................................... 30 Sherman and Pecoraro I.C.II .................................................. .................................... 32 Target complexes I.D. ............................. ................................................... ................... 35 Chapter II Experimental .................................................. ............................................ 41 General II.A .................................................. ................................................... ..................................... 41 Synthesis II.B..................................... ................................................... ......................... 42 Route 1 II.B.I .................................................. ................................................... ......... 42 Route 1a II.B.I.a .................................................. .................................................. 49 Route 1b II.B.I.b .................................................. ................................................. 55 Route 1c II.B.I.c .................................................. .................................................. 61 Route 2 II.B.II .................................................. ................................................... ....... 65 Route 2a II.B.II.a .................................................. ................................................ 65 Route 2b II.B.II.b .................................................. ................................................ 67 Route 3 II.B.III .................................................. ................................................... ...... 70 Computation II.C .................................. ................................................... ...................... 72 Chapter III Results and Discussion .................................................. ........................... 73 Synthesis III.A .................................................. ................................................... ................................ 77 Route 1 III.A.I .................................................. ................................................... ....... 77 Synthesis of iPr2TACN III.A.I.a ................................................... ......................... 78 Synthesis of Ph2TACN III.A.I.b .................................................. ......................... 79 Synthesis of Bn2TACN III.A.I.c .................................................. ......................... 82 Route 2 III.A.II .................................................. ................................................... ...... 86 Route 2a III.B.II.a .................................................. ............................................... 86 Route 2b III.B.II.b .................................................. ............................................... 89 Route 3 III.A.III .................................................. ................................................... .... 94 Evaluation of Routes III.A.IV .................................................. ................................. 97 Computation III.B ................................. ................................................... ...................... 99 Chapter IV Conclusions and future development .................................................. 104 References .................................................. ................................................... ................. 112 Appendix Spectra .................................................. ................................................... ... 121
iv List of Figures FIGURE 1. PROTONATION STATES OF OXALATE .................................................. .................. 1 FIGURE 2. THE MODEL OF CATALASE .................................................. ................................. 3 FIGURE 3. OXDC FROM B. SUBTILIS .................................................. ..................................... 7 FIGURE 4. ACTIVE SITE OF OXDC FROM B. SUBTILIS AND T. MARITIMA'S TM1287 ................. 8 FIGURE 5. OXALATE BOUND TO MANGANESE .................................................. ................... 10 FIGURE 6. DIFFERENTIAL BINDING MODES IN FOSA ................................................. .......... 11 FIGURE 7. STRUCTURE OF OXALATE OXIDASE FROM H. VULGARE ...................................... 18 FIGURE 8. ACTIVE SITE OF OXALATE OXIDASE FROM H. VULGARE ...................................... 19 FIGURE 9. EPR SPECTRA OF OXALATE OXIDASE FROM P. PASTORIS .................................... 20 FIGURE 10. UV-VISIBLE SPECTRA OF RECOMBINANT OXOX FROM P. PASTORIS .................. 21 FIGURE 11. EPR SPIN TRAPPING: CARBOXYLATE RADICAL IN OXOX .................................. 22 FIGURE 12. BINDING OF GLYCOLATE IN THE ACTIVE SITE OF OXOX. ................................... 25 FIGURE 13. SERINE, ALANINE, AND GLUTAMIC ACID. ................................................. ........ 26 FIGURE 14. THE OXALATE-BOUND AND HALIDE MN(II) COMPLEXES OF BPPPA. ................. 32 FIGURE 15. SHERMAN AND PECORARO LIGANDS. ................................................. .............. 33 FIGURE 16. STRUCTURES MN(II) COMPLEXES OF KBPZG AND TCMA ............................. 34 FIGURE 17. STRUCTURE OF [MN(IPR2TCMA)(H2O)2]2+ .................................................. .. 34 FIGURE 18. POSSIBLE LIGAND FRAMEWORKS .................................................. .................. 37 FIGURE 19. ACTIVE SITE OF HEMOCYANIN AND MODEL SYNTHESIS.................................... 38 FIGURE 20. VARIBALE ORIENTATION OF PHENYL RINGS AND O,O-DIMETHYL-PH2TCMA .... 39 FIGURE 21. TARGET COMPLEXES. ................................................. ..................................... 40 FIGURE 22. ACID TRAP. ................................................. ................................................... .. 46 FIGURE 23. FILTERING UNDER N2. ................................................. ..................................... 47 FIGURE 24. PROTON NMR SPECTRUM OFLIANG'S BN2TACN ........................................... 78 FIGURE 25.PROTON NMR SPECTRUM OF IPR2TACN .............................................. ........... 76 FIGURE 26. PROTON NMR SPECTRUM OF PH2TACNTS .................................................. ... 80 FIGURE 27. PROTON NMR SPECTRUM OF THE PRODUCT OF SULFURIC ACID DETOSYLAT ION OF PH2TACNTS. ................................................. ................................................... ..... 81 FIGURE 28. PROTON NMR SPECTRUM OF BN2TACN .............................................. ........... 84 FIGURE 29. CARBON NMR OF BN2TACN. ............................................. ............................ 85 FIGURE 30. PROTON NMR SPECTRUM OF POLYMERIC TACNTS ....................................... 87 FIGURE 31. PROTON NMR SPECTRUM OF ROUTE 2A TACNTS ............................................. 88 FIGURE 32. PROTON NMR SPECTRUM OF THE PRODUCT OF STEP 1 OF ROUTE 2B ............... 91 FIGURE 33. PROTON NMR SPECTRUM OF DEG ............................................... .................. 92 FIGURE 34. PROTON NMR SPECTRUM OF THE PRODUCT OF ROUTE 3 STEP 1 ...................... 96 FIGURE 35. IN PLANE AND OUT OF PLANE. ................................................ ................. 100 FIGURE 36. OPTIMIZATION OF THE LIGAND PH2TCMA. ............................................. ...... 101 FIGURE 37. OPTIMIZATION OF THE LIGAND O,O'-DIMETHYL-PH2TCMA .......................... 101
v FIGURE 38. WONG FE(III) COMPLEX AND PROPOSED COMPLEX ....................................... 106 FIGURE 39. MN COMPLEXES OF LIGANDS SIMILAR TO THE PROPOSED LIGA ND ................. 109 FIGURE 40. PROPOSED COMPLEXES BASED ON TACH .............................................. ........ 109 List of Schemes SCHEME 1. MECHANISTIC PROPOSAL FOR OXDC TANNER ET AL. 2001 ............................... 12 SCHEME 2. MECHANISTIC PROPOSAL FOR OXDC REINHARDT ET AL. 2003 ......................... 14 SCHEME 3. MECHANISTIC PROPOSAL FOR OXDC BURRELL ET AL 2007 .............................. 15 SCHEME 4. MECHANISTIC PROPOSAL FOR OXOX WHITTAKER 2002 ................................... 23 SCHEME 5. MECHANISTIC PROPOSAL FOR OXOX WHITTAKER 2006 ................................... 24 SCHEME 6. MECHANISTIC PROPOSAL FOR OXOX AND OXDC BORNEMANN 2007 ................ 27 SCHEME 7. MECHANISTIC PROPOSAL FOR OXOX WHITTAKER 2007 ................................... 28 SCHEME 8. THE SYNTHESIS OF THE MN(II) COMPLEX OF BPPPA ........................................ 30 SCHEME 9. ROUTE 1 ................................................. ................................................... ....... 42 SCHEME 10. ROUTE 2A. ................................................. ................................................... 65 SCHEME 11. ROUTE 2B. ................................................. ................................................... .. 67 SCHEME 12. ROUTE 3. ................................................ ................................................... ..... 70 SCHEME 13. SHERMAN LAB ROUTE TO TCMA. ................................................. ................... 73 SCHEME 14. LIANG ROUTE. ................................................. ............................................... 74 SCHEME 15. ROUTE 1 ................................................. ................................................... ..... 77 SCHEME 16. ROUTE 2A. ................................................. ................................................... 86 SCHEME 17. ROUTE 2B .................................................. ................................................... .. 89 SCHEME 18. MORPHOLONE FORMATION. ................................................. .......................... 93 SCHEME 19. ROUTE 3. ................................................ ................................................... ..... 94 SCHEME 20. WOLFGANG ROUTE .................................................. ...................................... 95 SCHEME 21. PRODUCTS OBSERVED BY HUANG .................................................. ................ 98 SCHEME 22. WONG SYNTHESIS .................................................. ...................................... 107 SCHEME 23. CRONIN SYNTHESIS. ................................................. .................................... 110 List of Tables TABLE 1. CALCULATED GEOMETRIC PARAMETERS LIGANDS... ................... 102
vi TACN AND JIBING TOWARD SYNTHETIC MODELS OF OXALATE DEGRADING METALLOENZYMES Erinn Brigham New College of Florida, 2011 ABSTRACT The enzymes oxalate oxidase and oxalate decarboxyla se degrade the toxic compound oxalate present in many edible plants. The se manganese-containing proteins are found in plants, bacteria, and fungi, but not i n mammals. Consequently, overexposure to oxalate can cause a variety of medical problems for mammals such as humans and livestock. Chief among these are painful edema due to the mechanical tissue damage caused by crystals of calcium oxalate, cardiologica l problems, and kidney and bladder stones. Frequently, oxalate poisoning is diagnosed using a bioassay of the urine which employs oxalate oxidase to determine the urinary ox alate concentration. The process of protein expression and purification for this proced ure is costly and time consuming. It would be of great advantage to replace the enzyme w ith a synthetic model. If the model is non-toxic, it might eventually be used to treat oxa late poisoning. Aside from these concerns, a catalytic model that i s also structurally similar to the active sites of these enzymes could provide insight into the mechanism of catalysis and even into the relationship between structure and fu nction of metalloproteins in general. In response to these interests and the previous wor k of Scarpellini et al., two new model complexes became the synthetic targets: manga nese(II) complexes of the N3O
vii donors 1,4-diphenyl-1,4,7-triazacyclononane-7-monoa cetate (Ph2TCMA) and 1,4dibenzyl-1,4,7-triazacyclononane-7-monoacetate (Bn2TCMA). In addition, a previously synthesized ligand was pursued in order to more ful ly explore the chemistry of its Mn(II) complex with oxalate: 1,4-diisopropyl-1,4,7-triazac yclononane-7-monoacetate (iPr2TCMA). iPr2TCMA and Bn2TCMA were brought to the penultimate step in their respective syntheses: 1,4-diisopropyl-1,4,7-triazac yclononane and the novel compound 1,4-dibenzyl-1,4,7-triazacyclononane. Ph2TCMAwas brought to the antepenultimate step in its synthesis: 1,4-diphenyl-7-tosyl-1,4,7-triaza cyclononane. It was discovered that the removal of the final tosyl protecting group in the case of the phenyl and benzyl substituted ligand was challenging, but was finally accomplished using a solution of sodium and naphthalene. Several alternate routes we re pursued in order to circumvent the use of tosyl protecting groups, but as yet none of them have been successful. ________________________________________ Dr. Suzanne E. Sherman Thesis Sponsor Natural Sciences Academic Division
1 Chapter I Introduction A. Oxalate Oxalate (figure 1) is a toxic compound produced by many plants, bacteria, and fungi. In most biological systems, it is predomina ntly found as an insoluble calcium oxalate salt.1 By the beginning of the 17th century, it was known that oxalic acid was responsible for the sour taste of plants of the gen us Oxalis which is Greek for acid. In fact, oxalic acid is one of the strongest organic a cids, with pKas of 1.3 and 4.3.1 Accordingly, it has been used to remove stains and rust. This practice, in addition to the widespread use of oxalate as a ligand, contributes to the high levels of oxalate in waste water, especially in the paper industry.2 Figure 1. Protonation states of oxalate: pKa1 = 1.3 and pKa2 = 4.3.1 Organisms that produce oxalate have various methods of oxalate metabolism, including the enzymes oxalate oxidase (oxox), oxala te decarboxylase (oxdc), and oxalyl CoA decarboxylase. However, mammalian organisms do not have these useful proteins.
2 Their limited means of oxalate elimination include excretion in the urine, solid excretion in the feces, and metabolism by oxalate degrading b acteria in the intestines.3 Consequently, when animals, particularly humans and livestock, consume high levels of oxalate, several cardiological and renal conditions can develop. In mammalian cells, needle-like calcium oxalate crystals mechanically c ause severe and painful tissue damage, whereas in higher plants crystals of calciu m oxalate, called raphides,4 are used as a method of storing calcium.1 Kidney and bladder stones can result from chronic oxalate consumption, but after one-time consumption of large quantities acute oxalate poisoning can occur. The minor and more common form of poisoning results in painful edema. In rare cases, extreme one-time consumption of oxalate can result in the crystallization of calcium oxalate monohydrate in r enal tissue leading to immediate renal failure and possibly death.5 The most prevalent sources of oxalate for humans i nclude rhubarb, spinach, and other leafy greens. Humans as ide, oxalate poisoning is often a problem for livestock grazing in pastures containin g grasses and other plants high in oxalate. Horses, sheep, and cows grazing in pasture s of Setaria anceps, Panicum maximum, Brachiaria humidicola, Penisetum purpureum, and species of the Oxalis genus have been reported to have enlarged faces due to ed ema.6 Diagnosis of oxalate poisoning commonly involves te sting urinary oxalate concentration using oxox in conjunction with catala se in a bioassay.7 The costly and time consuming nature of the expression and purification of these proteins has prompted research into less expensive methods. In 2009, Zuo et al.7 published a catalytic model of catalase (figure 2) that functions under the condit ions necessary for the assay, and is thus
3 capable of replacing catalase. One of our goals is to produce a catalytic complex to replace oxox in these assays. Figure 2. The catalytic complex developed by Zuo et al. The figure is taken from Scheme 1.7 Despite research efforts, the mechanisms of oxalate degrading metalloproteins oxdc and oxox continue to elude determination. It i s often difficult to study proteins because they are so large and complex. Mutagenesis studies can be extremely informative and are generally relatively straight forward proce sses, but it is sometimes difficult to gain information on all aspects of the protein thro ugh only biological methods. The field of biomimetic chemistry arose in response to the ne ed to simplify the system in order to understand it. In cases where there are many uncert ainties about the system, it is often helpful to create a small molecule model of the pro tein active site. The greater structural simplicity of the mimic allows for more convenient and reliable study of features that are difficult to probe using the protein itselfsuch a s the oxidation state of the metal center. There are two types of models: structural and funct ional. Using structural mimicry, we can compare the physical properties of model comple xes of varying oxidation states to those of the native enzyme. In this way, more infor med assumptions about structure and mechanism can be made. At one time, structural mode ls were integral tools for
4 elucidating the active site structures of metallopr oteins, but the need to create structural models has been diminished by major advances in pro tein crystallography. Functional mimics are models that perform the same catalysis a s the native enzyme, but are not necessarily similar in appearance. They are often l ess informative as they frequently use the 'wrong' metals and ligand sets, but they can ha ve medicinal, environmental, and industrial application. The ultimate goal of this p roject is to create small molecule complexes that mimic both the structure and functio n of the enzymes. The idea behind this objective is that chemical structure determine s reactivity. Often, structural mimics are not functional because of characteristics of th e protein outside of the first coordination sphere of the metal that are crucial t o the catalysis. Despite these complications, it is often possible to glean inform ation about which features of the protein are necessary to catalysis based on the beh avior of the structural model.8,9 In summary, it is the goal of this project to creat e a biologically inspired, synthetic, homogeneous catalyst capable of degradin g oxalate in response to these health and environmental concerns. These considerations as ide, we are also motivated by our desire to extend the boundaries of biological knowl edge. To further refine our ideas, we must examine the enzymes themselves.
5 B. Enzymes of oxalate degradation Oxalate oxidase (oxox), oxalate decarboxylase (oxdc ), and oxalyl CoA decarboxylase are the three enzymes known to metabo lize oxalate.10 Oxox and oxdc contain metal ions in their active sites, whereas o xalyl CoA decarboxylase does not. For this reason, we will focus on oxox and oxdc. Oxalat e oxidases catalyze the oxygen dependent conversion of oxalate into carbon dioxide and hydrogen peroxide, whereas oxdc converts oxalate into carbon dioxide and forma te,10 requiring only catalytic amounts of oxygen.11 These reactions are summarized below: Oxalate oxidase:10 Oxalate decarboxylase:10,11 The next two sections of the introduction elaborate on the properties of oxox and oxdc.
6 I. Oxalate decarboxylase Oxdc is a manganese containing protein12 found predominantly in fungi, and belongs to the cupin protein superfamily.10,13 Cupin proteins are characterized by conserved amino acid sequences as well as secondary structures. The name cupin comes from the Latin word cupa, which means small barrel.14 As such, their characteristic tertiary structure is a six-stranded -barrel. In oxdc and oxox, a manganese binding site is nestled in the barrel, giving rise to the term jel lyroll.10,13 The cupins are very functionally diverse, and range through Eukaryota, Bacteria, and Archaea. Oxdc is further classified as a bicupin because each subuni t contains two different Mn binding sites, one in each of the two structurally homologo us domains. The main difference between these two binding sites is the presence of Glu-333 in the second (C-terminus) domain binding site that is absent in the N-termina l binding site. However, in one conformation of the enzyme, a glutamate residue (16 2) is brought close to the N-terminal Mn binding site.15 This Glu is believed to provide a proton to form f ormate.16 Furthermore, each oxdc contains twelve Mn binding s ites because the functional form is a hexamer consisting of a trimer of dimers of bicupin subunits.16 The hexameric structure and the bicupin subunit are shown in figure 3.
7 Figure 3. The subunit of oxdc from B. subtilis (left) and the functional hexameric form (right) at 1.75 resolution. The gray spheres represent manganese. Figure taken from Mkel et al.13 PDB: 1J5816 The most important aspect of the proteins that was considered in the development of our mimics is the structure of the active sites. Although the two enzymes oxdc and oxalate oxidase (oxox) catalyze separate reactions, their active sites are almost identical. Thus, the discussion of the structure of the first coordination sphere of the active site of oxdc can be generalized to oxox. As can be seen in figure 4 (left), the manganese center (purple) is coordinated to four amino acid residues and two solvent molecules in a distorted octahedral geometry. Three histidines bin d facially, and a glutamate residue is bound trans to one of the histidines. The remaining coordinati on sites are two cis sites that are trans to the other two histidines.16,17 In figure 4 (left), these are occupied by solvent molecules water and formate.
8 Figure 4. Left: The structure of the active site of oxdc from B. subtilis determined at 1.75 resolution is shown. PDB: 1J58.16 Figure modified from Scarpellini et al.17 Purple: Mn, Green: C, Blue: N, Red: O. Hydrogens om itted for clarity. Right: In this crystal structure, the metal site of T. maritima's TM1287 (determined at a resolution of 1.95 PDB 1O4T) corresponds to the active site of oxdc. Notably, oxalate is endogenously bound in a bidentate manner Figure taken form Schwarzenbacher et al. 2004.18 Unfortunately, the structure of the enzyme is whe re our reliable assumptions end. The mechanism and the oxidation state of the restin g state of the enzyme are contested. Many research groups assume that the resting oxidat ion state is manganese(II) because ~95% of the isolated enzymes are divalent.19 Although the mechanism is significantly more mysteri ous than the oxidation state, it is still important to consider the proposed mech anisms when constructing a potential mimic. Many of the proposed mechanisms for oxdc ass ume that oxalate binds in a monodentate fashion, although the binding mode is n ot actually known. Proponents of bidentate binding cite the crystal structure of the putative oxdc of T. maritima It is assigned as an oxdc on the basis of structural simi larity to the oxdc from B. subtilis The protein, called TM12787, was expressed in E. coli and isolated. The crystals that formed
9 contained an endogenous molecule of oxalate bound i n a bidentate manner to a metal ion in the site that corresponds to the active site of oxdc from B. subtilis The oxalate was not added at any point, but was carried through isolati on from the cells themselves. The metal ion was assumed to be manganese because of the iden tification of the protein as an oxdc (pictured in figure 4(right)).18 However, there are two important limitations to thi s evidence for bidentate binding. Most importantly, the oxalate is still int act, indicating that this protein does not catalyze the degradation of oxalate. This impotence is important because if the protein is different enough to deactivate its oxdc function, t hen it could be different from oxdc in a way that might affect its binding mode. Furthermore the identity of the metal ion was not tested in any way. If the metal is not manganese, t hen the binding mode of oxalate in this structure is irrelevant.13 Oxalate is bound in a bidentate manner in the vast majority of small molecule complexes. In fact, it is often used as a bridging ligand to construct inorganic polymers and metal organic frameworks (MOFs). In these exten ded systems, oxalate is bound through all four oxygens.20-24 An example of oxalate bound to manganese is shown in figure 5.
10 Figure 5. The binding of oxalate to manganese. Figu re taken from Garcia-Couceiro et al.20 However, one of the key features of enzyme active s ites is that they control the geometry and orientation of their substrates. The d ifference in binding preference in the active site and in solution is well illustrated by the example of the manganese containing enzyme metalloglutathione transferase (FosA). The b inding modes of the substrate (fosfomycin), an inhibitor, and a phosphate group h ave been elucidated by 31P ENDOR (electron-nuclear double resonance) and ESE EPR (el ectron spin echo electron paramagnetic resonance) spectroscopy when they are bound both in the active site and to aqueous manganese.25 As can be seen clearly in figure 6, the binding mo des are not consistent between the aqueous and active site stru ctures.
11 Figure 6. Left: The binding modes of the substrate (S) fosfomycin, an inhibitor (Pf) phosphonoformate, and phosphate when bound to the p rotein (E) FosA. Right: The binding modes of the same ligands when manganese is solvated in water. The figure is taken from Walsby et al. 2005.25 The implication is that it is difficult to know whe ther oxalate binds bidentate in the active site of the enzymes, although it is well established that it is found very frequently bound bidentate in coordination complexe s and polymers.20-24 The steric requirements of the active site might even be criti cal to the catalysis. This uncertainty notwithstanding, many of the mecha nisms that have been proposed for oxdc assume that oxalate is bound in m onodentate fashion, presumably because the O2 dependence of the reaction suggests that O2 might occupy the other binding site in an octahedral complex. However, it is possible for manganese to expand its coordination number to 7, allowing for bidentat e binding of oxalate and the
12 coordination of dioxygen. Scarpellini et al., Tanas e et al., and Deroche et al. contain examples of seven-coordinate Mn(II) complexes.17,26,27 Based on the fact that isolated oxdcs contain primarily Mn(II) (~95%),12 most mechanistic proposals involve a Mn(II)/Mn(III) cycle, but some opt for a Mn(III)/Mn (II) cycle. The majority of mechanistic proposals for both enzy mes emerge in the context of these assumptions. Scheme 1 illustrates the first p roposals, which are based on EPR spectroscopy of the oxdc from B. subtilis expressed in E. coli .12 Scheme 1. These are the first mechanistic proposals for oxdc based on kinetic and EPR spectroscopy experiments and previous proposals for oxox. Part A illustrates the case where the resting state is Mn(II) and part B shows the case where the resting state is Mn(III). Scheme modified from Tann er et al. 2001.12 In part A of scheme 1, oxalate binds in a monodent ate fashion and O2 binds as well, oxidizing the Mn(II) to Mn(III) to form a sup eroxo species. Then CO2 is lost, reducing the Mn center and leaving a formyl radical The radical is not shown in A, but is identical to the formyl radical in B except that ox ygen is bound as superoxo in A. The formyl radical then combines with the radical O to form a percarbonate intermediate. At
13 this point, the mechanism diverges from the mechani sm proposed for oxox at the time. In the final step, the formate is lost and protonated, leaving Mn(III)-superoxo in equilibrium with Mn(II) and free O2. In Part B of scheme 1, Mn(III) binds oxalate and C O2 is lost, reducing Mn(III) to Mn(II) and leaving a formyl radical. Only after the loss of carbon dioxide does O2 bind, oxidizing Mn(II) to Mn(III). The rest of the cycle is the same as in part A. Part B is a less attractive mechanism than A because the catalytic c ycle could be completed without O2 in Part B, as oxygen binds only after the C-C bond has already been cleaved. Furthermore, the H atom abstraction is not detailed .12 Most importantly, there is little experimental evidence for these mechanisms. Following these proposals, kinetic studies employi ng the heavy isotope effect resulted in a new proposal by Reinhardt et al.,28 shown in scheme 2. Based on pH effects that they observed, Reinhardt et al. argue that mon oprotic oxalic acid is the actual substrate of oxdc (part A of scheme 2). Furthermore by isotopically labeling only one side of the normally symmetric oxalate, they determ ined that there must be a reversible slow step before the cleavage of the C-C bond (the fast step).
14 Scheme 2. The mechanism for oxdc proposed by Reinha rdt et al. 2003. Scheme taken from source.28 It should be noted that the oxidation states of Mn in this scheme were hypothetical. Additionally, H+ abstraction might be better termed a PCET (proton-coupled electron transfer).29 Arg-270 is included because it was believed to polarize the C=O bond to the Mn. The authors inferred from these two observations t hat the isotope-sensitive reversible step prior to decarboxylation is the pro ton transfer between oxalic acid and nearby Glu-333, shown in part B. Also shown in part B is the single electron transfer from the Oto the Mn(III) center, forming an oxygen-centered radical and reducing Mn(III) to Mn(II). The combination of these two ste ps to form the structure shown in B might be better termed a proton coupled electron tr ansfer. At this point, CO2 is lost from the previously protonated end (determined by isotop ic labeling), leaving a carbene intermediate (part C). The carbene then deprotonate s Glu-333 and Mn(II) transfers one electron to the bound oxygen, resulting in the Mn(I II)-formate intermediate in part D.28 Unfortunately, this mechanism fails to explain the importance of O2 in the catalytic cycle.
15 The main disadvantage of this proposal is the fact that it is based on the initial assumption that the C-terminal Mn site is the active site. Cur rently, it is believed the N-terminal site is the active Mn site, although some doubt has rece ntly been raised.30 This controversy is discussed later in this section. Reinhardt et al.s kinetic observations are still important and have been incorporated into the most current (2007) proposal for the mechanism by the Bornemann group, which is summarized in scheme 3. This propos al combines the mechanism initially put forth by Tanner et al. in 2001 with t he Reinhardt proposal. Scheme 3. The most recently proposed mechanism by t he Bornemann group. Scheme modified from Burrell et al.31 In this mechanism, monoprotic oxalic acid reversibl y binds to Mn(II), followed by the binding and reduction of oxygen to produce a Mn(III)-superoxo species. Then, the acidic H+ is removed from oxalate and Mn(III) is reversibly reduced to Mn(II) by the
16 bound oxalate O, while the electron density of the C=O bond is further polarized toward the O, as seen in the previous proposal by R einhardt et al. in 2003. This activated structure then decarboxylates, leaving a formyl rad ical bound to Mn(II)-superoxo. The next step is poorly understood, but it has been pro posed that the formyl radical forms a bond with a proton of Glu-162 that ultimately deriv es from the solvent.32 Glu-162 is used instead of Glu-333 because this mechanism assumes t hat catalysis occurs in the Nterminal Mn site instead of the C-terminal site. Fi nally, formate and oxygen dissociate, reducing Mn(III) to Mn(II) and the cycle is complet e. Svedruzic et al. also published a proposal earlier in 200729, but it also assumed that the C-terminal site was the active site, so it is not shown. The Bornemann mechanism is compelling because it a ccounts for the necessity of dioxygen and the surrounding amino acids, as well a s the available kinetic data. The argument that the N-terminal site is in fact the ac tive Mn site is quite convincing. There is a channel to the N-terminal manganese site that can be opened and closed via a conformational change in the lid peptides (161-16 5) to allow the substrate to access the N-terminal site and to close off the site during ca talysis. The mutation of Glu-162 of the N-terminal lid caused a drastic decrease in oxdc ac tivity and a simultaneous increase in oxox activity. Glu-162 is believed to provide the p roton necessary for the formation of formate instead of the CO2 produced by oxox, so it is logical that its replac ement would result in a conversion to oxalate oxidase behavior. Importantly, this amino acid is absent in the oxox active site. Additionally, Angerhofer e t al. performed an incisive multiplefrequency EPR experiment indicating that the two bi nding sites are in fact spectroscopically distinct, and that only one of th em is solvent accessible.33 Based on the
17 information about the mobile lid, it was assumed th at the solvent accessible site is the Nterminal site. However, more recent information fro m the Bornemann group suggests that the C-terminal site may also be capable of catalysi s, further complicating the situation.15,30 II. Oxalate oxidase Oxalate oxidase (oxox) is a germin protein primaril y found in higher plants, such as grains. Although the biological role is still no t fully understood, it is known as a germination marker protein and it produces hydrogen peroxide, which is known to be necessary for cross-linking in cell wall growth.34 Like oxdc, oxox is part of the cupin superfamily. As in oxdc, the active sites of oxox c ontain manganese and the protein is a trimer of dimers. The structure of the homohexamer is shown in figure 7.
18 Figure 7. The quaternary structure of oxalate oxida se from Hordeum vulgare (barley) (PDB 1FI2)35 determined at 1.6 resolution. The figure is modi fied from Woo et al. 2000.35 As can be seen, the familiar -jellyroll and helices characteristic of the cupin superfamily are present. Unlike oxdc, there are onl y six manganese ions in the homohexamer of oxox. Thus, the identity of the acti ve site is unambiguous, but the oxidation state of the resting enzyme is still unkn own. The manganese ions (green spheres in figure 7) are nestled in the center of t he jellyroll. The first coordination sphere of the active site is almost exactly the same as th at of oxdc. Despite these similarities, the mechanisms of oxalate degradation must be distinct because oxox converts oxalate into H2O2 and two molecules of CO2, whereas oxdc produces formate and CO2. This means that O2 is required in stoichiometric amounts in oxox, whe reas it is catalytic in oxdc. The mechanism of oxox is no better understood than that of oxdc, although oxox has been
19 studied more extensively. The structure of the acti ve site with two waters bound to manganese is shown in figure 8. Figure 8. The first coordination sphere and some ne arby amino acids of the active site of oxalate oxidase from Hordeum vulgare is shown at 1.6 resolution (PBD 1FI2)35. The figure is modified from Svedruzic et al. 2005 .10 As with oxdc, there are many proposed mechanisms fo r the catalysis of oxox. Only the most convincing and recent will be discuss ed. The first guess at the mechanism was made before the details of the structure were k nown so it is primarily based upon EPR data. However, it was echoed by those who eluci dated the structure some years later (Woo et al. 200035). Requena and Bornemann 199936 found that the optimal pH for oxox activity is 4.0, which was taken to mean that the m onoprotic form of oxalate is the substrate of the enzyme (recall that the pKas of ox alate are 4.3 and 1.27). Furthermore, they assert that Mn(II) is the resting oxidation st ate of the enzyme based upon EPR spectra of the purified enzyme, which display the c haracteristic peaks of Mn(II) and show
20 trace amounts of Mn(III). In accordance with their data, their proposal for the mechanism is as follows: monodentate binding of oxalate to Mn(II) causes an increase of oxidation potential. Then, O2 binds and is reduced to superoxide, oxidizing Mn(II) to Mn(III). The subsequent decarboxylation of oxalate leads to the formation of a peroxy species that could form hydrogen peroxide upon protonation.35,36 The way this peroxy species is produced is not discussed in detail by the authors, but the advantage of this explanation is that it avoids radical chemistry until both substrates are bound and it uses Mn(II) as the resting state. In 2002, Whittaker and Whittaker37 further confirmed the assumption that Mn(II) is the site of oxalate conversion using EPR spectro scopy. The X-band spectra (shown in figure 9) of the recombinant oxox isolated from Pichia pastoris indicate that the Mn(II) pattern is disturbed and the small signal from Mn(I II) disappears upon the addition of oxalate. When oxalate is added, there is a change i n the 55Mn nuclear hyperfine splitting, indicating the binding of oxalate to the manganese ion. Additionally, the UV-visible Figure 9. The anaerobic X-band EPR spectra of oxalate oxidase from P. pastoris are shown (A) in the absence of oxalate and (B) in the presence of oxalate. Figure taken from Whittaker and Whittaker 2002.37
21 spectrum (shown in figure 10) is affected by the ad dition of oxalate. There are peaks at ~725 nm and ~425 nm with a small sharp shoulder at 45 4 nm in the spectrum without oxalate. These peaks are consistent with Mn(III) in the active site. When oxalic acid is added, these peaks are silenced, indicating that Mn (III) reacts with oxalate to produce Mn(II). Figure 10. The UV-Visible spectra of recombinant ox ox isolated from P. pastoris are shown (A) in the absence of oxalate and (B) in the presence of oxalate. Inset: the difference spectrum. Figure taken from Whittaker an d Whittaker 2002.29 In addition to this characterization, Whittaker an d Whittaker also report spintrapping experiments that indicate the presence of a carboxylate radical in the sample of isolated enzyme and oxalate. The intensity of the p eak was relatively constant, with a half life of about six minutes. Furthermore, the peaks b ehaved almost identically under both aerobic and anaerobic conditions. This spectral fea ture can be explained by the small concentration of Mn(III), estimated to be ~5%, being reduced by oxalate to produce the formate radical. The actual spectrum and the predic ted spectrum of a DMPOformate radical spin adduct are shown in figure 11. (DMPO = 5,5-Dimethyl-1-Pyrroline N-Oxide)
22 Figure 11. EPR spin trapping experiment: A shows th e spectrum two minutes after oxalic acid was added to a solution of oxox and DMP O, and B shows the predicted spectrum of a Mn(II) bound formate radical, as simu lated using the program sim15 (Quantum Chemistry Program Exchanged QCPE265). The figure is taken from Whittaker and Whittaker 2002.37 Although the reaction is not catalytic, it does ind icate that the Mn(III)-oxalate complex is capable of effecting decarboxylation. Based on these data, they proposed the following m echanism (summarized in scheme 4): monoprotic oxalate binds to Mn(II), disp lacing one water molecule and raising the oxidation potential by stabilizing Mn(I II). Then, Mn(II) is oxidized to Mn(III) by oxygen, which creates a superoxide that deproton ates the remaining water molecule to form a nucleophilic peroxy radical. This radical co uld then attack oxalate to induce decarboxylation. The lack of evidence of other radi cals in the spin-trapping experiments indicates that the formation of the superoxide and the decarboxylation are closely coupled, possibly even concerted. After the loss o f CO2, there is a second decarboxylation to form CO2 and HOO, followed by the protonation of HOOto form hydrogen peroxide (not shown).
23 Scheme 4. (A) is the proposed explanation of the ob servation of carboxylate radical in EPR spectra of oxox in the presence of oxalate ( O2 independent) and (B) is the oxox mechanism proposed by Whittaker and Whittaker in 2002 based upon EPR and UV-Vis experiments. The scheme taken from Whitt aker and Whittaker 2002.37 Several years after their first proposal, the Whitt aker group proposed a radically different mechanism, which is summarized in scheme 5. In the first step, monoprotic oxalic acid binds in monodentate fashion to Mn(II), displacing one solvent molecule. The second solvent molecule is then displaced by the bi nding of O2. Oxygen is reduced by the manganese center to produce a Mn(III)-superoxo spec ies. This radical attacks the carbon, activating the substrate. In the fourth ste p, the oxygen-centered radical abstracts the carboxylic H atom to produce a distal carboxyla te radical. This unstable species decomposes through homolytic C-C cleavage, releasin g CO2 and reducing the manganese center to Mn(II). The authors propose that step six involves the hydrolysis of the percarbonate species to release hydrogen peroxide a nd another molecule of CO2.38
24 Scheme 5. The mechanism proposed by the Whittaker g roup in 2006 is shown. The closed circles connected to Mn(II) by dotted lines indicate coordinated water molecules. Figure taken from Opaleye et al. 2006.38 The presence of Asn-85 and Gln-139 is included in t he mechanism because Opaleye et al. concluded that they are integral to the binding of oxalate. Based on crystal structures of native, recombinant, and point-mutant oxalate oxidases, they found that Gln139 hydrogen bonds with both Asn-85 and a water mol ecule coordinated to the manganese. Additionally, Asn-75 and Asn-85 hydrogen bond with glycolate (a substrate analogue) bound to the Mn(II) center (figure 12). M ore importantly, they found that glycolate is bound monodentate in the crystal struc ture of oxox crystals soaked with glycolate solution.38 Because of glycolates structural similarity to ox alate, this structure is further evidence that oxalate binds to manganese through only one oxygen.38
25 Figure 12. Left: The binding of glycolate in the ac tive site of oxalate oxidase is shown. The figure is based on the X-ray crystal str ucture determined by Opaleye et al. 2006 at 1.81 resolution.38 The figure is taken directly from the source. Righ t: The structure of monoprotic oxalate is shown on the top, and the structure of glycolate is shown on the bottom. However, glycolate is not a perfect analogue. Figu re 12 (right) shows the structures of oxalate and glycolate side-by-side. I t is evident that glycolate does not possess the same planar shape as oxalate. Moreover, it is clear that the O-H bond of glycolate is much stronger than that of monoprotic oxalate. The difference in acidity is not as much of an issue as the difference in geomet ry. Many researchers favor mechanisms for oxox and oxdc that diverge at a single point. There are several reasons for this preferenc e, some of them are simply structural and some are evolutionarily based. It is commonly t hought that bicupins emerged from two separate complete gene duplication events that 1) produced plant seed storage proteins, and 2) created fungal and bacterial bicup ins (such as oxdc). In 2005, Escutia et al. discovered an isoform of oxalate oxidase with a bicupin structure reminiscent of the oxdc of B. subtilis In fact, it is believed that this bicupin oxox is related to the precursors of oxdc. Escutia et al. compared the bicupin oxox t o bacterial oxdc, and found that there
26 were important amino acid substitutions near the ac tive site. Most importantly, Glu-162 (oxdc numbering) was substituted for alanine in the bicupin oxox. Similarly, in the monocupin oxox of C. subvermispora the Glu is substituted with serine. It is believe d that the proton necessary to oxdcs catalysis is pr ovided by Glu-16215. As can be seen in figure 13, serine and alanine side chains do not do nate protons as readily as glutamic acid does. This substitution, in combination with the ex traordinarily similar structures of this bicupin oxox and oxdcs suggest that a divergent mec hanism might explain the catalysis of both proteins. Figure 13. The side chains of the amino acids serin e (Ser pKa 16), alanine (Ala pKa 50), and glutamic acid (Glu, pKa 4.1) which is deprotonated at physiological pH (glutamate).39 The Bornemann group has done significant research o n the active site of oxdc and its lid, which further supports a divergent mecha nism. The lid is a conserved sequence of amino acids including Glu-162 that forms a helix that covers the N-terminal site of oxdc in some conformations. In this conformation, G lu-162 is brought close to the active site. It is possible for Glu-162 to be protonated i n the beginning of the catalytic cycle because the site is closed to bulk solvent.
27 As previously discussed, the group31 systematically mutated the lid to determine that Glu-162 was in fact the residue that is primarily responsible for the specificity of oxdc, and that oxalate oxidase activity was observed upon the removal of the Glu proton source. Based on these findings and those of others, they propose the divergent mechanism shown in scheme 6.12,15,16,28,37 The divergent mechanism is very attractive because it e legantly accounts for evolutionary development and is based on the chemical and struct ural differences between the active sites to the best of contemporary knowledge. The next year (2007) brought another distinct propo sal from the Whittaker group. They performed EPR and burst kinetics experiments u sing recombinant oxox from P. pastoris containing Mn(II), Mn(III), and Mn(IV).40 They found that all three oxidation states were interconverting during turnover, and pr oposed the following mechanism based on their observations (see scheme 7). Interes tingly, the proposed mechanism does not include Mn(IV). Scheme 6. The divergent mechanisms for oxalate oxidase and oxalate decarboxylase proposed by the Bornemann group. Figure taken from Burrell et al. 2007.31
28 Scheme 7. The mechanistic proposal for the catalysi s of oxalate oxidase by Whittaker et al. in 2007 is shown. OxHrefers to monoprotic oxalic acid. The figure is taken directly from the source.40 The first step of the mechanism is the binding of oxalate to a Mn(III) center, assisted by Asn-75 and Asn-85. Then, oxalate reduce s Mn(III) to Mn(II) (which has been observed in anaerobic conditions), oxalate is depro tonated, and possibly binds in a bidentate fashion. The oxalyl radical that is forme d is highly unstable, and is known to undergo C-C bond fission rapidly in aqueous solutio n. It is plausible that a similar process occurs in the active site, releasing CO2. In step 4, the radical carboxylate reduces dioxygen to form another molecule of CO2 and superoxide, shown in the protonated hydroperoxyl radical form according to its pKa. Thi s oxygen-centered radical is a minor resonance form because the negative charge is place d on the carbon atom instead of the oxygen atoms. It is more reasonable to place a nega tive charge on one of the oxygens and represent the carbon as a carbene In the following step (5), Mn(II) reduces the hydroperoxyl. The reduction could be direct (throug h coordination) or through hydrogen atom transfer from a complexed water molecule. This mechanism differs from all
29 previous proposals because it uses the manganese to activate oxalate, whereas past ideas use the metal as a reductant for dioxygen, which pr ecipitates the rest of the reaction. Whittaker et al. also reasoned that the Mn(III)/Mn( II) reduction potential lies between +1.0 and +0.4 V versus NHE based on the free energy change of the half reaction oxalate CO2 + CO2 This method of estimation is based on the fact th at catalysts lower the activation energy of reaction b ut do not affect the thermodynamics of the reaction. Therefore, the change in energy of th e whole reaction must be negative. This allows for the estimation of the reduction potentia l because the reduction is the other half of the reaction. Moreover, this reduction potential is both biologically relevant and smaller than E of the free ion.40 As is evident, the mechanisms of both oxox and oxdc are poorly understood and there is still much to learn. In order to clarify t he ambiguities of the data so far, several research groups are approaching the puzzle with an alternative method, using small molecule models instead of the enzymes themselves. This research is discussed in the following section.
30 C. Research on small molecule models Thus far, two groups have attempted to prepare smal l molecule mimics of the oxox and oxdc active sites. The Berreau and Dunbar labs of Texas A & M and Utah State University work in collaboration, as do the Sherman and Pecoraro groups (of New College of Florida and the University of Michigan, respectively). The work of each group is discussed in turn. I. Berreau and Dunbar Fuller et al. developed a structural mimic of the a ctive sites of oxdc and oxox in 2005.41 The ligand is a tertiary amine with two pyridyl su bstituents, one of which has a sterically hindered amide substituent. Thus, the ligand provides an N3O donation sphere, and leaves two open coordination sites, similar to the active sites of the enzymes. Unfortunately, the use of an amide instead of a carboxylate decreases the possibility of generalizing the characteristics of the model Scheme 8. The synthesis of the Mn(II) complex of bpppa and the oxalate-bound complex that were developed as models of the active sites of oxalate oxidase and o xalate decarboxylase. Image taken from Fuller et al. 2005.41
31 complex to the enzyme active sites because an amide differs significantly from the carboxylate arm of glutamate. The Mn(II) complexes of this ligand, called bpppa (Nbenzyl-N-((6-pivaloylamido-2-pyridyl)methyl)-N-(2-p yridylmethyl)amine), are not well characterized in terms of what is desirable for a m odel of a metalloprotein that undergoes redox chemistry. The X-ray crystal structures are f ully described, but the reductionoxidation potentials were not reported. Scheme 8 sh ows the synthesis of the monomeric Mn(II) complex of the ligand and the dimeric comple x formed with oxalate. The monomeric complex shows the proper geometry for a s tructural mimic in that the nitrogen atoms are arranged facially and the solven t molecules are cis to each other. The crystal structure of the oxalate-bound complex is s hown in figure 14. The goal was to predict the binding mode of oxalate in the active s ite of the enzymes, but this effort was foiled by complexation with a second manganese. It may be an interesting project to develop a more sterically hindered ligand that woul d preclude this possibility, however there have been no published efforts in this direct ion thus far. The second structure in figure 14 is a representative view of the series of Mn(II)-halide complexes of the same ligand published the following year as models of th e binding of halides to the active sites of oxox and oxdc.42 These complexes are relevant because halide ions c an inhibit the catalysis of the enzymes by competitively binding t o manganese. However, these complexes were not characterized further than struc tural identification.
32 Figure 14. The oxalate-bound complex (left) and the Mn(II)-halide complex (right) are shown. This is the chloride complex, but comple xes with bromide and iodide have also been prepared. The images are modified fr om Fuller et al. 200541 (left) and 200642 (right). II. Sherman and Pecoraro The Sherman group at New College collaborated with the Pecoraro group at the University of Michigan to publish a new structural mimic of the active sites in 2008. They claim that the complex is the first accurate model.17 Two of the ligands that they used in their work are based on 1,4,7-triazacyclono nane (abbreviated TACN and pronounced tack-in). TACN and the ligands used ar e shown in figure 15.
33 Figure 15. The ligands used by Scarpellini et al.17 to model oxalate oxidase and oxalate decarboxylase, and TACN are shown. HTCMA is 1,4,7-triazacyclononaneN-acetic acid, KiPr2TCMA is potassium 1,4-diisopropyl-1,4,7-triazacy clononane-Nacetate, and KBPZG is potassium N,N-bis(3,5-dimethy lpyrazolyl methyl)glycinate. The main virtues of HTCMA and iPr2TCMA are that they utilize the TACN ring to enforce facial coordination, and they have a car boxylate arm that corresponds to the glutamate residue in the active sites. KBPZG has th e advantage of the pyrazole rings, which are more similar to histidines than the secon dary and tertiary amines of the two other ligands. Unfortunately, KBPZG does not coordi nate to Mn(II) in a facial manner with respect to the N donors, as shown in figure 16 Additionally, it crystallizes as an infinite polymeric chain with each carboxylate boun d to two manganese ions, although the X band EPR spectrum shows that it is a monomer in solution. These qualities decrease its relevance as a candidate for a structu ral mimic.
34 Figure 16. The crystal structures of the Mn(II) com plex of KBPZG (left) and Mn(II) complex of TCMA (right) are shown. The figure is ta ken from Scarpellini et al. 2008.17 The Mn(II) complex of TCMA also crystallizes as an infinite polymer (figure 16), but is a monomer in solution. In response to this p roblem, the other TACN-based ligand, iPr2TCMA, was made. The isopropyl groups introduce enou gh steric bulk to prevent the complex from crystallizing as a polymer. The crysta l structure of the Mn(II) complex is shown in figure 17. Figure 17. The crystal structure17 of [Mn(iPr2TCMA)(H2O)2]2+ is shown on the left, and the same complex sumperimposed on the structure15 of the active site of oxdc from B. subtilis, (PDB 1UW 8) is shown on the right. The images are taken from Scarpellini et al. 2008.17 The oxidation potential of [Mn(iPr2TCMA)(H2O)2]2+ was measured at 0.730 V versus NHE, which is within the range of the potential of oxox estimated by Whittaker et
35 al. in 2007.17,40 Thus far, none of the complexes have been shown to bind or degrade oxalate. It would be interesting to see how oxalate binds to these complexes, but this information has yet to be published. D. Target complexes Our ultimate goal is to create a structural and fun ctional model of oxox and/or oxdc, so there are restrictions on the ligand frame works that we might use. The most limiting is that the nitrogens must coordinate in a facial manner. Additionally, there must be a way to incorporate only one carboxylate arm into the coordination sphere. More over, there cannot be any other coordinating groups in th e ligand to occupy either of the two remaining coordination sites, which excludes some f unctional groups that might otherwise have been considered for their electronic contribution. Based on these restrictions, a limited number of possibilities exi st. The model complex made by Scarpellini et al. meets all of these requirements, but has yet to show oxalate binding or degrading activity. As such, one of the goals of th is project is to reproduce the iPr2TCMA based mimic of Scarpellini et al. to perform furthe r experiments. Some of the other ligands that were considered are pictured in figure 18 and will be discussed after additional guidelines are established. Proteins are exceedingly complex and there are a pl ethora of possible explanations for why a model complex is inactive as a catalyst. If the structure of the first coordination sphere is accurate, then improper redo x potential may be one of the causes.
36 Unfortunately, the potentials of the active sites a re poorly characterized. Whittaker and Whittaker et al. have estimated the reduction poten tial of oxox to be between +0.4 and 1.0 V vs. NHE, but this value is based on calculati ons of the free energy change that occurs in the oxidative cleavage of oxalate, not on experimental data.40 In this instance where direct measurement of a biological system is difficult, a successful model may provide some insight. If a catalytically active str uctural model of the active sites were made, then the potential of the complex might be re asonably assumed to be similar to the potential of the active site. One might imagine that the nature of the donor atom s, and thus their substituents, must match that of the active site donors in order to achieve both the same potential and coordination geometry of the active site. This is t rue to some extent. However, the magnitude of the effect of the difference in donors on similarity of the mimic to the enzymes is unclear. The outer coordination sphere of a metal ion contributes to its redox potential. The outer coordination sphere of a protein active site is highly ordered and relatively static. By contrast, the outer coord ination sphere of a model complex is comprised of solvent molecules and thus highly vari able. Fortuitously, redox potentials can be altered syste matically if a clever choice of model complex is made. By creating a series of comp lexes with different potentials, there is a chance that one of them might be similar enoug h to the active site that it can effect catalysis. The potential of the complex may be alte red by varying the -donation capacity of the ligand. Thus, among the ligands considered f or this project (figure 18) preference was given to ligands that have greater potential fo r alteration.
37 Figure 18. The ligand frameworks considered for thi s work are shown. In this respect, A, B, D, E, and F are all attracti ve options because the R groups could be altered systematically. A and B employ a f ramework commonly used to enforce facial coordination, called triaminocyclohexane (TA CH). There is precedence for the use of TACH derivatives in models of metalloproteins,43 but there are few examples of asymmetrically functionalized TACH,44,45 Asymmetry is necessary to provide only one carboxylate O atom donor. Additionally, TACH is oft en used for tetrahedral coordination,46,47 although it can accommodate octahedral donation sp heres.48-51 Promising asymmetric TACH derivatives described by Cronin et al.45 (scheme 23 and figure 40) were found after the project was already underway, and are discussed as options for future work in Chapter IV. C is exciting because the pyrazole rings are elect ronically similar to the histidine residues in the active sites of oxdc and oxox, and the structure enforces facial coordination. It was not pursued because it was exp ected to be synthetically challenging and the potential for alteration of the pyrazole ri ngs is limited.
38 D shows a ligand with exclusively tertiary amines a s N donors, which is a slight disadvantage because it is somewhat dissimilar from the active site. This is true of E and F as well as of previous structural mimics.17 The deciding factor for D was its flexibility; although it seems likely that it would assume a fac ial arrangement of the N atoms, it was not pursued because of the chance of meridonal coor dination. E and F are ligands based on 1,4,7-triazacylcononan e (TACN), which is a commonly used ligand that enforces facial coordinat ion and is well suited for octahedral geometry. There is also precedence for using it to model metalloprotein active sites.52 The TACN-based structural and functional models of hemocyanin are perhaps the best example of TACN being used successfully to model fa cially coordinated histidine residues.53-55 The structures of a model and the active sight are illustrated in figure 19. Figure 19. Top: An illustration of the first coordi nation sphere of hemocyanin. Bottom: The formation of a synthetic mimic. The ima ge is modified from Tolman 1997.55 It was shown that the binding of O2 to the model complex was reversible under certain conditions.54
39 E and F are also inspired by the successful structu ral model of oxox and oxdc by Scarpellini et al. using the manganese(II) complex of iPr2TCMA.17 The major disadvantage of these two ligands is that the N don ors are tertiary amines. E and F were chosen over A and B because the synthesis of TACN d erivatives similar to E and F were already partially developed in the Sherman lab. The phenyl substituents of E are intriguing because they could easily be used to tun e the potential by attaching electron withdrawing groups or electron donating groups. The benzyl substituents of F also offer this possibility, but to a lesser degree. There is concern that lone pairs on the nitrogens o f E might be too delocalized by conjugation with the phenyls for the ligand to bind a metal strongly enough to be useful. In order for the lone pair of the nitrogen atom to be sufficiently delocalized, it must be aligned properly with the system of the phenyl rings (figure 20 top left). I t may be the case that the system of the bare phenyl is actually nearly perpe ndicular to the lone pair of the Ns (figure 20 bottom left), which would elim inate this concern. Figure 20. Top left: An illustration of the case wh ere the lowest occupied p orbital of nitrogen is aligned with the system of a phenyl ring. Bottom left: An illustrat ion of the case where the lowest occupied p orbital of a n itrogen atom is in perpendicular to the system of a phenyl substituent. Right: The structu re of o,o'-dimethylPh2TCMA: (2-(4,7-bis(2,6-dimethylphenyl)-1,4,7-triazon an-1-yl)acetate).
40 If this problem occurs, one possible solution is to add substituents on the ortho positions of the phenyls to sterically restrict the orientation of the phenyl group (figure 20 right). In this way, the redox potential can still be varied directly, and metal binding would be expected to be stronger. These concerns ar e further addressed in the experimental and discussion sections of this thesis Based on the properties of the ligands shown in fig ure 18, the constraints of time, and precedence in the lab, it was decided that this project will focus on the synthesis of complexes with derivatives of ligands E and F. All of the target complexes are shown in figure 21. Much progress toward the synthesis of F was made by former thesis student Alexandria D. Liang. By the time that she left the Sherman lab, she had successfully synthesized 1,4-dibenzyl-7-tosyl-1,4,7-triazacyclon onane (a precursor of F). In her efforts toward F, she synthesized a ligand of uncertain ide ntity and had begun to explore metallation. The disputed nature of her putative F and further progress is discussed in each of the remaining sections of this thesis. Figure 21. The target complexes proposed to model t he active sites of oxox and oxdc. Left: [(iPr2TCMA)Mn(OH2)2]+, Middle: [(X2Ph2TCMA)Mn(OH2)2]+, Right: [(X2Bn2TCMA)Mn(OH2)2]+.
41 Chapter II Experimental A. General Unless otherwise noted, all syntheses were perform ed under dry nitrogen. Additionally, all reaction flasks were flushed with nitrogen following the addition of each reactant/reagent, or any sample taken from the flas k. Any reactions evolving strong acids were done in isolation on the schlenk line, and aft erwards the line was flushed with nitrogen. Solvents and reactants were purchased fro m either Acros Organics or Sigma Aldrich, and used without further purification unle ss otherwise indicated. All spectral data were recorded at room temperatur e, unless otherwise noted. Nuclear magnetic resonance spectra were recorded on a Bruker AC 250 MHz spectrometer. Fourier-transform infrared spectra we re recorded on a Nicolet Avatar 320 spectrometer using EZ OMNIC. Spectral predictions of proton NMR and carbon NMR were performed using ChemBioDraw Ultra 12.0 (2010).56 Geometry optimizations and energy calculations were performed using Gaussian 4.1.2. Dr. Steven Shi pman was instrumental in these calculations.
42 B. Synthesis In our attempts to synthesize the target ligands iPr2TCMA, Ph2TCMA and Bn2TCMA, three routes were explored to varying degrees Route 1 is by far the most developed, but the other two routes offer the possi bility of saving both time and chemical supplies if fully realized. I. Route 1 Scheme 9. Route 1 to the synthesis of R2TCMA. The route splits after step 6 into either step 7 or step 8. Reagents: 1: H2O/NaOH, TsCl, ether; 2: H2O/NaOH, TsCl, THF; 3: NaH, DMF; 4: HBr 33%wt in AcOH, Phenol; 5: iPr: Na2CO3, MeCN, isopropyl bromide; Bn: Benzyl chloride, Na2CO3, MeCN; Ph: Pd2(dba)3 (cat.), phosphine 7 (cat.), sodium t-Butoxide, toluene, bro mobenzene; 6: con. H2SO4 or HBr 33%wt in AcOH, Phenol or Na/naphthalene, THF; 7 ) a) ethyl bromoacetate, Na2CO3, MeCN b) ethyl bromoacetate, triethylamine, DCM; 8 ) Chloroacetic acid, LiOHH2O, EtOH/ H2O; 9) MeOH/KOH.
43 Synthesis of Tritosyldiethylenetriamine (DETAts3) (4-methyl-N,N-bis(2-(4methylphenylsulfonamido)ethyl)benzenesulfonamide) ( step 1) The procedure of this reaction was developed by Al exandria D. Liang and the author from the theses of past students.57-59 The procedure was repeated many times with consistent results. This reaction was left open to the atmosphere and ran at room temperature. In a two-neck 2 L round bottom flask, sodium hydroxide (29.1 g, 0.73 mol) was dissolved in water (188 mL) using a mechanical stirrer. Diethylenetriamine (24 mL, 221.54 mmol) was added to the reaction mixture. In an Erlenmeyer flask, tosyl chloride (133.578 g, 700.65 mmol) was dissolved in diethyl e ther (625 mL). The tosyl chloride solution was added to the reaction mixture through the second neck of the round bottom using a pressure equalizing addition funnel over on e hour and 25 minutes. The cloudy white solution was stirred for 22 hours and then fi ltered. The filtered solid was added to water (1.5 L) and stirred overnight in an Erlenmeye r flask. In the morning the milky solution was filtered and the cream colored solid w as dried in vacuo. The yield was 100.26 g (80%), but the product was contaminated wi th some tosyl chloride, which is included in the yield. 1HNMR: (CDCl3) 7.75 (d, 4H, Ts), 7.62 (d, 2H, Ts), 7.30 (d, ?Hoverlap with solvent peak, Ts), 5.30 (s, 1H, Ts-N-H ), 3.10-3.00 (m, 8H, N(CH2)2N), 2.45 (s, 9H, TsCH3), 1.80 (s, 4H (water)),1.30 (m, 2H, origin unknown ) ppm. 13CNMR: (CDCl3) 144.4, 136.9, 130.2, 130, 127.5, 127.4, 50.8, 42.9, 21.8 ppm.
44 Synthesis of ethane-1,2-diyl bis(4-methylbenzenesul fonate) (Ethylene glycol bistosylate = EGOts2) (step 2) The procedure of this reaction was developed by Al exandria D. Liang and the author from the theses of past students.57-59 The procedure was repeated many times with consistent results. This reaction was left open to the atmosphere. In a 2 L Erlenmeyer flask, sodium hydroxide (33.80 g, 0.85 mol) was dis solved in water (167 mL) using a mechanical stirrer. Ethylene glycol (16.5 mL, 0.296 mol) was added and the flask was placed on ice. In a separate Erlenmeyer flask, tosy l chloride (105.254 g, 0.552 mol) was dissolved in THF (200 mL). Using a separatory funne l, the solution of tosyl chloride was added slowly to the reaction mixture. The ice bath was allowed to come to room temperature and the solution was left to stir overn ight. The next morning, water (500 mL) was stirred in a separate Erlenmeyer flask. The EGO ts2 solution was slowly added to the stirring water over 30 minutes. The white solid was then filtered out of the yellow liquid. The solid was washed in the filter funnel with diet hyl ether (60 mL), followed by 0.1 M sulfuric acid (60 mL), water (60 mL), and then fina lly with enough ether to remove the majority of the water. The solid was then dried in vacuo. The yield was 76.75 g (70%). The proton NMR indicates that there was a small amo unt of ether, but the product was otherwise clean. 1HNMR: (CDCl3) 7.75(d, 4H, Ts), 7.39(d, 4H, Ts), 4.2(s, 4H, OCH2 CH2O), 2.5(s, 6H, TsCH3) ppm. 13CNMR: (CDCl3) 145.5, 132.5, 130.2, 128.2, 66.9, 21.9 ppm.
45 Synthesis of 1,4,7-tosyl-1,4,7-triazacyclononane (T ACNts3) (step 3) The procedure of this reaction was developed by Al exandria D. Liang and the author from the theses of past students.57-59 The synthesis was repeated many times with consistent results. In a 2L, 3-neck round bottom fl ask equipped with a water condenser, a thermometer, and a pressure equalizing addition fun nel, 60% NaH in mineral oil (unfiltered, 6 g, 150 mmol) was stirred with DMF (5 00 mL, anhydrous, used as purchased from Acros) and heated to 84oC. The heat was turned off while DETAts3 (42.430 g, 75 mmol) was added with vigorous stirrin g. The solution was heated to 110oC and went from cloudy white to yellow. Meanwhile in an Erlenmeyer flask, EGOts2 (27.783 g, 75 mmol) was added to 60% NaH in mineral oil (unfiltered, 0.126 g, 3.14 mmol) dissolved in DMF (200 mL) and the solution wa s stirred. The EGOts2 solution was added to the DETAts3 solution at room temperature over 2 hours. The sol ution was cooled to 90-95oC and left overnight. In the morning the solution h ad become orangeyellow. The heat was turned off and the solution wa s added to water (6 L) while stirring vigorously. The solution was left stirring overnigh t and a white precipitate formed. The solid was filtered off and washed in the filter wit h ice cold water, ethanol, and then ether. The light tan solid was dried in vacuo. The yield w as 33.14 g (56 mmol, 75% including impurities) of an off-white solid. 1HNMR: (CDCl3): 7.8(d, <1H, DETAts3), 7.75(d, 6H, Ts), 7.61(d, <1H, DETAts3), 7.39(d, 6H, Ts), 3.5(s, 12H, CH2s of ring), 2.4(s, 9H, TsCH3), and several impurity peaks: 5.2(0.5H, DCM), 3.2( 3H, CH2s of DETAts3),
46 3.0(0.1H, unknown), 2.9(0.1H), 1.6 (2H, water), 1.2 (0.8H, unknown), 0.8(0.3H, unknown) ppm. Detosylation of TACNts3: Synthesis of 1-tosyl-1,4,7-triazacyclononane (TAC Nts) (step 4) This procedure was adapted from Sessler et. al.60 The synthesis was repeated many times with consistent results. Impure TACNts3 from the previous step was used. The schlenk line was equipped with an acid trap con sisting of Teflon tubing attached to the exit of the bubbler, with a glass funnel insert ed in the other end. This funnel was hung over water in a beaker with a slightly larger mouth than that of the funnel. This apparatus is shown in figure 22. Figure 22. The acid trap apparatus.
47 The funnel must be secured above the water, as the water can be pulled into the line because HBr is so soluble in water. The reacti on can be done without the acid trap, but the hood (and lab) will likely fill with HBr fu mes. Then, a water condenser and thermometer were attach ed to a 250 mL round bottom flask with a thermometer adapter. TACNts (1,4,7-tritosyl-1,4,7triazocyclononane, 7.412 g, 12.5 mmol) was transfer red to the flask and put under nitrogen. Wet phenol (9.9 g, ~100 mmol) was filtered under nitrogen using the set up shown in figure 23 and added to the flask immediate ly afterward. The amount added to the flask was not known precisely. Figure 23. The assembly used to filter under nitrog en is shown. To operate this set up, first turn the nitrogen on high. Then, begin to slowly turn on the vacuum while watching the nitrogen bubbler. Do not allow air to travel into the line, this defeats the purpose and contaminates the atmosphere of the reaction flask, in addition to possibly getting mineral oil in the line. Before adding hydrogen bromide, the water condenser was attached to an ice bath using a water pump and a cooler. Then, 100 mL of 33 % wt HBr in glacial acetic acid was added slowly to the reaction flask while the reacti on was stirring and ice water ran through the water condenser. The flask was then flu shed with nitrogen and the joint
48 sealed with vacuum grease. The reaction was heated to 90C and stirred, resulting in a clear maroon solution with cream-colored foam aroun d the thermometer. The reaction was left in this state for 36 hours, during which t he ice bath was regenerated every 8 hours. The mixture was cooled to room temperature a nd the white precipitate was filtered off and washed with diethyl ether. The solid was dr ied in a vacuum desiccator to yield 4.396 g (9.9 mmol, 79%) of off-white powder. The so lid was then dissolved in 1M NaOH (60 mL). The magenta solution was extracted quickly with chloroform (8 x 12 mL) and dried over magnesium sulfate. The chloroform was re moved in vacuo to yield 2.503 g (8.83 mmol, 70.5%) of clear, slightly pink oil. Ses sler et al. reported clear, colorless oil, but in our repetitions of this procedure the produc t was occasionally a clear, slightly peach, crystalline solid. 1HNMR: (CDCl3) 7.7(d, 2H, Ts) 7.3(d, 2H, Ts) 3.3-3.1(m, 8H, TsNCH2CH2N) 2.9(s, 4H, CH 2 NH) 2.4(s, 3H, TsCH3) 1.8(s, 2H, H2O) ppm. 13CNMR: (CDCl3) 144.5, 141.3, 137.2, 130.8, 130.1, 129.2, 128.1, 127.9, 63, 56.4, 56.2, 41.5, 21.5 ppm.
49 a. Route 1a: Towards the synthesis of 1,4-diisopropyl1,4,7triazacylcononane-N-acetate (iPr2TCMA) 1,4-diisopropyl-7-tosyl-1,4,7-triazacyclononane (iPr2TACNts) (step 5): This synthetic procedure is modified from Mahapatra et al.61 The reaction was repeated several times with consistent results. Und er nitrogen, TACNts (1.148 g, 4.07 mmol), distilled acetonitrile (4.5 mL), and 2-bromo propane (16.27 mmol, 1.54 mL) were added to a round bottom flask. Anhydrous sodium car bonate (1.821 g, 17.2 mmol) was ground to a fine powder and irradiated in a microwa ve oven for 4 x 30 seconds to remove water. The solid was quickly added to the flask. Th e solution was stirred and heated to just below the boiling point of 2-bromopropane (59. 4oC) for 18 hours. The solution was cooled to room temperature and the white solid was filtered off twice, washing with acetonitrile. The solvent was removed by rotary eva poration and the yellow oil with some colorless crystals was redissolved in 4.5 mL of chl oroform. The chloroform layer was washed with 2.8 mL 1 M sodium hydroxide and the aqu eous layer was extracted with 3 x 2.5 mL chloroform. The combined organic layers were dried over magnesium sulfate and filtered. The chloroform was removed by rotary evap oration, yielding 1.176 g (3.199 mmol, 78.2%) of slightly yellow clear oil. When the sodium carbonate is finely ground there is more surface area available to react, but multiple filtrations are often necessary to remove all the fine powder after the reaction. The need to filter several times can be
50 avoided by increasing the available surface area by adding more sodium carbonate without grinding. 1HNMR: (CDCl3) 7.75(d, 2H, Ts), 7.39(d, 2H, Ts), 3.3(m, 4H, iPrNCH 2 ), 2.9(m, 4H, TsNCH 2 ), 2.8(sept, 2H, CH (CH3)2), 2.47(s, 4H, NCH2 CH2N), 2.4(s, 3H, Ts), 0.95(d, 12H, CH(CH 3 )2) ppm. 13CNMR: (CDCl3) 143, 137, 129.692, 127.362, 77.412, 53.927, 52.516, 50.575, 21.672, 18 .463 ppm. 1,4diisopropyl-1,4,7-triazacyclononane ( iPr2TACN) (step 6) This synthetic procedure is modified from Mahapatra et al.61 and was repeated several times with similar success. A round bottom flask was charged with iPr2TACNts (0.654 g, 1.78 mmol) and concentrated sulfuric acid (2.35 mL). The solution was stirred and heated under nitrogen to 95-105oC for 18 hours. In the morning, the dark solution was cooled to room temperature, and slowly brought to pH > 11 with 40% NaOH over an ice bath. Solid formed in the solution upon basifyi ng (presumably sulfate salts). This was overcome by the addition of either ice or water, fo llowed by more base. Then, the solution was washed with chloroform and extracted u ntil the organic layer was colorless. The chloroform solution was dried over anhydrous ma gnesium sulfate, filtered, and the chloroform was removed under vacuum. This yielded 0 .222 g (1.04 mmol, 58.5%) dark yellow oil. This procedure has been attempted at bo th higher and lower temperatures. At 120oC, the ring is degraded, and at 80-85oC the tosyl groups were not removed. 1HNMR: (CDCl3) 2.9(sept, 2H, CH (CH3)2), 2.75(t, 4H, HNCH 2 ), 2.6(t, 4H, iPrNCH 2 ), 2.5(s, 4H,
51 NCH2CH2N), 1.1(d, 12H, CH(CH 3)2) ppm. 13CNMR: (CDCl3) 53.1, 48.8, 47.5, 47, 18.9 ppm. Synthesis of ethyl 2-(4,7-diisopropyl-1,4,7-triazon an-1-yl)acetate (iPr2TACNEA) (step 7): This procedure is taken from Scarpellini et. al.62 Dichloromethane (1.5 mL) and iPr2TACN (90 mg, 0.422 mmol) were added to a three-neck round bottom flask. The flask was flushed with nitrogen, and the extra neck s covered with septa. Through one septum, ethyl bromoacetate (50 L) was added dropwi se. The solution was stirred for 25 minutes. Then the solution was placed on a water ba th at room temperature while triethylamine (60 L, 0.430 mmol) was added dropwis e through a septum. The solution was allowed to stir for 45 hours. The volume of the solvent was reduced, so more was added periodically. Then decolorizing charcoal was added. After one hour of stirring with charcoal, the solution was filtered through Celite and the filter cake was washed with dichloromethane and 2 mL of water. The aqueous laye r of the filtrate was then washed with dichloromethane 4 x 2 mL and the combined orga nic layers were dried over anhydrous magnesium sulfate. The solvent was remove d under vacuum. This yielded yellow oil. The proton NMR spectrum shows that the ring has been degraded. The yield was not calculated. This procedure should be attemp ted again without leaving the
52 reaction flask in the water bath. The water bath ma y have been heated enough by the stir plate to cause the ring to degrade, as the ring is unstable in dichloromethane. Synthesis of iPr2TACNEA (step 7): This procedure was modified from Warden et al.63 A round bottom flask was charged with acetonitrile (3.8 mL), anhydrous sodiu m carbonate (0.403 g), and iPr2TACN (93 mg, 0.43 mmol). While the yellow-orange mixture stirred, ethyl bromoacetate (50 L, 0.43mmol) was added slowly. The mixture was sti rred and lightly refluxed for 3 days. The mixture was cooled to room temperature an d the solid was filtered off. The acetonitrile was removed by rotary evaporation and the resulting orange oil was taken up in water (6.4 mL) and brought to pH = 12 with 5 M N aOH. This aqueous suspension was extracted with chloroform 6 x 6.5 mL. The organic l ayer was dried over magnesium sulfate and the chloroform was removed under vacuum The yield was 50 mg (0.185 mmol, 43%) of light orange oil. The 1HNMR spectrum shows that the ring was degraded.
53 Synthesis of iPr2TCMA (step 8): The procedure for this reaction is modified from a n undergraduate thesis by Duncan Steward, a previous student.64 The reaction was monitored by proton NMR. A screw-cap NMR tube was charged with iPr2TACN (54.0 mg, 0.253 mmol), chloroacetic acid (23.9 mg, 0.253 mmol), and LiOHH2O (12.6 mg, 0.300 mmol) in approximately 0.75 mL of deuterated methanol and 0.44 mL deuteriu m oxide. Not everything dissolved, but the mixture was heated under nitrogen at 60oC for 21 hours. The mixture was cooled to room temperature and placed in the fridge, under nitrogen overnight. The next day, the methanol-d4 was removed by rotary evaporation, and the remaini ng aqueous solution was extracted with chloroform. The water was removed fr om the aqueous layer by rotary evaporation to yield an oily, slightly cloudy solid The 1HNMR reveals that chloroacetic acid is still present. The 13CNMR shows all the peaks of the expected product, b ut there are also a few unexpected peaks, possibly due to ch loroacetic acid, acetic acid, and methanol. 1HNMR: (D2O) 4.1(s, 5.6H, ClAcOH), 3.4(s, 2H, OOCCH2N), 3.35(sept, 2H, CH (CH3)2), 3.1-2.7(m, 12H, ring CH2s), 1.1(d, 12H, CH(CH 3)2) ppm. 13CNMR: (D2O) 179.7, 175.1, 167.6, 57.6, 53.2, 48.0, 46.2, 44.7, 43.9, 17.2 ppm. The procedure above was repeated on a larger scale using a solvent system of 95% ethanol and water (same proportions as above) a s specified by Steward, and
54 refluxing with an external temperature of 51-61oC. In an attempt to reduce the amount of chloroacetic acid present in the product, a 3:4 rat io of chloroacetic acid to iPr2TACN was used. The yield was a mixture of brown oil and a cl oudy white oily solid, 43%. If the weight of the lithium hydroxide monohydrate is take n into account, 88 mg (0.325 mmol, 32%) of product remains. The 1HNMR shows that the addition of the pendant arm was successful, but impurity peaks are significant, and the ratio of chloroacetic acid to product was not significantly affected. Taking up t he solid in basic water and extracting with chloroform would likely improve the purity of the product. 1HNMR: (D2O) White solid: 8.4(s,0.5H), 4.1(s,7H), 3.95(d, 2H), 3.6(d, 0.6H), 3.4-3.2(m,4H), 3.0-2.8(m, 3H), 2.4(s, 0.4H),1.3-1.0(m, 12H) ppm. Brown oil: Integr ation unsure. 8.4s, 7.7d, 7.4d, 4.1s, 3.9s, 3.8s, 3.6q, 3.4-2.5m, 2.4s, 1.4-1.0m ppm. 13CNMR: (D2O) 177.9, 174.1, 132.4, 128.3, 69.5, 64.2, 56.2, 46.8, 17.1 ppm.
55 b. Route 1b: Towards the synthesis of 1,4-dibenzyl-1,4 ,7triazacyclononane-N-acetate (Bn2TCMA) Synthesis of 1,4-dibenzyl-7-tosyl-1,4,7-triazacyclo nonane (Bn2TACN) (step 5) The procedure is modified from Alexandria D. Liang s procedure,65 which was adapted from Mahapatra et al.61 The synthesis was repeated several times with cons istent results. A round bottom flask equipped with a stir bar and suspended in a polyethylene glycol bath was charged with fresh TACNts (0.845 g, 2.982 mmol), followed by 110 mL of acetonitrile. The acetonitrile can be dried over magnesium sulfate or distilled over calcium hydride. The flask was purged with nitrogen and then ground anhydrous sodium carbonate (2.663 g, 25.125 mmol) was added. The fla sk was purged again, and heating began. The external temperature rose to ca. 60C. A fter allowing to stir and heat for about 25 minutes, benzyl chloride (0.79 mL, 6.865 mmol) w as added dropwise. The boiling point of benzyl chloride is 65C, and the internal temperature of a flask is approximately ten degrees less than the external temperature. Thi s means that the external temperature must reach 75C before benzyl chloride will boil of f. The reaction was left heating and stirring under nitrogen for 19 hours. The temperatu re ranged between 61-71C. After 19 hours, the reaction mixture was filtered on a coars e frit (although a Buchner funnel with
56 filter paper has been used successfully as well). T he filtrate was dried over anhydrous magnesium sulfate and filtered. Coarse frits are no t sufficient for magnesium sulfateuse either a medium frit or a Buchner funnel. Approxima tely half of the solvent was removed from the filtrate by rotary evaporation. The reduce d filtrate was placed in the fridge overnight. By the next day, crystals had formed. Th e supernatant was removed by filtration on a Buchner funnel. The crystals on the filter pad were washed with cold pentane. Excess pentane on the crystals was removed by rotary evaporation. This yielded 0.361 g (0.86 mmol, 101%) of clear colorless crysta ls. Despite the yield being larger than is possible, the product was taken to be pure based on proton NMR and used with the assumption that it was not completely dry. Occasion ally, the product was more powdery and cream colored. 1HNMR: ((D3C)2CO) 7.67(d, 2H, Ts), 7.4-7.19(m, 12H, Ts + Ph), 3.67(s, 4H, NCH2Ph), 3.27(t, 4H, TsNCH2CH 2N), 3.03(t, 4H, TsNCH 2CH2N), 2.84(s, broad, 1.5H (water)), 2.68(s, 4H, NCH2CH2N), 2.39(s, 3H, TsCH3), 2.05(p, NMR solvent) ppm. 13CNMR: ((D3C)2CO) 143, 141.1, 130.6, 129.9, 129.1, 128, 127.7, 63 .0, 56.4, 56.1, 51.5, 21 ppm.
57 Synthesis of 1,4-dibenzyl-1,4,7-triazacyclononane ( Bn2TACN) (step 6) KOH Method The nucleophilic detosylation of Bn2TACNts was attempted once. In a round bottom flask with a reflux condenser, Bn2TACNts (44 mg, 0.1 mmol) was stirred in ethanol (5 mL). The solid partially dissolved, resu lting in an orange slurry. KOH (5.5 mg, 1.1 mmol) in H2O (0.16 mL) was added dropwise and the mixture was brought to reflux. The reaction was left under nitrogen for nine hours The pH was measured to be ca. 12.5. An aliquot was removed and reduced to dryness. The residue was taken up in CDCl3 and a proton NMR was taken. The spectrum indicates only starting material and ethanol. 1HNMR: (CDCl3) 7.63(d, 2H, Ts), 7.28(m, integrated obscured by s olvent, Ph + Ts), 3.67(s + q, 4.6H, CH2Ph + EtOH), 3.23(t, 4H, TsNCH2CH 2 ), 3.04(t, 4H, TsNCH 2 CH2), 2.7(s, 4H, unknown), 2.39(s, 3H, TsCH3), 1.23(t, 1H, EtOH) ppm.
58 Sodium Naphthalenide Method: This synthesis is adapted from Alonso and Andersson s general procedure for deprotecting sulfonyl aziridines.66 Several minutes before beginning the reaction, an oven-dried needle was attached to the nitrogen line and nitrogen was allowed to flow through to remove as much water and oxygen from the tubing as possible. In a scrupulously dry single-neck round bottom, naphthal ene (208 mg, 1.62 mmol) was dissolved in ~4mL of THF freshly distilled from sodi um/benzophenone. The flask was sealed with a new septum (dried in an oven at 140C for ~1 hour). The needle attached to the nitrogen line was inserted into the septum and the flask was flushed with nitrogen. Then, sodium metal (33 mg. 1.44 mmol) was added to the flask while maintaining a high flow of nitrogen from the needle and opening the ap paratus minimally. The flask was again sealed and flushed with nitrogen. Slowly, the sodium dissolved and the solution turned very dark green, indicating the presence of the radical anion, naphthalenide. The solution was allowed to stir at room temperature fo r three hours, but shorter time would probably be sufficient as long as there is no solid sodium remaining. The first time this synthesis was attempted, the solution was cooled wi th liquid nitrogen, but not immersed. This was excessive cooling. The literarture procedu re cooled the solution to -78C. The second time the synthesis was performed the solutio n was cooled to -78C, and the
59 desired product was obtained. Once the solution was chilled, Bn2TACNts (37 mg, 0.129 mmol) dissolved in dry THF (1 mL) was added dropwis e through the septum using a gastight syringe equipped with a luer lock needle. The solution was stirred for 45 minutes at -78C, and then added to 3M NaCl (10 mL) in a separ atory funnel. The dark green solution lost its color. The aqueous layer was then acidified to pH 1 with HCl(aq) and extracted with DCM 3 x 10 mL to remove naphthalene. The aqueous layer was then basified to pH 13 with 40% NaOH and extracted with DCM 3 x 10 mL and ethyl acetate 2x 10 mL. The organic layer was then reduced in vac uo to yield a yellow oil (27 mg, 0.087 mmol, 68%). 1HNMR: ((CD3)2CO) 7.8(dod, 0.8H, naphthalene), 7.36(dod, 0.8H, naphthalene), 3.6(s, 4H, CH 2 Ph), 2.8(t, 4H, HNCH 2 CH2NBn), 2.7(t, 4H, HNCH2CH 2 NBn), 2.42(s, 4H, BnNCH2CH2NBn), 1.7(s, 4H, unknown impurity), 1.16(s, 0.4H, unknown possibly NH) ppm. 1CNMR: ((CD3)2CO) 140.2 (ipso-Ph), 130.56(orthoPh), 129.6(meta-Ph), 127.08(para-Ph), 61.91(CH2Ph), 52.01(BnNC H2CH2NH), 49.41(BnNCH2CH2NBn), 44.65(BnNCH2C H2NH) ppm. HBr/Phenol Method: This synthesis was heavily modified from Sessler e t al.60 Dry, crystalline Bn2TACNts (0.361 g, 0.86 mmol) was added to a round bo ttom flask with a thermometer
60 adapter and stir bar. Then, phenol (0.907 g, 9.638 mmol) was filtered under nitrogen using the apparatus shown in the description of the detosylation of TACNts3. Once relatively dry, the phenol was quickly transferred to the round bottom flask, and the flask flushed with nitrogen. An acid trap was then assemb led, as described in the detosylation of TACNts3. Then, an ice bath with a water pump was attached to the water condenser. With stirring, and ice water running throught the c ondenser, HBr 33% wt in glacial acetic acid (8 mL) was slowly added to the flask. The solu tion was then heated to 90C. The temperature spiked to 110C, and was turned down. T he temperature then varied between 82 and 89C for 45 hours. During the course of the reaction the ice bath was regenerated every 8 hours. The reaction was followed by TLC. Af ter 45 hours, the dark reddish brown solution was cooled to room temperature, and some light precipitate formed. Then, the reaction mixture was added in portions to dieth yl ether cooled to just above its freezing point to precipitate the product. The flas k was continually swirled. It would have been better to use a stir bar. More tan precipitate formed, and the mother liquor became bright yellow. The slurry was filtered to remove th e liquid. The filter cake was washed with cool diethyl ether. There were brown shiney ch unks in the tan powder. By 1HNMR, the reaction degraded the ring (large multiplets in the the alkyl region). The solid weighed 0.364 g (1.176 mmol), which is more than 10 0% yield. The yield was not recalculated because the product was not useful for further synthesis. 1HNMR: (CDCl3 and DMSO-d6) 7.70-7.46(m, 13.5H, Ts + Ph), 7.36-7.26(m, solven t obscures integration), 5.3(s, broad, 22.1H), 4.06(s, 5H), 3.59-2.8(m, 111H degraded ring), 2.53(s, 7H), 2.37(s,7.5H), 1.13(q(?), 4.1H) ppm.
61 c. Route 1c: Towards the synthesis of 1,4-diphenyl-1,4 ,7triazacyclononane-N-acetate (Ph2TCMA) Synthesis of 1,4-diphenyl-7-tosyl-1,4,7-triazacyclo nonane (Ph2TACNts) (step 5) This procedure was modified from Nakanishi and Bolm67 and repeated several times with similar results. A round bottom flask with a water condenser and a stir bar was charged with tris(dibenzylideneacetone)dipalladium(0) (Pd2(dba)3) (9 mg, 0.0098 mmol), phosphine 7 (6 mg, 0.015 mmo l), monobromobenzene (40 L, 0.38 mmol), TACNts (42 mg, 0.19 mmol) in toluene (1.5 mL), and sodium tbutoxide (55 mg, 0.572 mmol) in order. Then, the re action was heated to 100C for 2 hours. The dark brown solution was then cooled to r oom temperature, and drum ethyl acetate (3.8 mL) was added with stirring. A precipi tate formed, and the slurry was filtered through Celite, giving an orange-brown liquid. The filtrate was dried over anhydrous sodium sulfate and filtered. The solvent was remove d by rotary evaporation, yielding 0.1
62 g (0.229 mmol) of orange-brown viscous oil. This is more than 100% yield, so the product is either impure, wet, or both. In a separa te instance of this synthesis, the product was purified by column chromatography using 2:1 pen tane: ethyl acetate on a silica column. This yielded approximately 10% of the start ing material weight. In the future, 1% triethylamine should be incorporated into the so lvent system for higher yield. 1HNMR: (CDCl3) 7.65(d, 2H, Ts), 7.29-7.20(m, 6H, Ts + Ph), 6.766.69(m, 6H, o-Ph + p-Ph), 3.63(s + t, 8H, NCH2CH2N + PhNCH2), 3.38 (t, 4H, TsNCH2), 2.41 (s, 3H, TsCH3) ppm. 13CNMR: (CDCl3) 147.7, 130, 129.6, 129.2, 128.4, 127.6, 125.5, 11 7.1, 112.7, 53.7, 51.9, 49.3, 25, 21.7 ppm. Synthesis of 1,4-diphenyl-1,4,7-triazacyclononane ( Ph2TACN) (step 6) HBr Method This procedure was adapted from Sessler et al.60 A round bottom flask was equipped with a stir bar and a water condenser atta ched to an ice bath. The acid trap described in the synthesis of TACNts was set up. Ph2TACNts (0.904 g, 2 mmol) was added to the round bottom flask. Phenol (1.9 g, 20. 2 mmol) was filtered under nitrogen using the procedure described in the synthesis of T ACNts and quickly transferred to the
63 flask. Very quickly, 33%wt HBr in glacial acetic ac id (12 mL) was added to the flask with positive pressure of nitrogen and stirring. Th e flask was flushed with nitrogen, and heated to 90-95C for 36 hours. The reddish purple reaction mixture with a very small amount of light-colored precipitate was cooled to r oom temperature and filtered on a Buchner funnel. The filter cake was washed with eth er, upon which a solid precipitated out of the filtrate. The filtrate was filtered agai n. The resultant powdery, purple solid quickly became black and tar-like on the filter pap er. Proton NMR of these solids revealed that the reaction degraded the ring, indic ated by the large multiplets in the alkyl regions. This reaction should be attempted again wi th purified starting material and possibly at lower temperature (approximately 80-90 C). Trace amounts of palladium from the coupling reaction might be wreaking havoc on this reaction. KOH Method The nucleophilic detosylation of Ph2TACNts was attempted once. In a round bottom flask with a reflux condenser, Ph2TACNts (101 mg, 0.23 mmol) was stirred in ethanol (5 mL). The solid partially dissolved, resu lting in an orange slurry. KOH (12.6 mg, 0.25 mmol) in H2O (0.36 mL) was added dropwise and the mixture was brought to reflux. The reaction was left under nitrogen for ni ne hours. The pH was measured to be
64 ca. 12.5. An aliquot was removed and reduced to dry ness. The residue was taken up in CDCl3 and a proton NMR was taken. The spectrum indicates mostly starting material. 1HNMR: (CDCl3) 8.1(d, 0.1H, free tosylate?) 7.65(d, 2H, Ts), 7.2 4(q, integration clouded by solvent, Ph + Ts), 6.73(d, 4H, o-Ph), 6.59(d, in tegration unsure, p-Ph), 4.68(s, 0.75H, H2O), 3.58(m, 14H, EtOH+TsNCH 2 CH2), 3.23(t, 3H, TsNCH2CH 2 ), 3.02(br s, 2H, unidentified impurity), 2.43+2.41(2s, 5H, TsCH3 + TsCH3?), 2.17(s, 0.5H, unknown impurityacetone?), 2.08(br s, 1H, H2O), 1.23(t, 5H, EtOH) ppm. Synthesis of 4,4'-(1,4,7-triazcyclononane-1,4-diyl) dibenzenesulfonic acid (SPh2TACN) (step 6) This procedure was heavily modified from Mahapatra et al.61 and attempted twice. A round bottom flask with a stir bar and a w ater condenser was charged with impure Ph2TACNts (87 mg, 0.199 mmol) and 3 mL concentrated su lfuric acid. The reaction was heated to between 95 and 105C for 18 hours. Following this, the reaction was cooled to room temperature and basified to pH 1 2 with 40% NaOH over an ice bath. A solid forms as the reaction mixture is basified. This was overcome by adding ice and water to the mixture. Then, the slurry was extracte d with chloroform 4 x 15 mL. The
65 organic layer was dried over anhydrous magnesium su lfate, filtered, and the solvent removed from the filtrate by rotary evaporation. Th is yielded 43 mg (0.1 mmol, ~50 %) clear oil. The 1HNMR shows a completely detosylated product. 1HNMR: (CDCl3) 7.8(d,2H, HO3SPh) 7.5(d,2H, HO3SPh) 3.6-0.9(large multiplets, integration unsure, degraded ring and solvents) ppm. II. Route 2 a. Route 2a Scheme 10. Route 2a to the synthesis of R2TCMA. Reagents: 1) TsCl, triethylamine, dichloromethane; 2) K2CO3, MeCN; 3) Na2CO3, MeCN, isopropyl bromide. Synthesis of 7-tosyl-1,4,7-triazacyclononane (TACNt s) (step 2)
66 This synthesis is very heavily modified from Huang et al.68 This procedure is a six-fold dilution of the concentrations specified b y Huang et al. A three-fold dilution was also performed, but with only 37% yield of the mono meric product, and a two-fold dilution yielded the polymer exclusively. This is f urther discussed in section III.A.II.a. A three-neck round bottom flask was equipped with a stir bar and a water condenser, and the remaining two necks were covered with septa. The septa were secured with rubber bands. Potassium carbonate (1.610 g, 11 .65 mmol) was finely ground and irradiated with microwaves to remove water, then pl aced in the round bottom flask and flushed with nitrogen. Acetonitrile (40 mL, distill ed from calcium hydride) was added to the flask, which was purged again and heated to ref lux. Then, (tosylazanediyl)bis(ethane2,1-diyl) bis(4-methylbenzenesulfonate) (DEAts3) (1.137 g, 2.00 mmol) synthesized by Alexandria Liang (see her 2010 thesis for synthesis65) was dissolved to 10 mL in distilled acetonitrile. Simultaneously, distilled ethylenedia mine (en) (134 L, 2.00 mmol) was diluted to 10 mL with distilled acetonitrile. These two solutions were quickly taken up into gastight syringes with luer lock needles. The syringes were degassed and each was inserted into one of the septa. Over the next 5 hou rs, the solutions were added dropwise to opposite sides of the flask with vigorous stirri ng. When adding these solutions, it is important to add only a drop at a time, and even mo re importantly, the solutions must be added simultaneously After the addition was complete, the reaction was heated at reflux overnight. The following morning, the reaction was cooled to room temperature and filtered several times to remove the potassium carb onate. Then, the filtrate was reduced by rotary evaporation to yield 0.486 g of pale yell ow oil (1.715 mmol, 85.7%). The 1HNMR indicates that the oil is the desired product, but is impure. It might be prudent to
67 do a simple acid-base extraction. 1HNMR: (CDCl3) 7.7(d, 2H, Ts), 7.3(d, 2H, Ts), 3.0(t, 4H, TsNCH2) 2.5(m, 4H, HNCH2), 2.2(s, 3H, TsCH3) ppm, plus many impurity peaks 2.8-1.0, see section III.A.II.a for spectra of both the monomer and the polymer. b. Route 2b Scheme 11. Route 2b to the synthesis of R2TCMA. Reagents: 1) DCM; 2) TsCl, ether, triethyl amine or pyridine; 3) and 8) K2CO3, MeCN; 4) and 9) iPr: Na2CO3, MeCN, isopropyl bromide; Bn: Benzyl chloride, Na2CO3, MeCN; Ph: Pd2(dba)3 (cat.), phosphine 7 (cat.), Na t-Butoxide, toluene, bromobenzene; 5) MeOH/KOH. This route was proposed by Eric Andreansky. The mer its and challenges of this route are further discussed in Chapter III. Synthesis of 2-(bis(2-hydroxyethyl)amino)acetic aci d (DEG (diethanolglycinate)) (step 6)
68 This procedure is an adaptation.69 Chloroacetic acid (5.1 g, 53.97mmol) was dissolved in deionized water (8.8 mL). Slowly, and over an ice bath, the solution was neutralized with 50% NaOH. Then, diethanolamine (5. 8 mL, 60.52 mmol) was added. The reaction flask was flushed with nitrogen and he ated to reflux for 4 hours. The clear solution was cooled to room temperature and placed in the fridge overnight. The solvent was partially reduced by rotary evaporation and the solution was placed in the fridge again, yielding clear, colorless, needle-like cryst als after a few hours. A crude proton NMR showed starting materials and products. The cry stals were washed with hot 80% methanol (40 mL, 55C). The funnel was kept hot wit h a heat gun. The filtrate was slowly cooled to room temperature and set in the fr idge overnight. This yielded clear colorless crystals. The supernatant was filtered of f and the crystals were washed with pentane and dried in vacuo to give 0.953 g (5.84 mm ol, 10.8% yield). The proton NMR shows the product, DEG, only. 1HNMR: (D2O) 3.95-3.88(m (t + s), 6H NCH2 CH 2O + NCH2COO), 3.45 (t, 4H NCH 2 CH2O) ppm. Synthesis of ethyl 2-(bis(2-hydroxyethyl)amino)acet ate (DEGEE (DEG ethyl ester)) (step 1) The procedure for this reaction was developed from Greenwald et al.70 This reaction was attempted at least six times, each in a slightly different way. The following is the most successful method used, but it has also been done similarly in ethanol.
69 Additionally, the reaction has been done using both potassium carbonate as an acid scavenger and with triethylamine as the base. Neith er was as effective as simply using an excess of diethanolamine. Dichloromethane (DCM, 100 mL) was dried over magnes ium sulfate and filtered. In a round bottom flask with a stir bar and a water co ndenser diethanolamine (DEA, 6.6 mL, 68.87 mmol) was added to DCM (50 mL). The graduated cylinder used to measure diethanolamine was washed with 25 mL of DCM, and th e rinse was added to the reaction flask. Then, ethyl bromoacetate (4.5 mL, 40.6 mmol) was added to the flask. The graduated cylinder used to measure ethyl bromoaceta te was rinsed with DCM (25 mL) and the rinse was added to the reaction flask. The slurry was stirred vigorously at room temperature. After a few minutes, the reaction unde rwent a significant exotherm and refluxed. The reaction was followed by TLC using 10 % methanol in DCM on aluminumbacked silica plates. These samples were visualized with UV lamp, iodine, and vanillin. After leaving the reaction for 18 hours (although t his is probably excessive), the cloudy white slurry was extracted with water 4 x 40 mL to remove diethanolamine. The organic phase was reduced by rotary evaporation without drying over a salt. The proton NMR of the resultant clear oil is somewhat convoluted, but seems to show a mixture of product and ethyl bromoacetate. Each time the organic phase was dried over a salt (either magnesium or sodium sulfate) the cloudy solution tu rned clear and there was no product present after rotary evaporation. 1HNMR: (CDCl3) 4.1(2.2H, q, EtOH and OOCCH 2 CH3), 3.9(0.1H, s, either OCH2COOEt or NCH2COOEt), 3.8(2H, s, unknown), 3.5(0.1H, q, OOCCH 2 CH3), 1.2(4H, t, EtOH and OOCCH2CH 3 )
70 III. Route 3 Scheme 12. Route 3 to the synthesis of R2TCMA. Reagents: 1) potassium carbonate and DMF; 2) TsCl, dimethylaminopyridine (cat), trie thylamine, DCM; 3) potassium carbonate, acetonitrile; 4) MeOH/KOH. Synthesis of 2,2'-(ethane-1,2-diylbis(benzylazanedi yl))diethanol (step 1) The procedure for this step is modified from Pulac chini and Watkinson.71 The synthesis was attempted four times. A round bottom flask was equipped with a stir bar and a water condenser. The apparatus was flame-drie d. To do this, first set up the vacuum pump and turn it on. Then, flush the apparatus with nitrogen. Turn the stopcock to the vacuum line and evacuate the flask. Wait until the vacuum pump stops gurgling. Then, start a Bunsen burner on a low flame. Starting at t he bottom of the EMPTY apparatus, heat the outside of the glass with the flame. Depen ding on how wet the glass is, one might see vapors rising. Keep the flame moving cons tantly, and work up the apparatus until you have reached the top. Be careful not to m elt the tubing. There should be no
71 water tubes attached to the condenser when flame dr ying. After drying is finished, close the stopcock, and take down the vacuum pump. Then, turn the nitrogen on high and open the line to the apparatus. Once the vacuum is relie ved, turn the nitrogen back down. Then, ethylene carbonate (3.527 g, 40.05 mmol) (dr ied in vacuo) was added to the flask under nitrogen flow, followed by DMF (20 mL). Next, finely ground and microwaved potassium carbonate (3.055 g, 22.1 mmol) was stirred into the slurry. N,NDibenzyl-ethylenediamine (2.67 mL, 11.33 mmol) was then added to the reaction flask. The flask was flushed with nitrogen, and brought to reflux at ~140-150C and left for 18 hours. Then, the reaction mixture was cooled to roo m temperature and the solvent was partially removed by rotary evaporation. Not all of the DMF could be removed in this way, so the remaining material was put under high v acuum with a warm water bath and stirring. This method resulted in a very thick yell ow paste. The flask was removed from vacuum, water was added (18.5 mL) and the slurry wa s stirred. This resulted in two phases, whereas a solid was expected to precipitate based on the literature procedure.71 The contents of the flask were sonicated for hours and then placed in the fridge in a fruitless attempt to cause a solid to crash out. Th e water layer was decanted off and the remaining residue was dried by rotary evaporation. The yield was not calculated, as the proton NMR revealed a mixture of reactants and degr aded reactants, with the possibility of a small amount of the desired product. This reac tion was repeated several times with similar results. Acetonitrile was also used as a so lvent, with similar results. Then, the reaction was performed neat, under nitrogen in an N MR tube and monitored by proton NMR. The product was seen amongst the reactant and degradation peaks, but it was not isolated. 1HNMR: (CDCl3) (Integration compromised) 7.33-7.24 (m, Ph), 4.47 (s),
72 4.41(s), 3.76-3.55(m), 3.40-3.32(m), 3.16(s), 2.95( s), 2.88(s), 2.75(s), 2.75-2.66(m), 2.02(s, broad), 1.22(t, ether) C. Computation The geometry of the ligands Ph2TCMA and o,o-dimethyl-Ph2TCMA were optimized using DFT with the basis set B3LYP/6-31++ (d,p) or higher in the modeling program Gaussian. In the case of each ligand, two s eparate calculations were performed: one starting with the phenyl ring perpendicular to the axis of the nitrogen atom p orbital, and one with the phenyl parallel to the same axis. The ligands were in their deprotonated state in all cases. The optimizations assume that t he molecule is in a vacuum. This is not the relevant environment, as steric considerations become much more heavily weighted in a vacuum than in solution, and electrostatic int eractions with the solvent are not considered. This may compromise the reliability of the data to some extent.
73 Chapter III Results and Discussion Progress toward the synthesis of all three target m odel complexes of oxalate oxidase and oxalate decarboxylase has been achieved to varying degrees. The target ligand iPr2TCMA is a known ligand, so the synthesis was more e asily accomplished than the other two target ligands, Ph2TCMA and Bn2TCMA, which are thus far unreported. The synthesis of iPr2TCMA was brought to the penultimate step with the s ynthesis of analytically pure iPr2TACN (figure 25). The synthesis of iPr2TCMA was accomplished using chloroacetic acid, but the product contained chloroacetic acid as an impurity (see appendix for spectra). At this point the project pa ssed into the hands of another student, Jaclynn Windsor. Upon relinquishing this goal, Ph2TCMA and Bn2TCMA became the new target ligands. In contrast to the isopropyl de rivative, the synthesis of these ligands involved a large amount of exploratory synthesis. B ecause the ligands differ only in the identity of the substituents, the syntheses can fol low common routes. Several previous students in the Sherman lab have developed the synt hesis of TCMA (scheme 13).58,65,72-75 Scheme 13. The synthetic route used by previous stu dents in the Sherman lab to synthesize TCMA.
74 Presumably, a similar approach could be used for th e proposed substituted TCMA ligands. The major difference in the route to the d isubstituted ligands is that only two tosyl protecting groups are removed from TACNts3 before the substituents are added, the final tosyl being removed afterward. Using the rout e shown in scheme 14, Alexandria Liang made significant progress toward the synthesi s of Bn2TCMA with the synthesis of Bn2TACNts. Scheme 14. The route used by Liang to synthesize Bn2TACNts. Liang used the same route as shown in scheme 13 to produce TACNts3. The main difficulty in synthesizing this target li gand was the removal of the final tosyl protecting group from Bn2TACNts without also removing the benzyl substituent s. This difficulty arises because the majority of proc edures used to remove N-Ts protecting groups are also used to remove N-Bn protecting grou ps.76 Many of these methods were attempted by Liang, including samarium iodide, HBr in glacial acetic acid with phenol, reduction with sodium naphthalenide, refluxing in c oncentrated sulfuric acid, and a photolytic cleavage. The route that seems to have w orked the best for Liang was a photochemical detosylation. However, this resulted in a product that has since eluded satisfactory characterization due to the splitting of the benzylic singlet into two singlets (figure 24 top). Additionally, the carbon NMR spect rum (figure 24 bottom) contains two peaks in the range expected for benzylic carbons at 64.6 and 58.8 ppm. This splitting of the peaks sheds doubt upon the identity of Liangs product because the two benzylic carbons of Bn2TACN and their attached hydrogens are expected to b e in identical
75 environments, resulting in only one chemical shift for each type of nucleus. Furthermore, the photochemical detosylation procedure is time co nsuming and requires a large amount of space and special equipment. Thus, other methods of detosylation were pursued by the author, including detosylation with HBr 33%wt in gl acial acetic acid with phenol and nucleophilic detosylation with KOH. The acidic meth od was unsuccessful, destroying the TACN ring. The KOH method yielded pure starting mat erial after 18 hours of reflux in ethanol. Finally, the selective detosylation of Bn2TACNts to yield Bn2TACN was successfully completed using the sodium naphthaleni de detosylation method adapted from Alonso and Andersson.66 Bn2TACN, 1,4-dibenzyl-1,4,7-triazacyclononane, is a novel compound; it is unreported in the literature to the knowledge of the author at the time of writing. Concurrent with these methods, oth er routes were pursued in an attempt to circumvent the necessity of deprotecting. Specif ically, routes 3 and 2b ( vide infra ) were pursued as alternative syntheses of Bn2TCMA. Unfortunately, they were unsuccessful in the hands of the author, but they m ay prove useful in the future. The synthesis of Ph2TCMA, a completely novel ligand, was developed to t he antepenultimate step (Ph2TACNts), and the same detosylation methods were att empted as those used for Bn2TACNts in addition to refluxing in concentrated sul furic acid. The acidic methods were unsuccessful, as was the KOH me thod. The results of detosylation are uncertain; it seems as though the Na/naphthalen e detosylation may have been successful, but the NMR spectra are not satisfactor ily clean. Simultaneously, route 2b ( vide infra ) was attempted in an effort to avoid detosylation. These routes have been explored to varying degrees and the merits and disa dvantages of each are discussed in section III.A.
Figure 24. Top: t he proton NMR detosylation of Bn2 TACNts two singlets at ~3.4 and ~3.8 ppm. putative Bn2 TACN in CD 63.6 and 58.8 ppm (circled he proton NMR spectrum of Liang's product of the photochemical TACNts in CD3CN Notably, the benzylic hydrogens are split into two singlets at ~3.4 and ~3.8 ppm. Bottom: the carbon NMR spectrum of Liangs TACN in CD 3 CN. Notably, there are two peaks in the benzylic ra nge at (circled ). The spectra are taken from Liangs thesis. n r r 76 of Liang's product of the photochemical Notably, the benzylic hydrogens are split into carbon NMR spectrum of Liangs CN. Notably, there are two peaks in the benzylic ra nge at taken from Liangs thesis. 65
77 A. Synthesis I. Route 1 Scheme 15. Route 1 to the synthesis of R2TCMA. Reagents: 1: H2O/NaOH, TsCl, ether; 2: H2O/NaOH, TsCl, THF; 3: NaH, DMF; 4: HBr 33%wt in AcO H, Phenol; 5: iPr: Na2CO3, MeCN, isopropyl bromide; Bn: Benzyl chloride, Na2CO3, MeCN; Ph: Pd2(dba)3 (cat.), phosphine 7 (cat.), Na t-Butoxide, toluene bromobenzene; 6: con H2SO4 or HBr 33%wt in AcOH, Phenol or Na/naphthalene, TH F; 7) a) ethyl bromoacetate, Na2CO3, MeCN b) ethyl bromoacetate, triethylamine, DCM; 8 ) Chloroacetic acid, LiOHH2O, EtOH/ H2O; 9) MeOH/KOH. The first step in this route (scheme 15) is the syn thesis of DETAts3. The procedure generally yielded 80% product with tosyl chloride impurities that did not negatively impact further steps. Step 2, the synthe sis of EGOts2, was similarly successful, with 70% yield. The ring closure to create TACNts3 resulted in 75% desired product with the solvents being the only major impurities. The s ubsequent detosylation (step 4) to produce TACNts gave 70.5% very pure product. At thi s point (step 5 of scheme 15), the synthesis diverges according to the identity of the R group. Each derivative is discussed in turn in sections III.A.I.a-c.
a. Synthesis of The addition of the substituents also went fairly w ell, with a yield of 78.2% for propylation. The literature isopropyl derivative was accomplished using concent rated sulfuric acid to obtain yields typically near 50% iPr2 TACN, 94%. The overall yield at this step is 16.9%. Figure 25 The proton NMR spectrum of Synthesis of iPr2TACN The addition of the substituents also went fairly w ell, with a yield of 78.2% for propylation. The literature 61 procedure reports a yield of 90%. The detosylation of the isopropyl derivative was accomplished using concent rated sulfuric acid to obtain yields TACN, the best being 58.5%. The literature61 indicates a yield of The overall yield at this step is 16.9%. The proton NMR spectrum of iPr2TACN in CDCl3. n r 78 The addition of the substituents also went fairly w ell, with a yield of 78.2% for The detosylation of the isopropyl derivative was accomplished using concent rated sulfuric acid to obtain yields indicates a yield of
79 b. Synthesis of Ph2TACN The addition of substituents to form Ph2TACNts (figure 26) was performed according to the literature procedure67 using a Buchwald-Hartwig catalyst, which consisted of a phosphine ligand and Pd2(dba)3 (dba = dibenzylideneacetone). The yield was over 100%. This is attributed in part to residu al solvent, as the proton NMR spectrum (figure 26) indicates that the product is otherwise pure. However, some excess mass may also be attributed to the continued presence of the palladium catalyst after work up. This palladium impurity only became evident during the w ork up of the sulfuric acid detosylationwherein a blue film formed on the fla sk when it was rinsed with water.
Figure 26 The proton NMR impurities at 2.04 (acetone), 2.10, 4.1, and 1.26 ( ethyl acetate), and 0.9 and 1.6 (H2O). The carbon NMR The detosylation of the phenyl substituted derivati ve similar to the detosylation of the tosyl groups may have been position ~50% yield (14.5% overall). In the spectrum of Ph distinguishable tosyl doublet appears at 7.5 ppm (f igure 26), while the doublet in the same ra nge of the product spectrum appears at 7.8 ppm (fig ure 27). The proton NMR spectrum of PH2TACNts in CDCl3 There are solvent impurities at 2.04 (acetone), 2.10, 4.1, and 1.26 ( ethyl acetate), and 0.9 The carbon NMR spectrum is available in the appendix. The detosylation of the phenyl substituted derivati ve wa s attempted in similar to the detosylation of iPr2TACN using concentrated sulfuric acid It appears may have been removed and the phenyl rings sulfonated in the para ~50% yield (14.5% overall). In the spectrum of Ph 2 TACNts (figure 26), the distinguishable tosyl doublet appears at 7.5 ppm (f igure 26), while the doublet in the nge of the product spectrum appears at 7.8 ppm (fig ure 27). This assignment is rn 80 There are solvent impurities at 2.04 (acetone), 2.10, 4.1, and 1.26 ( ethyl acetate), and 0.9 (pentane), is available in the appendix. s attempted in a fashion It appears that sulfonated in the para TACNts (figure 26), the distinguishable tosyl doublet appears at 7.5 ppm (f igure 26), while the doublet in the This assignment is
81 uncertain, it is also possible that the tosyl group s remain and are shifted for some other reason. Unfortunately, there are many impurities in the alkyl region, making it impossible to discern whether the desired product is present, and also indicating that the ring was degraded. Figure 27. Proton NMR spectrum of the product of th e attempted sulfuric acid detosylation of Ph2TACNts in CDCl3. HBr 33%wt in glacial acetic acid with phenol was al so used according to the method described by Sessler et al.60 in an attempt to detosylate Ph2TACNts. This
82 procedure was unsuccessful, resulting in a spectrum containing large multiplets in the alkyl range that resemble the Pyrenees. These peaks were interpreted as evidence of the degradation of the TACN ring. c. Synthesis of Bn2TACN The detosylation of Bn2TACNts was attempted using HBr in acetic acid with phenol and using KOH. The acidic procedure was enti rely unsuccessful, leading to a degraded product, while the KOH reaction yielded st arting material. The synthesis of Bn2TACN from Bn2TACNts (92% with impurities) was finally achieved u sing Na/naphthalene in THF according to the general proc edure for deprotecting sulfonyl aziridines described by Alonso and Andersson66. The first time this reaction was attempted, liquid nitrogen was used as a coolant fo r the reaction. This temperature was excessively low. When the procedure was repeated us ing dry ice and isopropanol instead, the reaction was successful. The proton and carbon NMR spectra with their respec tive predicted spectra can be seen in figures 28 and 29. The predicted spectra ag ree quite well with the observed spectra. Unfortunately, naphthalene was carried thr ough the work up and lingers on as an impurity. It may be possible to remove it in vacuo as it has a fairly high vapor density. Alternatively, it could be removed by column chroma tography or by taking up the product in aqueous acid and extracting with toluene as naphthalene is likely to be highly soluble in toluene. The removal of naphthalene to p roduce a clean spectrum of this unreported compound is currently underway.
83 Impurities aside, the most interesting aspect of th e proton spectrum is the benzylic peak at 3.6 ppm. It is a singlet integrating for fo ur hydrogens, as would be expected. This result is in contrast with Liangs putative Bn2TACN which produced a spectrum with two singlets integrating for equal amounts (2H) in the range expected for the benzylic hydrogens (figure 25). Additionally, there were two peaks in the region expected for a benzylic carbon in the carbon NMR spectrum of Liang s product (figure 25), but there is only one peak in the spectrum of Bn2TACN produced by sodium detosylation (figure 29). Because of the symmetry of the molecule, it is expe cted that there would be only on peak. This discrepancy indicates that Liangs product may not have been Bn2TACN, but some other substance or combination of substances.
Figure 28. Top: The 1 HNMR spectrum of t he Na/naphthalene method predicted spectrum. HNMR spectrum of Bn2TACN in CO(CD3)2 as produced by he Na/naphthalene method Impurities: 7.75, 7.35 (naphthalene ) ppm r n 84 as produced by ) ppm Bottom: The
Figure 29. Top: carbon NMR detosylation in CO(CD 3 carbon NMR spectrum of Bn2 TACN from the Na/naphthalene 3 )2. Bottom: predicted spectrum. n r 85 TACN from the Na/naphthalene
86 II. Route 2 a. Route 2a Scheme 16. Route 2a to the synthesis of R2TCMA. Reagents: 1) TsCl, triethylamine, dichloromethane; 2) K2CO3, MeCN; 3) R = iPr: Na2CO3, MeCN, isopropyl bromide; R = Bn: Na2CO3, MeCN, BnCl; R = Ph: sodium tert-butoxide, phosphi ne 7 (see synthesis of Ph2TACNts in experimental for structure), Pd2(dba)3, toluene. This route (scheme 16) was pursued because it yield s TACNts in two steps as opposed to the four steps necessary in route 1 (sch eme 15). At first, route 2a was carried out with 0.2 M DEAts3 (tritosyl diethanolamine) as prescribed by the lit erature preparation,68 but this procedure resulted in the formation of a very clean polymer (figure 30). The melting points of these polymers ranged be tween 250-320 C, while the literature reports the product as pale yellow oil. In order to avoid the formation of polymers, twofold (0.1 M), threefold (0.067M), and sixfold (0.033M) dilutions of the literature procedure were performed. The twoand t hreefold dilutions still yielded primarily polymeric material. The sixfold dilution, which is the procedure described in section III.A, appears to have been successful. How ever, the spectrum (figure 31) is distinct from the spectrum of TACNts obtained throu gh the detosylation of TACNts3 (see appendix). The product was golden oil, 85.7% crude yield. There were several
unidentified impurities in the alky Recrystallization from a number of solvents was att solution or was insoluble in all the solvents teste d indicates decomposition after sitting in insoluble in hexanes. Figure 30 The proton NMR spectrum of the polymeric product of the Huang method in CDCl3 The peaks at 2.05 and 1.6 ppm are the water, respectively. n unidentified impurities in the alky l range of the proton NMR spectrum (figure 31) Recrystallization from a number of solvents was att empted, but the product remained in solution or was insoluble in all the solvents teste d N otably, the proton NMR spectrum indicates decomposition after sitting in acetone overnight and the product is The proton NMR spectrum of the polymeric product of the Huang The peaks at 2.05 and 1.6 ppm are the result of acetone and r 87 of the proton NMR spectrum (figure 31) empted, but the product remained in otably, the proton NMR spectrum the product is completely The proton NMR spectrum of the polymeric product of the Huang result of acetone and
Figure 31. Top: t he proton NMR the sixfold dilution of the literature procedure reported by Huang et al. The unlabeled peaks are unidentified impurities. TACNts. n he proton NMR spectrum of the putative monomer resulting from the literature procedure reported by Huang et al. The unlabeled peaks are unidentified impurities. Bottom: the predicted spectrum of r 88 monomer resulting from the literature procedure reported by Huang et al. 68 in CDCl3. Bottom: the predicted spectrum of
89 b. Route 2b Scheme 17. Route 2b to the synthesis of R2TCMA. Reagents: 1) DCM; 2) TsCl in ether with triethylamine or pyridine; 3) K2CO3, MeCN; 4) R = iPr: Na2CO3, MeCN, isopropyl bromide; R = Bn: Benzyl chloride, Na2CO3, MeCN; R = Ph: Pd2(dba)3 (cat.), phosphine 7 (cat.), Na t-Butoxide, toluene, bromobenzene; 5) MeOH/KOH. Beginning from the top left 6) NaOH/H2O; 7) TsCl in ether with triethylamine or pyridine; 8) K2CO3, MeCN; 9) see step 4. The concept for this route (scheme 17) was develope d by Eric Andreansky in an effort to avoid tosyl protecting groups (although t osylates are used as leaving groups in steps 2 and 7). There are some doubts about its lik elihood of success. Primary among these concerns is the possible C-O bond formation i n the first step. This undesirable product would result from the alcoholic Os acting a s nucleophiles instead of the amines. Despite this doubt, there is precedent for success in the literature.70 Step 1 in scheme 17 has been attempted seven times, but the product of the reaction results in a convoluted proton NMR spectru m (figure 32). It appears that the major species is the starting material, ethyl bromo acetate, although there may be a very small amount of product indicated by the singlet at 3.99 ppm (NCH 2 COOEt). Ethyl
90 bromoacetate was used because it was available, ins tead of the more expensive and uncommon t-butyl bromoacetate that the reference pr ocedure70 employed. This substitution could have had a detrimental effect on the yield and purity of the product, as the reaction mixture is extracted with water to rem ove diethanolamine, and a t-butyl group would limit the water solubility of the produ ct more effectively than an ethyl group. Consequently, the procedure should be repeat ed with t-butyl bromoacetate, although it is significantly more expensive. Moreov er, the fact that the reference chose to use a relatively expensive and rare material may in dicate that the extra steric bulk was necessary to the success of the synthesis. Diethanol glycine was synthesized by the method des cribed in the literature69 and shown in step 6 of scheme 17. The synthesis utilize s chloroacetic acid instead of ethyl bromoacetate as in step 1, and yields 10.8% of very clean crystalline solid (figure 33).
Figure 32. Top: The 1 HNMR Inset: enlarged portion of s spectrum of DEGEE ( diethanol glycine ethyl ester of route 2b). HNMR spectrum of the product of step 1 of route 2b in CDCl Inset: enlarged portion of s pectrum from 3.3 to 4.3 ppm. Bottom: the predicted diethanol glycine ethyl ester the desired product of step 1 n 91 of the product of step 1 of route 2b in CDCl 3. Bottom: the predicted the desired product of step 1
Figure 33. Top: 1 HNMR predicted spectrum. HNMR spectrum of diethanol glycine (DEG) in D2 O. Bottom: 92 O. Bottom:
93 Despite the success of the synthesis of diethanol g lycine, this branch of the route was not pursued further because the utility of the product (diethanol glycine) in step 7 of scheme 17 is uncertain. Scheme 18 shows a possible side product. This product is expected to be prevalent because of the statistical favorability of the nucleophilic attack of the carboxylate and the stability of the resulta nt six-membered ring. Moreover, the reaction is favored by the pKa. In fact, another synthetic approach using glyo xal as the electrophile cited the equilibrium between the morp holone and the open chain compound as the primary reason for their difficulty in isola ting the product.77 It would be expected that the ring would be even more favored in the cas e where the leaving group is a tosylate. Scheme 18. Possible problem in route 2b step 6 (ref er to scheme 17). Though it is possible that the ring could open back up in the subsequent reaction with an ethylenediamine derivative (the carboxylate serving as a leaving group), this possibility was not tested. In light of these potential complic ations, this branch of route 2b was not emphasized.
94 III. Route 3 Scheme 19. Route 3 to the synthesis of R2TCMA. Reagents: 1) potassium carbonate and DMF; 2) TsCl, dimethylaminopyridine (cat), trie thylamine, DCM; 3) potassium carbonate, acetonitrile; 4) MeOH/KOH. Route 3 (scheme 19) is based upon the similar synth eses used by Pulacchini and Watkinson71 and used by former student Ellen Wolfgang.73 Wolfgangs synthesis is shown in scheme 20. Route 3 is extremely attractiv e because it would accomplish the synthesis of the target ligand in four steps, hopef ully resulting in a higher overall yield and a reduction of the time spent on synthesizing t he ligand. It also avoids the use of tosyl protecting groups, which was a major goal before th e detosylation of Bn2TACNts was realized. The route could also be generalized to th e other target ligands either by the use of the respectively substituted ethylenediamine der ivatives or by the use of ethylenediamine with subsequent derivatization.
95 Scheme 20. Part of the synthetic route used by Elle n Wolfgang,73 derived from the route developed by Pulacchini and Watkinson.71 Reagents: 1) potassium carbonate and DMF; 2) TsCl, dimethylaminopyridine (cat), trie thylamine, DCM; 3) potassium carbonate, acetonitrile. In 3, Pulacchini and Watki nson used several H2NR reactants, including R= Bn and Ts, but not methyl p ropionate as used by Wolfgang. Wolfgangs target ligand (TCMP 1,4,7-triazacyclonon ane-mono-propionate) was similar to ours, except that two of the nitrogen at oms were secondary amines and the third had a propionate arm instead of our proposed acetate. Unfortunately, Wolfgangs synthesis was hampered by the presence of the tosyl amine protecting groups (product of step 3 of scheme 20). When she attempted to remove them, the propionate did not survive the strongly reducing reaction conditions. However, it was thought that the decarboxylation of our acetate arm could be avoided if N,N-dibenzyl-ethylenediamine were used in step 1 instead of N,N-ditosyl-ethylen ediamine, circumventing the detosylation. Regrettably, the synthetic approach i n scheme 19 did not run as smoothly as the procedures used by Wolfgang or Pulacchini. The synthesis was repeated four times, none of them successful. A representative proton NM R spectrum is shown in figure 34. The spectrum indicates a large amount of N,N-diben zyl-ethylenediamine starting material, accompanied by many indistinct peaks that remain unidentified. The reason for
96 the outcome of this reaction is not known. The reac tion should be tried again in other hands. Additionally, the reaction using N,N-diR-et hylenediamine as the nucleophile in step 1 might be useful in the synthesis of other R2TCMA type ligands, and the synthesis may be more successful with different substituents. Figure 34. The proton NMR spectrum of the product m ixture that resulted from the attempted synthesis of 2,2'-(ethane-1,2-diylbis(ben zylazanediyl))diethanol.
97 IV. Evaluation of routes Route 1 is the most well-developed and reliable of all the approaches that were explored. This is perhaps because the synthesis of TACN and its derivatives has been the focus of much effort for quite some time in the She rman lab. Route 1 is based upon the methods that emerged as successful in the past.57,58,65,72-75 However, it may be advantageous to move towards a more efficient synth etic route. As it stands, the overall yield at the point of TACNts is 29%, whereas the yi eld of crude TACNts by route 2a is 85.7%. If the yield of the synthesis of the precurs or for route 2a, DEAts3 in the hands of Liang is considered, the overall yield is still mor e than twice the yield of route 1, at 66%. Though the product in route 2a is not purified, it is the product of only two steps whereas the product of route 1 takes four steps and require s special and time-consuming set ups. Purification of crude TACNts from route 2a either b y column chromatography, salting out with acidification, or a simple acid-base extra ction (highly recommended) would probably still yield more TACNts than route 1, and in considerably less time. This purification was not pursued for this thesis becaus e it led to the same product as route 1, TACNts, and thus had all the problems inherent in t he detosylation after the addition of substituents. Now that the detosylation of Bn2TACNts has been largely worked out, it would be highly advantageous for future students to return to this route and perfect it. It would be a logical thought to try to use substitute d ethylenediamine in order to shorten the synthesis even further. This approach may work, and is certainly worth a try given the availability of N,N-dibenzyl-ethylenediamine. Howe ver, there is precedent in the literature for the failure of this reaction. Huang et al.68 observed the formation of undesirable products when N,N-dimethyl-ethylenedia mine was used (scheme 21). Of
98 course benzyls are not the same as methyls, but it seems likely that they might have a similar or even worse effect. Scheme 21. The products observed by Huang et al. an d their rationalization.68 If route 2b were to work, it would be better than route 2a by virtue of the fact that it has fewer steps, although the difference in this respect is minimal (4-5 versus 5-6 steps). However, fewer steps does not necessarily m ean better yield. A functional route 2a might be preferable if the overall yield is high er, but the yield of 2b is difficult to predict. Route 3 was perhaps the most frustrating of all th e routes pursued. Its seeming simplicity is seductive. However, the first step is time consuming, and the high temperatures cause some of the contents of the PEG bath to evaporate and condense on all nearby glassware. More importantly, each of fou r attempts resulted in disastrous NMR spectra, indicating large amounts of starting mater ial in addition to the other broad and unidentified peaks, the significance of which are u nclear. The second step seems as though it would be relatively straightforward, but the third step requires six days of reflux according to the literature.
99 B. Computation The computational aspect of this thesis arose in re sponse to doubts that the target ligand Ph2TCMA would coordinate a metal strongly enough to be useful as a model complex. The reason for this doubt is the direct at tachment of the nitrogen atoms to the aromatic phenyl substituents, a problem that the be nzyl derivative avoids. This direct attachment might cause the lone pair of the nitroge n to be delocalized to a point such that its capacity as a sigma donating ligand is compromi sed. However, this effect could only be observed if the phenyl substituent is in a confo rmation such that the system of the phenyl groups can overlap with the p orbital of the nitrogens. The ligand Ph2TCMA was optimized using DFT with a B3LYP/6-31G++(d ,p) or higher basis set, and it was found that the phenyl system is aligned with the lone pair of the nitrogen, as can be seen in figure 36. However, when the dimethyl ortho substituted version of the ligand (o,o-dimethyl-Ph2TCMA, figure 20) was optimized at the same level of theory, the phenyl system is not aligned with the lone pair of the ni trogen (figure 37). The relevant information from the calc ulations are the C-N bond length, the three C-N-C bond angles, and the displacement from planarity ( vide infra ). The C-N bond length is expected to decrease with increasing interaction between the phenyl system and the nitrogen lone pair. Meanwhile, the C -N-C bond angles are expected to approach 120 with increased overlap between the or bitals because the nitrogen atoms orbital hybridization would approach sp2. In the case of no interaction between the orbitals, the C-N-C angle would be expected to appr oach 107, which is the C-N-C bond angle of ammonia. It would have been better to comp are the ligand C-N-C angles to those
100 of a tertiary amine such as trimethyl amine optimiz ed under the same conditions as the ligands, but this was only realized retrospectively The displacement from planarity describes the orientation of the phenyl ring with respect to the nitrogen atom lone pair. Because of the relative ease of measurement, the plane that is referred to is the p lane of the carbon atoms bonded to the nitrogen atom. This is an approximation.78 In plane (i.e. a displacement angle of 0) means that the plane of the phenyl ring is perpendicular to the nitrogen lone pair (figure 35) An in plane conformation allows the p orbital to overlap with the system. Conversely, a displacement angle of 90 would indicate that the p henyl plane contains the axis of the nitrogen 2p orbital, precluding the possibility of interaction between the p orbital and the system. Thus, increasing displacement angle indica tes decreasing interaction between the phenyl system and the nitrogen lone pair, and vice versa. The calculated values for all of these geometric aspects are reported in tabl e 1 for both compounds. In order to increase the reliability of these calcu lations, the optimization of each ligand was performed with two different initial con formations: 1) phenyl ring in plane and 2) phenyl ring out of plane (perpendicular). Th is method decreases the likelihood that the calculation finds a local minimum in energy ins tead of the lowest energy conformation. The results of both starting points a gree well with each other for both ligands (table 1). Figure 35. Top: "in plane" = 0 displacement from planarity. Bottom: "out of plane" = 90 displacement from planarity.
101 Figure 36. The optimization of the ligand Ph2TCMA. Notably, the phenyl groups are arranged so that their system is in line with the nitrogen atoms lone pa ir and the nitrogen is nearly planar. Figure 37. The optimization of the ligand o,o'-dime thyl-Ph2TCMA. Notably, the phenyl groups are arranged so that their systems are approximately perpendicular to the nitrogen atoms lone pair.
102 Table 1. The relevant geometric parameters of both ligand derivatives calculated using DFT with the basis set B3LYP/6-31G++(d,p) or higher. Ligand Starting positiona PhC-N length () Displacement from planarityb C-N-C1 C-N-C2 C-N-C3 Ph2TCMA In plane 1.38358 1.38793 14.730 7.034 117.131 121.955 120.159 116.850 120.823 118.765 Perpendicular (out of plane) 1.38505 1.38690 6.717 10.627 117.142 117.426 121.506 119.159 120.675 121.561 Average 1.38587 9.777 119.429 c Standard deviation 0.00193 3.747 1.930 c o,o-dimethylPh 2 TCMA In plane 1.43489 1.44206 87.591 80.623 117.148 115.021 117.481 112.987 111.894 115.299 Perpendicular (out of plane) 1.43770 1.44270 86.782 83.420 116.865 117.781 116.322 114.445 112.797 111.545 Average 1.43934 84.604 114.965 d Standard deviation 0.00370 3.210 2.228 d a: The position of the phenyl relative to the axis of the adjacent nitrogen atoms p orbital in the init ial input structure of the ligand. b: A measure of the angle between the plane of the phe nyl ring and the plane of the three carbon atoms bound to each nitrogen atom. An angle near 90 indi cates that the phenyl ring is near parallel to the nitrogen p orbital, which indicates that the phenyl system is nearly perpendicular to the N lone pair. An angle near 0 would indicate the opposite. c: The average or standard deviation of all the C-N-C angles for Ph2TCMA together. d: The average or standard deviation of all the C-N-C angles for o,o-dimethylPh2TCMA together. Notably, the C-N bond distances of the unsubstitute d Ph2TCMA are shorter than those of the dimethyl substituted derivative, with respective average lengths of 1.38587 and 1.43934 This difference indicates that the C -N bond is stronger in the unsubstituted ligand, which can be rationalized by the orbitals of the phenyl and the nitrogen atom interacting to a larger degree in the unsubstituted ligand. This conclusion is further supported by the calculated average C-N-C a ngles of the two ligands, with 119.429 for Ph2TCMA and 114.965 for o,o-dimethyl-Ph2TCMA, indicating that the nitrogens of the former are much closer to being sp2 hybridized than those of the latter ligand. Furthermore, the conformation of the phenyl rings in the unsubstituted ligand is more favorable for interaction with the nitrogen lo ne pair than the conformation of the phenyls of the unsubstituted derivative. The phenyl rings of the former are displaced by
103 an average of 9.777 while those of the latter are displaced by an average of 84.604. These angles indicate that the methyl groups force the system of the phenyl rings to be nearly perpendicular to the nitrogen lone pair. These calculations strongly suggest that the nitrog en lone pair is extensively delocalized by the phenyl substituents, but that th e N lone pair of the o,o-dimethylPh2TCMA is localized on the N. This result implies tha t the methyl substituted o, odimethylPh2TCMA would bind to metals more strongly than Ph2TCMA. However, it would be instructive to calculate how much the in p lane conformation of Ph2TCMA is favored over the out of plane conformation. It coul d be a small enough difference that both states are well populated, which would make me tallation more accessible. This possibility was not pursed because of the limitatio ns of time and expertise of the author. Another limitation of this information is that the calculations are performed in the gas phase, which is not the relevant environment fo r ligands. Resultantly, the effects of solvent molecules are not taken into account. In th e gas phase, steric considerations are amplified compared to the weight they would carry i n solution. This must be kept in mind when evaluating the importance of the calculated co nformation of o,o-dimethylPh2TCMA.
104 Chapter IV Conclusions and Future Development Progress has been made toward the synthesis of new complexes that are expected to model the structure of the almost identical acti ve sites of the manganese containing enzymes oxalate decarboxylase and oxalate oxidase. There is still work to be done toward the synthesis of these complexes. The syntheses of the two new target ligands Bn2TCMA and Ph2TCMA are still incomplete, but they are very clearl y within reach. The project is at a very exciting point in that the most difficult problems of ligand synthesis have been overcome. The only remaining step in the case of Bn2TCMA is the addition of the carboxylate arm, and then metallation can finally b e achieved. In terms of future work, great emphasis will be placed on the pursuit of the target ligand Bn2TCMA. Second in priority, the synthesis of an ortho substituted ver sion of Ph2TCMA is highly advisable. Geometry optimization of the ligand using Gaussian predicts that in the lowest energy conformation, the systems of the phenyl substituents of Ph2TCMA are aligned with the lone pairs of the ring nitrogens, further validatin g concern over the coordinative ability of Ph2TCMA. Ortho substituents might sterically constrain the ring to a conformation where the phenyls are no longer aligned, freeing the lone pair on the nitrogens to donate to a metal. It is also possible that the substituents mi ght congest the ligand and decrease metal
105 binding affinity, but this seems unlikely. Mesityl derivatives would also be interesting, and the required reactants are available at New Col lege. Once the syntheses of the target ligands Bn2TCMA and Ph2TCMA (or its ortho substituted derivative) are fully developed, the ma nganese(II) complexes should be characterized and their reactivity with and binding of oxalate studied. Once their chemistry is fully explored, the substituents on th e ortho and para positions of the aromatic rings could be systematically altered to t une the redox potential of the manganese center. It would be especially interestin g to use both donating and withdrawing substituents. Some good candidates migh t be: methoxy, nitrile, sulfate, methyl (any alkyl), nitro, halides, amines, etc. Be cause the potentials of the enzymes are not known, both directions should be explored. Thus there are two main considerations remaining: synthetic feasibility and the coordinati on ability of the substituents. The first speaks for itself. The coordinative ability of the substituent is important because it would be detrimental for the model to have a polymeric cr ystal structure. The TACN ring is highly desirable as a framework fo r the structural mimicry of oxox and oxdc, as is made very clear by the superpo sition of the structures of the model by Scarpellini et al.17 and the active site of oxdc (see figure 17 in sect ion I.C.2). However, using our current methods, the synthesis o f TACN is very time consuming. This inconvenience generally holds true for TACN de rivatives as well. Consequently, the target ligands are ambitious for the time constrain ts of an undergraduate thesis. This concern is especially pertinent if the student wish es to pursue inorganic chemistry, as the ligand synthesis leaves little time for experience in coordination chemistry. This is not to
106 say that the target ligands herein should not be pu rsued; quite the contrary. But in light of all of these considerations, it may be wise to put a portion of our eggs in another basket by devoting some time to a more accessible ligand a s a side project, or perhaps a separate project entirely. The following two proposed complexes have not been thought through well enough to start even as side projects. However, the y seem reasonable on the surface. The first is less promising as a model complex, but has the advantage of employing a more synthetically reliable ligand. The second complex i s a more promising model because the ligand enforces facial binding, rather than just fa voring it. However, parts of the ligand synthesis are uncertain. The iron(III) complex79 in figure 38 is one of the few examples of a synth etically simple N3O donating ligand that displays facial coordination of the nitrogens. Figure 38. Left: The crystal structure of a represe ntative complex from the series of Fe(III) complexes that Wong et al. report. Right: S uggested target complex for future exploration, based on the similar Fe(III) co mplex made by Wong et al. The image is modified from the source.79 Here, R = t-Bu and R = H or R = R = t-Bu. The ligand synthesis (scheme 22) is four consecutiv e steps, in addition to a one pot synthesis (without purification) to make bis(py ridin-2-ylmethyl)amine (one of the
107 reactants in the fourth step). Each step is a well established synthetic method, and only the last step requires a reaction time of longer th an one day (overnight). Furthermore, the reported overall yield is over 80%. The synthesis i s shown in scheme 22. Scheme 22. The synthesis used by Wong et al. R = tBu and R' = H or R = R' = t-Bu. The scheme is modified from the source.79 Furthermore, the metal reported is Fe(III), which i s isoelectronic and similar in size to Mn(II). However, the two cis ligands shown in green are chlorides, and the shap e of the ligand suggests that it is capable of adopti ng an undesirable mer -binding mode with respect to the nitrogen donors. This flexibili ty may cause complications in the synthesis of the target complex shown in figure 38 (right). If meridonal binding occurs when R = R = t-Bu, it might be prevented by even b ulkier substituents on the hydroxybenzyl arm. Obviously, the O donor is not a carboxylate like in the TACN-based ligands and the active sites of oxox and oxdc. There are analog ous ligands wherein the hydroxybenzyl arm is replaced by a carboxylate, but the manganese complexes of these tend to form dimers (see Jiang et al.80 for examples). Conversely, the pyridyl Ns of this
108 ligand are a closer approximation to the histidines of the active sites than the amine N donors of the TACN-based ligands. Its conceivable that the hydroxybenzyl substituent could be replaced with a carboxylate arm, but this ligand runs the risk of dinuclear or polynuclear complex formation. One of the main goal s is to systematically vary the potential of the model complex, and this objective could still be achieved in a similar way to that planned for the TACN-based ligands by varyi ng the substituents on the pyridyl and hydroxybenzyl substituents. Arguably, this trip odal ligand is even more facilely tuned. A brief investigation of manganese complexes of thi s and similar ligands yielded mixed results. There are many examples of similar l igands binding manganese in a meridonal fashion. A few examples are shown in figu re 39. Notably, the substituents on the hydroxybenzyl arm are either H or methoxy, whic h are both much smaller than the tbutyl groups that are proposed. Hopefully the t-but yl would be bulky enough to encourage facial coordination of the nitrogens.
109 Figure 39. Top: examples of manganese complexes of ligands similar to the proposed ligand bound in a meridonal fashion.81 Bottom: An example of a dimeric manganese complex with the proposed ligand where R = H.82 Figure modified from the sources. Figure 40. The proposed complexes based on the TACH ligand framework. The complex on the left corresponds to R = 5 in scheme 23, while the complex on the right corresponds to R = 6.
110 The second proposed new target complexes are based on the ligand cis,cis triaminocyclohexane (TACH) (figure 40). Cronin et a l.45 synthesized several of the uncommon monofunctionalized TACH derivatives, some variants of which could provide an N3O donation sphere. As previously discussed, TACH en forces facial coordination of the nitrogen atoms to the metal, which is a great a dvantage. The substituent contains a functionalized aromatic ring, so the substituents c ould be varied to change the potential of the metal center. Cronins synthesis and the propos ed syntheses are summarized in scheme 23. Scheme 23. The synthesis used by Cronin et al.45 is shown in black. The metalcontaining products were characterized by X-ray cry stallography. Green signifies proposed reactants and speculative products. TACH is reacted with three equivalents of an aromat ic aldehyde to produce the C3 symmetric Schiff base intermediate. Then, this comp ound is metalated, which increases the electrophilicity of the C=N bonds, allowing wat er to attack to eventually reform the aldehyde and the primary amine. Remarkably, this re action only occurs at two of the imines, leaving the third intact. This selectivity allows for the asymmetry necessary for our ligands. At this stage, scheme 23 only shows th e product observed by Cronin, but in
111 the case of our proposed ligand, one of the X ligan ds would be replaced by an oxygen donor from the R group. If it is possible to comple te this synthesis with manganese as the activating metal, then no additional steps would be necessary. If not, extra steps would be necessary to replace Ni, Cu, or Zn with Mn. The lig and could then be tuned by adding substituents at the amines (bottom right of scheme 23). Importantly, the proposed aldehyde R = 5 (salicylal dehyde) is commercially available (R = 6 is not) and TACH is fairly simply synthesized in two steps with 86% yield from commercially available starting material s.83 Thus, the synthesis of the ligand is only four steps. Using the R = 5 ligand, the comple x is unlikely to polymerize, so substituents on the other two amines are hopefully unnecessary. If they are necessary, it seems like it would be a fairly straight forward pr ocess to functionalize them (shown in green, scheme 23), although the metal may need to b e removed first. Regardless of whether these particular proposed lig ands are pursued, it would be prudent to incorporate a more accessible ligand int o plans for future research. Given that the current project is now at an exciting stage whe re the target ligands are extremely close to realization, these proposals belong in the realm of the relatively distant future.
112 References 1. Lane, B.G. Oxalate, germin, and the extracellula r matrix of higher plants. The FASEB Journal 8 294-301 (1994). 2. Cassland, P., Larsson, S., Nilvebrant, N.-O. & J onsson, L.J. Heterologous expression of barley and wheat oxalate oxidase in an E. coli t rxB gor double mutant. Journal of Biotechnology 109 53-62 (2004). 3. Abratt, V.R. & Reid, S.J. Oxalate-Degrading Bact eria of the Human Gut as Probiotics in the Management of Kidney Stone Diseas e. Advances in Applied Microbiology 72 63-87 (2010). 4. Watson, J.T., Jones, R.C., Siston, A.M., Diaz, P .S., Gerber, S.I., Crowe, J.B. & Satzger, R.D. Outbreak of Food-borne Illness Associ ated with Plant Material Containing Raphides. Clinical Toxicology 43 17-21 (2005). 5. McMartin, K. Are clacium oxalate crystals involv ed in the mechanism of acute renal failure in ethylene glycol poisoning? Clinical Toxicology 47 859-869 (2009). 6. Tokarnia, C.H., Do¨bereiner, J. & Peixoto, P.V. Poisonous plants affecting livestock in Brazil. Toxicon 40 1635-1660 7. Zuo, G., Jiang, X., Liu, H. & Zhang, J. A novel urinary oxalate determination method via a catalase model compound with oxalate oxidase. Anal. Methods 2 254-258 (2009). 8. Gray, H.B. Biological inorganic chemistry at the beginning of the 21st century. Proceedings of the National Academy of Sciences 100 35633568 (2003). 9. Lee, S.C. & Holm, R.H. Speculative synthetic che mistry and the nitrogenase problem. Proceedings of the National Academy of Sciences 100 35953600 (2003). 10. Svedruzic, D., Jonsson, S., Toyota, C.G., Re inhardt, L.A., Ricagno, S., Lindqvist, Y. & Richards, N.G.J. The enzymes of oxalate metabo lism: unexpected structures and mechanisms. Archives of Biochemistry and Biophysics 433 176192 (2005).
113 11. Shimazono, H. Oxalic Acid Decarboxylase, A New Enzyme from the Mycelium of Wood Destroying Fungi. The Journal of Biochemistry 42 321-340 (1955). 12. Tanner, A., Bowater, L., Fairhurst, S.A. & Born emann, S. Oxalate Decarboxylase Requires Manganese and Dioxygen for Activity OVEREX PRESSION AND CHARACTERIZATION OF BACILLUS SUBTILIS YvrK AND YoaN The Journal of Biological Chemistry 276 43627-43634 (2001). 13. Mkel, M.R., Hildn, K. & Lundell, T.K. Oxalat e decarboxylase: biotechnological update and prevalence of the enzyme in filamentous fungi. Applied Microbiology and Biotechnology 87 801-814 (2010). 14. Escutia, M.R., Bowater, L., Edwards, A., Bottri ll, A.R., Burrell, M.R., Polanco, R., Vicuna, R. & Bornemann, S. Cloning and Sequencing o f Two Ceriporiopsis subvermispora Bicupin Oxalate Oxidase Allelic Isofo rms: Implications for the Reaction Specificity of Oxalate Oxidases and Decarb oxylases. Applied and Environmental Microbiology 71 3608-3616 (2005). 15. Just, V.J., Stevenson, C.E.M., Bowater, L., Tan ner, A., Lawson, D.M. & Bornemann, S. A Closed Conformation of Bacillus subtilis Oxala te Decarboxylase OxdC Provides Evidence for the True Identity of the Active Site. The Journal of Biological Chemistry 279 1986719874 (2004). 16. Anand, R., Dorrestein, P.C., Kinsland, C., Begl ey, T.P. & Ealick, S.E. Structure of Oxalate Decarboxylase from Bacillus subtilis at 1.7 5 Resolution. Biochemistry 41 7659-7669 (2002). 17. Scarpellini, M., Gtjens, J., Martin, O.J., Kam pf, J.W., Sherman, S.E. & Pecoraro, V.L. Modeling the Resting State of Oxalate Oxidase and Oxalate Decarboxylase Enzymes. Inorg. Chem. 47 3584-3593 (2008). 18. Schwarzenbacher, R., Delft, F. von, Jaroszewski L., Abdubek, P., Ambing, E., Biorac, T., Brinen, L.S., Canaves, J.M., Cambell, J ., Chiu, H.-J., Dai, X., Deacon, A.M., DiDonato, M., Elsliger, M.-A., Eshagi, S., Fl oyd, R., Godzik, A., Grittini, C., Grzechnik, S.K., Hampton, E., Karlak, C., Klock, H. E., Koesema, E., Kovarik, J.S., Kreusch, A., Kuhn, P., Lesley, S.A., Levin, I., McM ullan, D., McPhillips, T.M., Miller, M.D., Morse, A., Moy, K., Ouyang, J., Page, R., Quijano, K., Robb, A., Spraggon, G., Stevens, R.C., van den Bedem, H., Vel asquez, J., Vincent, J., Wang,
114 X., West, B., Wolf, G., Xu, Q., Hodgson, K.O., Wool ey, J. & Wilson, I.A. Crystal Structure of a Putative Oxalate Decarboxylase (TM12 87) From Thermotoga maritima at 1.95 Ar Resolution. PROTEINS: Structure, Function, and Bioinformatics 56 392395 (2004). 19. Chang, C.H., Svedruzic, D., Ozarowski, A., Walk er, L., Yeagle, G., Britt, R.D., Angerhofer, A. & Richards, N.G.J. EPR Spectroscopic Characterization of the Manganese Center and a Free Radical in the Oxalate Decarboxylase Reaction. The Journal of Biological Chemistry 279 52840-52849 (2004). 20. Garca-Couceiro, U., Castillo, O., Cepeda,, J. Lanchas, M., Luque, A., Perez-Yanez, S., Roman, P. & Vallejo-Sanchez, D. Influence of th e Synthetic Conditions on the Structural Diversity of Extended Manganese-Oxalato1,2-bis(4-pyridyl)ethylene Systems. Inorganic Chemistry 49 11346-11361 (2010). 21. Ying-Feng Han, Wei-Guo Jia, Wei-Bin Yu & Guo-Xi n Jin Stepwise formation of organometallic macrocylces, prisms and boxes from I r, Rh and Ru-based halfsandwich units. Chemical Society Reviews 38 3419-3434 (2009). 22. Zhu, L.-H., Zeng, M.-H. & Ng, S.W. Poly[bis([mu ]2-4,4-bipyridyl[kappa]2N:N)bis(4,4-bipyridyl-[kappa]N)tris([mu]2 -oxalato[kappa]4O,O:OO)trimanganese(II)]. Acta Crystallographica Section E 61 m916-m918 (2005). 23. Krishnamurty, K.V. & Harris, G.M. The Chemistry of the Metal Oxalato Complexes. Chemical Reviews 61 213-246 (1961). 24. Youssef, M.A.M.A., Mautner, F.A. & Vicente, R. Two new 2D mu-azido, muoxalato manganese(II) complexes. Synthesis, structu re and magnetic properties. Inorganica Chimica Acta 361 2895-2900 (2008). 25. Walsby, C.J., Tesler, J., Rigsby, R.E., Armstro ng, R.N. & Hoffman, B.M. Enzyme Control of Small-Molecule Coordination in FosA as R evealed by 31P Pulsed ENDOR and ESE-EPR. Journal of the American Chemical Society 127 8310-8319 (2005). 26. Tanase, T., Tamakoshi, S., Doi, M., Mori, W. & Yano, S. Seven-coordinate manganese(II) complexes with a cage-type N-glycosid e ligand. Configurational switch depending on counter anions. Inorganica Chimica Acta 266 5-7 (1997).
115 27. Deroche, A., Morgenstern-Badarau, I., Cesario, M., Guilhem, J., Keita, B., Nadjo, L. & Houe-Levin, C. A Seven-Coordinate Manganese(II) Complex Formed with a Single Tripodal Heptadentate Ligand as a New Supero xide Scavenger. Journal of the American Chemical Society 118 4567-4573 (1996). 28. Reinhardt, L.A., Svedruzic, D., Chang, C.H., Cl eland, W.W. & Richards, N.G.J. Heavy Atom Isotope Effects on the Reaction Catalyze d by the Oxalate Decarboxylase from Bacillus subtilis. Journal of the American Chemical Society 125 1244-1252 (2003). 29. Svedruzic, D., Liu, Y., Reinhardt, L.A., Wroc lawska, E., Cleland, W.W. & Richards, N.G.J. Investigating the roles of putativ e active site residues in the oxalate decarboxylase from Bacillus subtilis. Archives of Biochemistry and Biophysics 464 36-47 (2007). 30. Moomaw, E.W., Angerhofer, A., Moussatche, P., O zarowski, A. & Garca-Rubio, I. Metal Dependence of Oxalate Decarboxylase Activity. Biochemistry 48 61166125 (2009). 31. Burrell, M.R., Just, V.J., Bowater, L., Fairhur st, S.A., Requena, L., Lawson, D.M. & Bornemann, S. Oxalate Decarboxylase and Oxalate Oxi dase Activities Can Be Interchanged with a Specificity Switch of up to 282 000 by Mutating an Active Site Lid. Biochemistry 46 12327-12336 (2007). 32. Muthusamy, M., Burrell, M.R., Thorneley, R.N.F. & Bornemann, S. Real-Time Monitoring of the Oxalate Decarboxylase Reaction an d Probing Hydron Exchange in the Product, Formate, Using Fourier Transform Infra red Spectroscopy. Biochemistry 45 10667-10673 (2006). 33. Angerhofer, A., Moomaw, E.W., Garcia-Rubio, I., Ozarowski, A., Krzystek, J., Weber, R.T. & Richards, N.G.J. Multifrequency EPR S tudies on the Mn(II) Centers of Oxalate Decarboxylase. The Journal of Physical Chemistry B Letters 111 50435046 (2007). 34. Lane, B.G., Dunwell, J.M., Ray, J.A., Schmitt, M.R. & Cuming, A.C. Germin, a Protein Marker of Early Plant Development, Is an Oxalate Oxidase. The Journal of Biological Chemistry 268 12239-12242 (1993).
116 35. Woo, E.-J., Dunwell, J.M., Goodenough, P.W., Ma rvier, A.C. & Pickersgill, R.W. Germin is a manganese containing homohexamer with o xalate oxidase and superoxide dismutase activities. Nature Structural Biology 7 1036-1040 (2000). 36. Requena, L. & Bornemann, S. Barley (Hordeum vul gare) oxalate oxidase is a manganese-containing enzyme. Biochemical Journal 343 185-190 (1999). 37. Whittaker, M.M. & Whittaker, J.W. Characterizat ion of recombinant barley oxalate oxidase expressed by Pichia pastoris. Journal of Biological Inorganic Chemistry 7 136-145 (2002). 38. Opaleye, O., Rose, R.-S., Whittaker, M.M., Woo, E.-J., Whittaker, J.W. & Pickersgill, R.W. Structural and Spectroscopic Stud ies Shed Light on the Mechanism of Oxalate Oxidase. The Journal of Biological Chemistry 281 6428-6433 (2006). 39. Berg, J.M., Tymoczko, J.L. & Stryer, L. Biochemistry (W. H. Freeman and Company: New York, 2007). 40. Whittaker, M.M., Pan, H.-Y., Yukl, E.T. & Whitt aker, J.W. Burst Kinetics and Redox Transformations of the Active Site Manganese Ion in Oxalate Oxidase. The Journal of Biological Chemistry 282 7011-7023 (2007). 41. Fuller, A.L., Watkins, R.W., Dunbar, K.R., Pros virin, A.V., Arif, A.M. & Berreau, L.M. Manganese(II) chemistry of a new N3O-donor che late ligand: synthesis, X-ray structures, and magnetic properties of solventand oxalate-bound complexes. Dalton Transactions 1891-1896 (2005). 42. Fuller, A.L., Watkins, R.W., Arif, A.M. & Berre au, L.M. Synthesis and characterization of a series of N3O-ligated mononuc lear Mn(II) halide complexes. Inorganica Chimica Acta 359 1282-1290 (2006). 43. Lewis, E.A., Khodr, H.H., Hider, R.C., Lindsay Smith, J.R. & Walton, P.H. A manganese superoxide dismutase mimic based on cis,c is-1,3,5triaminocyclohexaneElectronic supplementary informa tion (ESI) available: Synthetic details, determination of the physical properties o f ligands, SOD assays, electrochemistry and references. See http://www.rsc .org/suppdata/dt/b3/b314955b/. Dalton Trans. 187 (2004).doi:10.1039/b314955b 44. Archibald, S.J., Nairn, A.K., Timmins, P. & Wal ton, P.H. Mono-benzyl substituted cis,cis-1,3,5-triaminocyclohexanes. Tetrahedron Letters 46 6441-6443 (2005).
117 45. Cronin, L., Greener, B., Foxon, S.P., Heath, S. L. & Walton, P.H. Syntheses and Single-Crystal X-ray Structures of a Series of Mono substituted cis,cis-1,3,5Triaminocyclohexane-Based Complexes. Inorganic Chemistry 36 2594-2600 (1997). 46. Kajita, Y., Arii, H., Saito, T., Saito, Y., Nag atomo, S., Kitagawa, T., Funahashi, Y., Ozawa, T. & Masuda, H. Syntheses, Characterization, and Dioxygen Reactivities of Cu(I) Complexes with cis,cis-1,3,5-Triaminocyclohex ane Derivatives: A Cu(III)2O2 Intermediate Exhibiting Higher CH Activation. Inorganic Chemistry 46 3322-3335 (2007). 47. Ebrahimpour, P., Cushion, M., Haddow, M.F., Hal lett, A.J. & Wass, D.F. Cu(i) complexes based on cis, cis-1,3,5-tris(arylideneami no)cyclohexane ligands: synthesis, structure and CO binding. Dalton Trans. 39 10910 (2010). 48. Park, G., Dadachova, E., Przyborowska, A., Lai, S.-jen, Ma, D., Broker, G., Rogers, R.D., Planalp, R.P. & Brechbiel, M.W. Synthesis of novel 1,3,5-cis,cistriaminocyclohexane ligand based Cu(II) complexes a s potential radiopharmaceuticals and correlation of structure a nd serum stability. Polyhedron 20 3155-3163 (2001). 49. Park, G., Przyborowska, A.M., Ye, N., Tsoupas, N.M., Bauer, C.B., Broker, G.A., Rogers, R.D., Brechbiel, M.W. & Planalp, R.P. Steri c eects caused by N-alkylation of the tripodal chelator N,N,N-tris(2-pyridylmet hyl)-cis,cis-1,3,5triaminocyclohexane (tachpyr): structural and elect ronic properties of the Mn(II), Co(II), Ni(II), Cu(II) and Zn(II) complexes. Dalton Transactions 2003 318-324 (2003). 50. Yang, J.Y., Shores, M.P., Sokol, J.J. & Long, J .R. High-Nuclearity MetalCyanide Clusters: Synthesis, Magnetic Properties, and Inclu sion Behavior of Open-Cage Species Incorporating [(tach)M(CN)3](M = Cr, Fe, Co ) Complexes. Inorganic Chemistry 42 1403-1419 (2003). 51. Childers, M.L., Su, F., Przyborowska, A.M., Bis hwokarma, B., Park, G., Brechbiel, M.W., Torti, S.V., Torti, F.M., Broker, G., Alexand er, J.S., Rogers, R.D., RuhlandtSenge, K. & Planalp, R.P. Pyridine-Ring Alkylation of Cytotoxic r-1,c-3,c-5-Tris[(2pyridylmethyl)amino]cyclohexane Chelators: Structur al and Electronic Properties of
118 the MnII, FeII, NiII, CuII and ZnII Complexes. European Journal of Inorganic Chemistry 2005 3971-3982 (2005). 52. Wieghardt, K. The Active Sites in Manganese-Con taining Metalloproteins and Inorganic Model Complexes. Angewandte Chemie International Edition 28 11531172 (1989). 53. Kitajima, N., Fujisawa, K., Fujimoto, C., Moroo ka, Y., Hashimoto, S., Kitagawa, T., Toriumi, K., Tatsumi, K. & Nakamura, A. A new model for dioxygen binding in hemocyanin. Synthesis, characterization, and molecu lar structure of the .mu..eta.2:.eta.2 peroxo dinuclear copper(II) complexes [Cu(HB(3,5-R2pz)3)]2(O2) (R = isopropyl and Ph). Journal of the American Chemical Society 114 1277-1291 (1992). 54. Mahapatra, S., Halfen, J.A., Wilkinson, E.C., Q ue, L. & Tolman, W.B. Modeling Copper-Dioxygen Reactivity in Proteins: Aliphatic C -H Bond Activation by a New Dicopper(II)-Peroxo Complex. Journal of the American Chemical Society 116 97859786 (1994). 55. Tolman, W.B. Making and Breaking the Dioxygen O O Bond: New Insights from Studies of Synthetic Copper Complexes. Accounts of Chemical Research 30 227-237 (1997). 56. CambridgeSoft ChemBioDraw Ultra 12.0. (2010). 57. Karylowski, O.J. Undergraduate Thesis, New Coll ege of Florida. (1999). 58. Schmidt, R.W. Undergraduate Thesis, New College of Florida. (2008). 59. Novaes-Card, S.E. Undergraduate Thesis, New COl lege of Florida. (2007). 60. Sessler, J.L., Sibert, J.W. & Lynch, V. Model S tudies Related to Hemerythrin. Synthesis and Characterization of a Bridged Tetranu clear Iron( 111) Complex. Inorganic Chemistry 29 4146-4148 (1990). 61. Mahapatra, S., Halfen, J.A., Wilkinson, E.C., P an, G., Wang, X., Jr., V.G.Y., Cramer, C.J., Jr., L.Q. & Tolman, W.B. Structural, Spectros copic, and Theoretical Characterization of Bis(u-oxo)dicopper Complexes, N ovel Intermediates in CopperMediated Dioxygen Activation. Journal of the American Chemical Society 118 11555-11574 (1996).
119 62. Scarpellini, M., Wu, A.J., Kampf, J.W. & Pecora ro, V.L. Corroborative Models of the Cobalt(II) Inhibited Fe/Mn Superoxide Dismutase s. Inorganic Chemistry 44 5001-5010 (2005). 63. Warden, A., Graham, B., Hearn, M.T.W. & Spiccia L. Synthesis of Novel Derivatives of 1,4,7-Triazacyclononane. Organic Letters 3 2855-2858 (2001). 64. Steward, D. Undergraduate Thesis, New College o f Florida. (2004). 65. Liang, A.D. Undergraduate Thesis, New College o f Florida. (2010). 66. Alonso, D.A. & Andersson, P.G. Deprotection of Sulfonyl Aziridines. Journal of Organic Chemistry 63 9455-9461 (1998). 67. Nakanishi, M. & Bolm, C. Palladium-Catalyzed NArylations of 1,4,7Triazacyclononanes. Adv. Synth. Catal. 348 1823-1825 (2006). 68. Huang, J., Zhou, Z. & Chan, T.H. An Improved Sy nthesis of 1,4,7Triazacyclononanes (tacns) and 1,4,7,10-Tetracyclod odecanes. Synthesis 14 23412344 (2009). 69. Khromov-Borisov, N.V. & Remizov, A.L. Synthesis and transformations of some derivatives of N,N-bis(2-hydroxyethyl)aminoacetic a cid [bis(hydroxyethyl)glycine]. I. Dehydration of [bis(hydroxyethyl)amino]acetc aci d. Zhurnal Obschei Khimii 23 598-605 (1953). 70. Greenwald, R.B., Zhao, H., Yang, K., Reddy, P. & Martinez, A. A New Aliphatic Amino Prodrug System for the Delivery of Small Mole cules and Proteins Untilizing Novel PEG Derivatives. Journal of Medicinal Chemistry 47 726-734 (2004). 71. Pulacchini, S. & Watkinson, M. An Efficient Rou te to Symmetrically and Unsymmetrically Substituted Azamacrocyclic Ligands. European Journal of Organic Chemistry 4233-4238 (2001). 72. Odom, D. Undergraduate Thesis, New College of F lorida. (1994). 73. Wolfgang, E. Undergraduate Thesis, New College of Florida. (2002). 74. Novaes-Card, S.E. Undergraduate thesis, New Col lege of Florida. (2007). 75. M. George Undergraduate thesis, New College of Florida. (1998). 76. Wuts, P.G.M. & Greene, T.W. Greenes Protective Groups in Organic Synthesis (John Wiley and Sons: Hoboken, New Jersey, 2007).
120 77. Farfan, N., Cuellar, L., Aceves, J.M. & Contrer as, R. High-Yield Syntheses of N-(2Hydroxyethyl)-N-alkylglycine Derivatives by Reactio n of Ethanolamines with Glyoxal. Synthesis 10 927-929 (1987). 78. This is an approximation because the C-N lengt hs and the C-N-C angles are not exactly equivalent, but they are close enough to pr ovide meaningful information. 79. Wong, Y.-L., Mak, C.-Y., Kwan, H.S. & Lee, H.K. Mononuclear iron(III) complexes supported by tripodal N3O ligands: Synthesis, struc ture and reactivity towards DNA cleavage. Inorganica Chimica Acta 363 1246-1253 (2010). 80. Jiang, X., Liu, H., Zheng, B. & Zhang, J. Coord ination modes of bridge carboxylates in dinuclear manganese compounds determine their ca talase-like activities. Dalton Trans. 8714 (2009).doi:10.1039/b907687e 81. Reddig, N., Pursche, D., Krebs, B. & Rompel, A. Mononuclear manganese(III) catechol compounds as substrate adduct complexes fo r manganese-substituted intradiol cleaving catechol dioxygenases. Inorganica Chimica Acta 357 2703-2712 (2004). 82. Iikura, H. & Nagata, T. Structural Variation in Manganase Complexes: Synthesis and Characterization of Manganese Complexes from Carbox ylate-containing Chelating Ligands. Inorganic Chemistry 37 4702-4711 (1998). 83. Bowen, T., Planalp, R.P. & Brech, M.W. An Impro ved Synthesis fo cis,cis-1,3,5triaminocyclohexane. Synthesis of Novel Hexadentate Ligand Derivatives fro the Preparation of Gallium Radiopharmaceuticals. Bioorganic & Medicinal Chemistry Letters 6 807-810 (1996).
121 Appendix Spectra DETAts 3 (CDCl 3 )
122 DETAts 3 (CDCl 3 )
123 DETAts 3 (CDCl 3 )
124 EGOts 2 (CDCl 3 )
125 EGOts 2 (CDCl 3 )
126 EGOts 2
127 TACNts 3 (CDCl 3 )
12 8 TACNts 3
129 TACNts (HBr Method) (CDCl 3 )
130 TACNts (HBr Method)
131 i Pr 2 TACNts (CDCl 3 )
132 i Pr 2 TACNts (CDCl 3 )
133 Bn 2 TACNts ((CD 3 ) 2 CO)
134 Bn 2 TACNts ((CD 3 ) 2 CO)
135 Bn 2 TACNts ((CD 3 ) 2 CO)
136 Ph 2 TACNts (CDCl 3 )
137 Ph 2 TACNts (CDCl 3
138 i Pr 2 TACN (CDCl 3 )
139 Ph 2 TACN (Na/naphthalene Method) (CDCl 3 )
140 Ph 2 TACN (Na/naphthalene Method) (CDCl 3 )
141 iPr 2 TCMA (Steward Method) in D 2 O
142 iPr 2 TCMA (Steward Method) in D 2 O
143 TACNts (Huang Method Polymer) (CDCl 3 )
144 TACNts (Huang Method Monomer) (CDCl 3 )
145 Route 2b Step 1 Product (CDCl 3 )