Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Biologically relevant physical studies of insulin-enhancing vanadium complexes Liboiron, Barry D. 2002

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

Item Metadata

Download

Media
831-ubc_2002-75040X.pdf [ 8.65MB ]
Metadata
JSON: 831-1.0061319.json
JSON-LD: 831-1.0061319-ld.json
RDF/XML (Pretty): 831-1.0061319-rdf.xml
RDF/JSON: 831-1.0061319-rdf.json
Turtle: 831-1.0061319-turtle.txt
N-Triples: 831-1.0061319-rdf-ntriples.txt
Original Record: 831-1.0061319-source.json
Full Text
831-1.0061319-fulltext.txt
Citation
831-1.0061319.ris

Full Text

B i o l o g i c a l l y Re levan t P h y s i c a l Studies o f I n s u l i n - E n h a n c i n g V a n a d i u m C o m p l e x e s by Barry D. Liboiron B. Sc. (Hons.), University of Guelph, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA JUNE 2002 © Barry D. Liboiron, 2002 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Ci^6^ S T"R>j The University of British Columbia Vancouver, Canada Date /\U&. & 2DOZ DE-6 (2/88) Abstract Investigations of the in vivo transport and accumulation of insulin-enhancing vanadium complexes are presented. Detailed spectroscopic studies of bis(maltolato)oxovanadium(IV) (BMOV), bis(ethylmaltolato)oxovanadium(IV) (BEOV) and the inorganic salt vanadyl sulfate (VOSO4) were carried out on in vitro solutions of serum proteins, and in vivo tissue samples from animals previously treated with a vanadium complex. Serum proteins apo-transferrin and albumin are both capable of effecting the decomposition of B M O V and BEOV under physiological conditions (pH 7.4, 0.16 M NaCl) to form vanadyl-protein adducts. Interactions of these complexes with the proteins were studied by continuous wave electron paramagnetic resonance (EPR) and difference ultra-violet spectroscopies. Apo-transferrin can bind up to two equivalents of B M O V at the Fe(III) binding sites, but (bi)carbonate (or another suitable synergistic anion) must be present for vanadyl binding to take place. The inability of maltolate to act in this role is demonstrated. Chelated vanadyl sources show no preference for either the N - or C-terminus binding site. Albumin binds B M O V only at the strong Cu(II) binding site; the presence of maltol imparts a site selectivity to the system in that B M O V will not bind at exposed carboxylates. Through consideration of the active equilbria in solution, formation of a ternary maltol-vanadyl-albumin complex is proposed and discussed in terms of reactivity differences between B M O V and V O S 0 4 and transport of chelated vanadyl sources in the bloodstream. Pulsed EPR methods - electron spin echo envelope modulation (ESEEM) and hyperfine sublevel correlation (HYSCORE) - were used to study the in vivo coordination of vanadyl ions in rat kidney, liver and bone samples, which had been previously taken from animals chronically administered BEOV via drinking water. The chelated vanadyl source ii becomes ligated by amines in the kidney and liver, and by as many as three different phosphate groups in bone mineral. Model studies of the vanadyl-triphosphate and vanadyl-hydroxyapatite systems were also studied to gain structural insights into the in vivo coordination state of the vanadyl ions in bone. Both systems proved to be very good models of the in vivo complex. Based on the number and relative magnitudes of the isotropic and anisotropic 3 I P and ' H coupling constants, a proposed solution structure, consistent with all spectroscopic data, is presented. i i i Table of Contents Abstract i i Table of Contents iv List of Figures x List of Tables xvi List of Equations xvii List of Schemes xviii List of Abbreviations xix Acknowledgments xxv Dedication xxvi j Biologically Relevant Physical Studies of Insulin Enhancing j p Vanadium Complexes 1.1. Medicinal Inorganic Chemistry 1 1.1.1. Metallopharmaceuticals in Clinical Use 2 1.2. Vanadium 5 1.2.1. Biologically Relevant Aqueous Chemistry of Vanadium 5 1.2.2. Vanadium in Biological Systems 8 1.2.2.1. Naturally-Occurring Vanadium 8 1.2.2.2. Medical, Toxic and Biochemical Effects of Vanadium 9 1.3. Vanadium and Diabetes Mellitus 11 1.3.1. Diabetes Mellitus 11 j ^ 2 Vanadium Pharmaceuticals: A New Class of Anti-Diabetic ^ Agents? iv 1.3.2.1. The Second Generation: Complexes of Vanadium 13 1.3.2.2. Insulin Enhancement by Vanadium Species 16 j Understanding the Metabolism of Vanadium Pharmaceutical ^ Agents 1.4.1. Absorption and Transport 20 1.4.2. Biodistribution and Accumulation 21 1.4.3. Metabolism of Chelated Vanadium Sources 22 1.5. References 22 ^ 2 Interaction of Insulin-Enhancing Vanadyl Complexes with ^ Serum Proteins 2.1. z Introduction 30 2.1.1. Proteins as Drug Transport Vehicles 30 2.1.2. Apo-transferrin 31 2.1.2.1. Interactions with Vanadium 3 3 2.1.3. Serum Albumin 34 2.1.3.1. Interactions with Vanadium 35 2.1.4. Chelated V(IV) Complexes and Serum Proteins 37 2.2. Experimental 37 2.2.1. Materials 37 2.2.2. Solution Preparation 38 2.2.3. Protein Purification and Manipulation 38 2.2.4. General Procedure for Preparation of Protein Samples 39 2.2.5. Preparation of Carbonate-Free Apo-transferrin Solution 40 2.2.6. Synthesis of (Imidazole)bis(maltolato)oxovanadium(IV) 40 2.2.7. Stability of the (1-Methylimidazole)-BMOV Adduct 41 2.2.8. Instrumental Parameters 43 2.2.8.1. Potentiometric Titration 43 2.2.9. Spectrophotometric Titration 44 2.2.9.1. Electron Paramagnetic Resonance Spectroscopy 45 2.2.9.2. Ultraviolet-Visible Spectroscopy 46 2.2.9.3. Other Instruments 47 2.2.10. Computations 47 2.3. ' Results and Discussion 47 2.3.1. Stability of Bis(ethylmaltolato)oxovanadium(IV), B E O V 47 2.3.2. Use of EPR for Protein Binding Studies 51 2.3.3. Interaction of VOSO4 with Apo-transferrin 53 2.3.4. Interaction of B M O V with Apo-transferrin 56 2.3.4.1. Preferential Site Binding 59 2.3.4.2. Maltolate as a Synergistic Binding Anion 62 2.3.5. Interactions o f V O S 0 4 with Serum Albumin 64 2.3.6. Interactions of B M O V with Serum Albumin 65 2.3.7. Strong Versus Weak Site Binding 67 2.3.7.1. Speciation Analysis of Vanadyl Binding by HSA 73 2.3.7.2. Interactions of Imidazoles with B M O V 79 2.4. References 90 vi Pulsed EPR Studies of Chapter 3 Bis(ethylmaltolato)oxovanadium(IV): Biodistribution and 96 In Vivo Coordination Chemistry 3.1 Introduction 96 3.1.1. In Vivo Structural Determination 97 3.1.2. Limitations of Electron Paramagnetic Resonance 99 Electron Nuclear Double Resonance (ENDOR) Spectroscopy ^ j ^ Electron Spin Echo Envelope Modulation (ESEEM) j ^ Spectroscopy 3.1.4.1. Development of Pulsed Paramagnetic Resonance Methods 104 3.1.4.2. Theory of ESEEM 105 3.1.5. Interpretation of ESEEM Spectra 110 3.2. Experimental 112 3.2.1. Animal Dosing Procedures 113 3.2.2. Preparation of Spectroscopic Samples 113 3.2.3. Spectroscopic Methods 114 3.3. Results and Discussion 115 3.3.1. Liver and Kidney 116 3.3.2. Bone 125 3.3.2.1. Contour Lineshape Analysis of Bone HYSCORE Spectra 130 3.3.3. Muscle 135 3.4. References 137 vii Chapter 4 Model Studies of In Vivo Vanadyl Coordination in Bone 143 4.1 Introduction 143 4.1.1 The Vanadyl-Triphosphate System 145 4.1.2. Hydroxyapatite and Vanadyl Interactions 147 4.2. Experimental 148 4.2.1. V02+-triphosphate (VO-TPH) 148 4.2.2. Vanadyl-Hydroxyapatite (VO-HA) 149 4.2.3. EPR Measurements 149 4.3. Results and Discussion 150 4.3.1. Characterization of V O - H A 150 4.3.2. EPR/ESEEM 151 4.3.3. HYSCORE 155 4.3.4. ' Analysis of HYSCORE Spectra 160 4.3.5. Sum Combination Lines in ESEEM Spectra 164 4.3.6. Phosphorus Couplings 166 4.3.7. Proton Couplings 168 4.3.8. Comparison Between the Models 172 4.3.9. Structure of the Complex 173 viii 4.3.10. Comparison to In Vivo Coordination in Bone 179 4.4. References 181 Chapter 5 Conclusions and Future Work 186 Protein Transport of Insulin-Enhancing Vanadium Complexes 5.1.1. The Ligand Taxi Mechanism 190 5.2. Tissue Accumulation of Vanadyl Species 192 5.3. References 195 ix List of Figures Figure 1.1 Sample metallopharmaceuticals currently in clinical use: cisplatin 4 (1), carboplatin (2), sodium bis(thiosulfato)aurate(I) (3), auranofin (4), radioactive technetium-99m complexes for SPECT imaging: Ceretec (5), Myoview (6), Cardiolite (7). Magnevist, a Gd(III)-DTPA (DTPA = /vr,Ar,A^',A^",Ar'"-diethylenetriaminepentaacetic acid) complex used as an MRI contrast agent (8). Figure 1.2 Speciation diagram for V(V), vanadate V 0 2 + , [V0 2 + ] = 1.0 mM. 7 Figure 1.3 Speciation diagram for V(IV), vanadyl (V0 2 + ) , [ V 0 2 + ] 0 =10.0 7 mM. Figure 1.4 Structure of amavadin, the V(IV) natural product of Amanita 9 muscaria mushrooms. Figure 1.5 Structure of maltolate-derived V(IV) complexes (R = CH3: 14 bis(maltolato)oxovanadium(IV), B M O V ; R = C H 2 C H 3 : bis(ethylmaltolato)oxovanadium(IV), BEOV). Figure 1.6 Representative anti-diabetic V(IV) complexes of different donor 15 atom configurations: (a) VO(acetylacetonato)2 (0,0); (b) VO(picolinato)2 (0,N); (c) VO(N-octylcysteineamino)2 (N,S); (d) VO(pyirolidine-Af-carbodithioato)2 (S,S). Figure 1.7 Structural motifs of insulin enhancing vanadium pharmaceuticals 16 for potential treatment of Type 2 diabetes mellitus: (a) Vanadyl cation in acidic aqueous solution; (b) Vanadate anion in aqueous solution; (c), (d) Peroxovanadate complexes, where n = 1-3 and m = 0, 1; (e) Representative V(IV) chelate complex, B M O V . Figure 1.8 Structure of streptozotocin (STZ), an antibiotic toxic to the [3 cells 17 of the pancreas. Figure 2.1 Schematic diagram of transferrin metal ion and anion binding site 33 (adapted from Baker 10). Figure 2.2 Schematic diagram of Cu(II) binding site of HSA (adapted from 35 Harford and Sarkar^). The amino acid sequence starting at the N -terminus is aspartic acid, alanine and histidine. Figure 2.3 (a) Variable pH U V spectrophotometric titration of VO 2 * (1.205 49 mM) and Hema (0.1408 mM) at 298 K, 0.16 M NaCl; (b) Experimental (points) and calculated (lines) absorbance values versus pH for determination of log K | . Figure 2.4 Speciation diagram for the V0 2 + -Hema system ([V0 2 + ] = 0.01 50 mM; [Hema] = 0.02 mM). x Figure 2.5 EPR spectra of aqueous (VO)x-transferrin solutions (298 K, pH 7.4, 0.16 M NaCI, 5 m M Na 2 C0 3 ) : (dark) V O S 0 4 (0.22 mM), apo-Tf (0.23 mM); (light) V O S 0 4 (0.41 mM), apo-Tf (0.22 mM). Figure 2.6 Difference U V titration of apo-Tf by V 0 2 + ([apo-Tf] = 0.0186 m M , [ V 0 2 + ] = 1.55 mM, 300 K , pH 7.4, 0.16 M NaCI): (top) difference U V spectra recorded during titration; (bottom) change in absorbance versus molar equivalents of VO2*", monitored at 253 nm. Figure 2.7 EPR spectra of VO-transferrin (298 K, pH 7.4, 0.16 M NaCI): (top) [V0 2 + ] = 1.11 mM, [apo-Tf] = 0.29 mM; (middle) [BMOV] = 0.112 mM, [apo-Tf] = 0.134 mM; (bottom) [BMOV] =1.12 mM. Figure 2.8 EPR spectra of BMOV-apo-Tf solutions showing displacement of V 0 2 + ions in (VO)2-transferrin by Fe 3 + (298 K, 0.16 M NaCI). Figure 2.9 EPR spectra (-7/2,, and -5/2,, peaks) of (VO)x-transferrin (T = 130 K , pH 7.40, 0.16 M NaCI).' Figure 2.10 EPR spectra of decarbonated apo-Tf and B M O V before and after addition of N a 2 C 0 3 (T = 298 K, pH 7.4, 0.16 M NaCI). Figure 2.11 EPR spectra of aqueous (VO) x-HSA solutions (T = 298 K, pH 7.4, 0.16 M NaCI): (dark) [V0 2 + ] = 0.459 mM, [HSA] = 0.56 mM; (light) [V0 2 + ] = 2.18 mM, [HSA] = 0.52 mM. Figure 2.12 Difference U V titration of HSA with V 0 2 + ( [V0 2 + ] = 1.495 mM, [HSA] = 4.64 x 10"2 mM, 300 K, pH 7.4, 0.16 M NaCI). Figure 2.13 EPR spectra of VO-HSA (T = 298 K, pH 7.4, 0.16 M NaCI): (top) [V0 2 + ] = 0.285 mM, [HSA] = 0.274 mM; (middle) [BMOV] = 0.084 mM, [HSA] = 0.521 mM; (bottom) [BMOV] = 1.22 mM. Figure 2.14 Frozen solution EPR spectra of V O S 0 4 or B M O V with HSA (T = 130 K, pH7.40, 0.16 M NaCI). Figure 2.15 Frozen solution EPR spectra of V O S 0 4 or B M O V with HSA: enlargement of high field parallel resonances +3/2, +5/2 and +7/2 from Figure 2.14 (T = 130 K, pH 7.40, 0.16 M NaCI). Figure 2.16 S-band EPR spectra of V O S 0 4 or B M O V with HSA (T = 130 K, pH7.4, 0.16 M NaCI). Figure 2.17 Magnification of low field region of S-band EPR spectra from Figure 2.16 (T = 130 K, pH 7.4, 0.16 M NaCI). Figure 2.18 Difference UV titration of HSA with B M O V (300 K, pH 7.4, 0.16 M NaCI, [BMOV] = 1.467 mM; [HSA] = 0.085 mM). 54 55 56 58 62 63 65 65 66 68 69 70 71 72 xi Figure 2.19 Solution thermodynamics simulation of the B M O V - H S A system, 74 using stability constants previously reported;27,38 (top) percent concentration (of V 0 2 + ) of B M O V and [ma-VO-HSA] versus ternary complex stability; (bottom) titration simulation of a 1:1 solution of V 0 2 + and HSA ([V0 2 + ] = [HSA] = 0.01 mM; log p\ = 15.0). Figure 2.20 Frozen solution EPR spectra from a titration of a 1:1 solution of 77 V 0 2 + and HSA with maltol: (A) frozen solution of B M O V for comparison; (B) initial VO-HSA spectrum; (C) low equivalents of maltol removes V 0 2 + from the weak sites; (D) excess maltol has no effect on the EPR spectrum; (E) 2:1 B M O V to HSA solution for comparison (dashed lines indicate line positions for B M O V ; T = 130 K, pH 7.40, 0.16 M NaCl). Figure 2.21 Structure of L-histidine, with the imidazole ring component 79 highlighted. Figure 2.22 Trans (a) versus cis (b) coordination (relative to V=0) structures 81 of the BMOV-ImH adduct. Figure 2.23 Frozen solution EPR spectra of B M O V (top) and a 2:1 solution of 83 ImH: B M O V (middle), along with its simulation (bottom) ([BMOV] = 2.63 mM, [ImH] = 5.23 mM, T = 130 K, pH 7.4). Figure 2.24 Titration of B M O V (0.122 mM) with ImMe (25.1 mM) monitored 84 by UV-Vis spectroscopy, from 0 to 2 molar equivalents ImMe (200 uL) (V 0 = 25.5 mL, T = 298 K, pH 7.40, 0.16 M NaCl). Figure 2.25 Theoretical (solid line) versus experimental absorbance (data 85 points) values for the determination byUV-vis titration of log K i for the interaction of ImMe with B M O V . Figure 2.26 Structures of L-lysine (a) and 1-aminopropane (b). 88 Figure 3.1 Schematic diagram depicting the underlying principles of ENDOR 102 spectroscopy, for an S = lA, I = lA system. A : Normal allowed EPR transition (ms - ±1, mi = 0), where excitation of an electron (solid arrow) leads to rapid relaxation (dashed arrow) back to the ground state. B: Saturation condition for EPR. Application of high power leads to destruction of the Boltzmann distribution. EPR transition intensity decreases as there is not net change in populations of the higher and lower spin states. C: Application of radio frequencies to a saturated EPR system alleviates the condition by excitation of nuclear transitions (labelled NMR). An increase in the EPR signal strength is recorded as a function of the frequency of the applied radio frequency (adapted from Makinen et al.14). xu Figure 3.2 Schematic diagram depicting the effective spectroscopic domain 105 of the three main paramagnetic probes (M = EPR-active transition metal ion) (adapted from Garner et al.29). Figure 3.3 Energy level diagram for a S = Vi spin system coupled to a single I 107 = Vi nucleus. Dashed arrows depict the allowed Am s = ±1 transitions, while solid double headed arrows the nuclear transitions (v a and v p ) between the energy levels of each M s spin manifold that couple with the EPR transitions to create the modulation effect of ESEEM spectroscopy (adapted from Deligiannakis et al. ^ ) . Figure 3.4 Energy level diagram for S = lA spin system coupled with a single 109 I = 1 nucleus. Single quantum (sq) transitions are indicated by dashed arrows while double quantum (dq) transitions are shown with solid double headed arrows. Note that the energies of the two sq transitions within a spin manifold need not necessarily be equal (adapted from Goldfarb et al.31). Figure 3.5 First derivative field sweep ESE spectrum of rat kidney tissue, 117 taken from an animal previously administered BEOV via drinking water; dashed lines indicate field positions used in E S E E M studies (T = 20 K, v = 9.3966 GHz, x = 200 ns). Figure 3.6 Modulus two-pulse E S E E M spectra of rat kidney, taken from an 118 animal previously treated with BEOV via drinking water (T = 30 K, v = 9.3977 GHz). Figure 3.7 Stacked modulus three-pulse ESEEM spectra of BEOV-treated rat 118 kidney (B = 3333 G, T = 30 K, v = 9.3966 GHz, step = 16 ns). Figure 3.8 HYSCORE spectrum of BEOV-treated rat kidney (B = 3 333 G, T 122 = 30 K, v = 9.3966 GHz, T = 200 ns, 256 x 256 points, step = 16 ns; sq = single quantum 1 4 N cross peak; dq = double quantum l 4 N cross peak; H = cross peaks of *H). Figure 3.9 Modulus two-pulse E S E E M spectra of BEOV-treated rat bone (T 127 = 30 K, v = 9.3038 GHz). Figure 3.10 Stacked modulus three-pulse ESEEM spectra of bone taken from 127 BEOV-treated animal (B = 3301 G, T = 30 K, v = 9.3038 GHz, step = 16 ns). Figure 3.11 HYSCORE spectrum of rat bone, measured at m v =-1/2 peak of 129 the FS-ESE spectrum (B = 3 315 G, T = 30 K, v = 9.3252 GHz, x = 120 ns, 256 x 256, step 16 ns; 1, 2, 3: cross-peaks arising from 3 1 P ; P: 3 I P matrix peak, 5.7 MHz. H: ' H matrix peak, 14.1 MHz). xiii Figure 3.12 Stacked HYSCORE spectrum of BEOV-treated rat bone depicting 129 peak shapes of cross peaks 2 and 3 from Figure 3.11 after suppression of 3 1 P matrix peak (x = 280 ns). Figure 3.13 Contour lineshape analysis plot of squares of frequencies of CP 1 - 131 3 from Figure 3.1 land Figure 3.12 for determination of A and T. Figure 3.14 Possible coordination mode of V 0 2 + ions in bone mineral (m,n = 134 1, 2 independently). Figure 4.1 Proposed solution structures of the vanadyl-triphosphate system: 146 (a) [VO(HL)] 2" and [VO(L)]3" monoligand complexes ^,19; (b) [VO(HL)(L)]7" and [VO(L) 2] 8" bisligand complexes 1 9; (c) tridentate [VO(L)]3" species proposed by Buglyo et al.}^ based on potentiometric and EPR data. (L = P3O10 5 ") Figure 4.2 IR spectra of hydroxyapatite (HA) and amorphous calcium 151 phosphate (ACP) (in KBr). Figure 4.3 Modulus two-pulse ESEEM spectrum of VO-TPH (B = 3479 G, T 153 = 3 0 K , v = 9.713 GHz). Figure 4.4 Modulus two-pulse E S E E M spectrum of V O - H A (B = 3460 G, T 155 = 20 K, v = 9.717 GHz). Figure 4.5 (a) HYSCORE spectrum of VO-TPH, in the region of cross-peaks 156 P,-P 3 (B = 3479 G, T = 30 K, x = 120 ns, v = 9.713 GHz), (b) Enlarged stacked plot presentation of the (+,+) quadrant to visualize the P 2 and P 3 cross peaks. Figure 4.6 HYSCORE spectrum of VO-TPH, in the region of cross-peaks 158 H1-H3 (spectral parameters as in Figure 4.5). Figure 4.7 HYSCORE spectrum of VO-HA, in the region of cross-peaks PA- 158 P c (B = 3465 G, T = 30 K, x = 128 ns, v = 9.710 GHz). Figure 4.8 HYSCORE spectrum of VO-HA, in the region of cross-peaks H A 159 and H B (B = 3465 G, T = 30 K, x = 192 ns, v = 9.710 GHz). Figure 4.9 Contour lineshape analysis plot for 3 1 P cross-peaks observed in 160 HYSCORE spectra of VO-TPH (e.g. Figure 4.5) recorded at various x values. The larger point was arbitrarily chosen as v a and the smaller coordinate as v p . The solid lines depict the linear regressions of the experimental data points, while the dashed curve represents a plot of | v a + v p | = 2v P with v P = 6.00 MHz. xiv Figure 4.10 Contour lineshape analysis plot for *H cross-peaks observed in 162 HYSCORE spectra of VO-TPH of various x values. The larger point was arbitrarily chosen as v a and the smaller coordinate as vp. The solid lines depict the linear regressions of the experimental data points, while the dashed curve represents a plot of | v a + vp | = 2v H with v H = 14.81 MHz. Figure 4.11 Proton sum combination peaks in the ID four-pulse E S E E M 165 spectrum (magnetic field 3479 G, x = 120 ns, v = 9.713 GHz). Figure 4.12 Proposed solution structure of VO-TPH at pH 5.0, based on 174 E S E E M and HYSCORE data. Figure 4.13 Structure of [(tacn)CoP30ioH2], based on the X-ray crystal 179 structure reported by Haight, Jr. et al.^ (tacn = 1,4,7-tri azacyclononane). xv List of Tables Table 2.1 Comparison of stability constants for insulin-enhancing vanadyl 51 complexes. Table 2.2 Spin Hamiltonian parameters for B M O V and BMOV-ImH (pH 82 7.4, 0.16 M NaCl). Table 3.1 Comparison of the isotropic coupling constant magnitudes for 124 amine and imine equatorial l 4 N donors to V 0 2 + by E S E E M spectroscopy. Table 3.2 Parameters derived from contour lineshape analysis of H Y S C O R E 132 spectra of BEOV-treated rat bone. Table 4.1 Spin Hamiltonian values for V O model complexes and in vivo 152 V 0 2 + in bone mineral (20 K, v = 9.7170 GHz). Table 4.2 Parameters3 derived from contour lineshape analysis of 163 HYSCORE spectra of VO-TPH and VO-HA. Table 4.3 Isotropic (| A |) and anisotropic (| T |) *H coupling constants of 170 various V0 2 +complexes compared to the proton couplings of V O -TPH and VO-HA. Table 4.4 Comparison of hyperfine coupling constants obtained by pulsed 180 EPR study of VO-TPH, V O - H A and vanadyl-treated bone. List of Equations Equation 2.1 42 Equation 2.2 42 Equation 2.3 42 Equation 2.4 42 Equation 2.5 42 Equation 2.6 52 Equation 3.1 106 Equation 3.2 119 Equation 3.3 119 Equation 3.4 122 Equation 3.5 130 Equation 3.6 130 Equation 4.1 166 XVII List of Schemes Scheme 2.1 Equilibrium equations for binding of V 0 2 + ions to HSA (298 K, pH 5.0). 3 5 (VO) s-HSA = strong site bound vanadyl-HSA; (VO) x (VO) s -HSA = strong and weak site bound vanadyl-HSA, where x = 1-5. Scheme 2.2 Scheme 2.3 Scheme 2.4 Preferential loading of Fe(III) to the N and C-terminus sites of apo-Tf. 5 6 Possible equilibria involved in a solution o f V 0 2 + , HSA and maltol (Equilibrium constants K i , K 2 , p 2 and K sare taken from previous reports.27,38 KaSsoc, IQis and Piemary are proposed constants). 36 50 61 75 xviii List of Abbreviations Abbreviation Definition micro (10"6) T interpulse time (in pulse sequences) n excitation pulse J _ perpendicular £ extinction coefficient (M" 1 cm"1) V frequency X wavelength P density A change A peak shift in sum combination harmonics (two and four pulse experiments) V a / p nuclear transition frequency V | nuclear Zeeman frequency (generic nucleus) I  parallel +LSIMS liquid secondary ion mass spectrometry, positive detection mode ID one dimensional 2D two dimensional 2Fe-2S ferredoxin two iron, two sulfur centre A hyperfine coupling constant A electron nuclear hyperfine tensor A absorbance A Angstrom (10"'°m) A A S atomic absorption standard acac acetylacetonate anion ACP amorphous calcium phosphate ADP adenosine diphosphate xix ^ i s o isotropic hyperfine coupling constant apo-Tf apo-transferrin Arg arginine Asp aspartic acid atm atmosphere ATP adenosine triphosphate ATPase adenosine triphosphatase ax axial B, excitation magnetic field B E O V bis(ethylmaltolato)oxovanadium(IV) B M O V bis(maltolato)oxovanadium(IV) BSA bovine serum albumin B j total moles B M O V c centi(10"2) Caco-2 endothelial cell line calc calculated CC control untreated (rat treatment group) cm"1 wavenumber COSY correlation spectroscopy (NMR) CP cross-peak (bone sample, Chapter 3) CT control treated (rat treatment group) cw continuous wave Cys cysteine d day Da Dalton (1 gram per mole) DC diabetic untreated (rat treatment group) dd distilled and deionized DFT density functional theory DPPH diphenylpicrylhydrazide XX dq double quantum DT diabetic treated (rat treatment group) DTPA N,N,N\N' \N'' '-diethylenetriaminepentaacetic acid e.g. for example E A elemental analysis edda ethylenediamminediacetate ema" ethylmaltolate anion ENDOR electron nuclear double resonance EPR electron paramagnetic resonance EPRI electron paramagnetic resonance imaging eq equatorial E S E E M electron spin echo envelope modulation et al. and others FS-ESE field swept electron spin echo FT Fourier transform g gram g Lande splitting factor, g-factor G gauss (magnetic field unit) G giga(109) g electron Zeeman tensor Ga/p intercept of contour lineshape analysis plot gly glycine h Planck's constant H # 'H cross-peak (triphosphate, Chapter 4) H2salen A ,^A^'-bis(salicylidene)ethylenediamine * HA hydroxyapatite Hema ethylmaltol HEPES 2-hydroxyethyl-1,4-piperazine-ethanesulfonic acid hf hyperfine xxi Himac 4-imidazoleacetic acid His histidine Hma maltol H 0 static magnetic field HPLC high performance liquid chromatography HSA human serum albumin H x *H cross-peak (hydroxyapatite, Chapter 4) H Y S C O R E hyperfine sub-level correlation HySS Hyperfine simulation software Hz Hertz (one cycle per second) I nuclear spin operator i.d. inner diameter i.e. that is IDDM insulin-dependent diabetes mellitus ImH imidazole ImMe 1 -methylimidazole ip intraperitoneal IR infrared I T total moles 1 -methylimidazole k kilo (103) K Kelvin (temperature unit) K equilibrium constant K w ionization constant for water (10"13'74) L liter L generic ligand (in chemical formulae) L-band 1.8 GHz (EPR frequency) L C liquid chromatography L N 2 liquid nitrogen M molar, moles per litre (mol L" 1) xxii m milli (IO"3) m metre M generic electron paramagnetic resonance active metal centre M generic transition metal ion ma" maltolate anion M A L D I matrix-assisted laser desorption ionization meox 2-methylquinolin-8-olate m i secondary quantum number for nuclear spin angular momentum mol mole MRI magnetic resonance imaging m s secondary quantum number for electron spin angular momentum M s electron spin manifold MS mass spectrometry M W molecular weight n nano(10"9) N M R nuclear magnetic resonance NOE nuclear overhauser effect o.d. outer diameter °C degrees Celsius P# 3 1 P cross-peak (triphosphate, Chapter 4) pH -log[H +] pic picolinate ppm parts per million PTPase protein tyrosine phosphatase P x 3 I P cross-peak (hydroxyapatite, Chapter 4) Q nuclear quadrupolar coupling tensor Q nuclear quadrupolar coupling constant Q quality factor for a resonant cavity Q a / p slope of contour lineshape analysis plot xxiii Q-band 35 GHz (EPR frequency) r radius S electron spin angular-momentum vector operator s second S-band 4 GHz (EPR frequency) SHQ super high Q SPECT single photon emission computed tomography sq single quantum STZ steptozotocin T anisotropic coupling constant T temperature T|_ perpendicular anisotropic hyperfine coupling constant Tj longitudinal relaxation time (pulsed N M R , EPR) T 2 transverse relaxation time (pulsed NMR, EPR) Thr threonine TM001 transverse magnetic field resonant cavity TOF time-of-flight tp [PaO.of Tyr tyrosine U B C University of British Columbia U V ultraviolet V volt vis visible V O - H A vanadyl-hydroxyapatite model sample VO-TPH vanadyl-triphosphate model sample W Watt w/v weight/volume W-band 95 GHz (EPR frequency) x, y, z Cartesian coordinates X-band 9 GHz (EPR frequency) xxiv Acknowledgments I would like to thank first and foremost the Lord Jesus Christ for giving strength and guidance to complete my project. I cannot extend enough gratitude to my wife and best friend Rosaleen, whose love and support I could not do without. Thank you and love forever. This thesis is the product of input and guidance from many sources, due to the interdisciplinary nature of the project. Professionally, the assistance of the following people is most gratefully acknowledged: Dr. Chris Orvig, my research advisor, for his tireless optimism and encouragement. Dr. Graeme Hanson (Centre for Magnetic Resonance, U . Queensland), for the crash course in EPR instrumentation and software, Linux OS and a needed sojourn to Australia. Dr. Sergei Dikanov (Illinois EPR Research Centre, U . Illinois at Urbana-Champaign), for excellent E S E E M spectra, along with detailed answers to my many questions regarding them. Dr. Kathie Thompson, for many thoughtful discussions on vanadium biochemistry, paper proofreading, and insightful, challenging questions regarding my conclusions. Mr. Edmond Lam (Chem 449 student), for his important contribution of determining the stability of imidazole interactions with B M O V in Chapter 2. The many past and present members of Equipe Orvig, especially Dave, Tim, Peter, Simon, Marco, Leon, Mike, Ashley, Nico, Song Bin, Ika and Alex. The work of the support staff of the department is greatly appreciated, most notably the work of Zoltan, Milan, Mike, Judy, Lani, Diane, Sheri, Dave and Jason. The Bricklayers basketball team, particularly the forwards who never get the ball. Lynn and Bob MacDonald for the best carpool in the Lower Mainland. xxv To Albert Aldege Liboiron, on the 10 anniversary of your passing, this thesis honours your intelligence and gentle nature. You are sorely missed. xxvi Chapter 1 Biologically Relevant Physical Studies of Insulin-Enhancing Vanadium Complexes 1.1. Medicinal Inorganic Chemistry Metals play a central role in biological systems, yet their medicinal applications are, at present, surprisingly underdeveloped, considering the importance of endogenous metal centers such as iron in hemoglobin, calcium in bone mineral and copper in ceruloplasmin. The chemistry of metal ions leads to a striking variety of functions, both chemical and structural, in virtually every system in the body. It is obvious that nature has selected various metal ions for use in numerous metalloprotein systems to take advantage of metals' diverse redox activity and coordination structures. Bioinorganic chemistry, itself a relatively new field, has made great progress in the comprehension of roles of metal ions in biological molecules and how these roles are modulated by the ligand set, typically an organic scaffold. It seems logical that development of metallopharmaceuticals would therefore lead to entirely new approaches to the treatment of various diseases, through the exploitation of different metabolic targets and pathways, and novel in vivo activities. To date, however, only a small number of metal-based drugs are approved for use in Canada and the United States. Most of these drugs are currently used for diagnostic purposes only, and while important, are not for the actual treatment of disease. When one speaks of medicinal inorganic chemistry as a discipline, the field is bounded by discussions of specific chemical systems, previously shown to have regular and reproducible effects within the body that can be or have been optimized and improved upon 1 References begin on page 22 Chapter 1 for treatment or diagnosis of a disease state. 1 Examples of such treatments could include alleviation of an excess or deficiency of a natural endogenous metal ion such as iron or copper, or administration of a metal complex for inhibition or promotion of native biochemical systems. The scope of the field is focused on the discovery of inorganic or organometallic complexes for the treatment or diagnosis of disease, coupled with mechanistic investigations into the effects (including toxicity) of these complexes in biological systems. While metals have been used as anecdotal medicines for thousands of years, modern medicinal inorganic chemistry has its origins only in the last 30 years. The state of the field was the focus of a recent issue of Chemical Reviews; 1 active areas of research were subsequently reviewed by Guo and Sadler.2 1.1.1. Metallopharmaceuticals in Clinical Use As mentioned above, surprisingly few metal complexes have been approved for medical use. Structures of a representative sample are shown in Figure 1.1. The field of medicinal inorganic chemistry could be considered to have started with cisplatin, possibly the most successful, entirely inorganic metallopharmaceutical. Cisplatin, c/5-pPt(NH3)2Cl2], was serendipitously discovered by Rosenberg et al. in 1965.3 This simple platinum(II) complex (Figure 1.1,1), containing no atoms of carbon, was approved for medical use in 1978 for the treatment of testicular and ovarian cancer. Cisplatin has since become one of the three most widely used antitumor drugs^ and has reduced the fatality rate of testicular cancer from 90% to less than 10%.5 Through extensive study of the mechanism, thousands of prospective platinum(II) and platinum(IV) complexes have been analyzed, leading to the approval of the less nephrotoxic carboplatin (Figure 1.1, 2).4 2 References begin on page 22 Chapter 1 Platinum's precious metal congener, gold, has also been approved for use for a small number of complexes, as a treatment of rheumatoid arthritis. Gold(I) complexes (Figure 1.1, 3), including the second-generation drug auranofin, (Figure 1.1,4) have been in clinical use for decades, but their mechanism of action is not well understood.^ Both platinum and gold complexes provide fine examples of a common theme in medicinal inorganic chemistry, in which the administered complexes are in fact pro-drugs. Biodistribution and mechanistic studies have shown that, within the body, the coordination structures of both Au-based anti-inflammatories and Pt-based antitumor drugs are altered post-administration, leading to generation of the active species in vivo. The most common examples of metal-based drugs, however, are those used for diagnostic purposes. Perhaps the most prevalent, and well known, is the use of barium sulfate (BaS0 4) as a contrast agent for X-ray imaging. Several other imaging agents are also in widespread use. Several 99m-Tc and Gd complexes have been approved for use for in vivo imaging, by single photon emission computed tomography (SPECT) and magnetic resonance imaging (MRI), respectively.^ These complexes, through their radioactive ( 9 9 m Tc) or magnetic (Gd) properties, have allowed for detailed in vivo images to be obtained through non-invasive means. The quality and ease of obtaining such pictures have led to earlier detection and diagnosis of dozens of diseases. Technetium scans account for over 90% of all diagnostic nuclear medicine procedures worldwide, due to the success of imaging agents such as Ceretec, Myoview and Cardiolite (Figure 1.1, 5-7, respectively).9 3 References begin on page 22 Chapter 1 8 Figure 1.1: Sample metallopharmaceuticals currently in clinical use: cisplatin (1), carboplatin (2), sodium bis(thiosulfato)aurate(l) (3), a u r a n o f i n (4), radioactive technetium-99m complexes for SPECT imaging: Ceretec (5), Myoview (6), Cardiolite (7). Magnevist, a Gd(III)-DTPA (DTPA = N,N,N',N' \N'' '-diethylenetriaminepentaacetic acid) complex used as an MRI contrast agent (8). The prominence of these radiopharmaceuticals, however, does not diminish the importance of gadolinium-based contrast agents for MRI. Figure 1.1 depicts the structure of 4 References begin on page 22 Chapter 1 Magnevist (8), one example of a series of Gd(III) complexes approved for clinical use. 10 The structures of these drugs highlight another central concept in medicinal inorganic chemistry: the judicious choice of ligands for the modulation of compound properties. Adjustable properties include increasing the biological target specificity or decreasing the inherent toxicity of the metal ion (i.e. unchelated Gd(III) is a toxic metal ion in humans, but administration as a tightly bound chelate results in no adverse effects).8 As in any medical or pharmacological field, the search continues for new treatments, or improvements on current practices. Current research in medicinal inorganic chemistry has led to the exploration of several metals for potential use in medicine. 1 One such metal receiving considerable attention over the past 20 years is vanadium. 1.2. Vanadium Vanadium, element number 23, was first discovered in Mexico in 1802, but not fully characterized. In 1831, it was rediscovered by Sefstrom, and named after Vanadis, the Norse goddess of beauty, for the bright and varied colours of its solutions. 11 Vanadium, a minor constituent of the Earth's crust (100 ppm, greater than Cu, Zn, Pb and Sn),12 has found considerable industrial application in the alloying of steel and in the construction of aerospace components. 11 1.2.1. Biologically Relevant Aqueous Chemistry of Vanadium Vanadium aqueous chemistry is complicated due to a rather intricate hydrolysis series, which includes formation of polyoxometallates for V(V), and mono- and divanadylhydroxy products for V(IV). Speciation diagrams for V(V) and V(fV) are depicted 5 References begin on page 22 Chapter 1 in Figure 1.2 and Figure 1.3, calculated from stability constants stated in Baes Jr. and Mesmeric and Chasteenl^, respectively. Of importance to biological applications is the formation of oligovanadates ( V 2 - V | 0 ) at higher concentrations (> 1 mM) and insoluble hydroxide complexes for V(IV) in the physiological pH range (pH 6-8). Vanadium(V) and (IV) are the most important redox states in biological systems, 15 although V(III) systems are known (vide infra)}^ Vanadium(V) is typically found as V0 3 " , HVO4 2 " or H 2 V0 4 ~ , while dissolution of vanadium in acidic aqueous solution results in the formation of the vanadyl cation, V 0 2 + (actually [VO(H 2 0) 5 ] 2 + ) , the most stable oxocation of the first row of transition metals. 13 Conversion between the two is pH-dependent, with vanadyl being stable in acidic aqueous solution, but rapidly oxidized to V(V) above pH 3. The stability of V 0 2 + beyond the acidic pH range can be modulated by complexation, with some V(IV) complexes possessing oxidation half-lives several times greater than the common V(IV) salt, vanadyl sulfate (VOSO4). 17,18 Conversely, conversion back to the starting V(IV) species is usually facile, and almost any reducing agent can be used. 19 Vanadium is similar to Mo in that its anionic and cationic forms can participate in biological processes. Tetrahedral vanadate ([HXV04](3"X)") is structurally similar to phosphate while V(IV) and V(III) ions compete with other endogenous metal ions for coordination to endogenous ligands.20 These similarities lead to considerable in vivo activity, dependent on oxidation state. Within the body, the main oxidation states are +IV and +V,21>22 balanced by an equilibrium between the two species due to the presence of molecular oxygen and biological reducing agents such as ascorbic acid^ 3 and cysteine.24 6 References begin on page 22 Chapter 1 2 4 6 8 10 12 pH Figure 1.2: Speciation diagram for V(V), vanadate V 0 2 + , [V0 2 + ] = 1.0 mM. 2 4 6 8 10 12 pH Figure 1.3: Speciation diagram for VflV), vanadyl (V0 2 + ) , [V0 2 + ] 0 = 10.0 mM. 7 References begin on page 22 Chapter 1 1.2.2. Vanadium in Biological Systems From a biological perspective, vanadium species and activities can be divided into two general groups: species in which the presence of vanadium is deliberate and natural, and instances where exposure to and accumulation of vanadium has occurred, leading to the presence of vanadium in systems in which it is not normally present in elevated concentrations. 1.2.2.1. Naturally-Occurring Vanadium Despite a myriad of studies, the essentiality of vanadium to mammalian systems is still debated.25,26 Deficiency diseases have so far only been identified in the chick and the rat, when vanadium is excluded from the diet.25,27 Reintroduction of vanadium into the diet ameliorates the detrimental effects. The exact role of vanadium in these organisms has not been determined. Vanadium has proven essential for a number of simpler organisms. Curiously, vanadium is accumulated to molar concentrations in ascidians (commonlyknown as sea squirts), mainly as the oxygen-sensitive V(ffl), after reduction of vanadate from sea water. The purpose of this high concentration, typically 107 times greater than the environmental level,28 is not known.29 Amanita muscaria, a species of mushroom, contains significant amounts of the element as the natural product amavadin, one of most stable V(IV) complexes known, and a rare example of a non-oxo-containing V(IV) compound (Figure 1.4).30 Higher animals generally possess an intracellular vanadium concentration of less than 0.2 uM, thought to be too low to exert any biochemical effect.^ 1 Vanadium is also utilized in some eucaryotic enzymes, namely haloperoxidases (algal) and nitrogenases (bacterial).32 The former are thought to utilize a V(V) center to 8 References begin on page 22 Chapter 1 coordinate peroxide, followed by oxidation of a halide ion to the hypohalous acid.-" N O evidence exists for concurrent redox activity at the V(V) center.34 Figure 1.4: Structure of amavadin, the V(IV) natural product of Amanita muscaria mushrooms. The vanadium congener of the usual FeMo cofactor of nitrogenase enzymes is thought to have a similar structure, with the Mo atom replaced with a vanadium in the Fe-S cluster. The oxidation state is unknown. Curiously, the metallic substitution imparts superior cold temperature activity to the enzyme compared to the Mo-containing variant.35 1.2.2.2. Medical, Toxic and Biochemical Effects of Vanadium Shortly after its discovery, vanadium was purported to have medical benefits in the treatment of several diseases. By the late 1800s, it was thought to have medicinal value in the treatment of arteriosclerosis, anemia, and infections, particularly tuberculosis and syphilis. 11 Interestingly, French physicians also reported beneficial effects on lipid metabolism and in the treatment of diabetes mellitus.15,36 As w a s common for the period, vanadium enjoyed a short period of "wonder drug" status; by the 1920s it was believed that 5 mg of vanadium over 24 hours was beneficial for treatment of anemia, tuberculosis, chronic 2-9 References begin on page 22 Chapter 1 rheumatism and diabetes. 11 These anecdotal treatments were used despite a burgeoning understanding of the toxicity of vanadium compounds. Priestley first reported, in 1876, the intense toxicity of sodium vanadate to various small animals-^ and several others made additional contributions at the turn of the century. 11 Increased vanadium levels in the biosphere due to human activity have raised concerns over the toxic effects of vanadium exposure. 11 Vanadium intake through the diet is low, 10-20 ug d" 1, 3^ but does result in detectable levels (i.e. by graphite furnace atomic absorption spectroscopy) of the metal in most tissues.39 The toxicity of metal ions varies according to oxidation state, route of exposure and, of course, the amount. In accordance with established trends, vanadium(V) species possess greater toxicity than V(TV) and V(III) species of a homologous series.40 Accidental exposure to vanadium is usually through the respiratory tract or by mouth, and vanadium is known to be far more toxic by the former route.41 Through ingestion, the most commonly observed effect is gastrointestinal distress, resulting in diarrhea and, at higher doses, loss of body weight .42 Vanadium may have more severe toxic effects on reproduction. Recent studies have demonstrated teratogenic effects and decreased fertility in both genders of rats.43,44 Biochemically, vanadium species alter a wide variety of metabolic processes. The effects of vanadate on biological systems are particularly well documented. As a likely result of the aforementioned similarity to phosphate, vanadates have inhibitory, stimulatory and/or regulatory effects on a host of phosphate-utilizing enzyme systems, due to the formation of enzyme-vanadate intermediates.45 Vanadyl ions have also been found to promote or inhibit enzyme activities.46 The most significant effect of either species, however, is the strong 10 References begin on page 22 Chapter 1 inhibition of the N a + - K + ATPase pump by vanadate anions, observed by Cantley et al.^ This discovery spurred an interest in the biochemistry of vanadium in general, subsequently leading to the discovery of in vitro insulin-like effects of sodium metavanadate. This V(V) salt was shown to increase glucose uptake and stimulate glycogen synthesis in rat diaphragm, liver and fat cells,48,49 a n c j enhance glucose transport and oxidation in rat adipocytes.50 Ironically, these effects are thought to be independent of V(V) inhibition of sodium-potassium pumps; numerous studies showed that V(V) is reduced to V(IV) intracellularly,51>52 and sodium transport was unaffected by the administered vanadate s a j t 50,53,54 j n e increased understanding of the effects of vanadium on metabolic processes, particularly those concerned with energy production and storage, led researchers to hypothesize a potential role for vanadium compounds in the treatment of diabetes mellitus. 1.3. Vanadium and Diabetes Mellitus 1.3.1. Diabetes Mellitus Diabetes mellitus is a heterogeneous disorder of lipid and carbohydrate metabolism. It is actually two diseases, the first insulin-dependent diabetes mellitus (IDDM, commonly referred to as Type 1) and the second non-insulin dependent (NIDDM, Type 2). The primary difference between the sub-diseases is the circulating concentration of the signaling hormone insulin. Insulin is a pancreatic hormone that stimulates cellular uptake and usage of glucose, as well as lipogenesis. It is also responsible for inhibiting gluconeogenesis or usage of glycogen energy stores in the liver. In IDDM, non-functioning (3 cells of the pancreas (responsible for the secretion of insulin in normal individuals) leads to an absolute lack of the minimum amounts of insulin required. NIDDM is associated with either peripheral and 11 References begin on page 22 Chapter 1 hepatic tissue resistance to insulin action, or a relative lack of circulating insulin that is insufficient for normal regulation.55,56 Type 2 diabetes mellitus is much more prevalent than Type 1, and has been associated with age and obesity.^7 Approximately 90% of diabetes mellitus patients are Type 2, with the total number of diabetic patients estimated to be approximately 3-6% of the total Canadian population.58,59 The disease is strikingly prevalent in Aboriginal communities, with greater than 30% of the Aboriginal population diagnosed with Type 2 diabetes mellitus by age 65; it is truly an epidemic in progress.60 Either type of diabetes mellitus leads to poorly regulated blood glucose levels, in addition to severe disruptions of glucose and lipid utilization.61 Lack of treatment, or poor maintenance of blood glucose values within a certain range, can lead to debilitating secondary effects, such as heart disease, hypertension, limb amputations and blindness. Control of diabetes mellitus, obligatory for Type 1, is typically achieved through regular intramuscular (im) injection of natural or synthetic insulin. Long-term Type 2 diabetics frequently require exogenous insulin therapy as well, due to the progressive failure of current NIDDM oral drug therapies.62 Insulin is a protein and hence cannot be administered orally; insulin treatment requires a direct im injection to avoid destruction of the hormone in the GI tract. Orally-active agents that could replace (Type 1) or at least improve the effectiveness of circulating insulin (Type 2) would be quite useful as a new avenue for treatment of diabetes mellitus.62 1.3.2. Vanadium Pharmaceuticals: A New Class of Anti-Diabetic Agents? The discovery of oral, anti-diabetic activity of sodium metavanadate in diabetic rats by Heyliger et al. in 1985 generated considerable interest.63 Previous studies had 12 References begin on page 22 Chapter 1 demonstrated in vitro insulin-like activity of vanadium compounds in rat adipocytes;49,50,54 establishment of reproducible action by oral gavage represented the first report of systemic pharmacological activity in an in vivo system. Sodium metavanadate was shown to reduce blood glucose levels^ 3 and partially or completely restore enzymatic glycolysis, lipolysis and glycogenesis in muscle and liver tissue^4-66 without a concomitant increase in circulating insulin levels.67,68 These effects were later shown to be long-lasting, with amelioration of blood glucose levels continuing for some time after withdrawal of vanadium treatment.69 Vanadium(V) salts, however, were not viewed as pharmacologically viable, as the therapeutic index (i.e. the ratio between the minimum toxic dose over the minimum dose required for therapeutic effect) for such compounds was low; the required dose for consistent observation of anti-diabetic effects typically resulted in gastrointestinal toxicity in laboratory animals. These difficulties were attributed to the low absorption of vanadium from the gastrointestinal tract; vanadium(IV) salts, less toxic than the V(V) sources, were tested and also found to be poorly absorbed (less than l%),70-72 hut possibly more effective.73 Modification of the gastrointestinal absorption of vanadium compounds was viewed as a method to allow for decreased doses and hence ameliorate toxicity problems. 1.3.2.1. The Second Generation: Complexes of Vanadium In 1992, the in vivo testing of a promising chelated V(IV) source in diabetic rats was first reported. 7 4 The compound, bis(maltolato)oxovanadium(rV) (BMOV) (see Figure 1.5), was found to be absorbed better from the gastrointestinal tract, and 2-3 times more effective at lowering blood glucose levels than its parent V(fV) salt, vanadyl sulfate.75,76 13 References begin on page 22 Chapter 1 R O O. R Figure 1.5: Structure of maltolate-derived V(IV) complexes (R = CH 3 : bis(maltolato)oxovanadium(rV), BMOV; R = CH 2 CH 3 : bis(ethylmaltolato)oxovanadium(IV), BEOV). The discovery of the favourable properties of BMOV is attributed to the judicious choice of the proligand maltol, an approved food additive in Canada and the United States, with acceptable water solubility and a single ionizable proton. Two maltolate anions combine with the vanadyl ion to form a neutral V(FV) complex of high hydrolytic stability and water solubility.^ "7 Since the discovery of BMOV, dozens of V(1V) and V(V) complexes have been synthesized and characterized. Modifications of the maltolato ligand structure have been attempted (such as lengthening of the pyrone alkyl group, as in bis(ethylmaltolato)oxovanadium(rV), BEOV, in Figure 1 .5)J8,79 a s w e u a s different ligand donor sets, such as 0,0 (e.g. acetylacetonate),80 o,N (e.g. picolinate),^ ! N,S (e.g. N-octylcysteineamide)82 a n d S,S (e.g. pyrrolidine-Af-carbodithiolate).83 The structures of these V(rV) complexes, all of which have been tested in recent years for anti-diabetic activity, appear in Figure 1.6. 14 References begin on page 22 Chapter 1 Figure 1.6: Representative anti-diabetic V(IV) complexes of different donor atom configurations: (a) VO(acetylacetonato)2 (0,0); (b) VO(picolinato)2 (0,N); (c) VO(N-octylcysteineamino)2 (N,S); (d) VO(pyrrolidine-/V-carbodithioato)2 (S,S). In general, many of these complexes demonstrate roughly similar anti-diabetic activity, but comparisons are difficult between studies of different compounds (vide infra). Some complexes, such as the peroxovanadate(V) complexes reported by Shaver and co-workers,^4 have demonstrated high anti-diabetic activity in vitro, but in vivo activity has not been reported. Further development of vanadium salts, and subsequently, complexes of vanadium(III),85 _(iv)74 a n d _(v)86,87 have led to the identification of three general classes of anti-diabetic vanadium compounds: 1) inorganic salts of V(IV) (vanadyl sulfate, VOSO4) and V(V) (vanadates, e.g. sodium orthovanadate, NasVCU); 2) peroxovanadates, formed from combinations of vanadium(V) with hydrogen peroxide (mono- and diperoxovanadates 15 References begin on page 22 Chapter 1 of chemical forms [VO(0 2)(H 20) 2(L-L ')r (n = 0,1) and [VO(02)2(L-L')]n" (n = 1, 2, 3)) and 3) chelated V(IV) and V(V) complexes.88 These structural types are depicted in Figure 1.7. (e) Figure 1.7: Structural motifs of insulin enhancing vanadium pharmaceuticals for potential treatment of Type 2 diabetes mellitus: (a) Vanadyl cation in acidic aqueous solution; (b) Vanadate anion in aqueous solution; (c), (d) Peroxovanadate complexes, where n = 1-3 and m - 0, 1; (e) Representative V(IV) chelate complex, BMOV. 1.3.2.2. Insulin Enhancement by Vanadium Species Vanadium compounds, both inorganic salts and coordination complexes, have been extensively studied in efforts to understand their insulin-mimetic effects. Animal models of diabetes mellitus are most frequently used, the most common of these being the streptozotocin-diabetic Wistar rat. Streptozotocin (STZ, see Figure 1.8) is an antibiotic that selectively destroys the (3 cells of the pancreas, resulting in compromised insulin production 16 References begin on page 22 Chapter 1 and a diabetic state. Despite the reduced insulin levels, STZ-diabetic rats can survive for extended periods.89 OH Figure 1.8: Structure of streptozotocin (STZ), an antibiotic toxic to the P cells of the pancreas. It is important to note, however, that STZ-diabetic rats are insulinopenic, but not strictly insulin-free.89 This model is therefore appropriate for studies of vanadium action as these drugs do not exhibit anti-diabetic effects in the absence of insulin.90 Hence, the effects of vanadium species are best described as insulin-enhancing, not insulin-mimetic, as vanadium drugs can never fully replace insulin. 4^ in the literature, however, these two terms are frequently used interchangeably. Studies of insulin-enhancing vanadium compounds are complicated due to the wide variety of their effects, administration method, biological variability, and disagreements in the literature regarding the most appropriate indicator of anti-diabetic effect.86 ft has become clear, however, through compilation of all cellular, in vivo and biochemical studies, that vanadium (administered as either vanadyl sulfate or a vanadate salt) exerts its effect downstream of the insulin receptor,91 increases glucose consumption by non-oxidative 17 References begin on page 22 Chapter 1 means (i.e. glycogen synthesis)9^ and has effects on both peripheral and hepatic tissue.42 Biochemically, differences exist between the action of V(V) species and V ^ O species. Vanadate is a potent inhibitor of cellular Na +-K +-ATPase pumps (vide supra) and protein tyrosine phosphatases (PTPases),9^ however, a link between protein phosphorylation and vanadium anti-diabetic action has not been unequivocally demonstrated.86 Vanadyl sulfate has been shown to be much more effective than vanadate at activating cytosolic protein tyrosine kinases in rat adipocytes.46 Ultimately, however, vanadyl sulfate has been shown to be only 25-35% as effective as vanadate or insulin in stimulating glucose metabolism and inhibiting lipolysis in vzYro.50,94 The s \^ e 0 f action of vanadium species is unknown. Researchers attribute this gap to the incomplete knowledge of insulin action and its myriad of effects, and the complicated aqueous and redox chemistry of the main biologically significant vanadium species (+1V and +V).86 Vanadyl sulfate and sodium metavanadate have been previously studied in chronic dosing experiments in humans. 15,62,92 Two weeks of treatment with sodium metavanadate (125 mg d"1) improved insulin sensitivity in Type 2 patients, as well as some Type 1 subjects, due to increased non-oxidative glucose disposal. Anti-diabetic effects in all trials, however, were marginal. Vanadium dose levels were up to 100 times less than those used in diabetic rat studies, and in a number of cases, vanadium treatment had to be scaled back or withdrawn to ameliorate gastrointestinal toxicity, despite the overall reduced dose. 15,62,92 Que to the body size versus body weight differences between rats and humans, the actual dose levels for humans are typically 40-80 times less (on a mg doseper kg body weight basis) than they are for rats. This significant pharmacokinetic difference suggests that in fact, the above studies 18 References begin on page 22 Chapter 1 utilized a realistic therapeutic dose.9^ The lone chelated V(IV) complex to be clinically tested is BEOV, which completed Phase 1 trials in 1999. 1.4. Understanding The Metabolism of Vanadium Pharmaceutical Agents The beneficial effects of vanadium compounds in the treatment of diabetes mellitus have been well documented, but considerable resistance to their potential use as pharmaceutical agents remains. Most objections are based upon two concerns: (1) the toxicity of vanadium compounds and the narrow therapeutic index for vanadium salts,86 and (2) incomplete knowledge of biochemical effects and metabolism of vanadium species. While chelated vanadium complexes such as B M O V have been shown to possess increased efficacy, leading potentially to a lower total vanadium dose, the metabolism of these species is even less understood than the well-studied V(V) and V(IV) salts. Some chelated V(IV) complexes exhibit greater efficacy over V O S O 4 regardless of administration route (i.e. intraperitoneal injection versus oral gavage).80,96 \[ n a s been proposed, therefore, that this superior anti-diabetic activity cannot be a result of greater gastrointestinal absorption alone.86 At some point post-administration, therefore, there is (are) a clear difference(s) between the transport, bioaccumulation and/or activityof vanadium(rV) complexes over V(IV) salts. This thesis seeks to investigate this possibility, by studying the transport and in vivo coordination chemistry of the benchmark V(IV) complexes, B M O V and BEOV, against those of VOSO4. This hypothesis requires the assumption that V(IV) chelates are absorbed into the bloodstream from the gastrointestinal tract as intact complexes, and hence could be expected to: (1) interact with endogenous 19 References begin on page 22 Chapter 1 ligands in ways dissimilar to that of VOSO4, and (2) change the bioaccumulation pattern and coordination chemistry of vanadium ions within organs of the body. 1.4.1. Absorption and Transport Transport of vanadium species in vivo is thought to be carried out by two sets of endogenous ligands. Proteins, belonging to the high molecular weight fraction of blood serum, carry out most of this function. Low-molecular weight donors, such as citrate, ascorbate and phosphate species are thought to play only a minor ro le . 9 7 F o r V O S 0 4 , chelation by either fraction is highly desirable to protect the metal ion from hydrolysis at physiological pH. It is decidedly unlikely that vanadium hydroxide products would be the active species as these complexes possess very slow aqueous substitution kinetics and are very insoluble in water. 1 3 A significant metal-binding capability is maintained in the bloodstream, likely to serve a scavenging role against free metal ions (e.g. Fe(III) and Cu(II)) that might otherwise lead to toxic effects. Two common metal-binding species in serum are apo-transferrin and serum albumin. Previous studies have shown that both proteins will bind V 0 2 + , and have been implicated in the transport of vanadyl ions.throughout the bloodstream. 14 Electron paramagnetic resonance (EPR), coupled with chromatographic techniques, has demonstrated a co-elution of V 0 2 + ions with the two protein fractions from blood samples. Chelation by either protein likely protects the metal ion from hydrolysis as well as from oxidation to V(V). It is unknown, however, whether these proteins serve to transport VO 2 * ions directly to the active site, or, in the case of apo-transferrin, assist in intracellular transport. 1 is conceivable, however, that these vanadyl-protein adducts are the active species themselves, or at least a preliminary form of the active complex that leads to the observed anti-diabetic effects. 20 References begin on page 22 Chapter 1 Protein adduct formation, therefore, can be viewed as a biotransformation pathway leading to the generation of the active metabolite; administered vanadium would hence be a pro-drug Chapter 2 of this thesis will study the interactions of B M O V and VOSO4 with human apo-transferrin and human serum albumin using spectroscopic techniques. The differences between chelated and inorganic V(PV) sources will be highlighted and discussed in relation to final generation of an active species and delivery to peripheral tissues. 1.4.2. Biodistribution and Accumulation Accumulation of vanadium has been shown to be independent of administered oxidation state,98,99 and takes place primarily in bone, liver and kidney tissue.98-101 The distribution of chelated versus ionic sources appears to be similar, with an overall greater accumulation of vanadyl complexes.102 j0 further investigate potential differences in in vivo chemistry of B M O V versus VOSO4, the determination of the coordination state of vanadyl ions within the primary bioaccumulatory organs would be useful. To gain insight into the coordination structure of BEOV-sourced vanadyl ions in vivo, advanced pulsed EPR techniques were applied directly to tissue samples taken from vanadium-treated normal and STZ-diabetic rats. Chapter 3 provides detailed spectroscopic descriptions of the coordination environment in bone, liver and kidney samples. Comparisons against previous studies using VOSO4, as well as studies using other spectroscopic probes, are made to allow for conclusions to be drawn regarding the ultimate accumulated form of vanadyl within the three tissues. Due to the uniqueness of the spectroscopic results for bone, the synthesis and study of model vanadyl complexes are described in Chapter 4, in an attempt to gain further insight into the structure of the vanadyl ions in bone mineral. 21 References begin on page 22 Chapter 1 1.4.3. Metabolism of Chelated Vanadium Sources The interdependence of the metabolic processes studied in this thesis will be highlighted in Chapter 5. The final chapter of the thesis will examine how the transport and bioaccumulation functions of the body work in concert to generate an insulin-enhancing compound, and what implications the results of this thesis have on the understanding of the action of this class of pharmaceuticals. The current state of the metabolic model, taking into account results reported herein, will be described. Lastly, some thoughts on future studies regarding the metabolism of vanadium complexes will be suggested, to further the understanding of this promising new line of anti-diabetic therapeutics. 1.5. References 1) Orvig, C ; Abrams, M . J. Chem. Rev. 1999, 99, September issue. 2) Guo, Z. J.; Sadler, P. J. Adv. Inorg. Chem. 2000, 49, 183. 3) Rosenberg, B.; Van Camp, L.; Krigas, T. Nature 1965, 205, 698. 4) Wong, E.; Giandomenico, C. M . Chem. Rev. 1999, 99, 2451. 5) Bosl, G. J.; Motzer, R. J. N. Engl. J. Med. 1997, 337, 242. 6) Shaw (III), C. F. Chem. Rev. 1999, 99, 2589. 7) Reichert, D. E.; Lewis, J. S.; Anderson, C. J. Coord. Chem. Rev. 1999,184, 3. 8) Thunus, L.; Lejeune, R. Coord. Chem. Rev. 1999,184, 125. 9) Dilworth, J. R.; Parrott, S. J. Chem. Soc. Rev. 1998, 27, 43. 10) Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem. Rev. 1999, 99, 2293. 11) Nriagu, J. O. in Vanadium in the Environment. Part 1: Chemistry and Biochemistry; Nriagu, J. O., Ed.; John Wiley and Sons, Inc.: New York, 1998, p 1. 22 References begin on page 22 Chapter 1 12) Willsky, G. R. in Vanadium in Biological Systems: Physiology and Biochemistry; Chasteen, N . D., Ed.; Kluwer Academic Publishers: Dordrecht, 1990. 13) Baes Jr., C. F.; Mesmer, R. E. The Hydrolysis of Cations; Robert E. Kreiger Publishing Company: Malabar, Fl. , 1986. 14) Chasteen, N . D. Biol. Mag. Res. 1981, 3, 53. 15) Goldfine, A . B.; Patti, M.-E. ; Zuberi, L.; Goldstein, B. J.; LeBlanc, R.; Landaker, E. J.; Jiang, Z. Y . ; Willsky, G. R.; Kahn, C. R. Metabolism 2000, 49, 400. 16) Thompson, K. H.; Orvig, C. Coord. Chem. Rev. 2001, 219-221, 1033. 17) Sun, Y . ; James, B. R.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1996, 35, 1667. 18) Amin, S. S.; Cryer, K. ; Zhang, B.; Dutta, S. K.; Eaton, S. S.; Anderson, O. P.; Miller, S. M . ; Reul, B. A. ; Brichard, S. M . ; Crans, D. C. Inorg. Chem. 2000, 39, 406. 19) Crans, D. C ; Amin, S. S.; Keramidas, A. D. in Vanadium in the Environment. Part 1: Chemistry and Biochemistry; Nriagu, J. O., Ed.; John Wiley and Sons, Inc.: New York, 1998, p73. 20) Baran, E. J. Inorg. Biochem. 2000, 80, 1. 21) Rehder, D.Angew. Chem. Int. Ed. Engl. 1991, 30, 148. 22) Rehder, D. Met. Ions Biol. Sys. 1995, 31, 1. 23) Song, B.; Aebischer, N . ; Orvig, C. Inorg. Chem. 2002, 41, 1341. 24) Sakurai, H. ; Shimomura, M . ; Ishizu, K. Inorg. Chim. Acta 1981, 55, L67. 25) Nielsen, F. H. ; Uthus, E. O. in Vanadium in Biological Systems: Physiology and Biochemistry; Chasteen, N . D., Ed.; Kluwer Academic Publishers: Dordrecht, 1990, p 51. 26) Nielsen, F. H. Met. Ions Biol. Sys. 1995, 31, 543. 27) Hopkins Jr., L. L.; Mohr, H. E. Fed. Proc. 1974, 33, 1773. 23 References begin on page 22 Chapter 1 28) Michibata, H.; Uyama, T.; Kanamori, K. in Vanadium Compounds. Chemistry, Biochemistry, and Therapeutic Applications; Tracey, A. S. and Crans, D. C , Eds.; American Chemical Society: Washington, D. C , 1998; Vol . 711. 29) Smith, M . J.; Ryan, D. E.; Nakanishi, K. ; Frank, P.; Hodgson, K. O. Met. Ions Biol. Sys. 1995, 31, 423. 30) Matoso, C. M . M . ; Pombeiro, A. J. L.; da Silva, J. J. R. F.; da Silva, M . F. C. F.; da Silva, J. A . L.; Baptista-Ferreira, J. L.; Pinho-Almeida, F. in Vanadium Compounds. Chemistry, Biochemistry, and Therapeutic Applications; Tracey, A. S. and Crans, D. C , Eds.; American Chemical Society: Washington, D. C , 1998; Vol . 711, p 241. 31) Liu, J.; Reuland, D. J.; Rosenhein, L.; Cao, Z. X . ; Franklin, L. A. ; Ganguli, S. Metabolism 1997, 46, 562. 32) Wever, R.; Hemrika, W. in Vanadium in the Environment. Part 1: Chemistry and Biochemistry; Nriagu, J. O., Ed.; John Wiley and Sons, Inc.: New York, 1998; Vol . 30, p 285. 33) de Boer, E.; Wever, R. J. Biol. Chem. 1988, 263, 12326. 34) de Boer, E.; Boon, K. ; Wever, R. Biochemistry 1988, 27, 1629. 35) Riittimann-Johnson, C ; Chatterjee, R.; Shah, V . K. ; Ludden, P. W. in Vanadium Compounds: Chemistry, Biochemistry and Therapeutic Applications; Tracey, A . S. and Crans, D. C , Eds.; American Chemical Society: Washington, D. C , 1988; Vol . 711, p 228. 36) Nechay, B. R.; Nanninga, L. B.; Nechay, P. S. E.; Post, R. L.; Grantham, J. J.; Macara, I. G.; Kubena, L. F.; Timothy, D. P.; Nielson, F. H. Fed. Proc. 1986, 45, 123. 37) Preistley, J. Phil. Trans. Part 2 1876, 495. 38) Harland, B. F.; Harden-Williams, B. A. J. Am. Diet. Assoc. 1994, 94, 891. 24 References begin on page 22 Chapter 1 39) French, R. J.; Jones, P. J. H. Life Sci. 1993, 52, 339. 40) Venugopal, B.; Luckey, T. D. Chemical Toxicity of Metals and Metaloids; Plenum Press: New York, 1978; Vol . 2. 41) Boyd, D. W.; Kustin, K. in Advances in Inorganic Biochemistry; Eichhorn, C. L. and Marzilli, L. G., Eds.; Elsevier: Amsterdam, 1984; Vol . 6, p 311. 42) Thompson, K. H.; McNeill, J. H.; Orvig, C. Chem. Rev. 1999, 99, 2561. 43) Leonard, A. ; Gerber, G. B. in Vanadium in the Environment. Part 2: Health Effects; Nriagu, J. O., Ed.; John Wiley and Sons, Inc.: New York, 1998; Vol . 2, p 39. 44) Domingo, J. L. Reprod. Toxicol. 1996,10, 175. 45) Stankiewicz, P. J.; Tracey, A. S.; Crans, D. C. Met. Ions Biol. Sys. 1995, 31, 287. 46) L i , J., Elberg, G., Crans, D . C , Shechter, Y . Biochemistry 1996, 35, 8314. 47) Cantley Jr., L. C ; Josephson, L.; Warner, R.; Yanagisawa, M . ; Lechene, C.; Guidotti, G. J. Biol. Chem. 1977, 252, 7421. 48) Tolman, E. L.; Barris, E.; Burns, M . ; Pansini, A. ; Partridge, R. Life Sci. 1979, 25, 1159. 49) Tamura, S.; Brown, T. A. ; Dubler, R. E.; Larner, J. Biochem. Biophys. Res. Commun. 1983,113, 80. 50) Shechter, Y . ; Karlish, S. J. D. Nature 1980, 284, 556. 51) Macara, I. G.; Kustin, K.; Cantley, J., L. C. Biochim. Biophys. Acta 1980, 629, 95. 52) Cantley Jr., L. C ; Aisen, P. J. Biol. Chem. 1979, 254, 1781. 53) Dubyak, G. R.; Kleinzeller, A . J. Biol. Chem. 1980, 255, 5306. 54) Degani, FL; Gochin, M . ; Karlish, S. J. D.; Shechter, Y . Biochemistry 1981, 20, 5795. 55) DeFronzo, R. A. ; Simonson, D.; Ferrannini, E. Diabetologia 1982, 23, 313. 25 References begin on page 22 Chapter 1 56) Butler, P. C ; Kryshak, E. J.; Schwenk, W. F.; Haymond, M . W.; Rizza, R. A . Diabetes 1990,39,217. 57) Choi, B. C ; Shi, F. Diabetologia 2001, 44, 1221. 58) "Diabetes in Canada: National Statistics and Opportunities for Improved Surveillance, Prevention, and Control," Diabetes Division, Health Canada, 1999. 59) James, R.; Young, T. K.; Mustard, C. A. ; Blanchard, J. Health Reports 1997, 9, 47. 60) Young, T. K. ; Reading, J.; Elias, B.; O'Neil, J. D. Can. Med. Assoc. J. 2000, 163, 561. 61) Lienhard, G. E.; Slot, J. W.; James, D. E.; Mueckler, M . M . Sci. Am. 1992, 267, 86. 62) Cohen, N . ; Halberstam, M . ; Shlimovich, P.; Chang, C. J.; Shamoon, H.; Rossetti, L. J. Clin. Invest. 1995, 95, 2501. 63) Heyliger, C. E.; Tahiliani, A . G.; McNeill, J. H. Science 1985, 227, \A14. 64) Brichard, S. M . ; Ongemba, L. N . ; Henquin, J.-C. Diabetologia 1992, 35, 522. 65) Blondel, O.; Simon, J.; Chevalier, B.; Porfha, B. Am. J. Physiol. 1990, 258, E459. 66) Gi l , J.; Miralpeix, M . ; Carreras, J.; Bartrons, R. Biol. Chem. 1988, 263, 1868. 67) Bendayan, M . ; Gingras, D. Diabetologia 1989, 32, 561-567. 68) Ramanadham, S.; Cros, G. H.; Mongold, J. J.; Serrano, J. J.; McNeill , J. H . Can. J. Physiol. Pharmacol. 1990, 68, 486. 69) Cam, M . C ; Faun, J.; McNeill , J. H. Metabolism 1995, 44, 332. 70) Shechter, Y . Diabetes 1990, 39, 1. 71) Shechter, Y . ; Meyerovitch, J.; Farfel, Z.; Sack, J.; Bruck, R.; Bar-Meir, S.; Amir, S.; Degani, H. ; Karlish, S. J. D. in Vanadium in Biological Systems; Chasteen, N . D., Ed.; Kluwer: Dordrecht, 1990, p 129. 26 References begin on page 22 Chapter 1 72) Orvig, C ; Thompson, K. H.; Battell, M . ; McNeill, J. H. Met. Ions Biol. Sys. 1995, 31, 575. 73) Dai, S.; Thompson, K. H.; Vera, E.; McNeill, J. H. Pharmacol, and Toxicol. 1994, 75, 265. 74) McNeill , J. H. ; Yuen, V . G.; Hoveyda, H. R.; Orvig, C. Med. Chem. 1992, 35, 1489-1491. 75) Yuen, V . G.; Orvig, C ; Thompson, K. H.; McNeill, J. H. Can. J. Physiol. Pharmacol. 1993, 71, 270. 76) Dai, S.; Yuen, V . G.; Orvig, C ; McNeill, J. H. Pharmacol. Commun. 1993, 3, 311. 77) Caravan, P.; Gelmini, L.; Glover, N . ; Herring, F. G.; L i , H. ; McNeill , J. H. ; Rettig, S. J.; Setyawati, I. A. ; Shuter, E.; Sun, Y . ; Tracey, A. S.; Yuen, V . G.; Orvig, C. J. Am. Chem. Soc. 1995, 117, 12759. 78) Dikanov, S. A . ; Liboiron, B. D.; Thompson, K. H. ; Vera, E.; Yuen, V . G.; McNeill , J. H. ; Orvig, C. J. Am. Chem. Soc. 1999, 121, 11004. 79) Thompson, K. H. ; Liboiron, B. D.; Sun, Y . ; Bellman, K. D. D.; Setyawati, I. A. ; Patrick, B. O.; Karunaratne, V. ; Rawji, G.; Wheeler, J.; Sutton, K.; Bhanot, S.; Cassidy, C.; McNeill , J. H. ; Yuen, V. G.; Orvig, C. J. Biol. Inorg. Chem. 2002, in press. 80) Reul, B. A. ; Amin, S. S.; Buchet, J.-P.; Ongemba, L. N . ; Crans, D. C.; Brichard, S. M . Br. J. Pharmacol. 1999, 126, 467. 81) Sakurai, H. ; Fujii, K.; Watanabe, H.; Tamura, H. Biophys. Res. Commun. 1995, 214, 1095. 82) Cam, M . C ; Cros, G. H.; Serrano, J.-J.; Lazaro, R.; McNeill, J. H. Diab. Res. Clin. Pract. 1993, 20, 111. 27 References begin on page 22 Chapter 1 83) Watanabe, H. ; Nakai, M . ; Komazawa, K.; Sakurai, H. J. Med. Chem. 1994, 37, 876. 84) Posner, B. I.; Faure, R.; Burgess, J. W.; Bevan, A. P.; Lachance, D.; Zhang-Sun, G.; Fantus, I. G.; Ng, J. B.; Hall, D. A. ; Soo Lum, B.; Shaver, A . J. Biol. Chem. 1994, 269, 4596. 85) Melchior, M . ; Rettig, S. J.; Liboiron, B. D.; Thompson, K. H. ; Yuen, V . G.; McNeill , J. H. ; Orvig, C. Inorg. Chem. 2001, 40, 4686. 86) Crans, D. C. J. Inorg. Biochem. 2000, 80, 123. 87) Shaver, A. ; Ng, J. B.; Hall, D. A. ; Soo Lum, B.; Posner, B. I. Inorg. Chem. 1993, 32, 3109. 88) Thompson, K . H. ; McNeill , J. H. ; Orvig, C. in Topics in Biological Inorganic Chemistry; Clarke, M . J. and Sadler, P. J., Eds.; Springer-Verlag: Heidelberg, 1999; Vol . 2, p 139. 89) Bell Jr., R. H. ; Hye, R. J. J. Surg. Res. 1 9 8 3 , 35, 433. 90) Posner, B. I.; Yang, C. R.; Shaver, A. in Vanadium Compounds. Chemistry, Biochemistry, and Therapeutic Applications; Tracey, A. S. and Crans, D. C , Eds.; American Chemical Society: Washington, D. C , 1998; Vol. 711, p 316. 91) Bhanot, S.; Girn, J.; Poucheret, P.; McNeill, J. H. Mol. Cell Biochem. 1999, 202, 131. 92) Goldfme, A. B.; Simonson, D. C ; Folli, F.; Patti, M.-E. ; Kahn, C. R. J. Clin. Endocrinol. Metah. 1995, 80, 3311. 93) Swarup, G.; Cohen, S.; Garbers, D. L. Biophys. Res. Commun. 1982,107, 1104. 94) Shechter, Y . ; Shisheva, A. ; Lazar, R.; Libman, J.; Shanzer, A . Biochemistry 1992, 31, 2063. 95) McNeill, J. H. , personal communication. 96) Yuen, V . G.; Orvig, C ; McNeill, J. H. Can. J. Physiol. Pharmacol. 1995, 73, 55. 28 References begin on page 22 Chapter 1 97) Kiss, T.; Kiss, E.; Garribba, E.; Sakurai, H. J. Inorg. Biochem. 2000, 80, 65. 98) Hopkins Jr., L. L.; Tilton, B. E. Am. J. Physiol. 1966, 211, 169. 99) Parker, R. D. R.; Sharma, R. P. J. Environ. Pathol. Toxicol. 1978, 2, 235. 100) Wiegmann, T. B.; Day, H. D.; Patak , R. V. J. Toxicol. Environ. Health 1982, 10, 233. 101) Al-Bayati, M . ; Raabe, O. G.; Giri, S. N . ; Knaak, J. B. Amer. Coll. Toxicol. 1991, 10, 233. 102) Setyawati, I. A. , Thompson, K .H . , Sun, Y . , Lyster, D.M. , Vo., C , Yuen, V . G . , Battell, M . , McNeill , J.H., Ruth, T.J., Zeisler, S., Orvig, C. Appl. Physiol. 1998, 84, 569. 29 References begin on page 22 Chapter 2 Interaction of Insulin-Enhancing Vanadyl Complexes with Serum Proteins 2.1. Introduction For any pharmaceutical agent, the transport and biodistribution of the compound are of considerable importance. The method by which and the locations to which a substance is transported can influence the activity, biotransformation and specificity of the drug. Ultimately, virtually all drugs enter into the bloodstream and are then transported by various means to the site(s) of action. Since vanadium complexes are now thought of as pro-drugs that undergo endogenous transformation to a putative active species, I'2 insights into the delivery of the original compound or its metabolites would be useful in delineating the mechanism of action of this class of potential drugs.3 A 2.1.1. Proteins as Drug Transport Vehicles Many serum proteins serve a transport role in the blood. The uptake and delivery of endogenous molecules throughout the body are important functions, and many proteins serving in this role are highly specialized. Proteins are responsible for carrying a wide variety of compounds, from those that are strongly hydrophobic, (e.g. fatty acids) to those that are potentially toxic (e.g. transition metal ions). The transport and binding of metal ions are important capabilities within the blood and cells of the body themselves, as unchelated metal ions, particularly redox-active ions such as Fe(III) or Cu(II), can quickly lead to Fenton-based radical generation and cytotoxicity associated with this process. The body 30 References begin on page 90 Chapter 2 therefore has developed intricate transport and storage mechanisms to avoid the release of free metal ions. For Fe(III), the blood transport protein is transferrin while for Cu(JJ), several proteins work in tandem, with serum albumin carrying a significant portion of the copper load in the blood.5 Both of these proteins are designed in such a way as to maximize binding capability to their respective target metal ions, while at the same time, maintaining a degree of reversibility once the target tissue is reached. Of equal weight, however, is the ability of these proteins to act as metal scavengers, as both are capable of binding a wide variety of metal ions, both those that are naturally present or those present from exogenous sources. The interaction of serum proteins with drug compounds is an active field of pharmaceutical research. Serum albumin has been shown to react with several classes of drugs, and many drug designs today take interactions with this common protein into account. A recent study found that a large percentage of drugs, across many classes, interact in some way with albumin due to two regions of hydrophobic residues of the protein. These interactions affect the drug's final biodistribution, and in some cases, biotransformation.$ For coordination complexes as drugs, however, it is the metal-binding capabilities of serum proteins that are of interest. 2.1.2. Apo-transferrin The main Fe(III)-transport and storage protein in the bloodstream is transferrin. It is a remarkable macromolecule, capable of binding tightly, yet reversibly, two hydrolytically-unstable Fe(III) ions. Transferrin is a single chain, bilobal glycoprotein of 80 kDa,9J0 with each near-homologous lobe responsible for binding of one metal ion each. 11 Its typical serum concentration in human blood is ~ 37 uM, and transferrin is approximately 30% Fe-31 References begin on page 90 Chapter 2 saturated in the bloodstream. 10 This resting Fe load leaves a free transferrin (apo-transferrin, i.e. no prosthetic group, in this case, Fe(III)) concentration of ~ 25 uM (or 50 uM in metal binding sites) open for binding of other metabolites, typically hard metal ions. 12 Stability constants for the formation of several metal-transferrin complexes have been determined; while the most stable complex is formed with Fe(III), a wide variety of metal ions can be bound, with the stability of the interaction determined by the metal ion's similarity to Fe(III). Accordingly, group 13 metal ions Al(III), Ga(III) and In(III) bind with high stability (log K i = 13.5, 18.1 and 18.7 respectively) due to their trivalence and similar ionic radius to Fe(III). 13-15 Transferrin requires the presence of an anion to bind most metal ions. Endogenously, this role is served by either bicarbonate or carbonate, 16 but other anions have been shown to also act in a synergistic manner, such as oxalate, lactate, glycine and nitrilotriacetate.'7 In the absence of an appropriate binding anion, metal binding is very weak, and conversely, negligible anion binding takes place without metal ions. The metal binding site consists of one histidine (His), two tyrosines (Tyr), an aspartate (Asp), along with the (bi)carbonate anion bound to the metal ion in a bidentate fashion. A n arginine (Arg) residue, a threonine (Thr) and amine donors of an adjacent a-helix, collaboratively H-bond to the anion to hold it in close proximity to the metal ion (Figure 2.1). 10 32 References begin on page 90 Chapter 2 , A s p Tyr Figure 2 . 1 : Schematic diagram of transferrin metal ion and anion binding site (adapted from Baker 1 0 ) . 2.1.2.1. Interactions with Vanadium The reactions of vanadyl and vanadate ions with apo-transferrin have been previously studied, chiefly by N .D. Chasteen and by W. R. Harris. Chasteen and his co-workers used V 0 2 + ions as spin probes for the electron paramagnetic resonance (EPR) spectroscopic investigation of the metal binding sites of the protein in the 1970s and 1980s before the transferrin structure had been solved. It was shown that apo-transferrin will bind up to two vanadyl ions at the Fe(III) binding sites. The two binding sites are distinguishable by EPR. Vanadyl binding was shown to require the presence of a synergistic binding anion; (bi)carbonate, malonate, nitrilotriacetate and oxalate were demonstrated to act synergistically to bind vanadyl ions.18 Transferrin-bound vanadyl ions are unstable to oxidation, with an oxidative half-life of 8.1 minutes at pH 7.5 and 0.2 atm 0 2 . 1 9 Conversely, vanadate binds only relatively weakly to apo-transferrin, and the interaction does not require participation of a synergistic anion (log K, = 6.5, pH 7.4, 298 K ) . 2 0 33 References begin on page 90 Chapter 2 Transferrin has been implicated as one of the main transport routes for vanadyl ions in the bloodstream. Vanadium (as either vanadyl sulfate or ammonium metavanadate) administered intravenously to rats or dogs separates with the transferrin fraction of serum during electrophoresis,21 as it does when administered to rats orally. 2 2 Recent computer model studies predicted that as much as 70% of absorbed vanadyl ions become bound in blood serum by apo-transferrin, although the stability constants for vanadyl binding to apo-transferrin have not yet been determined.23 2.1.3. Serum Albumin Serum albumin is the most prevalent protein in the blood, with a concentration of approximately 0.6 m M . 2 ^ It serves many roles, such as fatty acid transport, metal ion scavenging (particularly soft metal ions such as Hg(II)25 and Cd(II)26), and maintenance of blood osmotic pressure.2^ Albumin is a globular, single chain protein of 585 amino acids and has a molecular mass of 66 kDa.5 From a bioinorganic perspective, metal ion transport is one of its key roles, and it is implicated in the transport of several metal ions, including Cu(II),l2>28"30 Zn(II)3 132 a n ( j Ni(Il). 24 Albumin has also been shown to interact strongly with Au(I) antiarthritic drugs through its lone free thiol of Cys-34.33,34 In human serum albumin (HSA), Cu(II) ions are specifically bound in a site near the N-terminus of the protein, by the N-terminal free amine, two peptide nitrogens and a histidine. The Cu(II) ions are bound in a square planar geometry (Figure 2.2). 34 References begin on page 90 Chapter 2 i Protein Figure 2.2: Schematic diagram of Cu(II) binding site of HSA (adapted from Harford and Sarkar^). The amino acid sequence starting at the N-terminus is aspartic acid, alanine and histidine. 2.1.3.1. Interactions with Vanadium Chasteen and Francavilla, using EPR spectroscopy, conducted the most definitive study of the interaction of vanadyl sulfate with HSA.35 Plots of EPR peak intensity versus the concentration of vanadyl sulfate were biphasic, indicating two distinct types of binding sites. The strong binding site would take up one equivalent of V 0 2 + , and the vanadyl ions could be displaced from the protein byNi(II) ions (which bind at the Cu(II) site).24 The weakly bound vanadyl ions were determined to be non-specific interactions, probably with exposed carboxylate groups. Up to 5 equivalents of V 0 2 + could be bound. Analysis of the rotational correlation times of the two types of bound vanadyl ions showed that strong site bound vanadyl ions rotated with a frequency similar to that of the protein, while the weak site ions had a greater degree of freedom, and hence their EPR spectra were less anisotropic at room temperature. This result supported the assignment of the weak sites as non-specific 35 References begin on page 90 Chapter 2 interactions on the protein surface. Using EPR peak intensities, the authors determined binding constants for both sites at pH 5.0 and 298 K (Scheme 2. l ) . 3 ^ V 0 2 + + HSA - (VO)s-HSA log K s = 6.41(5) x V 0 2 + + (VO)s-HSA . (VO)x(VO)s-HSA log K w = 4.38(2) Scheme 2.1: Equilibrium equations for binding of V 0 2 + ions to HSA (298 K, pH 5.0).35 (VO) s -HSA = strong site bound vanadyl-HSA; (VO) x (VO) s -HSA = strong and weak site bound vanadyl-HSA, where x = 1-5. The binding constants are thought to be greater in physiological conditions,3 5 but at least 6 times weaker overall relative to apo-Tf. 1 9 Recent electrophoretic (gel and capillary) determinations of stability confirm the presence of two types of binding sites, with only modestly stronger binding at pH 7.2 (log K s = 7.84-8.08; log K w = 5.81-5.93) but up to 20 vanadyl ions bound per protein molecule.27 HSA-bound V 0 2 + ions are more rapidly oxidized than apo-Tf bound V 0 2 + ions (6.5 ± 1.4 minutes). * 9 The spin Hamiltonian values for strong site bound vanadyl ions are consistent with only histidine binding from the Cu(II) site, with no participation likely from the peptide nitrogens of the N-terminus amine,3 6 although this result has been disputed.27 The longer rotational correlation time of strong site bound V 0 2 + is probably due to a near encapsulation of the paramagnetic ion in the Cu(II) site. A recent report detailed the insulin-enhancing effects of a vanadyl complex of bovine serum albumin ( B S A ) . 3 7 Surprisingly, the VO-BSA complex prolonged the oxidative stability of vanadyl ions during incubation with rat adipocytes. The complex was also shown 36 References begin on page 90 Chapter 2 to activate enzyme systems involved in lipid metabolism in vitro. Thus, this endogenous reaction may have important implications in the action of vanadium insulin-enhancing agents. 2.1.4. Chelated V(IV) Complexes and Serum Proteins This chapter will examine the interactions of two types of vanadium pharmaceuticals (see Chapter 1) with HSA and apo-Tf. Through the use of EPR and difference U V spectroscopy, interactions between inorganic vanadyl sulfate or chelated V(IV) species (BMOV, BEOV) with apo-Tf and HSA will be demonstrated. Due to the potential biological ramifications of these interactions, the products of these reactions were examined in detail. Comparison between free and bound source protein adducts indicate some similarities and some differences, depending on the protein. The biological implications of these interactions are considered, in particular the likely effects on biodistribution and transformation of V(IV) species in vivo. The chapter concludes with a model study of the interaction of imidazoles with B M O V , as a model of both non-specific interactions with endogenous histidine or as a potential binding mode between B M O V and HSA. 2.2. Experimental 2.2.1. Materials Water was distilled (Corning MP-1 Megapure still) and deionized (Barnstead D9802 and D9804 cartridges) prior to use. Apo-transferrin and albumin were obtained from Sigma-Aldrich. Apo-transferrin was 97% iron-free, and obtained as a lyophilized powder (Sigma #T4283). Albumin (98%) was globulin-free, crystallized and lyophilized (Sigma #A8762). Sodium carbonate, vanadyl sulfate trihydrate, sodium chloride, citric acid, imidazole, 1-37 References begin on page 90 Chapter 2 methylimidazole, 1,10-phenanthroline and 2-hydroxyethyl-1,4-piperazine-ethanesulfonic acid (HEPES) were of the highest grade available from Sigma-Aldrich and used as received. Atomic absorption standard (AAS) solutions of vanadium and iron were obtained from Aldrich. Maltol and ethylmaltol were obtained from Pfizer. BMOV-^S and B E O V 1 were synthesized according to established procedures. The structure and purity of the isolated complexes were confirmed by FT-IR, MS (+LSIMS) and EA. 2.2.2. Solution Preparation Due to the strong metal binding properties of the proteins under study, particularly apo-transferrin, glassware was demetallated prior to use. Glassware, sample tubes and storage tubes were placed either in a 1 M HNO3 bath for extended periods, or thoroughly rinsed with Chelex-100 resin in distilled and deionized (dd) H2O. A l l solutions were made from reagent grade materials. A l l bulk dialysis solutions and buffers were stored with ~ 20 mL of Chelex-100 resin (in H 2 0) to remove any adventitious metal ions. A l l pH readings were obtained with a glass combination electrode interfaced with a Fisher potentiometer, previously calibrated against pH 4.0 and pH 7.0 standard buffer solutions (Aldrich). 2.2.3. Protein Purification and Manipulation Both proteins were dialyzed using Sigma cellulose dialysis tubing with a molecular weight exclusion limit of 10 000 Da. Apo-transferrin was dissolved in -10 mLO.l M pH 6.0 citrate buffer and dialyzed exhaustively at 277 K against several charges of citrate buffer, 0.1 M NaC10 4 in 0.05 M HEPES buffer (pH 7.40) to remove protein-bound citrate, and finally against 0.05 M HEPES buffer (pH 7.40) to remove perchlorate anions. Standardization of the protein concentration was carried out by either absorbance at 280 nm (s = 9.23 x 104 M" 38 References begin on page 90 Chapter 2 ' cm" 1 ) 3 9 or by spectrophotometric titration against a carefully prepared 1:3 solution of FeCl3 and nitrilotriacetic acid (at pH 4.0). Albumin samples were dissolved in -10 m L o f 0.01 M 1,10-phenanthroline (pH 6.0), then dialyzed over several days at 277 K against several charges of the 1,10-phenanthroline buffer, followed by an identical NaClCXt solution and finally against 0.05 M HEPES buffer (pH 7.40). Concentration was determined spectrophotometrically (X = 278 nm; s = 4.2 x 104 M" 1 cm"').40 Protein solutions were stored at 277 K in sterilized polythene bottles or serum tubes. 2.2.4. General Procedure for Preparation of Protein Samples Conversion of V(IV) to V(V) species was minimized through the use of anaerobic environments throughout all preparative procedures. Pure Ar (99.8%) was humidified by passage through a gas dispersion tube into a volume of water before introduction into reaction vessels where volumetric amounts were required. A l l degassing of protein solutions was achieved by a gentle flow of Ar over the surface of the solution, to avoid denaturation due to frothing, for 30-40 minutes prior to sample manipulations. Degassing of analytical solutions (e.g. buffers, B M O V , ligand) was completed by Ar sparge through a 20 gauge needle in a septum-sealed vessel, with an outlet needle to ensure a constant Ar flux. For all protein solutions, a small aliquot between 100-500 uL was reserved to determine protein concentration after dilution, for accurate analytical results. For samples prepared for low temperature study, glycerol was added in varying percentages as a cryoprotectant to ensure good glass formation. Glycerol concentration was set at either 33 or 50% by volume. 39 References begin on page 90 Chapter 2 2.2.5. Preparation of Carbonate-Free Apo-transferrin Solution Apo-Tf, in a solution of 0.05 M HEPES buffer (pH 7.4), was placed in a 25 mL three-neck pear flask equipped with a combination glass electrode for pH measurement. The flask ports were sealed with septa and the electrode port fitted with a cored stopper to maintain an anaerobic environment. The solution was acidified, drop-wise by syringe, with 1 M HC1 to pH 3.5 with gentle stirring. Humidified Ar, previously bubbled through a 40% w/v NaOH solution to remove carbon dioxide, was passed over the surface of the solution for 40 minutes. The pH was returned to 7.4 by blowing N H 3 gas over the solution surface. Ammonia, in a carrier Ar gas, was generated by bubbling Ar through a 1:1 solution of 25% w/v NaOH and concentrated aqueous ammonium hydroxide. 2.2.6. Synthesis of (Imidazole)bis(maltolato)oxovanadium(IV) A Schlenk vessel was charged with B M O V (0.360 g, 1.13 mmol) and imidazole (ImH, 0.075 g, 1.10 mmol) and flushed with Ar. To this vessel was added 50 mL of degassed (3 cycles freeze/pump/thaw) CHC1 3 and the solution was stirred magnetically for 25 minutes. Degassed diethyl ether (30 mL) was added anaerobically, resulting in immediate precipitation of a dull brown solid in the reaction solution. The mixture was stirred for 5 minutes, then Schlenk-filtered under Ar. The solid was dried in vacuo over P2O5 for 12 hours. E A calc. (found) for C, 2 H,o0 7 V: C: 46.77 (44.73), H 3.66 (3.72), N 7.27 (7.87). MS (+LSIMS): 386 (M+l), 369 (M-oxo), 318 (M-ImH). IR (cm"1): 3439 (v N . H); 1592, 1553, 1469 (vc=o and v c = c ) ; 961 ( v v = 0 ) . EPR (298 K, CHC1 3): 8 line pattern, giso = 1.9712; ^ i s o = 90.40 x 10"4cm"'. 40 References begin on page 90 Chapter 2 2.2.7. Stability of the (l-Methylimidazole)-BMOV Adduct The general procedure for titration of B M O V with 1-methylimidazole (ImMe) follows. A stock B M O V solution was made using degassed 0.05 M HEPES buffer with 0.16 M NaCl, pH 7.4. An aliquot of the stock solution was diluted to the desired concentration with additional buffer, in Ar-purged flasks. This dilute B M O V solution served as the analytical solution for the titration. A stock solution of ImMe (MW = 82.10 g mol"1, p = 1.03 g mL"1) was prepared in degassed 0.05 M HEPES buffer with 0.16 M NaCl, pH 7.4. This solution was the titrant solution. Starting at zero equivalents of titrant, an sample of the analytical solution was drawn by syringe and injected into an Ar-flushed cuvette, fitted with a septum. The U V absorbance at 275 nm was then measured (an average of 5 readings taken over 90 seconds, at 298 K, maintained by a circulating water bath), and the sample returned, by syringe, to the analytical solution. An aliquot of titrant was added, and the solution allowed to stir at 298 K for 15 minutes, after which time the absorbance was remeasured. This process was repeated until the conclusion of the experiment, typically when a 2 equivalent excess of ImMe over B M O V was reached. Absorbance data were corrected for dilution effects. Modeling of the experimental absorbance data by non-linear regression analysis was carried out using an iterative procedure contained with a locally written SigmaPlot 2001 program. Utilizing mass balance equations, analyte concentrations were evaluated for each experimental point, and multiplied by extinction coefficients to obtain theoretical absorbance values. These points are then used to generate a theoretical data set. The equilibrium equation and mass balance equations governing the system are given in Equations 2.1-2.3. 41 References begin on page 90 Chapter 2 Equation 2.1 K\ = [BMOV - ImMe] [BMOV][ImMe] Equation 2.2 BT = [BMOV] + [BMOV - ImMe] Equation 2.3 IT = [ImMe] + [BMOV-ImMe] (where B T represents the total concentration of BMOV, I T the total concentration of ImMe and [BMOV], [BMOV-ImMe] and [ImMe] the free concentrations of the analytes in the titration solution.) Manipulation of the above equations places all variables in terms of the complex of interest (BMOV-ImMe) and its formation constant Ki , which can be solved using the quadratic formula, as shown in Equation 2.4. The total absorbance A for the titration solution is defined as the sum of the absorbances of all absorbing species in solution: B M O V , ImMe and BMOV-ImMe (Equation 2.5, where £ is the extinction coefficients for the three analytes, respectively). The extinction coefficients for [BMOV] and [ImMe] were previously determined by Beer's Law plots. The concentration of these analytes can be put in terms of B T , IT and [BMOV-ImMe] using the mass balance and equilibrium equations. Iteration of 8j and Ki leads to the Equation 2.4 [BMOV - ImMe] = K\(BT + Ir + l)± V ^ i 2 (BT + IT +1)2 - 4K\2 (BT)(IT) 2K\ Equation 2.5 A = £BMOV[BMOV] + £,mMe[ImMe] + £,[BMOV-ImMe] 42 References begin on page 90 Chapter 2 least-squares fitting of the theoretical absorbance curve to the experimental, and the accurate determination of the stability constant for adduct formation. 2.2.8. Instrumental Parameters 2.2.8.1. Potentiometric Titration Potentiometric measurements of ethylmaltol in the presence and absence of vanadyl ion were made with a Metrohm glass combination electrode, interfaced to either a Fisher ion Analyzer or Accumet 921 meter. The electrode was calibrated daily by titrating a known amount of HC1 with a known concentration of NaOH. A plot of mV (measured) versus pH (calculated) gave a working slope and intercept so that pH could be read directly as -log [H +]. The meter and a Dosimat 665 autoburet (Metrohm) were interfaced with a PC computer such that all titrations were automated. Solutions were titrated in covered, jacketed cells; temperature was regulated at 298 ±0.1 K using a Julabo circulating water bath. Ar, prepurified by passage through 10% w/v NaOH to exclude CO2, was bubbled through the titration solution to maintain an anaerobic environment. Ionic strength of all solutions was maintained at 0.16 M NaCI. Ethylmaltol was checked for purity by elemental analysis prior to titration. Stock vanadyl solutions were prepared by analytical partial dilution of its atomic absorption standard (in 5% HC1, Aldrich) and excess acid determined via titration from pH 2 to 3 via Gran's method.41 Vanadyl concentrations of stock solutions were periodically reconfirmed by titration against potassium permanganate. Sodium hydroxide (~ 0.1 M) solutions were prepared by dilution of saturated NaOH solution (~ 50% w/v) with freshly-boiled dd H 2 0 . 43 References begin on page 90 Chapter 2 Standardization of NaOH solutions was conducted potentiometrically against potassium hydrogen phthalate. Ethylmaltol titrations were conducted between pH 2 to 11, while vanadyl :ethylmaltol experiments were conducted from pH 2 to 9. Typically 80-100 data points were acquired per titration. Ligand-to-metal ratios were varied between 4:1 to 1:1. Protonation and stability constants were determined using the program SUPERQUAD.42 The program operates by formulating separate mass balance equations and calculating the instantaneous concentration of all species for each point in the titration, based on the system model (i.e. stability constants) and the total concentrations of each component. The hydrolysis series for vanadyl was also included in the model. 4 3 2.2.9. Spectrophotometric Titration Due to significant formation of the 1:1 vanadyl:ethylmaltol complex at the starting pH, variable pH spectrophotometric titration was employed to obtain a more accurate determination of log K | . Since the total acid concentration at low pH is high, any error in the amount is reflected in the 1:1 stability constant. Spectrophotometric titration, however, is independent of excess acid. High metal to ligand ratio solutions (-9:1) were titrated between the range 1.8 < pH < 3.8 and the absorbance changes monitored between 300-330 nm. Due to the high M : L ratio, it is assumed only the 1:1 complex is formed and hence absorbance changes are due to the formation of [VO(ema)]+. Evaluation of spectrophotometric data was conducted using a least-squares curve-fitting routine within Igor Pro 3.1, operating on PC workstation. Since the absorbance changes observed in the titration are due entirely to the formation of the 1:1 complex, the 44 References begin on page 90 Chapter 2 extinction coefficient 8 and log K i could be determined by an iterative process. Six wavelengths were monitored and 16 data points were obtained per titration. The experimental absorbance values at each pH value were fitted with calculated curves, the differences between the two minimized by iteration of the values of S and log K i . 2.2.9.1. Electron Paramagnetic Resonance Spectroscopy EPR spectra were acquired on a number of spectrometers. A brief description of each instrument configuration is included. For all instruments, variable temperature (100-298 K) regulation was achieved through the use of a L N 2 flow-through cryostat coupled with a Eurotherm B-VT-2000 controller unit. Most spectra were acquired on a Bruker ECS-106 X -band spectrometer (Department of Chemistry, UBC), equipped with an EIP 562A microwave frequency counter. A Varian E500 gaussmeter or alternatively, a solution of DPPH in toluene was used for magnetic field calibration. Spectra were recorded in a TMoi i cavity. Multifrequency experiments (S- and X-band) were conducted on a Bruker Elexsys E500 instrument (Centre for Magnetic Resonance, University of Queensland, Brisbane, Australia), equipped with a EIP 548B frequency counter and Bruker ER 035M N M R gaussmeter. A flexline resonator and a Bruker Super High Q (SHQ) cavity were used for S-band and X -band experiments, respectively. Finally, some X-band spectra were acquired on a Bruker Elexsys E580 spectrometer (Centre for Magnetic Resonance), equipped with a Bruker SHQ cavity and L N 2 flow-through cryostat. Onboard instruments carried out microwave and magnetic field calibration. A Eurotherm 4131 V T was used for regulation of low temperatures. 45 References begin on page 90 Chapter 2 Spectra were acquired in either 4 mm i.d. quartz tubes (frozen solutions) or 1 mm i.d. quartz capillaries placed in 4 mm tubes (solution), all from Wilmad Glass. Spectra were acquired with 4096 points, with gain and offset optimized for maximum relative signal intensity. Typical acquisition parameters were: modulation frequency 50 kHz (ECS-106) or 100 kHz (E500, E580), modulation amplitude 1-5 G, conversion time 20.48 ms, time constant 5.12 ms, center field 3350 G and sweep width 1200-2000 G. The typical acquisition time per scan was 80.96 seconds. Microwave power was set over a wide range, depending on the overall sensitivity of the cavity and the acquisition temperature. In general, frozen solution or the use of the SHQ cavity required lower power settings < 5 mW. Room temperature spectra could be acquired at 20-50 mWwith no signal distortion. Spectra were processed using WinEPR 1.0, Xepr (1.2) or XeprView 1.2.15. Direct comparison between spectra acquired at different frequency values was achieved by converting the magnetic field data to g-factor by the equation g=(\0h ve)/([3B) where h is Planck's constant, v e the resonance frequency (in MHz), p e the Bohr magneton and B the field value (in G) at each point. 2.2.9.2. Ultraviolet-Visible Spectroscopy Ultraviolet-visible spectroscopy was conducted using a Hewlett-Packard 8543 diode array spectrophotometer, equipped with a flow-through temperature regulation cell interfaced with a Fisher IsoTemp 1016D circulating bath. Difference ultraviolet spectra were obtained using a double beam Shimadzu UV-2100 scanning spectrophotometer equipped with a flow-through sample holder and a Haake circulating water bath. 46 References begin on page 90 Chapter 2 2.2.9.3. Other Instruments Infrared spectra (IR) were obtained as KBr disks in the range 4000-400 cm"1 on a Galaxy Series 5000 FT-IR spectrophotometer. Mass spectra (+ ion detection) were obtained on a Kratos Concept II H32Q instrument for Cs + liquid secondary ion mass spectrometry (LSIMS). Elemental analysis was performed by Mr. Peter Borda of this department, on a Carlo Erba instrument. 2.2.10. Computations Simulation of EPR spectra was carried out using Bruker SimFonia 1.25 or XSophe 1.0.4. Simulations were conducted on a PC or Linux workstation, respectively. Spectra were simulated with axial or slightly rhombic spin Hamiltonian parameters. Lineshapes were simulated by accounting for mrdependent linewidths and, where appropriate, g and A strain equations included in the XSophe package. Solution chemistry simulations were calculated using the program HySS 2.0.44 2.3. Results and Discussion 2.3.1. Stability of Bis(ethylmaltolato)oxovanadium(IV), BEOV 2"F Potentiometric and spectrophotometric titration of solutions of V O and Hema were conducted to determine the stability constants for formation of the 1:1 and 2:1 species in aqueous solution. The stability of these complexes were then compared to those determined for B M O V , as well as estimates of stability previously determined for the interaction of vanadyl ions with the two proteins of interest. Potentiometric titrations were conducted from pH 2 to 9, and were modeled with the vanadyl hydrolysis series included. At no point during 47 References begin on page 90 Chapter 2 the titrations were precipitates observed, and the experimental curves were smooth and regular, without any anomalous points, indicating that production of insoluble hydrolysis products was not a factor. Spectrophotometric titrations were used to accurately determine the log K i constant, which was subsequently used as a fixed value in the fitting of potentiometric data. Through the use of high metal-to-ligand ratios (~ 9:1), the formation of the mono [VO(ema)]+ complex can be observed in the U V spectrum. Complexation to vanadyl ions results in a decrease in absorbance at 275 nm and a broad increase at ~ 317 nm. Absorbance data at 6 wavelengths were evaluated by computer modeling to obtain a log K i value of 8.87(2). A sample U V titration curve and calculated fit to the experimental data is shown in Figure 2.3. The system was adequately modeled with formation of a 1:1 ([VO(ema)+]), 2:1 (VO(ema)2) and a hydroxo complex, [VO(ema)2(OH)]~. The equilibrium reactions and corresponding stability constants are shown in Scheme 2.2. Based on these values, a speciation diagram for the VO-Hema system was calculated (Figure 2.4). Conducting the titrations to higher pH values allowed for the detection of the first hydrolysis species [VO(ema)2(OH)]" and its stability constant log K h , which subsequently has been shown to exist in the B M O V system as well.45 The first hydrolysis product becomes a significant species at pH 7-8 (depending on total metal ion concentration), indicating that at physiological pH, a small amount of BEOV may be in the form of an anionic complex, which may affect its cellular absorption and transport properties. 48 References begin on page 90 Chapter 2 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 PH Figure 2.3: (a) Variable pH U V spectrophotometric titration of VO 2 * (1.205 mM) and Hema (0.1408 mM) at 298 K, 0.16 M NaCI; (b) Experimental (points) and calculated (lines) absorbance values versus pH for determination of log Ki. 49 References begin on page 90 Chapter 2 ema" + H + V 0 2 + + ema" [VO(ema)] + + ema" V 0 2 + + 2 ema" • V O ( e m a ) 2 + "OH -Hema [VO(ema)] + V O ( e m a ) 2 V O ( e m a ) 2 [VO(ema) 2 (OH)]" log K a = 8.53(2) log K-i = 8.87(2) log K 2 = 7.54 (3) log B 2 = 16.41(3) log K h = 7.74(4) Scheme 2.2 100 80 40 H 20 / / VO(ema)2 \ \ [VO(emaf / ^ / 60 -I \ / \ / K I I \ I \ I \ / 2+ \ / \ / / y[VO(ema)2(OH)]" / \ / / \ i \ 1 \ \ \ [VO(OH)3] \ 1 H i - * " 1 (™ 2 3 4 5 6 7 10 PH Figure 2.4: Speciation diagram for the V0 2 + -Hema system ([V0 2 + ] = 0.01 mM; [Hema] 0.02 mM). In comparison to related complexes (Table 2.1), BEOV is slightly more stable than B M O V , although the two systems are, as expected, very similar. Maltol-based complexes are more stable than a number of insulin-enhancing complexes previously studied, however, thermodynamic stability is not a reliable indicator of anti-diabetic efficacy.46 it does seem 50 References begin on page 90 Chapter 2 likely though, that the stability of the bis complex would affect the level of absorption from the GI, into the bloodstream and finally into target tissue (assuming the complex remains intact). Passive diffusion through biological membranes is superior for neutral species. Table 2.1: Comparison of stability constants for insulin-enhancing vanadyl complexes. Complex Constant B M O V BEOV VO(acac)2 VO(pic) 2 L o g K a 8.46(2) 8.53(2) 8.83 5.23 Log K | 8.80(2) 8.87(2) 8.59 6.68 Log p 2 16.31(3) 16.41(3) 16.10 11.99 Conditions 298 K 0.16 M NaCl 298 K 0.16 M NaCl 298 K 0.1 M N a C 1 0 4 298 K 0.1 M N a N 0 3 Reference 38 This work 47 48 acac = acetylacetonate; pic - picolinate. 2.3.2. Use of E P R for Protein Binding Studies Due to the d 1 electron configuration of V(IV) (i.e. V 0 2 + ) , electron paramagnetic resonance (EPR) spectroscopy is a powerful structural method for the determination of coordination geometries and electronic structure. EPR is virtually analogous to nuclear magnetic resonance, in that it utilizes the unpaired spin of an electron, versus a spin-active (i.e. non-zero, non-integer mi) nucleus. The technique relies entirely upon the Zeeman effect, where a particle with a non-zero, non-integer spin, and hence a magnetic field, can orient its magnetic vector either in line with an externally applied magnetic field (low energy state), or 51 References begin on page 90 Chapter 2 against it (high energy). The energy difference between the two levels is proportional to the magnitude of the external field. For EPR, transition energies lie in the microwave range. EPR has a number of advantages over NMR. Due to the stronger magnetic moment of the inherently more sensitive. EPR is routinely applied to a wider variety of systems, including powders, frozen solutions, single crystals and conventional solutions. A major disadvantage is that unpaired electron spin is required for observation of a spectrum, which limits application of the technique to radicals, paramagnetic metal ions and instantaneously generated spin systems. This restriction can be advantageous, however, in situations where diamagnetic or spin-paired impurities are also present, as they are silent in the EPR spectrum. The vanadyl ion is a particularly good paramagnetic ion to study by EPR. Its favourable properties include a lack of accessible excited states that usually enables room temperature experiments, well behaved paramagnetism arising almost entirely from spin angular momentum and an ability to substitute for a wide variety of divalent metal ions in metalloproteins. Consequently, its use as a spin label is well documented. EPR spectra of V 0 2 + species can be described by the use of the spin Hamiltonian equation which includes electron Zeeman and 5 1 V nuclear hyperfme interactions, as shown in Equation 2.6. Where p e is the Bohr magneton, H the applied magnetic field, g the electronic Zeeman interaction tensor, S and I the electronic and nuclear spin operators, respectively, h is Planck's constant and A the electron-nucleus hyperfine interaction tensor. Due to its extensive use as a spin probe, g and A parameters for vanadyl complexes have been electron (resulting in a relative susceptibility 1800 times greater than that of H) EPR is Equation 2.6 52 References begin on page 90 Chapter 2 correlated to many structure types. Vanadyl complexes, usually possessing a square pyramidal structure, are axial systems by EPR. The parallel components (relative to the external magnetic field, generally assigned to the molecular z-axis) of the g and A tensors are most sensitive to changes in the coordination sphere of the vanadyl ion. The interaction of the complexed vanadyl sources B M O V and BEOV with serum proteins is most easily demonstrated with EPR. Due to the large size of the proteins (vide supra), the rotation or tumbling of these macromolecules in solution is slow, approaching the time scale of EPR spectroscopy. If an EPR-active metal ion is bound to the protein, frequently the binding is of sufficient strength that the motion of the metal ion is bound to that of the larger protein. Thus, even at room temperature, the slow tumbling of the metal-protein adduct results in a highly anisotropic spectrum as the magnetic inequivalence of the molecular axes are not averaged. In room temperature spectra of vanadyl complexes, the anisotropy is nearly averaged out, with only mi-dependent line broadening of the EPR signals, due to coalescence of the parallel and perpendicular features. For vanadyl salts such as VOSO4, the use of EPR has an additional advantage; uncomplexed vanadyl ions at physiological pH form an EPR-inactive spin-paired dimer. The only detectable V 0 2 + ions at pH 7.4, therefore, are those that are bound in a coordination complex that protects the metal ion from hydrolysis. 4 9 2.3.3. Interaction of VOSO4 with Apo-transferrin The interaction of an inorganic V(TV) source with apo-Tf can be observed with EPR. Figure 2.5 shows an overlay of EPR spectra of VOSO4 and of apo-transferrin solutions. 53 References begin on page 90 Chapter 2 i 2.5 2^ 4 23 22 2 J 2 U J 7j ?.i g-tactor Figure 2.5: EPR spectra of aqueous (VO)x-transferrin solutions (298 K, pH 7.4, 0.16 M NaCl, 5 mM Na 2 C0 3 ): (dark) V O S 0 4 (0.22 mM), apo-Tf (0.23 mM); (light) V O S 0 4 (0.41 mM), apo-Tf (0.22 mM). The spectra indicate that more than one equivalent can bind to apo-transferrin, in agreement with earlier results of Cannon and Chasteen.50 The binding of vanadyl ions to the two Fe(III) binding sites can also be demonstrated with difference U V spectroscopy (Figure 2.6). This method has been used in several studies of metal ion binding to apo-Tf. Upon binding of the metal ion, deprotonation of tyrosine residues leads to a strong positive peak at 253 nm and a weaker positive peak at 295 nm. The plot of the change in relative extinction coefficient versus added V 0 2 + (Figure 2.6, bottom) shows a virtually straight region to r = 2 (r = ratio of molar equivalents V 0 2 + to apo-Tf), followed by a precipitous decrease in slope to near zero until hydrolysis and formation of a fine precipitate (which adversely affects the accuracy of the U V absorbance values as shown by the final two data points). 54 References begin on page 90 Chapter 2 240 260 280 300 Wavelength (nm) 320 E o CO <1 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Molar Equivalents V 0 2 + Figure 2.6: Difference U V titration of apo-Tf by V 0 2 + ([apo-Tf] = 0.0186 mM, [V0 2 + ] = 1.55 mM, 300 K, pH 7.4, 0.16 M NaCI): (top) difference U V spectra recorded during titration; (bottom) change in absorbance versus molar equivalents of V O , monitored at 253 nm. 55 References begin on page 90 Chapter 2 2.3.4. Interaction of BMOV with Apo-transferrin The interaction of the chelated complexes with apo-Tf is also easily demonstrated with EPR. Figure 2.7 compares the spectra of VOSO4 and B M O V with apo-Tf. 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor Figure 2.7: EPR spectra of VO-transferrin (298 K, pH 7.4, 0.16 M NaCl): (top) [VO z ] = 1.11 mM, [apo-Tf] = 0.29 mM; (middle) [BMOV] = 0.112 mM, [apo-Tf] = 0.134 m M ; (bottom) [BMOV] = 1.12 mM. Again, up to 2 equivalents of B M O V can be bound to apo-Tf. Comparison between the EPR spectra of the reaction solutions containing either B M O V or VOSO4 with apo-Tf shows them to be virtually identical, suggesting that the reaction proceeds similarly for both free and chelated metal sources. The middle spectrum shows no evidence of any remaining B M O V . These results suggest that apo-Tf is a strong binder of V 0 2 + ; due to the high stability of B M O V , apo-Tf must possess a very strong affinity for V 0 2 + ions to be able to demetallate the complex at comparable concentrations. 56 References begin on page 90 Chapter 2 Vanadyl ions have been shown to bind specifically in the Fe(III) sites, and EPR can be used to demonstrate that BMOV-sourced V 0 2 + bind in the same positions. Addition of Fe(III) to a solution of B M O V and apo-Tf causes a change in the observed anisotropic spectrum to the isotropic spectrum of B M O V (Figure 2.8). Thus, Fe(III) displaces V O ions from the protein, regardless of the vanadyl source. The displacement is not quantitative under the conditions of Figure 2.8. This observation is most likely due to the presence of free maltol, which competes effectively for Fe(III) through the formation of a highly stable tris complex (log p 3 = 28.45)51. Secondly, rapid addition of Fe(III) to a solution at pH 7.4 will result in hydrolysis of a portion of the added metal ion, due to the very strong hydrolysis reactions of Fe(III) and subsequent slow conversion to soluble Fe complexes due to the presence of chelating ligands. In Figure 2.8(b), up to 1.8 equivalents of free maltol are present in the solution, assuming apo-Tf causes complete decomposition of B M O V . The free maltol binds either Fe 3 + or V 0 2 + , but forms the more thermodynamically stable complex with Fe(III). Therefore, even after addition of an equivalent amount of Fe(III), only a percentage of the V 0 2 + ions was displaced from the protein. Upon addition of excess Fe(III), virtually all of the vanadyl ions are displaced and form BMOV. 57 References begin on page 90 Chapter 2 [apo-Tf] = 0.213mM; [BMOV] = 0.395 mM i 1 1 1 1 1 1 1 1 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor 2_j_ Figure 2.8: EPR spectra of BMOV-apo-Tf solutions showing displacement of V O ions in (VO)2-transferrin byFe 3 + (298 K, 0.16 M NaCI). In the bloodstream, assuming complexes such as B M O V arrive intact, it is reasonable to expect that free apo-Tf binding sites would rapidly cause complex decomposition, with concomitant free ligand release. Such a reaction serves to highlight the importance of using a non-toxic ligand (such as maltol) in potential vanadium pharmaceuticals. Endogenous chelating agents such as apo-transferrin are fully capable of causing decomposition of an administered vanadyl complex. Ligands such as acetylacetone would therefore be unsuitable as a component of a vanadyl pharmaceutical agent due to exposure to free acetylacetone. Apo-Tf is therefore a viable transport mechanism for both free and bound vanadyl ions. Attempts were also made to study the interaction of B M O V with apo-Tf by difference ultraviolet spectroscopy, based on methods developed by Harris and Pecoraro. ^  While this method was useful for the study of the interaction of VOSO4, the release of free maltolate after reaction with the protein resulted in strong absorbance across the U V range, thereby 58 References begin on page 90 Chapter 2 obscuring the weaker difference peaks observed during VOSO4 experiments. Evaluation of difference absorbance data (As versus molar equivalents B M O V to apo-Tf) yielded a straight line segment, the slope of which correlated quite well to the expected extinction coefficients for maltol at pH 7.4. The spectra presented for both B M O V and VOSO4 with apo-Tf are virtually identical, strongly suggesting that the metal-protein complexes made by both vanadyl sources are the same. This result is inconsistent, however, with data recently reported by other researchers. Crans, Willsky and co-workers suggest that, based on EPR data, the reaction products resulting from the interactions of VOSO4 or B M O V with apo-Tf are different.52 Room temperature EPR solution spectra also demonstrated the strong binding of apo-Tf for B M O V -sourced V 0 2 + ions, but spectra of either vanadyl source with apo-Tf were slightly different, with A0 values varying by approximately 9 G. Frozen solution spectra of solutions prepared by the methods reported in this thesis are identical for both VOSO4 and B M O V . No simulation data or frozen solution spectra were reported, but Willsky et al. concluded that there could be "no doubt" that the two vanadyl-protein adducts were different. Given what is currently known about metal binding by apo-Tf (vide supra), a further investigation into the binding of chelated vanadyl ions was attempted in order to rationalize this result. Potential sources for this difference could be either preferential site loading of apo-Tf, or formation of a ternary maltol-vanadyl-protein complex, with maltolate acting as the synergistic anion. Experiments investigating these possibilities are described in the next sections. 2.3.4.1. Preferential Site Binding Chasteen and co-workers have demonstrated by X and Q-band EPR that the two Fe(III) binding sites are spectroscopicallyinequivalent.50>53-55 Frozen solution EPR 59 References begin on page 90 Chapter 2 spectra of (VO)-transferrin and (VO)rtransferrin showed either two (X-band)50 or three (Q-band)55 pH-dependent resonance sets (A, Bi and B2),53,55 potentially corresponding to the distinct N - and C-terminus binding sites. Chasteen also found that at physiological pH, X -band EPR peaks, particularly the parallel resonances, were split into two resonance sets, corresponding to two different binding sites in the protein. At higher pH (> pH 8), EPR spectra showed formation of a new resonance set out of the A set, possibly indicating a conformational change of the protein in response to the increased pH that affected the metal binding at the original A site. These results were unexpected as the high sequence homology of the two binding sites had been previously determined.1 0 In the late 1970s the accurate determination of the stability of Fe(III) binding to apo-Tf (log K i = 20.7; log K 2 = 19.4; pH 7.4, ambient [HC0 3 "]) 9 was also determined. These values further demonstrated the inequivalency of the two binding sites. This stability difference has been exploited in several studies to produce pure N - or C-terminus Fe-loaded transferrin. Pure C-terminus (Fe)-transferrin is obtained i f the metal ion is presented to the protein as a complex, for example as one equivalent of [bis(nitrilotriacetato)iron(III)]3~, while a mixture of C and N-terminus loaded (Fe)-transferrin is obtained i f an iron salt, such as FeCl 3 is used instead. N-terminus loaded (Fe)-transferrin can be obtained by first loading both sites using any iron source in excess, followed by dialysis of the metalloprotein against pyrophosphate or other chelating agent (Scheme 2.3).56 Thus, chelated Fe sources are preferentially loaded at the C-terminus site, which is potentially inequivalent to the other binding site.53,55 Recent work has shown that for lanthanide cations (M 3 + ) , the C-terminus site showed a degree of selectivity for smaller lanthanides, while the N-terminus site did not possess the same specificity.57 60 References begin on page 90 Chapter 2 NaHC0 3 Dialysis FeCI3 + pH <4.5 pyrophosphate NaHC0 3 [Fe(NTA)2] + pH 7.4 Scheme 2.3: Preferential loading of Fe(III) to the N and C-terminus sites of apo-Tf. 5 o The differences between the VOSO4 and B M O V spectra obtained by Willsky et al. could possibly arise from preferential loading of the chelated source (BMOV) to either binding site, or the possibility that binding of B M O V requires a conformational change in the protein that is not required to bind V 0 2 + . An EPR study was conducted to investigate this possibility. To increase spectral resolution, frozen solution spectra were obtained. Figure 2.9 displays an overlay of three spectra recorded at 130 K, to study the loading pattern of VOSO4 and B M O V to apo-Tf. The similarity of the three spectra is easiest to observe in the low field -7/2|| and -5I2\\ peaks (Figure 2.9). The VOSO4 spectrum (black line) clearly shows two resonance sets that, despite near coincidence at X-band, are resolved. Loading of either one (red line) or two equivalents (blue line) of B M O V results in spectra that are nearly superimposable each other and the VOSO4 spectrum. If either of the preferential binding or conformational change mechanisms were operative, one would expect addition of one equivalent of B M O V to result in observation of only one resonance set in the EPR spectrum, corresponding to the A , Bi or B 2 spectra previously reported, or a new metal-transferrin complex not previously observed for any other metal ion source. Clearly, the single equivalent loads in the same manner as the second; accordingly two distinct resonance sets for both V 0 2 + binding sites are observed. 61 References begin on page 90 Chapter 2 2.40 2.35 2.30 2.25 2.20 g-f actor Figure 2.9: EPR spectra (-7/2B and -5/2,| peaks) of (VO)x-transferrin (T = 130 K , pH 7.40, 0.16 M NaCl). 2.3.4.2. Maltolate as a Synergistic Binding Anion In agreement with Fe binding, vanadyl binding to apo-Tf also requires a synergistic binding anion, such as bicarbonate, or others previously shown to be active in Fe(III) binding (vide supra). The common elements between these anions are: (1) a carboxylic acid functionality; (2) a potential donor atom one or two carbons away from the carboxylate and (3) a negative or dinegative charge state.54 Willsky et al. postulate that different reaction products could be produced if one of the maltolato ligands of B M O V was acting as the synergistic anion, in place of (bi)carbonate.52 Given the commonality between the anions previously determined to act synergistically with apo-Tf, it would seem unlikely that maltolate, a uninegative ion lacking a true carboxylic acid group, would be able to interact with both the positive arginine and threonine residues and simultaneously stabilize the V O 62 References begin on page 90 Chapter 2 ion through coordination. To confirm this hypothesis, the binding capabilities of maltolate were investigated. Figure 2.10 (blue) depicts the room temperature EPR spectrum of a 1.6:1 solution of B M O V and apo-Tf (at pH 7.4), previously decarbonated at pH 3.5 to remove nearly all bicarbonate present in the solution. The spectrum is largely isotropic and identical to the room temperature spectrum of BMOV. A set of weak resonances appears in the spectrum as well; some residual bicarbonate remains in solution. Determination of the peak heights of this resonance set indicates that approximately 10% of the vanadyl ions are bound by apo-Tf due to the presence of remaining HC0 3~ ions. This value is in good agreement with the pH-dependent residual bicarbonate levels observed by Campbell and Chasteen. 54 [BMOV] = 0.228 mM, [apo-Tf] = 0.134 mM » BMOV and apo-tf, decarbonated || after addition of Na^COj |L 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor Figure 2.10: EPR spectra of decarbonated apo-Tf and BMOV before and after addition of N a 2 C 0 3 (T = 298 K, pH 7.4, 0.16 M NaCl). Upon addition of N a 2 C 0 3 (Figure 2.10, black), an anisotropic spectrum corresponding to the previously observed (VO)2-transferrin complex is produced. If maltolate was capable of 63 References begin on page 90 Chapter 2 facilitating V 0 2 + binding to the protein, anisotropic peaks corresponding to a (VO)-transferrin adduct would be more intense. Instead, maltolate cannot act as a synergistic anion, and in the absence of any such anion, the B M O V resonance set is observed. Upon provision of carbonate, protein binding of V 0 2 + takes place immediately and completely; the B M O V resonance set can no longer be seen in Figure 2.10 (black). These observations are in agreement with earlier work, which stated that any synergistic binding anion must possess a carboxylate, which maltolate is lacking.^4 The studies reported in this section, therefore, strongly suggest that the reaction products of either VOSO4 or B M O V with apo-Tf are in fact identical. Neither ternary complex formation nor preferential site loading can account for the observations of Willsky et al. Potential sources of the differences in EPR spectra could be variable g strain due to a lack of cryoprotectant in their sample solutions, or through rapid addition of large amounts of vanadyl sulfate, followed by quick freezing of the solution. Such a procedure would result in hydrolysis of the metal ion, which subsequently slows metal ion uptake by apo-Tf. 2.3.5. Interactions of V O S O 4 with Serum Albumin EPR spectra and U V difference titration clearly demonstrate binding of vanadyl ions to HSA. The EPR spectra shown in Figure 2.11 are consistent with those obtained earlier by Chasteen and Francavilla.35 Difference U V titration confirms the presence of a single strong binding site and at least 5 weaker binding sites on the protein (Figure 2.12). 64 References begin on page 90 Chapter 2 1.9 1.8 1.7 1.6 g-lactar Figure 2.11: EPR spectra of aqueous (VO) x-HSA solutions (T = 298 K, pH 7.4, 0.16 M NaCI): (dark) [V0 2 + ] = 0.459 mM, [HSA] = 0.56 mM; (light) [V0 2 + ] = 2.18 mM, [HSA] 0.52 mM. 12000 -i 10000 8000 4000 A 2000 A Molar equivalents V O Figure 2.12: Difference U V titration of HSA with V 0 2 + ( [V0 2 + ] = 1.495 m M , [HSA] = 4.64 x 10" 2mM, 300 K, pH 7.4, 0.16 M NaCI). 2.3.6. Interactions of B M O V with Serum Albumin The interaction of B M O V with serum albumin does initially appear to result in reaction products slightly different from those with VOSO4. Comparison of the EPR spectra 65 References begin on page 90 Chapter 2 of 1:1 V O S 0 4 : H S A and 1 :6 B M O V : H S A (Figure 2.13) solutions demonstrate that the predominant resonance sets are different from those of VOSO4 and HSA at room temperature. It can be suggested that the B M O V : H S A spectrum, through comparison to the B M O V spectrum, is composed of two sets of resonances, one of which is similar to that of B M O V . This result agrees with results reported by Willsky et al.^ T I I I I I I I I 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor Figure 2.13: EPR spectra of VO-HSA (T = 298 K, pH 7.4, 0.16 M NaCl): (top) [V0 2 + ] = 0.285 mM, [HSA] = 0.274 mM; (middle) [BMOV] = 0.084 mM, [HSA] = 0.521 mM; (bottom) [BMOV] = 1.22 mM. Willsky et al. suggest that perhaps a ternary maltolate-vanadyl-albumin complex is formed but beyond comparison of RT EPR spectra, no analysis of the reaction products is given. More detailed examination of the high and low field parallel peaks from frozen solution EPR 66 References begin on page 90 Chapter 2 spectra, however, reveal that the differences in the EPR spectra are simply due to preferential strong versus weak site binding of B M O V and VOSO4. 2.3.7. Strong Versus Weak Site Binding Due to the weak binding associated with the non-specific binding sites of HSA, it would seem unlikely from a thermodynamic perspective that they could compete for V 0 2 + with the tightly bound maltolate ligand. In their work, Willsky et al. did not take into account the two different types of binding sites (strong and weak) in HSA, despite previous characterization by Chasteen et a/. 35 Figure 2.14 shows overlaid EPR spectra of two solutions of VOSO4 and HSA (r = 4.18 (A) and 0.82 (B) where r = rVOS0 4 ] / [HSA]) , a solution of B M O V and HSA (r = 1.82, (C) and 3.04 (D)) and a solution of B M O V (E). Figure 2.15 shows the corresponding high field parallel peaks (+3/2, +5/2 and +7/2). The low ratio B solution exhibits the resonance set corresponding to, predominately, strong site VO-HSA. Spectrum A displays both strong and weak site VO-HSA. Weak site peaks are broadened, likely due to a number of unresolved peaks coincident on each other. Comparison of these two spectra to C reveals a resonance set in BMOV-HSA that does not correspond to either strong or weak site VO-HSA. The other resonance set does, however, match well to B, or strong site VO-HSA. The first resonance correlates to E, the frozen solution spectrum of B M O V . Lastly, in D, with excess B M O V present in solution, no increase in peak intensity due to weak site bound V O species is observed. 67 References begin on page 90 Chapter 2 i 1 1 1 1 1 1 1 1 1 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor Figure 2.14: F rozen solut ion E P R spectra o f V O S 0 4 or B M O V w i t h H S A (T = 130 K , p H 7.40, 0.16 M N a C l ) . 68 References begin on page 90 Chapter 2 ( 1 1 1 1 1 1 1 1 1 1 1 1.83 1.81 1.79 1.77 1.75 1.73 1.71 1.69 1.67 1.65 1.63 1.61 g-factor Figure 2.15: Frozen solution EPR spectra of VOSO4 or B M O V with HSA: enlargement of high field parallel resonances +3/2, +5/2 and +7/2 from Figure 2.14 (T = 130 K , pH 7.40, 0.16 M NaCl). 69 References begin on page 90 Chapter 2 These observations are further confirmed through the use of S-band EPR at 4.2 GHz. The use of the lower frequency is advantageous because it lowers g strain, a second order effect that causes peak shifts and line broadening.58 S-band EPR spectra, labeled in the same manner as those shown in the X-band spectra, allow for differences to be observed in the perpendicular region as well. The S-band spectrometer, however, due to the use of a flexline resonator versus a SHQ resonance cavity for X-band, is not as sensitive and hence the same interpretation cannot be performed on the parallel peaks. The perpendicular region, with comparatively narrower line width, shows similar differences between the frozen solution spectra, confirming the observations in X-band. These differences in the perpendicular resonance sets are not easily observed in X-band. [VOSCg/[HSA] = 0.83 _ • [VOS04]/[HSA] = 4.16 [BMOV]/[HSA] = 1.83 i 1 1 1 1 1 1 i 1 1 1 1 1 1 1 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 g-factor Figure 2.16: S-band EPR spectra of V O S 0 4 or B M O V with HSA (T = 130 K, pH 7.4, 0.16 M NaCI). 70 References begin on page 90 Chapter 2 i 1 1 1 1 1 1 1 1 1 1 2.40 2.35 2.30 2.25 2.20 2.15 2.10 2.05 2.00 1.95 1.90 g - f a c t o r Figure 2.17: Magnification of low field region of S-band EPR spectra from Figure 2.16 (T = 130 K, pH7.4, 0.16 M NaCl). These results demonstrate that BMOV-sourced vanadyl ions can only be bound at the strong binding site of HSA, and do not interact appreciably with the weak non-specific sites. Samples containing vanadyl sulfate, with no competing chelating agent present, exhibit coordination to the weak sites. This observation is corroborated by difference UV titration of HSA with BMOV (Figure 2.18). Despite the strong absorbance of maltol species in solution, a clear biphasic plot is obtained, with the inflection point occurring close to 1 equivalent of BMOV. Beyond this inflection point, the slope is proportional to the extinction coefficient of BMOV at 253 nm. 71 References begin on page 90 Chapter 2 250 300 350 Wavelength (nm) E o 03 < 400 0.5 1.0 1.5 2.0 Molar Equivalents BMOV 2.5 Figure 2.18: Difference U V titration of HSA with B M O V (300 K, pH 7.4, 0.16 M NaCI, [BMOV] = 1.467 mM; [HSA] = 0.085 mM). This result offers some support to the formation of a ternary complex; The clear observation of biphasic nature of B M O V binding to HSA is apparent, yet this experiment cannot be 72 References begin on page 90 Chapter 2 reproduced for the binding of B M O V to apo-Tf. Perhaps the release of free ligand that takes place with V 0 2 + binding to apo-Tf obscures the difference U V peaks to such a degree that protein binding cannot be observed. If free ligand release were taking place with HSA binding, one would expect to observe a similar result. Figure 2.18 in fact demonstrates that two distinct binding periods take place during the titration, suggesting that one B M O V molecule interacts with HSA to form a BMOV-HSA adduct. 2.3.7.1. Speciation Analysis of Vanadyl Binding by HSA While the EPR spectra do confirm the observation of Willsky et al. that the reaction products of B M O V and VOSO4 with HSA are different, the correct interpretation is that in fact, the presence of the chelating maltolate ligands imparts site selectivity to the system, as the weak binding sites are not strong enough to abstract the metal ion from the B M O V chelation complex. Consideration of the stability constants of both B M O V and VO-HSA, however, makes it unlikely that the HSA strong site would be able to abstract the metal ion either. The most probable conclusion is the formation of a ternary complex (i.e. one maltolato ligand, one vanadyl ion and one protein). Spectroscopic analyses of previous sections demonstrate that such a complex cannot be observed directly by EPR, due to either a lack of resolution or coincidence of g and A tensors. It is possible to make an argument for the likely existence of a ternary complex for the reaction of B M O V with HSA. . Chasteen and Francavella determined the stability constants for both strong and weak site binding of V 0 2 + cations by H S A ; 3 ^ these were later refined by Purcell et al?^ Using a speciation simulation program, the concentrations of species relative to the total concentration of vanadyl ion were calculated. If a hypothetical solution of one equivalent of vanadyl ion, 20 equivalents of HSA and two equivalents of maltol at pH 7.4 were made, 84% 73 References begin on page 90 Chapter 2 of the total V 0 2 + would be in the form of B M O V (using log K s = 7.97; log K , = 8.80; log (32 = 16.31; log K a = 8.80 as stability constants, initial [V0 2 + ] = 0.01 mM). Interaction of B M O V with HSA is observed, however, at ratios far below 20:1. Another reaction must be taking place. 100 80 ro 60 ro > ro £ 40 A 20 \ / \ / \ / \ / \ / \ / V / B M O V A ma-VO-HSA / \ / \ / \ / \ / \ 10 12 14 16 log p t (Ternary complex) 20 0.5 1.0 1.5 Molar Equivalents Maltol Figure 2.19: Solution thermodynamics simulation of the B M O V - H S A system, using stability constants previously reported;27,38 (top) percent concentration (of VO 2*) of B M O V and [ma-VO-HSA] versus ternary complex stability; (bottom) titration simulation of a 1:1 solution of V 0 2 + and HSA ([V0 2 + ] = [HSA] = 0.01 mM; log p\ = 15.0). 74 References begin on page 90 Chapter 2 If a ternary complex is invoked, as in a reaction detailed in Scheme 2.4, HSA is able to compete with the maltol ligand for coordination to the metal ion at ambient molar ratios. Indeed, Figure 2.19 shows a simulation of the solution chemistry taking place, given the stability constants in Scheme 2.4, a constant pH of 7.4 and a molar ratio of 1:1:2 vanadyl:HSA:maltol. s- VO(ma) 2 P2 +ma K 2 -ma K-2 +HSA, K a [ma-VO(HSA)] [VO(HSA)] -HSA, K. ; Scheme 2.4: Possible equilibria involved in a solution of V 0 2 + , HSA and maltol (Equilibrium constants K i , K 2 , p 2 and K s are taken from previous reports.27,38 K a s s o c , Kdjs and Ptemary are proposed constants). As the strength of ternary complex formation increases (top plot), the formation of a ternary maltol-vanadyl-HSA complex is favoured over B M O V formation. At log p t = 15, formation of the ternary complex and of B M O V are equally favoured. Taking this value as a theoretical stability constant for ternary complex formation, the second plot simulates a titration of a 1:1 solution of V 0 2 + and HSA with maltol. The model predicts that 75 References begin on page 90 Chapter 2 approximately 80% of the total V 0 2 + would be bound at the strong site, with ~ 10% at the weak site. The remainder would be present as hydrolyzed products. The stability constants for VO-HSA complex formation are insufficient to prevent hydrolysis at pH 7.4. As the maltol concentration increases, formation of the ternary complex takes place. At 1 equivalent, however, B M O V formation takes over and begins to abstract the metal ion from the ternary complex. Thus, inclusion of a ternary species with a thermodynamic stability comparable to the log p 2 value for B M O V (16.31)38 provides an explanation of how a B M O V - H S A complex is observed at low ratios, despite the drastically lower stability constants for the interaction of vanadyl ions with HSA. To investigate this hypothesis, a 1:1 solution of V 0 2 + and HSA was titrated with maltol and monitored by EPR, the results of which are shown in Figure 2.20. The base spectrum (B) displays the resonance sets belonging to strong and weak site VO-HSA. As maltol is added, the resonance set for weak site VO-HSA disappears rapidly, but no new resonance set emerges to take its place (C). This observation would be consistent with the formation of a ternary complex that by EPR is coincident with strong site VO-HSA. Beyond this point, however, the data are inconsistent with the ternary complex hypothesis. Further addition of maltol to the solution, even to large excess, does not result in the expected observation of the B M O V resonance set (A). At a 14:1 excess of maltol to V 0 2 + , the only resonance set observed belongs to that of strong site VO-HSA (D). If a ternary ma-VO-HSA complex was forming in solution, excess maltol would eventually abstract the V 0 2 + to form the thermodynamically stable B M O V complex due to the increasing concentration of free maltol (Scheme 2.4). 76 References begin on page 90 Chapter 2 [BMOV] = 1.02 mM i 1 1 1 1 1 1 1 1 1 1 1.82 1.80 1.78 1.76 1.74 1.72 1.70 1.68 1.66 1.64 1.62 g-factor Figure 2.20: Frozen solution EPR spectra from a titration of a 1:1 solution of V 0 2 + and HSA with maltol: (A) frozen solution of B M O V for comparison; (B) initial VO-HSA spectrum; (C) low equivalents of maltol removes V 0 2 + from the weak sites; (D) excess maltol has no effect on the EPR spectrum; (E) 2:1 B M O V to HSA solution for comparison (dashed lines indicate line positions for B M O V ; T = 130 K, pH 7.40, 0.16 M NaCl). A number of explanations can be proposed. Thermodynamic models of the solution chemistry indicate that at least ternary complex formation must take place for any interaction of BMOV-sourced V 0 2 + to take place. Further along this argument, it is possible that the intact B M O V complex reacts with HSA directly, through the vacant sixth coordination position. Such an interaction would not be expected to be highly stable. The second 77 References begin on page 90 Chapter 2 proposal considers the order of addition in preparation of the VO-HSA complex. In previous figures (Figure 2.14 and Figure 2.15), B M O V was added directly to the protein solution, while in this experiment, VOSO4 was used as the vanadyl ion source. A potential kinetic barrier exists for the subsequent binding of maltol to V 0 2 + ions bound at the strong site. Protein folding around the binding site is a possible mechanism by which abstraction of the metal ion by maltol could be inhibited. Previous studies of (Fe)2-transferrin have demonstrated that pure thermodynamic arguments are not necessarily accurate predictors of metalloprotein chemistry with chelating ligands. 1° Several studies have observed slow Fe(III) release from (Fe)2-transferrin to chelators that by thermodynamic arguments should easily abstract the metal ion from the binding sites.59-63 Frequently, the main limiting factor is a slow conformation change of the diferric protein to allow access to the metal ion by the chelating ligand. Typically, saturation kinetics are observed.59-63 2+ Through the use of V O S 0 4 , the protein was presented with an unchelated V O source; removal by or subsequent binding of maltol molecules could be inhibited by the presence of the protein. The addition of maltol does, however, result in a redistribution of bound vanadyl ions from the weak to the strong site. It is likely that these vanadyl ions are bound in some manner by a combination of maltol and protein donors, however such a complex cannot be detected by cw EPR. Potentially, high resolution EPR methods may be able to resolve the resonance set for the proposed maltol-VO-protein species. Application of specialized techniques such as electron nuclear double resonance (ENDOR) or electron spin echo envelope modulation (ESEEM) (see Chapters 3 and 4), combined with l 7 0 labeling of the maltol ligand, could detect the ternary complex predicted to be present by thermodynamic simulation. 78 References begin on page 90 Chapter 2 2.3.7.2. Interaction of Imidazoles with B M O V In the absence of a definitive experiment demonstrating formation of a ternary complex between B M O V and HSA, attention turned to probing the possibility by studying a model system. Vanadyl compounds have been previously shown to bind well to imidazole nitrogens, and considering that the Cu(II) binding site of HSA contains a histidine residue, investigating the interaction of an imidazole-containing moiety could shed light on the feasibility of ternary complex formation. Histidine, however, possesses up to four possible binding sites (two nitrogen donors on the amino acid side chain, a carboxylate and a free amine) that would complicate any rational attempt to study its reaction with B M O V . Additionally, within a protein, the carboxylate and free amine functionalities are taken up in formation of peptide bonds to neighbouring amino acids and hence are either unavailable for bonding, or exhibit impaired coordination abilities. The reaction was therefore studied using the imidazole ring compound as a model ligand for histidine coordination (Figure 2.21). Coordination to the vacant sixth position of the B M O V complex would represent a reasonable mechanism by which albumin could interact with an intact B M O V molecule, binding it to the protein resulting in the previously observed anisotropic spectrum, with spectral parameters similar to those from VO 2 * ions bound at the strong site of HSA. Figure 2.21: Structure of L-histidine, with the imidazole ring component highlighted. 79 References begin on page 90 Chapter 2 Attempts were made to synthesize the adduct using anaerobic Schlenk techniques. B M O V and ImH were dissolved in chloroform, a non-donor solvent, to limit competition for the open coordination site of B M O V with ImH. Under these conditions, however, it has been shown that V(IV) complexes, including B M O V , are air-sensitive and rapidly oxidize to a V(V) compound in solution. A l l synthetic procedures were therefore carried out under anaerobic conditions. While these efforts were partly successful, an impure product was obtained. A l l efforts to recrystallize or otherwise purify the complex failed. The solid complex, isolated as a dull brown powder, was observed to be air-stable, but unstable to donor solvents such as methanol or water (i.e. loss of ImH). The difficulty encountered in attempting to recrystallize the product is likely due to cis/trans isomers in the solid state, which have previously hampered similar efforts with pentacoordinate VO(IV) complexes such as B M O V . 3 8 Mass spectra and IR data are consistent with formation of the adduct. A shift in the characteristic V=0 peak from 993 cm"1 for B M O V to 961 cm"1 is strongly supportive for coordination to the position trans to the strong oxo bond. It could not be determined, however, whether this represents a nitrogen (imidazole) or oxygen (maltolate) donor to the trans position. The IR also shows a broad peak at 3439 cm"1 characteristic of an N - H bond stretch. EPR spectra recorded at ambient temperature in aqueous solution do not demonstrate any major differences from the isotropic B M O V spectrum. A very slight contraction in the Als0 constant and a slight increase in the isotropic g value are the only discernable differences. In CHCI3, only a small change in isotropic hyperfine coupling was observed (-90.05 x 10"4 cm"1 for B M O V versus -90.40 x 10"4 cm"1 for the adduct). 80 References begin on page 90 Chapter 2 Figure 2.22: Trans (a) versus cis (b) coordination (relative to V=0) structures of the BMOV-ImH adduct. Thegiso was virtually unchanged (1.9707 for B M O V versus 1.9712 for BMOV-ImH). In general, the binding of an imidazole moiety causes line broadening and a slight change in g and A values compared to the B M O V spectrum. An increase of nearly 2 G at v = 9.6002 GHz was observed in the linewidth at half-height of the integrated -7/2 line. The line broadening could be due to unresolved 1 4 N hyperfine coupling (1= 1, 99.63% abundant) from coordination to an equatorial position (Figure 2.22(b)), as well as the destruction of the near axial symmetry of B M O V in non-donor solvents. These anisotropic effects are of course averaged in the room temperature spectrum, but the addition of a ternary ligand to the complex results in unique x, y and z axes that lead to a greater range of line positions, and hence broader lines, in the solution spectrum. This effect is commonly referred to as g and A strain. It is also possible that an equilibrium between the associated and dissociated species is taking place, but the individual species cannot be resolved in the spectrum. Frozen aqueous solution spectra demonstrate a decrease in the Az coupling constant, although the 81 References begin on page 90 Chapter 2 expected increase in rhombicity is not observed. Spin Hamiltonian parameters for B M O V and BMOV-ImH in aqueous solution and 33% glycerol:water are shown in Table 2.2. Table 2.2: Spin Hamiltonian parameters for B M O V and BMOV-ImH (pH 7.4, 0.16 M NaCI) (from spectra shown in Figure 2.23). Lw complex gy" b Siso Aza A/ A a A- b (-7/2)b B M O V 1.938 .979 1.975 1.9663 -171.0 -62.3 -65.0 -95.5a 30.6 G BMOV-ImH a r n i n n tr 1.944 b r r t *\ f~\ T r .980 1.977 1.967 -164.8 -60.0 -55.5 -92.2 32.4 G T = 298 K ; b T = 130 K The spin Hamiltonian parameters for frozen solutions are similar to those observed for the interaction of pyridine with BMOV.64 i n general, strong donor groups cause a decrease in the observed .4 values: Az below -170 x 10"4 cm"1 and A^y below -60 x 10"4 cm"1 compared to the B M O V spectrum. The decrease in the A\ hyperfine coupling constant, together with an increase in the gz value, is consistent with the addition of a nitrogen donor in the first coordination sphere.49 Only slight rhombicity is observed, however, which suggests coordination of the imidazole trans to the oxo with the increase in linewidth due to increased g and A strain. 64 82 References begin on page 90 Chapter 2 i 1 1 1 1 1 1 1 1 1 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 g-factor Figure 2.23: Frozen solution EPR spectra of B M O V (top) and a 2:1 solution of ImH: B M O V (middle), along with its simulation (bottom) ([BMOV] = 2.63 mM, [ImH] = 5.23 mM, T = 130 K, pH 7.4). While synthetic methods demonstrate that formation of a ternary imidazole-maltolate-vanadyl adduct is possible, the question remained whether this reaction was likely to proceed under physiological conditions. The interaction of 1-methylimidazole (ImMe) with B M O V was studied by UV-vis spectroscopy. Titration results were inexplicably more consistent experiment to experiment using ImMe as the titrating ligand versus ImH. Titration of a solution of B M O V with ImMe resulted in an increase in absorbance around 275 nm, with a broad decrease observed at -325 nm and two clear isosbestic points at 252 and 297 nm. A typical stacked absorbance plot is shown in Figure 2.24. 83 References begin on page 90 Chapter 2 O cz CO J 3 i O w < 2.00 1.75 1.50 1.25 -I 1.00 0.75 H 0.50 0.25 A 0.00 increased [ImMe] increased [ImMe] 250 275 300 325 350 375 Wavelength (nm) 400 425 Figure 2.24: Titration of BMOV (0.122 mM) with ImMe (25.1 mM) monitored by UV-vis spectroscopy, from 0 to 2 molar equivalents ImMe (200 (V0 = 25.5 mL, T = 298 K, pH 7.40, 0.16 M NaCI). Extinction coefficients for BMOV, ImMe and BMOV-ImMe were determined by Beer's Law plots. For the adduct, large molar excesses (> 25) of ImMe were added to solutions of BMOV and the absorbance measured for several different solutions. An average of these values yielded an experimental Ei value (at 275 nm) of 17.3 ± 0.7 x 103 M"1 cm*1 for the 1:1 adduct, compared to a value of 14.3 ± 0.3 x 103 M"1 cm"1 for the extinction coefficient of BMOV. ImMe was found to have negligible (S|m Me < 3 M"1 cm"1) absorbance at 275 nm. 84 References begin on page 90 Chapter 2 Least-squares curve fitting analysis was applied to the absorbance data to obtain a value for the stability constant (log Ki) for adduct formation between ImMe and B M O V . The fit of the theoretical absorbance data to the experimental values is shown in Figure 2.25. 2.0 1.5 J , i 1 1 1 1 0 5e-5 1e-4 1e-4 2e-4 Total Concentration ImMe (M) Figure 2.25: Theoretical (solid line) versus experimental absorbance (data points) values for the determination by UV-vis titration of log K | for the interaction of ImMe with B M O V . The fit to the experimental data was very good (r2 > 0.995) and yielded reasonably accurate log K i values. Some difficulty was encountered in that K i and Si were found to be inversely covariant, leading to a greater error in calculated K ( values than those usually reported for determinations made using this method. To alleviate this problem, 8 i was allowed to vary with a certain constraint limit (typically < 2000 M" 1 cm"1), however, this proved to be unnecessary. The calculated Si value (over 3 determinations) agreed quite well (within 85 References begin on page 90 Chapter 2 experimental error) to the experimental extinction coefficient determined by a Beer's Law plot of excess ImMe with B M O V . The final results for the 3 determinations are reported in Table 2.3. Table 2.3: Log K | and determined Si values for the titration of B M O V with ImMe (T = 298 K, pH = 7.40, 0.16 M NaCl). Trial logK, Calculated Si 1 4.65 16640 2 4.46 18250 3 4.41 15970 Mean 4.51 16950 a 2 0.13 1170 The stability constant for adduct formation between B M O V and ImMe was found to be log K i =4.51 ± 0.13. As noted, the error associated with this determination is likely due to covariance in the two experimental parameters, and lack of greater accuracy in the determination of Si. The magnitude of the stability constant suggests that ternary complex formation between imidazole functionalities and the free coordination position of B M O V and its analogues could be biologically significant. Previous studies of adduct formation between vanadium(IV) complexes have also determined reasonable thermodynamic stability for formation of adduct complexes. The coordinating ability of certain solvents to B M O V was found to be pyridine > water > methanol > dichloromethane.38 Coordination of the sixth 86 References begin on page 90 Chapter 2 position ligand can be either cis or trans to the V=0 bond,64 j , u t f o r strong donors such as pyridine the cis isomer is dominant.64-68 Various ternary V(V) derivatives of B M O V also have the ternary ligand coordinated in the equatorial plane.38,69 For the interaction of pyridine to VO(acac)2 (in benzene), a smaller log K, value of 1.89 at 18°C was found.^O in this study however, the stability constant was determined in H2O, itself a competing ligand for the open coordination position of BMOV. , The coordinating ability of ImMe would therefore likely be greater than that of pyridine. Earlier studies utilized conditions that would greatly favour ternary complex formation: either formation of the adduct in neat solutionis or in a non-coordinating (i.e. non-competing) solvent system.64,70 Under biological conditions, this result suggests that a histidine side-chain can easily replace a water molecule in the sixth coordination site of B M O V , especially considering the rapid exchange of the water molecule in the trans position. In the absence of direct spectroscopic evidence, it is possible that the complex formed between B M O V and HSA could in fact be a ternary complex, with the B M O V molecule tethered to the protein via the coordinating histidine residue in the Cu(II) binding site. Additionally, i f B M O V arrives in target tissue as the intact complex, this interaction could be one method by which BMOV-sourced vanadyl ions become bound by intracellular proteins, via non-specific interactions with dangling histidine side chains accessible on protein surfaces. In order to investigate the potential interaction of B M O V with free amines, attempts were made to study the interaction of 1-aminopropane in water at pH 7.4. Some spectral evidence exists for vanadyl ions interacting with intracellular free amines to form vanadyl -protein complexes.^ Using specialized EPR techniques, Fukui et al. measured the magnitude of superhyperfine coupling to l 4 N donors and found that the coupling was due to 87 References begin on page 90 Chapter 2 free amine nitrogens, likely from the free amine within a peptide, or the -NH2 group of a lysine side chain.71 To investigate this potential in vivo interaction with lysine residues, 1-aminopropane was selected as a model ligand. H 2 N (a) (b) Figure 2.26: Structures of L-lysine (a) and 1-aminopropane (b). The solubility of aminopropane in water, however, was insufficient to obtain consistent U V -vis spectra as the potential ligand formed a finely dispersed precipitate in solution, or possibly a colloidal solution that interfered with accurate spectral measurement. One incomplete data set was obtained which would suggest that 1-aminopropane does react with B M O V , but product characterization and stability constant determination could not be completed. The BMOV-ImH interaction therefore supports the characterization of the B M O V -HSA system as a ternary reaction. Without additional structural data (e.g. X-ray diffraction), however, the composition and coordination structure of the BMOV-imidazole adducts cannot be determined unambiguously. In addition to the cisltrans isomers described previously, it is possible that imidazoles cause dissociation of one of the maltolato ligands, however, fitting of the U V data was not improved by including 1:1:1 or 1:1:2 VO:maltol:ImMe species in the model. Makinen and Brady recently reported electron nuclear double resonance (ENDOR) 88 References begin on page 90 Chapter 2 data on the VO(acac)2:BSA system that suggested that the complex remained intact during reaction with BSA, with an overall 1:1 stoichiometry.72 The results reported in this thesis support such an assignment. The stability of the interaction, assuming it takes place at the free imidazole of the coordinating histidine of Cu(II) site of BSA, would have to be lower than the stability determined in this model study. Perhaps the size of the complex complicates the interaction of its free coordination position with the Cu(II) site, resulting in a decreased stability. Clearly, the stability of the B M O V : H S A adduct cannot be additive (i.e. log p 2 + log K | = 20.81) as EPR spectra show a mixture of B M O V and protein adduct species below 1 equivalent. It seems more likely, therefore, that loss of one maltolato ligand must take place before interaction with the protein can take place. The stability of this adduct must be comparable to the stability of the B M O V complex, as discussed earlier (vide supra). The B M O V : H S A interaction may lose its importance in the biological environment, due to the presence of apo-Tf and low molecular weight chelating agents. Previous competition studies have shown that V 0 2 + ions bind preferentially to apo-Tf over H S A , 1 9 and computer models of serum solutions also support the predominant role of apo-Tf for vanadyl ion transport.23 Current work on the VO-transferrin system suggests a stability constant for binding one V 0 2 + of between 10 1 2 5 to 10 l4,73 although no details on the method and conditions are currently available. It is interesting to note, however, that while V O S O 4 associates preferentially to apo-Tf over HSA in competition studies, the provision of the vanadyl source as a chelate may in fact augment HSA binding to such a degree that that interaction may indeed be dominant. It is clear from the EPR spectra that apo-Tf interacts strongly with B M O V , but ternary complex formation is not likely operative. Whether ternary complex formation between B M O V and HSA would allow for in vivo competition 89 References begin on page 90 Chapter 2 between the two proteins for V 0 2 + coordination and transport remains to be elucidated. Potentially, chromatographic methods could be used to quantify the association of vanadyl ions to either protein to determine which would be the dominant transport mechanism for chelated vanadyl sources in the bloodstream. Such methods were previously used in several studiesJ 9,74,75 Suffice to say, both serum proteins interact with BMOV and BEOV, and likely other chelated V(IV) sources as well. This interaction, particularly with HSA, recently took on a greater importance, as a recent report found that the presence of BS A actually augmented the in vitro activity of VO(acac)2.72 Protein binding of chelated vanadyl sources could in fact hold greater importance than initially realized. 2.4. References 1) Thompson, K. H.; Liboiron, B. D.; Sun, Y.; Bellman, K. D. D.; Setyawati, I. A.; Patrick, B. O.; Karunaratne, V.; Rawji, G.; Wheeler, J.; Sutton, K.; Bhanot, S.; Cassidy, C.; McNeill, J. H.; Yuen, V. G.; Orvig, C. J. Biol. Inorg. Chem. 2002, in press. 2) Setyawati, I. A.; Thompson, K. H.; Sun, Y.; Lyster, D. M. ; Vo, C.; Yuen, V. G.; Battell, M . ; McNeill, J. H.; Ruth, T. J.; Zeisler, S.; Orvig, C. J. Appl. Physiol. 1998, 84, 569. 3) Crans, D. C. J. Inorg. Biochem. 2000, 80, 123. 4) Baran, E. J. J. Inorg. Biochem. 2000, 80, 1. 5) Harford, C ; Sarkar, B. in Handbook of'Metal-Ligand Interactions in Biological Fluids: Bioinorganic Chemistry; Berthon, B., Ed.; Marcel Dekker: New York, 1995; Vol. 1, p 405. 90 References begin on page 90 Chapter 2 6) Rang, H. P.; Dale, M . M . ; Ritter, J. M . Pharmacology; Churchill Livingstone: Edinburgh, 1999. 7) Herve, F.; Urien, S.; Albengres, E.; Duche, J . -C; Tillement, J. Clin. Pharmacokinet. 1994, 26, 44. 8) Colmenarejo, G.; Alvarez-Pedraglio, A. ; Lavandera, J.-L. J. Med. Chem. 2001, 44, 4370. 9) Aisen, P.; Leibman, A. ; Zweier, J. J. Biol. Chem. 1978, 253, 1930. 10) Baker, E. N . Adv. Inorg. Chem. 1994, 41, 389. 11) MacGillivray, R. T. A. ; Mendez, E.; Sinha, S. K.; Sutton, M . R.; Brew, K. Proc. Natl. Acad. Sci. USA 1982, 79, 2504. 12) Taylor, D. M . in Perspectives in Bioinorganic Chemistry; Hay, R. W., Dilworth, J. R. and Nolan, K. B., Eds.; JAI Press: London, 1993; Vol. 2, p 139. 13) Harris, W. R.; Sheldon, J. Inorg. Chem. 1990, 29, 119. 14) Harris, W. R.; Pecoraro, V . L. Biochemistry 1983, 22, 292. 15) Harris, W. R.; Chen, Y . ; Wein, K. Inorg. Chem. 1994, 33, 4991. 16) It is not currently known definitively whether bicarbonate or carbonate serves in this role, however, the current literature suggests carbonate is the synergistic binding anion. 17) Schlabach, M . R.; Bates, G. W. J. Biol. Chem. 1975, 250, 2182. 18) Campbell, R. F.; Chasteen, N . D. J. Biol. Chem. 1977, 252, 5996. 19) Chasteen, N . D.; Grady, J. K.; Holloway, C. E. Inorg. Chem. 1986, 25, 2754. 20) Harris, W. R.; Carrano, C. J. J. Inorg. Biochem. 1984, 22, 201. 21) Harris, W. R.; Friedman, S. B.; Siberman, D. J. Inorg. Biochem. 1984, 20, 157. 22) Chasteen, N . D.; Lord, E. M . ; Thompson, H. J. in Frontiers in Bioinorganic Chemistry; Xavier, A. , Ed.; V C H : Weinheim, Germany, 1985. 91 References begin on page 90 Chapter 2 23) Kiss, T.; Kiss, E.; Garribba, E.; Sakurai, H. J. Inorg. Biochem. 2000, 80, 65. 24) Sarkar, B. Life Chem. Rep. 1983,1, 165. 25) Lau, S.-J.; Sarkar, B. / . Toxicol. Environ. Health 1979, 5, 907. 26) Goumakos, W.; Laussac, J.-P.; Sarkar, B. Biochem. Cell Biol. 1991, 69, 809. 27) Purcell, M . ; Neault, J. F.; Malonga, H.; Arakawa, H. ; Tajmir-Riahi, H. A . Can. J. Chem. 2001, 79, 1415. 28) Lippard, S. J.; Berg, J. M . Principles of Bioinorganic Chemistry; University Science Books: M i l l Valley, CA, 1994. 29) Lippard, S. J. in Bioinorganic Chemistry; Bertini, I., Gray, H. B., Lippard, S. J. and Valentine, J. S., Eds.; University Science Books: Mi l l Valley, CA, 1994, p 508. 30) Sadler, P. J.; Viles, J. H. Inorg. Chem. 1996, 35, 4490. 31) Giroux, E. L.; Duieux, M . ; Schechter, P. J. Bioinorg. Chem. 1976, 5, 211. 32) Foote, J. W.; Delves, H. T. Analyst 1984, 109, 709. 33) Shaw (III), C. F. Comments Inorg. Chem. 1989, 8, 233. 34) N i Dhubhghaill, O. M . ; Sadler, P. J.; Tucker, A. J. J. Am. Chem. Soc. 1992,114, 1118. 35) Chasteen, N . D.; Francavilla, J. J. Phys. Chem. 1976, 80, 867. 36) Chasteen, N . D.; DeKoch, R. J.; Rogers, B. L.; Hanna, M . W. J. Am. Chem. Soc. 1973, 95, 1301. 37) Motoyashiki, T.; Miyake, M . ; Yoshida, A. ; Morita, T.; Ueki, H. Biol. Pharm. Bull. 1999, 22, 780. 38) Caravan, P.; Gelmini, L.; Glover, N . ; Herring, F. G.; L i , H. ; McNeill , J. H. ; Rettig, S. J.; Setyawati, I. A . ; Shuter, E.; Sun, Y . ; Tracey, A . S.; Yuen, V . G.; Orvig, C. J. Am. Chem. Soc. 1995,117, 12759. 92 References begin on page 90 Chapter 2 39) Chasteen, N . D. Coord. Chem. Rev. 1977, 22, 1. 40) Reynolds, J. A. ; Gallagher, J. P.; Steinhardt, J. Biochemistry 1970, 9, 1232. 41) Gran, G. Acta. Chem. Scand. 1950, 4, 559. 42) Gans, P.; Sabatini, A. ; Vacca, A. J. Chem. Soc. Dalton Trans. 1985, 1195. 43) Chasteen, N . D. Struct. Bond. (Berlin) 1983, 53, 105. 44) Alderghi, L.; Gans, P.; Ienco, A. ; Peters, D.; Sabatini, A. ; Vacca, A . Coord. Chem. Rev. 1999, 184, 311. 45) Buglyo, P.; Kiss, E.; Fabian, I.; Kiss, T.; Sanna, D.; Garribba, E.; Micera, G. Inorg. Chim. Acta. 2000, 306, 174. 46) Crans, D. C ; Mahroof-Tahir, M . ; Kermidas, A. D. Mol. Cell. Biochem. 1995, 153, 17. 47) Imura, FL; Suzuki, N . Bull. Chem. Soc. Jpn. 1986, 59, 2779. 48) Duma, T.; Hancock, R. J. Coord. Chem. 1994, 31, 135. 49) Chasteen, N . D. Biol. Mag. Res. 1981, 3, 53-119. 50) Cannon, J. C ; Chasteen, N . D. Biochemistry 1975,14, 4573. 51) Gerard, C ; Hugel, R. J. Chem. Res. 1980, 314. 52) Willsky, G. R.; Goldfine, A . B.; Kostyniak, P. J.; McNeill , J. H. ; Yang, L. Q.; Khan, H. R.; Crans, D. C. J. Inorg. Biochem. 2001, 85, 33. 53) Chasteen, N . D.; White, L. K.; Campbell, R. F. Biochemistry 1977,16, 363. 54) Campbell, R. C ; Chasteen, N . D. J. Biol. Chem. 1977, 252, 5996. 55) White, L. K. ; Chasteen, N . D. Phys. Chem. 1979, 83, 279. 56) Baldwin, D. A. ; DeSousa, D. M . R. Biochem. Biophys. Res. Commun. 1981, 99, 1101. 57) Harris, W. R.; Yang, B.; Abdollahi, S.; Hamada, Y . Inorg. Biochem. 1999, 76, 231. 93 References begin on page 90 Chapter 2 58) Weil, J. A. ; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance. Elementary Theory and Practical Applications; John Wiley and Sons, Inc.: New York, 1994. 59) Baldwin, D. A. Biochim. Biophys. Acta 1980, 623, 183. 60) Cowart, R. E.; Swope, S.; Loh, T. T.; Chasteen, N . D.; Bates, G. W. J. Biol. Chem. 1986, 261, 4607. 61) Kretchmar, S. A. ; Raymond, K. N . Inorg. Chem. 1988, 27, 1436. 62) Harris, W. R.; Bali, P. K. Inorg. Chem. 1988, 27, 2687. 63) Bali, P. K. ; Harris, W. R.; Nesset-Tollefson, D. Inorg. Chem. 1991, 30, 502. 64) Hanson, G. R.; Sun, Y . ; Orvig, C. Inorg. Chem. 1996, 35, 6507. 65) Atherton, N . M ; Gibbon, P. J.; Shohoji, M . C. B. J. Chem. Soc. Dalton Trans. 1982, 2289. 66) Sawant, B. M . ; Shroyer, A. L. W.; Eaton, G. R.; Eaton, S. S. Inorg. Chem. 1982, 21, 1093. 67) Yordanov, N . D.; Zdravkova, M . Polyhedron 1993,12, 635. 68) Da Silva, J. J. R. F.; Wootton, R. J. Chem. Soc, Chem. Commun. 1969, 3175. 69) Sun, Y . ; James, B. R.; Rettig, S. J.; Orvig, C. Inorg. Chem. 1996, 35, 1667. 70) Walker, F. A. ; Carlin, R. L.; Rieger, P. H. J. Chem. Phys. 1966, 45, 4181. 71) Fukui, K. ; Ohya-Nishiguchi, H. ; Nakai, M . ; Sakurai, H.; Kamada, H. FEBS Lett. 1995, 368, 31. 72) Makinen, M . W.; Brady, M . J. J. Biol. Chem. 2002, 277, 12215. 73) Kiss, T., personal communication to C. Orvig. 74) De Cremer, K. ; De Kimpe, J.; Cornells, R. Fresenius J. Anal. Chem. 1999, 363, 519. 94 References begin on page 90 Chapter 2 75) De Cremer, K. ; Cornelis, R.; Strijckmans, K.; Dams, R.; Lameire, N . ; Vanholder, R. J. Chromato. B 2001, 757, 21. 95 References begin on page 90 Chapter 3 Pulsed E P R Studies of Bis(ethylmaltolato)oxovanadium(IV): Biodistribution and In Vivo Coordination Chemistry 3.1. Introduction The previous chapter investigated likely serum interactions of potential vanadium pharmaceutical agents within the bloodstream. It has been shown through the use of EPR and difference U V spectroscopies that two common serum proteins, apo-transferrin and albumin, are likely to be significant factors in the transport and biotransformation of chelated vanadium sources in the bloodstream. These studies also confirmed the first detectable reactivity difference between bis(maltolato)oxovanadium(IV) and its inorganic congener vanadyl sulfate with biological metal-chelating agents under in vivo like conditions. Whether this reactivity difference is a factor in the increased activity of chelated vanadium sources cannot be determined from these studies, although a recent study of VO(acac)2 and bovine serum albumin (BSA) interactions observed an augmented insulin-enhancing effect for the VO(acac) 2-BSA adduct.1 The previous chapter therefore further developed the current understanding of one part of the vanadium metabolic model, that of transport to various body tissues. The uptake and storage of vanadium species within specific tissues is of considerable interest both for the development of a complete metabolic model and for detection of an active species ultimately responsible for insulin enhancement. The biodistribution and organ accumulation of exogenous vanadium species have been widely studied, particularly after the discovery of vanadium's insulin-enhancing effects. Bone, kidney and liver tissues appear to be the 96 References begin on page 137 Chapter 3 primary sites of vanadium accumulation, irrespective of administered oxidation state and administration route.2-6 These studies, however, utilized 4 8 V radioactive labeling that is insensitive to both the coordination and the oxidation state. Thus, important information regarding the chemical form of accumulated vanadium is lacking while such knowledge could give some understanding of the structure of the putative active species. Spectroscopic methods, however, can be used to gain insight into solid, solution and in vivo chemical structures. These methods must possess high selectivity to avoid saturation of the spectrum with unwanted features. For detection of V 0 2 + in vivo, the paramagnetism of its complexes can be used to considerable advantage. Due to the strong electron spin magnetic moment (~ 1800 times greater than that of the proton), detection limits in biological samples can reach the micromolar range. Additionally, paramagnetic resonance methods are naturally selective for paramagnetic species only; any interference from competing diamagnetic substances (e.g. V(V) complexes generated from in vivo oxidation of an administered V(IV) complex) is completely obviated. Thus, it would seem clear that EPR methods should be exceedingly useful in the study of the in vivo coordination structure of V(IV) species. 3.1.1. In Vivo Structure Determination Despite the spectroscopic advantages detailed above, application of EPR methods to in vivo systems has been relatively rare. A recent advancement in magnetic resonance imaging using a modified L-band (1.4 - 1.9 GHz) EPR spectrometer was reported by Berliner and Fujii in 1985.7 7 n e L-band has several advantages over the conventional X -band systems in that larger sample sizes can be accommodated, and a dramatic reduction in dielectric signal loss is afforded due to the lower frequency. An in vivo spatial resolution of 97 References begin on page 137 Chapter 3 up to 0.3 mm could be achieved through addition of a paramagnetic imaging agent to the capillary system of two celery stalk samples.7 Electron paramagnetic resonance imaging (EPRI) has since developed into a specialized technique for studying the in vivo distribution of radicals; it finds considerable application in medical science for whole organ and in some cases whole animal studies, often on living subjects. While magnetic resonance imaging (MRI) is superior for obtaining whole organ images, EPRI provides finer detail due to its selectivity for paramagnetic species versus the ubiquitous proton in MRI.8>9 Direct medical application of EPRI techniques however is restricted due to a number of instrumental limitations. 8 A novel application of EPR methods for the in vivo study of insulin-enhancing V(IV) complexes was reported by Sakurai and co-workers. 1 0 Rat subjects were anesthetized and cannulated in such a way as the rat's own circulation pumped blood through a 0.5 mm i.d. length of tubing, and returned to the rat through a second cannulation. The tubing was connected to a flow-through aqueous cell for real-time monitoring of the X-band EPR spectrum. The workers used this technique to track the appearance and elimination of paramagnetic V species from the bloodstream, and calculated pharmacokinetic parameters for the absorption and transport of an orally-administered V(IV) complex. The results concurred with the earlier ^ V - B M O V study in that the V(IV) drug was rapidly eliminated from the bloodstream.6 While the quality of the pharmacokinetic data was comparable to that obtained by other methods,5>6 n 0 real time structural information could be derived. Understandably, the spectra were of low intensity, likely due to the necessity of acquisition at ambient temperatures and a biologically-imposed maximum V concentration in the bloodstream. Additionally, the presence of other paramagnetic metal 98 References begin on page 137 Chapter 3 ions, particularly Mn(II), partially obscured the spectra. These difficulties limit the use of blood circulation monitoring EPR for in vivo structural applications. 3.1.2. Limitations of Electron Paramagnetic Resonance While the previous chapter demonstrated the utility of conventional cw EPR techniques for the detection, quantification and characterization of paramagnetic species in (frozen) solution, the spectra do not yield all the structural detail possible. Contained within the characteristic 8-line 5 1 V-coupled spectrum is a great deal of information regarding the distance, angle and number of donor and neighbouring atoms in the immediate vicinity of the paramagnetic ion. This information is typically lost in the cw spectrum due to inhomogeneous line-broadening. These losses are not limited to the vanadyl cation but can also be observed with other paramagnetic transition metal ions such as Cu(II), low and high spin Fe(III) and Co(II). Inhomogeneous line-broadening is a result of four factors. 11 The first is a consequence of imperfections, or inhomogeneity, in the external magnetic field. In this case, the individual spins experience inequivalent fields due to their instantaneous position in the sample tube, resulting in slightly different resonant frequencies. Secondly, the number of (super)hyperfine components within a particular peak maybe so great that an envelope of all signals is observed. Contained within the broadened peaks, therefore, are additional splittings from spin-bearing nuclei such as 'H , l 4 N or 3 I P . If these splittings could be resolved and quantified, a new source of structural information could be realized. Anisotropic interactions in disordered samples can also contribute to line broadening Anisotropic g and hyperfine interactions lead to slight differences in local magnetic fields, potentially leading to highly unsymmetrical lineshapes. Lastly, dipolar interactions with 99 References begin on page 137 Chapter 3 other fixed paramagnetic centres can result in the imposition of a random local field at the observed paramagnetic centre. 11 Several methods exist for alleviating line broadening effects. Dynamic effects, such as tumbling, can reduce this type of broadening, but for samples in which rotation is restricted (i.e. powder or frozen solution), other methods must be used to overcome these problems. In the case of the vanadyl ion, molecular tumbling is never rapid enough to eliminate all anisotropy, and so mi-dependent linewidths are observed. Isotopic substitution with more favourable nuclei is frequently used to reduce EPR linewidths. 12 The most common example of this procedure is the use of deuterium in the place of hydrogen atoms. The application of other frequency bands can also be advantageous. After X-band, the three most commonly applied microwave frequencies are S, Q and W. S-band, typically around 4 GHz, usually results in an alleviation of g and A strain, which leads to peak narrowing proportional to the weaker external field. Application of Q (35 GHz) or W (95 GHz) band frequencies greatly increases spectral dispersion, spreading overlapping resonances (e.g. from perpendicular and parallel resonance sets) across a greater range. Higher frequencies, however, can result in greatly increased g and A strain i f the molecule under study is of low symmetry. Spectroscopic techniques to extract information from inhomogeneously-broadened EPR spectra are dominated by two experiments: electron nuclear double resonance (ENDOR), both continuous wave and pulsed, and electron spin echo envelope modulation (ESEEM) spectroscopies. Both procedures are capable of determining hyperfine coupling constants contained within peaks of a conventional EPR spectrum, through either direct (ENDOR) or indirect (ESEEM) monitoring of the nuclear transitions. 13 100 References begin on page 137 Chapter 3 3.1.3. Electron Nuclear Double Resonance (ENDOR) Spectroscopy ENDOR is based on the coherent excitation of electron and nuclear transitions through simultaneous application of radio and microwave fields. A schematic diagram depicting the electron and nuclear transitions involved is shown in Figure 3.1. After optimization of an EPR transition signal at a fixed magnetic field value and low power, the signal is saturated through an increase in microwave intensity. A sweep of radio frequencies leads to excitation of forbidden transitions involving simultaneous electron and nuclear "spin flips" when the energy of a nucleus interacting with the electron spin is achieved. The spin populations of the high and low energy states are desaturated, leading to an increase in the observed intensity of the EPR signal. 14 ENDOR spectra are usually presented as radio frequency relative to the Larmor frequency of the nucleus of interest (e.g. -5 to +5 MHz) versus EPR transition intensity (recorded as a function of microwave absorption). ENDOR frequently finds application in the study of paramagnetic metalloproteins,15-18 where its ability to accurately locate coupled protons in the vicinity of the paramagnetic metal center provides complimentary data to single crystal X-ray diffraction studies. Recently, ENDOR was used to differentiate between hydroxyl and water coordination in the Mn(II) binding site of concanavalin A, based on the spectroscopic differences of the protons of the two ligands. 15 101 References begin on page 137 Chapter 3 f- m, = +l/2 -| N M R m = ± 1 / 2 : r m s = +l/2 m, = -l/2 m, = -1/2 >• m s = -l/2 m, = +1/2 A B C Figure 3.1: Schematic diagram depicting the underlying principles of ENDOR spectroscopy, for an S = Vi, I = Vi system. A: Normal allowed EPR transition (m s = ± 1 , mi = 0 ) , where excitation of an electron (solid arrow) leads to rapid relaxation (dashed arrow) back to the ground state. B: Saturation condition for EPR. Application of high power leads to destruction of the Boltzmann distribution. EPR transition intensity decreases as there is not net change in populations of the higher and lower spin states. C: Application of radio frequencies to a saturated EPR system alleviates the condition by excitation of nuclear transitions (labelled NMR). An increase in the EPR signal strength is recorded as a function of the frequency of the applied radio frequency (adapted from Makinen and Mustafi 1^). 3.1.4. Electron Spin Echo Envelope Modulation (ESEEM) Spectroscopy E S E E M techniques differ from ENDOR spectroscopy in that E S E E M is always a pulsed, time-domain experiment, whereas continuous wave ENDOR resides in the frequency domain. Pulsed experiments, upon which E S E E M is based, are a relatively recent development in paramagnetic resonance. The modulation effect observed in E S E E M experiments was first described by Mims and co-workers. ^  Further work developed a 1 0 2 References begin on page 137 Chapter 3 theoretical and empirical library with which to interpret ESEEM spectra and extract useful structural information not previously available in the predominant cw experiment.20-22 T h e development of E S E E M and other pulsed EPR techniques has lagged behind that of N M R due to imposing instrumental requirements.23 While the pulsed EPR instrument is analogous in virtually every manner to the modern N M R spectrometer, the energies and consequent bandwidth required to observe the modulation effect are far greater in EPR. N M R spectrometers excite transitions over a kilohertz range of energies, while for EPR, a 3 G band corresponds to approximately 3 MHz.23 Resonant cavities for cw experiments, being highly tuned instruments of narrow bandwidth, are therefore not suitable for pulsed FT EPR experiments. Related to the bandwidth problem is pulse length. Due to the need to excite all nuclear transitions concurrently over a bandwidth of up to 100 MHz, a very short (and therefore broadband), powerful microwave ( B i ) pulse generator is required. Such instrumentation is much more difficult to produce in the microwave range than for the radiowave domain. The second major factor is timescale. EPR transitions relax at a much faster rate compared to nuclear transitions, on the order of microseconds or lower. N M R relaxation times are typically in the millisecond range, or longer. The instrumentation required to work on these very short timescales has only been recently developed. The combination of these two factors required the manufacture of highly specialized pulsed EPR instruments. Despite the many advances in the field, pulsed EPR instruments are currently limited to a bandwidth of approximately 70 G.23 Fortunately for E S E E M and pulsed ENDOR methods, only a narrow band is required (vide infra). 103 References begin on page J37 Chapter 3 3.1.4.1. Development of Pulsed Paramagnetic Resonance Methods N M R spectroscopy was revolutionized by the application of pulsed techniques, which moved the technique from the frequency domain to the time domain. The advantages of this development were two-fold: 1) continuous wave techniques are obligated to scan the entire magnetic field region, including baseline regions which contain no signals of interest. Pulsed techniques allow for the excitation and observation of all transitions simultaneously, greatly increasing the acquisition rate; 2) development of specialized pulse patterns that allow for the observation of specific nuclei (e.g. through Nuclear Overhauser Enhancement (NOE) spectroscopy) and the correlation of resonances through the application of two-dimensional techniques such as Correlation Spectroscopy (COSY). Since the popularization of pulsed N M R methods, equivalent paramagnetic resonance techniques have been of considerable interest. The application of pulsed FT techniques to the realm of EPR spectroscopy has been an important development in bioinorganic, inorganic and biochemistry. The development of E S E E M and pulsed ENDOR has resulted in the delineation of several bioinorganic processes, mainly due to the high selectivity of these methods and the ability to resolve small hyperfine coupling constants. ESEEM data confirmed the number and relative geometries of coordinating histidine and azide nitrogens in reduced methemerythrin,24 the composition of the Mn(II) coordination site in cytochrome c oxidase of Rhodobacter sphaeroides, 25 and the direct coordination of a purine 1 4 N to Mo(V) in xanthine oxidase.26 ENDOR, on the other hand, has seen wide application in the study of [2Fe-2S] proteins, * 8 nitroxide spin labels of several proteins,^7 and in the accurate description of the solution structure of the pentacoordinate vanadyl ion. 16" 104 References begin on page 137 Chapter 3 E S E E M and ENDOR are highly complementary; the former is superior for determination of weak coupling constants, while ENDOR is better suited for strong coupling. '3 In a recent example, ENDOR provided higher resolution of l 4 N couplings for donor nitrogen atoms in a copper-histidine complex, while E S E E M was capable of determining the much weaker coupling constants of the remote 1 4 N of the imidazole ring.28 Figure 3.2 depicts the spectroscopic realms of the three paramagnetic resonance techniques described in this thesis. It is important to mention, however, that these realms are not exclusive, but indicate the general types of interactions that are (usually) most conveniently studied by the method in question.29 EPR ENDOR ESEEM Figure 3.2: Schematic diagram depicting the effective spectroscopic domain of the three main paramagnetic probes (M = EPR-active transition metal ion) (adapted from Garner et a/29). 3.1.4.2. Theory of E S E E M An E S E E M experiment involves the generation of a spin echo, which is modulated due to interferences between allowed and forbidden transitions of adjacent nuclei. A n analysis of the echo modulation can be conducted to obtain important structural parameters 105 References begin on page 137 Chapter 3 such as hyperfine coupling, g tensors and, in the case of nuclei with I > Vi, quadrupolar couplings. Application of ESEEM methods for the determination of small coupling constants is most easily explained for the simplest possible case, that of an S = 'A, I = Vi system. The spin Hamiltonian equation that describes the system is: Equation3.1 H = p S • g • H - p.,,1 • g„H +1 • A-S e where g (g„) defines the electronic (nuclear) g tensor, //the external magnetic field, S and I the electronic and nuclear spin vector operators, p the Bohr magneton, p n the nuclear magneton and A the electron-nuclear hyperfine tensor. The major question is how E S E E M spectroscopy provides accurate determinations of spin Hamiltonian parameters where conventional cw EPR spectroscopy fails. ESEEM spectroscopy, as mentioned, indirectly detects nuclear transitions via the unpaired electron spin, through simultaneous excitement of both allowed and forbidden transitions. Powerful, short pulses of microwave radiation, applied to a paramagnetic system under an external magnetic field, are used to effect these transitions. An energy level diagram for the S = Vi, I = Vi system is shown in Figure 3.3. Allowed EPR transitions are determined by the selection rule Am s = ±1, Ami = 0. The intense B i pulses excite forbidden transitions in which electrons are excited from one electron spin manifold to another, with a concurrent change in nuclear spin. 106 References begin on page 137 Chapter 3 +1/2 -1/2 -1/2 + 1/2 Figure 3.3: Energy level diagram for a S = Vi spin system coupled to a single I = Vi nucleus. Dashed arrows depict the allowed Am s = ±1 transitions, while solid double headed arrows the nuclear transitions (v a and v p ) between the energy levels of each M s spin manifold that couple with the EPR transitions to create the modulation effect of E S E E M spectroscopy (adapted from Deligiannakis et a/J3). E S E E M spectroscopy is conducted a particular magnetic field value, which usually coincides to an EPR transition observed in the conventional cw spectrum. The bandwidth problem encountered with FT-EPR, therefore, is usually avoided, and pulse times can become (comparatively) longer. Since the energy of the electron Zeeman interaction is much greater than the nuclear interaction, only one electronic (or EPR-allowed) transition is studied in any one experiment. Intense broadband pulses of microwaves excite all transitions concurrently, including the basic nuclear transitions v a and v p . The interactions between the allowed and forbidden (Ami = ±1) transitions generate the modulation effect. For the S = Vi, I = 1 system, 107 References begin on page 137 Chapter 3 the situation is more complicated - an energy level diagram for such a system is depicted in Figure 3.4. The I = 1 nuclear coupling results in three energy levels per electron spin manifold, the energies of which are dependent on ,4, T and nuclear quadrupolar coupling (Q). Transitions that are observed in ESEEM experiments are assigned as either single or double quantum, which refers to the net change in nuclear spin, either ±1 or ±2, both of which are normally forbidden in the cw experiment. The basis of the ESEEM method is the generation of a spin echo. Simple spin echoes are produced through the application of a two-pulse sequence, such as the classic Hahn echo (n/2 -T-JT-T- echo)}® This sequence comprises the pulse pattern of the common two-pulse E S E E M experiment. The amplitude of the echo is measured, and plotted as a function of the interpulse time x. The resulting time-domain plot appears similar to a free induction decay observed in NMR, but the amplitude of the "decay" signal is not smooth but modulated by nuclear transition frequencies excited (through EPR-forbidden transitions) by the strong broadband microwave pulses. Nuclei possessing a spin greater than zero produce a weak magnetic field that is observed by the electron(s) through a change in the local effective field. This interaction causes the spin moment to lag or lead by the frequencies of the nuclear transitions, giving rise to interference that are observed in the echo. 13,31 Purely isotropic interactions, however, do not produce modulations; an anisotropic nuclear component or quadrupolar interaction must be present for the modulation effect to be observed.32 While some interpretations can be made from the time-domain spectrum, Fourier transformation to the frequency domain is most commonly performed.23 The transform results in the deconvolution of the time domain signal into its component frequencies responsible for the echo modulation. 108 References begin on page 137 Chapter 3 /+1/2 i i '~r~ dq sq -1/2 sq Figure 3.4: Energy level diagram for S = Vi spin system coupled with a single I = 1 nucleus: single quantum (sq) transitions are indicated by dashed arrows while double quantum (dq) transitions are shown with solid double headed arrows (note that the energies of the two sq transitions within a spin manifold need not necessarily be equal (adapted from Goldfarb et a / . 3 3 ) ) For the simple two-pulse sequence, addition and difference frequencies of the fundamental transitions may also appear. Various pulse sequences have been devised to augment or suppress the observation of various fundamental and combination frequencies, such as the three-pulse stimulated echo (suppression) and the one-dimensional four-pulse experiment (augmentation) that can assist in the assignment of peaks and subsequent interpretation of the spectrum.34 Careful selection of x delay times can also suppress specific spectral features to allow for observation of weaker signals.3^ Recent developments 109 References begin on page 137 Chapter 3 in the field have seen pulse sequences as high as six, but the two and three-pulse sequences remain the most widely used.3^ 3.1.5. Interpretation of E S E E M Spectra The development of ESEEM spectroscopy has been greatly assisted by theoretical analyses of peak frequencies and lineshapes. Typically, the lineshapes of E S E E M spectra do not permit the determination of isotropic and anisotropic coupling constants through the maxima of the observed signals; numerical simulation is used to obtain these values. The interpretation of E S E E M spectra depends mainly on the magnitude of the isotropic coupling in relation to the nuclear Zeeman frequency, i.e. AiS0 to V ] . * 3 When^j s o = 2vi, a "matching condition" exists which results in dramatic peak narrowing and observation of strong peaks in the frequency domain that are directly proportional to the coupling parameters. For a S = Vi ,1= Vi system, the matching (sometimes called cancellation) condition results in observation of a single highly intense peak for one of the nuclear signals, and a broader, weaker peak for the other. Since the Larmor frequency of the nucleus is known at any particular magnetic field value, the spectroscopic parameters can be easily obtained without any additional numerical analysis. The matching condition can frequently be met by careful choice of the magnetic field, which alters the value of V | . In some instances, however, particularly for 1 4N (I = 1 with strong quadrupolar effects) coupling, the cancellation condition is not active, and so interpretation of the spectrum is based on either the ^ j s o > 2vi or Axso < 2vi cases. Deviation from the cancellation condition has dramatic effects on the observed spectrum. For the ^ i s o > 2vi case, most commonly observed for 5 1 V- I 4 N systems,3^ the two double quantum transitions are prominent but somewhat broadened.38,39 p o r 110 References begin on page 137 Chapter 3 systems in which the principal magnetic axes are coincident, single quantum transitions are not observed. In many systems, however, observation of weak sq transitions and their combinations can be discerned, which indicates some deviation from coincidence of the g, A and T axes.40 One of the first applications of pulsed EPR techniques to a true biological sample was reported by Fukui et al. in 1995.37 wistar rats were treated by intravenous injection of V O S 0 4 for 4 days, after which they were sacrificed. Kidney and liver samples were then recovered and analyzed by cw EPR and 2-pulse ESEEM spectroscopies. For both kidney and liver tissue, peaks identified with 1 4 N coupling and inner sphere protons from coordinated water molecules were observed. After assigning each peak, and processing using first order pertubation theory, the isotropic l 4 N couplings were compared to model VO(histidine) solutions, leading to the conclusion that the paramagnetic vanadyl centers had become ligated by one or more amines in these tissues. While previous studies of in vivo vanadium distribution have been conducted,4-6 little information was available regarding the chemical state, namely the oxidation and coordination states, of the metal ions. This study was the first to provide in vivo structural data on vanadyl ions within tissue samples, possibly indicating a short-term storage form, or 48 possibly a vanadyl-protein complex responsible for insulin enhancement. The V -VOSO4 and 4 8 V - B M O V study reported in 19986 revealed that, at least for BMOV, orally administered, chelated vanadyl sources were preferentially accumulated in bone, liver and kidney. The Fukui study was limited in that it used a first-generation (i.e. inorganic) anti-diabetic agent (VOSO4) coupled with an intravenous route of administration. I l l References begin on page 137 Chapter 3 This chapter will describe pulsed EPR studies of three tissues taken from male Wistar rats after chronic oral administration of BEOV in drinking water. In accordance with earlier biodistribution work,6 bone, liver and kidney samples were removed from treated rats and analyzed by E S E E M and HYSCORE spectroscopies. In addition, preliminary cw EPR measurements were attempted on rat muscle tissue. While the former three represent the major pooling sites for exogeneously added vanadium, muscle tissue represents one likely site of action of the active vanadium species ultimately resulting in amelioration of the diabetic state.41-43 Spectroscopic measurement of vanadyl ions in bone and muscle had never been reported and this work provided the opportunity to compare the hepatic and renal coordination state of vanadyl ions from an administered BEOV or VOSO4 source. This chapter deals with our published E S E E M spectroscopic studies of tissue samples, taken from rats previously chronically treated with BEOV in drinking water.44 Previous application of cw EPR to tissue samples to detect V 0 2 + ions suffered from extensive line broadening, precluding virtually any structural characterization of the in vivo paramagnetic ions.45 3.2. Experimental Ethylmaltol (Pfizer) and vanadyl sulfate, VOS04-3H 20 (Aldrich), were used as received. EPR tubes (4 mm o.d., clear fused quartz) were obtained from Wilmad Glass. B E O V was synthesized according to established procedures and checked for purity by EA, MS (+LSIMS) and FT-IR.46,47 A H analytical instruments were identical to those reported in Chapter 2. 112 References begin on page 137 Chapter 3 3.2.1. Animal Dosing Procedures A l l animal studies were conducted by Ms. Violet Yuen in the laboratories of Prof. John McNeill , Faculty of Pharmaceutical Sciences, UBC, in accordance with procedures and guidelines set out by the Canadian Council on Animal Care. Male Wistar rats (University of British Columbia Animal Care Unit), weighing between 190 and 220 g, were acclimatized for 7 days, then divided into 4 groups: control (CC), control-treated (CT), diabetic control (DC) and diabetic treated (DT). Rats were rendered diabetic by injection of streptozotocin via the tail vein. Animals were kept in polycarbonate cages and allowed ad libitum access to food (Purina rat chow, Ralston Purina, St. Louis, MO) and tap water. The room was kept at 22-25 °C in a 12:12 h lightdark cycle. For treated animals (CT and DT groups), B E O V was dissolved in the drinking water, resulting in a typical dose of 0.26-0.29 mmol kg"1 day"1. The dosing was conducted for 6 weeks, at which time the animals were sacrificed by pentobarbital injection (100 mg kg"1). Tissue samples (long bone, muscle, liver and kidney) were recovered and frozen at -20 °C. 3.2.2. Preparation of Spectroscopic Samples Soft tissue samples (liver, muscle and kidney) were cut into small pieces, quickly frozen in LN2 and placed into EPR tubes previously flushed with argon. Bone samples were prepared by rapid freezing of the long bone in LN2 and smashing it with a hammer into pieces small enough to fit into the sample tube. Care was taken to select bone mineral pieces only and not the soft tissue such as marrow. Samples were sealed with parafilm, then re-frozen in dry ice and kept on dry ice or at -20 °C prior to spectroscopic study. 113 References begin on page 137 Chapter 3 3.2.3. Spectroscopic Methods EPR spectra (both cw and pulsed) were obtained with a Bruker Elexsys E580 X-band spectrometer, interfaced with a Bruker ER035M teslameter for field calibration and a built-in Bruker frequency counter for microwave frequency measurement. Liquid He temperature studies were carried out using an Oxford Instruments CF 935 cryostat and an ITC 502 temperature controller. The samples were studied by several different ESEEM methods, the aforementioned two-pulse spin-echo sequence, a three-pulse stimulated echo experiment, and a two-dimensional hyperfine sublevel correlation (HYSCORE) four-pulse sequence. The three-pulse spectrum has the distinct advantage over the two-pulse spin echo experiment in that the modulation depth is much less dependent on the phase memory time. In addition, the spectrum is simplified due to the absence of nuclear sum and difference frequencies. The pulse sequence of 7t/2-x-7i/2-T-7i/2-x-ec/zo leads to formation of a stimulated echo observed time x after the third pulse. In the 2D four-pulse experiment (7r/2-x-7r/2-ti-7t-t2-7t/2-T-echo)4% also called HYSCORE, the intensity of the stimulated echo after the fourth pulse is measured with i2 and t| varied, and x constant. This two-dimensional (2D) set of echo envelopes gives, after complex Fourier transformation, a 2D spectrum with equal resolution in each direction. This procedure differs from the ID version of the four-pulse experiment where the time x between first and second pulses is kept constant and the times ti = 12 = T/2 are increased stepwise.49 The basic advantage of the HYSCORE technique is the creation in 2D spectra of off-diagonal cross-peaks whose coordinates are nuclear frequencies from opposite electron spin manifolds. The cross-peaks significantly simplify the analysis of congested spectra by correlating and spreading out the nuclear frequencies.50 Jn addition, the HYSCORE experiment separates overlapping peaks along a second dimension and 114 References begin on page 137 Chapter 3 enhances the signal-to-noise ratio through the application of a second Fourier transform.48,50 H Y S C O R E spectra are sensitive to the relative signs of the correlated frequencies and are usually presented as two quadrants (+,+) and (+,-)• A nucleus with 1= lA, such as ' H and 3 I P as depicted in Figure 3.3, has two hyperfine frequencies, v a and vp, which may produce one pair of cross-features, (v a, vp) and (vp, v a ) , in the (+,+) quadrant as well as another pair (v a , -Vp) and (vp, -v a) in the (+,-) quadrant. Peaks in the (+,-) quadrant appear primarily for strong hyperfine interactions, i.e. | T + 2A |» 4vi, while the peaks in the (+,+) quadrant appear predominantly for | T + 2A |« 4vi (where A is the isotropic hyperfine coupling, T is the perpendicular component of the axial anisotropic hyperfine tensor, and V| is the nuclear Zeeman frequency). Peaks may appear in both quadrants simultaneously i f the hyperfine and Zeeman couplings are comparable.^l Orientationally-disordered (i.e. powder or frozen solution) spectra of I = Vi nuclei also allow for the visualization of the interdependence between v a and vp belonging to the same orientations through cross-peak contour projection. The analysis of the contour allows for the direct, simultaneous determination of the nuclear isotropic and anisotropic hyperfine coupling constants, a procedure commonly known as contour lineshape analysis.50 3.3. Results and Discussion The tissue samples were studied with three pulse sequences commonly used in E S E E M spectroscopy. Early in the study it was observed that no spectroscopic differences existed between the CT and DT tissue samples, and so all further discussion refers to spectra of tissue taken from CT rats treated with BEOV. The dose administered in this study is much 115 References begin on page 137 Chapter 3 greater than those recently administered to human subjects (of VOSO4, 0.031 mmol kg"1 d"1, for a 60 kg individual).52 3.3.1. Liver and Kidney Field-sweep electron spin echo (FS-ESE) spectra were first acquired. This spectrum is the pulsed EPR equivalent to the scanned conventional cw experiment, and displays the complete V 0 2 + spectrum across a wide magnetic field range. The FS-ESE experiment collects an echo from a two-pulse sequence at various magnetic field values. Fourier transformation of the collected decay patterns yields the usual V 0 2 + spectrum acquired by the cw method. A typical spectrum from rat kidney is shown in Figure 3.5. These spectra, or the equivalent cw experiment, are generally acquired to ascertain resonant peak positions that can be used for E S E E M studies. The spectrum is typical for axial, magnetically dilute vanadyl ions, and has spin Hamiltonian values of §| = 1.945 + 0.005, A\\ = 168 ± 1 x 10"4 cm" \ g j _ = 1.98 ±0.01 and^_L=58± 1 x 10"4cm"'. These values are consistent with both N 2 0 2 or O 4 coordination spheres.53 Not surprisingly, no ligand superhyperfine coupling is observed. Short dotted lines indicate the field positions used for pulsed experiments. The four field positions correspond to the ±7/2\\, -3/2± and the +l/2± peaks. Three-pulse spectra were acquired on the large central, orientationally non-selective my = -1/2 peak (which contains coincident resonances of the parallel and perpendicular sets). Two and three pulse E S E E M spectra were acquired on the kidney and liver samples. The spectra were remarkably similar, suggesting a comparable coordination sphere in the two tissue samples. Spectra were more intense from the kidney sample, therefore analysis will be carried out only on the kidney sample. 116 References begin on page 137 Chapter 3 ! 1 1 1 1 1 1 1 1 2800 3000 3200 3400 3600 3800 4000 4200 Field (G) Figure 3.5: First derivative field sweep ESE spectrum of rat kidney tissue, taken from an animal previously administered BEOV via drinking water; dashed lines indicate field positions used in E S E E M studies (T = 20 K, v = 9.3966 GHz, t = 200 ns). The two-pulse E S E E M spectra at the four field positions are shown in Figure 3.6. Three-pulse E S E E M spectra (Figure 3.7) were recorded at the m v = -1/2 peak only. Acquisition at other field values resulted in a noisy spectrum in which clear peaks could not be readily identified. The dominant features of the two-pulse spectra are the signals from neighbouring protons, in the ranges of 12.0 MHz to 17.5 MHz and 25 M H z to 35 M H z (low field green spectrum to the high field black spectrum, respectively). The more intense, lower frequency peak corresponds to the N M R signals of protons in the vicinity of the paramagnetic ion, but not close enough for an actual hyperfine coupling to be observed. From the fundamental N M R equation, at a field value of 2828 G, the proton signal would be 117 References begin on page 137 Chapter 3 expected at 12.04 MHz, which correlates very well to the observed matrix peak of 12.0 M H z in the low field green spectrum. 4080 G I I I I I I I I I i 0 5 10 15 20 25 30 35 40 45 Frequency (MHz) Figure 3.6: Modulus two-pulse E S E E M spectra of rat kidney, taken from an animal previously treated with BEOV via drinking water (T = 30 K, v = 9.3977 GHz). i 1 1 1 1 i 1 1 1 1 i 1 1 1 1 i 1 1 1 1 i 1 1 1 1 i 0 5 10 15 20 25 F2:[MHz] Figure 3.7: Stacked modulus three-pulse ESEEM spectra of BEOV-treated rat kidney (B = 3333 G, T = 30 K, v = 9.3966 GHz, step = 16 ns). 118 References begin on page 137 Chapter 3 Thus, the E S E E M spectrum shows all N M R signals of the uncoupled protons essentially in the same manner as a spectrum from a 12.0 MHz N M R spectrometer. At this low frequency, no resolution of these peaks can be observed. The second weaker peak in the high frequency region is the sum combination peak of the uncoupled protons, or 2vH. Sum combination peaks arise from either the addition or subtraction of the basic nuclear frequencies in the system. For *H coupling, the two basic frequencies can combine (v a + vp ) or subtract (v a -vp) and are frequently observed as narrow features in a two-pulse and ID four-pulse spectrum. Combination peaks can be useful for analysis due to their narrow peak width and the fact that i f any splitting occurs, it is proportional to the anisotropic coupling constant. Matrix peaks correspond to nuclei with zero isotropic coupling (i.e. no contact interaction, A - 0), therefore any splitting of the combination peaks is due to anisotropic coupling, T.54 This value is directly proportional to the distance separating the paramagnetic ion and the nucleus under study, which for peaks measured at a parallel component of the spectrum is related to the combination peak shift by Equation 3.2.49 Where Ay = |v a + Vp - V||, T\_ is the anisotropic coupling constant for the x,y axes, and 0 is the angle between the principal axis of the electron g and axial hyperfine tensors. The magnitude of the axial anisotropic component can be related to distance by Equation 3.3:13 Equation 3.2 4^ Equation 3.3 hr2 where r is the distance between the paramagnetic species and the coupled nucleus. 119 References begin on page 137 Chapter 3 Of the four spectra shown, only the high field spectrum, corresponding to the +7/2 peak of the field-sweep ESE spectrum, exhibits any splitting of the sum combination peak. In this case, however, A|| for the bottom spectrum (4080 G) of Figure 3.6 is ~ 1.8 MHz. Calculation of T± using Equation 3.2 over a range of 9 values leads to a minimum anisotropic coupling constant of 7.2 MHz at 9 = 45°, which is simply much too large compared to all other previously reported anisotropic 'H coupling constants.54 it is therefore unlikely that the peak observed in the spectrum corresponds to an actual splitting of the 2v H peak. The spectrum has the weakest intensity of the four and so it must be assumed that this combination peak is in fact a noise artifact. Additional support for this conclusion is found in the first (green) spectrum, taken at the opposite parallel peak of the FS-ESE spectrum, which does not exhibit any splitting of the sum combination peak. The most important features of these spectra however are observed below 10 M H z The other peaks at 3.5 and 7.3 MHz correspond to the double quantum transitions of l 4 N nuclei coupled to the vanadium nucleus (vide infra). Changes in orientation appear to make little difference in the appearance of the acquired spectrum, save field-dependent variance in the proton matrix frequencies. This observation is consistent with l 4 N coupling as such interactions are predominately isotropic with only very minor anisotropic components.40 The frequencies of the dq transitions would therefore be insensitive to orientation selection. Lastly, a very weak peak is observed in two of the spectra below 3 MHz. This peak likely arises from a single quantum transition. The peak is sensitive to orientation selection; the observation of sq transitions is dependent on the angle between the principal g axis and the T tensor.32,40 120 References begin on page 137 Chapter 3 The three-pulse spectra shown in Figure 3.7 have the distinct advantages of improved resolution and decreased complexity. These spectral improvements result from the absence of sum and difference combination peaks because of the suppression of these signals from the pulse pattern, and also the independence of three-pulse spectra from phase memorytime limitations. Since sum and difference peaks are absent from the three-pulse experiment, superior resolution of the dq transition peaks is achieved. The three-pulse spectra also exhibit the same weak signal below 3 MHz previously noted, providing evidence that the peak must correspond to an sq transition as it appears in both two and three-pulse experiments. The other expected sq transition is not resolved in either Figure 3.6 or Figure 3.7. Both sets of ID spectra belong to the characteristic group of spectra typically observed for nitrogen-coupled signals in axial V 0 2 + complexes. In this case, the isotropic coupling constant is large compared to the nuclear Zeeman interaction (A > 2vi) and so the cancellation condition is not active. Several experimental and theoretical studies describing the characteristics of these spectra and the abstraction of coupling parameters have been reported.40>55-59 in the condition of A > 2v\ and in systems where the g, A and Q tensors are coaxial, single quantum transitions are very weak, or not observed at all.32,40 j w 0 strong double quantum transitions are expected below 10 MHz, and in the three pulse spectra (Figure 3.7) they appear at 3.5 and 7.3 MHz. Additionally, one sq transition is observed at ~ 1.5 MHz. Assignment of these peaks was confirmed by the two dimensional H Y S C O R E experiment. The HYSCORE spectrum, shown in Figure 3.8, shows two pairs of cross-peaks of different intensity. The intense peaks at (± 7.2, + 3.4) correspond to the strong dq transitions easily observed in the ID experiments. Cross-peaks also appear in the (+,-) 121 References begin on page 137 Chapter 3 quadrant at (± 7.1, + 1.5), corresponding to the weak resonance observed in the two and three pulse spectra at 1.5 MHz. The HYSCORE spectrum, therefore, allows for the resolution of two sq cross-peaks, one of which is not observed in the one-dimensional spectra due to the overlap with the strong dq transition at 7.2 MHz. According to simulation, the second sq V, [MHz]-15 -10 -5 0 5 10 15 Figure 3.8: HYSCORE spectrum of BEOV-treated rat kidney (B = 3333 G, T = 30 K, v = 9.3966 GHz, x = 200 ns, 256 x 256 points, step = 16 ns; sq = single quantum l 4 N cross peak; dq = double quantum l 4 N cross peak; H = cross peaks of 'H). Using previous analyses of S = Vi, I = 1 systems, it is possible to determine accurate coupling parameters from the peak positions of the double quantum transitions (v+, v.), according to Equation 3.4.60 2 2 Equation 3.4 Aiso = — Sv, 122 References begin on page 137 Chapter 3 From this equation, and using the frequencies of the peaks in the three-pulse spectra (B = 3333 G), an isotropic coupling constant of A\so = 4.9 MHz is obtained. The magnitude of this coupling constant is characteristic for direct equatorial coordination of W N , the coupling for which ranges between 4-8 MHz (Table 3.1). This magnitude is significantly different from remote nitrogen coupling, such as that to the remote non-binding l 4 N of an imidazole to VC* 2 + ( ^ i S 0 = 0.3 MHz).61 The value of the isotropic coupling can be compared against those obtained for other nitrogen donor types; several constants are reported in Table 3.1. Amine donors generally result in an isotropic coupling magnitude of ~ 5 MHz, whereas imine donors are higher at approximately 6-7 MHz. The HYSCORE spectrum shows an additional cross-peak at high frequencies, corresponding to *H coupling. The peaks appear at (12.7, 15.9) corresponding to a spread of ~ 3 MHz. Figure 3.3 shows that A0 is proportional to the difference in energies of the v a and vp transitions. The HYSCORE spectrum, therefore, demonstrates a proton coupling of A0 ~ 3 MHz. This coupling constant is strong evidence for the presence of water or hydroxide in the equatorial plane.61,62 The results are consistent with those reported by Fukui et al. for rat kidney and liver samples from rats previously treated with V O S 0 4 (| A\so | = 5.2 and 4.9 MHz, respectively).3 7 Clearly, the coordination state has changed from the BEOV complex at some point post-administration; for the vanadyl ions become bound equatorially by l 4 N . Additionally, chelated vanadyl sources appear to share the same biological fate as the vanadyl aqua ion, at least in kidney and liver tissue. The magnitude suggests equatorial binding by an amine, such as s-amine of a lysine residue or N-terminal amine of a protein. Isotropic 1 4 N coupling constants have been determined in a number of proteins where amine donation to the metal 123 References begin on page 137 Chapter 3 ion has been confirmed by other methods. Lysine coordination to V 0 2 + ions has been observed in pyruvate kinase (| ^ j s o | = 4.9 MHz),55 S-adenosylmethinine synthetase (| A:IS0 | = 4.3-4.8 MHz)59 and FI ATPase of spinach chloroplasts (\Ako | = 4.75 MHz).63,64 L a s t i y ; direct axial coordination by i 4 N can be ruled out, because several studies have demonstrated coupling constants in a much lower range, typically 2-3 MHz. 13 Table 3.1: Comparison of the isotropic coupling constant magnitudes for amine and imine equatorial 1 4 N donors to V 0 2 + by ESEEM spectroscopy. Complex3 1 4 N Donor Type | Aiso |, MHz Reference VO(gly) 2 Amine 5.0-5.1 55,60 V O - N H 3 b Amine 4.7 65 VO(edda) Amine 4.98 60 VO(meox) 2 Imine 6 40 VO(salen) Imine 5.8 60 VO(Himac)2(ImH) Imine 6.5 66 a gly = glycinate; edda = ethylenediaminediacetate; meox = 2-methylquinolin-8-olate; H2salen = A",A^-bis(salicylidene)ethylenediamine; Himac = 4-imidazoleacetic acid; ImH = imidazole. b on silica supported vanadium oxide. The results are also consistent with a more recent report from Fukui et alP' of in vivo vanadyl coordination from a chelated V 0 2 + source. The behaviour of bis(picolinato)oxovanadium(IV) (VO(pic)2), another potential insulin-enhancing complex, was examined in liver and kidney following intraperitoneal injection in rats. The study is of interest considering the structure of the complex, which contains two imine l 4 N donors. 124 References begin on page 137 Chapter 3 Their spectra in both organs, however, led to determinations of | A{so | = 4.9 MHz, in good agreement with results reported in this chapter and much different from the |^ 4jS01 value for VO(pic) 2 of 6.15 MHz.67 in the liver sample, however, some overlap of two sets of resonances was observed, one of which correlated well to that of VO(pic) 2. It would seem likely therefore that some of the original complex was still intact in the tissue. If such a situation was occurring with the BEOV-treated samples in this work, it could potentially explain the decreased signal intensity observed in liver samples. Intact BEOV, with an 0 4 coordination set, would be expected to be predominately ESEEM-silent, due to the extremely low concentration of bound spin active nuclei to couple with the unpaired electron spin ( 1 70, I = 5/2, 0.04% natural abundance). 3.3.2. Bone Bone is of particular interest because it is the major long-term storage site for vanadium.5,6,68 The FS-ESE spectrum was similar to those acquired for liver and kidney samples. The spectrum was typical for isolated V 0 2 + ions in an axial configuration. Simulation of the spectrum yielded spin Hamiltonian values of g| = 1.93 ± 0.01, g± = 1.996 ± 0.005, A\\ = 190 ± 0.5 G and A± = 80 ± 0.5 G. Such values are consistent with vanadyl ions in an O 4 donor set.53 The values are also significantly different from those of BEOV in physiological solution (pH 7.4, 0.16 M NaCl), suggesting that the coordination environment of the complex has changed post-administration. Two and three-pulse experiments were acquired at the four spectral positions used in the liver and kidney studies. Matrix peaks, as observed in the liver and kidney samples, similarly dominated the two-pulse spectra (Figure 3.9). 125 References begin on page 137 Chapter 3 The proton signal and its sum combination harmonic appear in the same range as in previous spectra. A new matrix peak appears in a lower frequency range, identified to be that o f 3 ' P . Its sum combination peak can occasionally be discerned at slightly lower frequency than the primary proton matrix peak (red line: 11.4 MHz versus 14.1 MHz for 'H). Several broad features appear below 5 MHz. The three-pulse E S E E M spectra of a bone sample acquired at the m y = -1/2 peak contain a narrow peak at vp = 5.7 MHz and an intense line at v H = 14.1 M H z equal to the Zeeman frequencies of 3 1 P and ' H , respectively, in an applied magnetic field of 3301 G. Extended peaks starting at ~1.5 MHz with shoulders to ~4 MHz (and suggestions of low intensity peaks between 5 and 10 MHz) were also observed. Spectra acquired on parallel and perpendicular components of the hyperfine structure show the same matrix peaks (the frequency of which varies, of course) in addition to a low frequency peak. The form and intensity of this latter peak vary with the orientation selection. Multinuclear species such as dimers and trimers as well as mixed valent species are not consistent with the observed EPR spectrum and the relaxation time of the echo delay, which are typical for isolated V O complexes. 126 References begin on page 137 Chapter 3 i 1 1 1 1 1 1 1 1 0 5 10 15 20 25 30 35 40 Frequency (MHz) F igure 3.9: Modulus two-pulse ESEEM spectra of BEOV-treated rat bone (T = 30 K , v = 9.3038 GHz). I i i i ' l i i i i I ' ' ' ' I ' ' ' ' I ' ' ' ' i ' ' ' ' I ' 0 5 10 15 20 25 V[MHz] Figure 3.10: Stacked modulus three-pulse ESEEM spectra of bone taken from BEOV-treated animal (B = 3301 G, T = 30 K, v = 9.3038 GHz, step = 16 ns). 127 References begin on page 137 Chapter 3 The broad nature of these new peaks precludes any meaningful interpretation. It is highly likely though that important coupling information is contained within the signal. To increase resolution and assist in the deconvolution of the multiple peaks contributing to the broad signals, the two-dimensional HYSCORE technique was applied. The contour H Y S C O R E spectrum is shown in Figure 3.11. Along the diagonal of the (+,+) quadrant lie the matrix peaks for 3 1 P and ' H observed in the three-pulse spectrum. In addition to the diagonal peaks, three pairs of cross features can be distinguished, designated as CP1, CP2 and CP3. CP1 appears as extended, weak arcs along the diagonal of the (+,-) quadrant, centered at ~7 MHz. The two coordinates of each point of the arc differ by close to 2v P = 11.4 MHz. Two other cross features are located in the (+,+) quadrant, centred symmetrically around the diagonal point v P = 5.72 MHz. CP2 has maxima at (10.3, 1.5) MHz. CP3 overlaps with the high intensity diagonal peak at the time x used in Figure 3.11; however, adjustment of the x delay can be used to suppress the intensity at the diagonal to allow observation of CP3 as a pair of extended ridges (even partially overlapping with CP2) with maxima at (~7.4, -4.3) MHz, as shown in Figure 3.12 (where x = 280 ns). The difference between the frequencies of the cross-peaks leads to their preliminary assignment as three 3 1 P nuclei with estimated hyperfine coupling constants of 14, 9, and 3 MHz, respectively. More detailed analysis of the isotropic and anisotropic hf coupling was achieved through the use of contour lineshape analysis. 128 References begin on page 137 Chapter 3 V , [MHz] Figure 3.11: HYSCORE spectrum of rat bone, measured at m v = -1/2 peak of the FS-ESE spectrum (B = 3315 G, T = 30 K, v = 9.3252 GHz, x = 120 ns, 256 x 256, step 16 ns; 1, 2, 3: cross-peaks arising from 3*P; P: 3*P matrix peak, 5.7 MHz. H: ! H matrix peak, 14.1 MHz). I 1 1 1 1 I r ~ 1 1 ' I 1 1 1 ' I 1 ' 1 1 I ' 1 1 1 I 0.0 2.5 5.0 7.5 10.0 V,[MHz] Figure 3.12: Stacked HYSCORE spectrum of BEOV-treated rat bone depicting peak shapes of cross peaks 2 and 3 from Figure 3.11 after suppression of 3 I P matrix peak (x = 280 ns). 129 References begin on page 137 Chapter 3 3.3.2.1. Contour Lineshape Analysis of Bone HYSCORE Spectra Previous work by Dikanov and Bowman has demonstrated that isotropic and anisotropic coupling constants can be obtained simultaneously through analysis of the HYSCORE peak contour lineshape as the delay time x is varied.69 For frozen or orientationally disordered samples, the contour lineshape, or the outline of the cross peaks on a contour plot, is characteristic of the relative magnitudes of the (an)isotropic coupling constants and the nuclear Zeeman interaction. For I = Vi systems, the relationship between the lineshape and the coupling values has been solved analytically. In this work, the cross peaks provide simultaneous determinations of the isotropic A and anisotropic T couplings of 3 1 P nuclei (spin I = Vi) when plotted v a 2 vs. v p 2 (Figure 3.13). In the figure, the squares of the frequencies of each cross peak (a and (3) are related by straight line segments. Dikanov and Bowman demonstrated that the slope and intercept of the line segments are directly proportional to the hyperfine coupling values, as shown in Equation 3.5 and Equation 3.6, due to the interdependence of the basic nuclear frequencies.69 Equation 3.5 V a = { Q a V 2 p + Ga} 1/2 T + 2A-4vt f 4v2 -A2 +2T 2 -AT Equation 3.6 Qa = and T + 2A + 4v V T + 2A + 4v J 130 References begin on page 137 Chapter 3 350.0 300.0 250.0 200.0 ri " a 1.50.0 > 100.0 50.0 0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 v p 2 M H z 2 Figure 3.13: Contour lineshape analysis plot of squares of frequencies of CP 1-3 from Figure 3.11 and Figure 3.12 for determination of .4 and T. The coordinates (v a , v p ) of arbitrary points along the top ridge forming each peak were measured from HYSCORE spectra recorded with different x values. Echo measurement at several values of x is necessary to avoid spectroscopic blind spots.32 The slopes and intercepts of the contour lineshape plot are given in Table 3.2, and yield two sets of A and T with opposite relative signs for each nucleus, and with the same 12A + T | value. The relative sign of these values cannot be determined in a non-selective experiment, although additional multifrequency experiments would lead to observation of the angular dependence of the relative magnitudes of A and T and hence the signs of the coupling constants. 131 References begin on page 137 Chapter 3 Table 3.2: Parameters derived from contour lineshape analysis of HYSCORE spectra of BEOV-treated rat bone. Cross-peak Q a M H Z 2 ^ ' MHz T, M H z 1 13.22 (0.5) 117.5 (5) -14.25/ 12.38 (0.25) 1.87 (0.01) 2 -6.67 (0.9) 123.1 (5) -9.19/7.72 (0.12) 1.47 (0.5) 3 -1.74(0.12) 86.1 (2) -3.78/2.40(0.35) 1.39(0.02) The determination of these coupling constants allows some conclusions to be made regarding the coordination structure of the vanadyl ions in bone. Two basic interpretations can be proposed. The first is that the vanadyl ions are uniformly coordinated in the bone by three magnetically-inequivalent phosphorus-containing ligands. In the bone, these ligands are highly likely to be from phosphates of the bone mineral phase, accounting for the spin Hamiltonian parameters consistent with 0 4 coordination. 3 I P couplings in bones correlate with the phosphorus couplings observed in V02 +-adenosine phosphate complexes.64 EPR titration has shown a 1:2 stoichiometry of metahligand binding for ADP and ATP, with metal chelation occurring via the phosphate groups.64 EPR spectra of [VO(ADP)2]~ in frozen solution exhibit additional hf structure assigned to four equivalent phosphorus nuclei with a coupling of 18.54 MHz. This result suggests that the coordination of the phosphate groups of two ADP molecules is geometrically equivalent with respect to the metal center. Regular structure is not observed for the ATP complex due to inequivalent couplings with phosphorus. ENDOR spectra of these complexes have shown, however, features assigned to 132 References begin on page 137 Chapter 3 the 3 I P coupling of 20.6 MHz only . 6 4 In contrast, HYSCORE spectra of [VO(ATP) 2] 2" contain cross-peaks from 3 I P with estimated couplings of 14.8 and 9.0 M H z assigned to the (3- and y-phosphorus atoms of ATP.64 The mineral phase of bone is composed of a roughly crystalline calcium phosphate polymer hydroxyapatite (HA), with a unit cell of Caio(P04)6(OH)2.70 Biological H A is characterized by numerous ion vacancies as well as two monolayers of water molecules on the bone surface. It is unclear from the ESEEM spectra whether the vanadyl ions are incorporated into the crystal lattice or are interacting with phosphates and hydroxides on the mineral surface, although recent data suggest a surface interaction only.^l The similar values of maximum phosphorus hf couplings in bones (i.e. -14.25 and -9.19 MHz) to the couplings in [VO(ADP)2]"and [VO(ATP) 2] 2" suggest at least partial coordination of V 0 2 + v i a phosphate oxygens, which leads to the appearance of unpaired spin density on phosphorus atoms. 2 This indicates that the vanadyl interacts with the H A fraction of bone and is located at least between the hydration shell and the bone surface. Consistent with this premise, E S E E M spectra show an intense peak at VH , which splits into two cross-peaks in the H Y S C O R E spectrum with A = ~3 MHz, indicating the presence of weakly coupled protons of a hydrogen-containing ligand such as water or hydroxide.61,62 The interaction of vanadyl ions with the mineral fraction of bone initially generated some concern over its potential use as an anti-diabetic treatment.^ Recent results suggest that short-term exposure to realistic anti-diabetic dose levels of B M O V has no negative effect on bone density or its mechanical properties.73 V 0 2 + ions may in fact have beneficial effects 133 References begin on page 137 Chapter 3 on the bone matrix and are currently under investigation as a potential treatment for senile osteoporosis.74 The E S E E M observations of V02+-phosphate oxygen coordination in bones can be interpreted in two ways. Firstly, all of the complexes could have the same structure, that is, all three phosphorus nuclei belong to ligands from the same complex. The three couplings could potentially be explained via a facial V 0 2 + tris(phosphate) complex, with two other ligands (i.e. water and hydroxyl) present, resulting in the overall octahedral structure shown roughly in Figure 3.14. Figure 3.14: Possible coordination mode of V O ions in bone mineral (m,n = 1 , 2 independently). Alternatively, vanadyl ions form several complexes with different structures but minor variations in g tensor and 5 1 V hf tensor, distinguished by the phosphorus couplings in which case the E S E E M spectra show the superposition of all of them. These proposed structures could be clarified through high-field EPR experiments or E S E E M experiments with samples prepared at different times after B E O V administration (assuming that the complexes do not all form at the same rate). This study represents the first observation of in vivo 3 1 P coupling to the vanadyl ion. It clearly demonstrates that ESEEM, particularly its two-dimensional extension, can be effectively used to examine the structural transformation of paramagnetic vanadium in bone o p H n O I O—P 134 References begin on page 137 Chapter 3 mineral. Complete resolution of the phosphorus couplings and the first separate measurement of the isotropic and anisotropic hyperfine couplings of the coordinated phosphates to V 0 2 + were achieved from two-dimensional spectra. Clearly, the paramagnetic fraction of vanadium present in bones after BEOV administration, as observed for liver and kidney, changes from its initial coordination. The vanadyl interacts with the H A fraction of bones and is at least partially coordinated by phosphate oxygens; the three distinct 3 1 P couplings indicate at least three different V-O-P binding moieties, from one or several complexes. The use of model complexes could assist in the elucidation of the structure(s) of the V02 +-phosphate complex(es), and these studies are presented in the following chapter. 3.3.3. Musc le While only a relatively low percentage of the administered dose of VOSO4 or B M O V accumulates in muscle tissue, the large percentage of total body mass comprised of muscle represents a significant compartment for vanadium accumulation.6 Most importantly, it is this tissue that is thought to be a main site of peripheral insulin-enhancing effect of vanadium complexes.75,76 These properties, however, did not allow for a sufficient concentration of paramagnetic V(IV) ions in the tissue for spectroscopic study. Muscle samples prepared in the same manner as for liver, kidney and bone tissues did not yield any detectable V(IV) signal. Thorough attempts were made to observe a signal. Several tissue samples from a number of rats were studied by cw EPR in an SHQ cavity, at temperatures ranging from 130 to 1.8 K . At 130 K , the signal was predominately due to what appears to be an ascorbate radical; decrease in temperature to 1.8 K led to saturation of this signal with a concurrent 135 References begin on page 137 Chapter 3 increase in intensity of background Cu(II) and Mn(II) signals. As a result, no subsequent pulsed EPR studies were conducted. The difficulties encountered in the muscle study could potentially be alleviated by a number of alterations in animal dosing procedure and sample preparation. Due to the overall low concentration of V(JV) in muscle tissue, all attempts must be made to achieve the highest possible non-lethal level in rat subjects. Potentially, direct intravenous administration, followed by tissue excision approximately 2 hours later, would result in a more readily detectable signal. Several pharmacokinetic studies indicate the rapid clearance of vanadium species from the bloodstream; recovery of muscle samples shortly after this short clearance time would likely yield the highest possible level of V(IV) in tissueAlO Alternatively, intramuscular (im) injection could be used, followed by removal of the treated muscle tissue after a suitable time period. The simplest solution would be to revert back to an in vitro experiment using muscle cell lines. The benefits are, of course, complete control over experiment conditions and dose levels. These procedures, however, have their shortcomings. The gastrointestinal tract is bypassed in both methods and hence any biotransformations or absorption effects (as yet not completely understood for chelated V(IV) drugs) are removed. Additionally, im injection of V(IV) species avoids most of the effects in the bloodstream, which previous work has demonstrated to be significant (see Chapter 2). Thus, many of the biological transformations are lost, limiting the potential use of the study and calling into question the applicability of the results versus studies that utilized oral administration. Excision of tissue, regardless of the administration route, should also be carried out in inert atmosphere, as was done by Chasteen et al., to prevent oxidation of V(IV) to V(V) before freezing of the samples.45 136 References begin on page 137 Chapter 3 For now, however, the spectroscopic characterization of vanadyl ions in muscle tissue remains an interesting and tantalizing problem. The acquisition of structural data for in vivo vanadyl ions at the likely site of peripheral action would provide valuable insight into the mechanism of insulin enhancement. Moreover, the use of paramagnetic resonance methods alleviates current difficulties in delineating the effects of V(IV) vs. V(V) species. The power of this selectivity has been demonstrated for kidney, liver and bone tissue and successful application of these techniques to muscle would also likely yield fruitful data. Such an experiment will likely require modifications in dosage and tissue recovery, and possibly, a decrease in the detection limit for pulsed EPR methods. 3.4. References 1) Makinen, M . W.; Brady, M . J. J. Biol. Chem. 2002, 277, 12215. 2) Sabbioni, E.; Marafante, E.; Alantini, L.; Ubertalli, L.; Birattari, C. Bioinorg. Chem. 1978, 8, 503. 3) Wiegmann, T. B.; Day, H. D.; Patak, R. V . J. Toxicol. Environ. Health 1982,10, 233. 4) Parker, R. D. R.; Sharma, R. P. J. Environ. Pathol. Toxicol. 1978, 2, 235. 5) Al-Bayati, M . A. ; Raabe, O. G.; Giri, S. N . ; Knaak, J. B. J. Am. Coll. Toxicol. 1991,10, 233. 6) Setyawati, I. A. ; Thompson, K. H. ; Yuen, V . G.; Sun, Y. ; Battell, M . ; Lyster, D. M . ; Vo, C.; Ruth, T. J.; Zeisler, S.; McNeill, J. H. ; Orvig, C. Appl. Physiol. 1998, 84, 569. 7) Berliner, L. J.; Fujii, H . Science 1985, 227, 517. 8) Kuppusamy, P.; Chzhan, M . ; Zweier, J. L. / . Magn. Res. B 1995,106, 122. 9) Sato, T.; Oikawa, K. ; Ohya-Nishiguchi, H. ; Kamada, H. Rev. Sci. Instrum. 1997, 68, 2076. 137 References begin on page 137 Chapter 3 10) Yasui, H. ; Takechi, K.; Sakurai, H. J. Inorg. Biochem. 2000, 78, 185. 11) Weil, J. A. ; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications; John Wiley and Sons, Inc.: Toronto, 1994. 12) Albanese, N . F.; Chasteen, N . D. J. Phys. Chem. 1978, 82, 910. 13) Deligiannakis, Y . ; Louloudi, M . ; Hadjiliadis, N . Coord. Chem. Rev. 2000, 204, 1. 14) Makinen, M . W.; Mustafi, D. Metal Ions Biol. Syst. 1995, 31, 89. 15) Carmieli, R.; Manikandan, P.; Kalb (Galboa), A. J.; Goldfarb, D. J. Am. Chem. Soc. 2001,123, 8378. 16) Grant, C. V. ; Cope, W.; Ball, J. A. ; Maresch, G. G.; Gaffney, B. J.; Fink, W.; Britt, R. D. J. Phys. Chem. B 1999,103, 10627. 17) Petersen, J.; Hawkes, T. R.; Lowe, D. J. J. Inorg. Biochem. 2000, 80, 161. 18) Cammack, R.; Gay, E.; Shergill, J. K. Coord. Chem. Rev. 1999,190-192, 1003. 19) Rowan, L. G.; Hahn, E. L.; Mims, W. B. Phys. Rev. 1965, 137, 61. 20) Mims, W. B. in Electron Paramagnetic Resonance; Geshwind, S., Ed.; Plenum Press: New York, 1972, p 263. 21) Mims, W. B.; Peisach, J. Biol. Magn. Reson. 1981, 3, 213. 22) Dikanov, S. A. ; Yudanov, V . F.; Tsvetkov, Y . D. J. Magn. Reson. 1979, 34, 631. 23) Eaton, S. S.; Eaton, G. R. in Vanadium in Biological Systems: Physiology and Biochemistry; Chasteen, N . D., Ed.; Kluwer Academic Publishers: Dordrecht, 1990, p 199. 24) Dikanov, S. A. ; Davydov, R. M . ; Graslund, A. ; Bowman, M . K. Am. Chem. Soc. 1998,120, 6797. 25) Espe, M . P.; Hosier, J. P.; Ferguson-Miller, S.; Babcock, G. T.; McCracken, J. Biochemistry 1995, 34, 7593. 138 References begin on page 137 Chapter 3 26) Lorigan, G. A. ; Britt, R. D.; Kim, J. H. ; Hille, R. Biochim. Biophys. Acta 1994,1185, 284. 27) Makinen, M . W.; Mustafi, D.; Kasa, S. Biol. Mag. Res. 1998,14, 181. 28) Mims, W. B.; Peisach, J. in Advanced EPR: Applications in Biology and Biochemistry; Hoff, A . J., Ed.; Elsevier: Amsterdam, 1989. 29) Garner, C. D.; Collison, D.; Mabbs, F. E. Metal Ions Biol. Syst. 1995, 31, 617. 30) Hahn, E. L. Phys. Rev. 1950, 80, 580. 31) Dikanov, S. A. ; Tsvetkov, Y . D. Electron Spin-Echo Envelope Modulation (ESEEM) Spectroscopy; CRC Press: Boca Raton, FL, 1992. 32) Goldfarb, D. in Specialist Periodical Report on Electron Spin Resonance; Atherton, N . M . , Davies, M . J. and Gilbert, B. C , Eds.; Royal Society of Chemistry: Cambridge, 1996; Vol . 15. 33) Goldfarb, D.; Kofman, V. ; Libman, J.; Shanzer, A. ; Rahmatouline, R.; Van Doorslaer, S.; Schweiger, A . J. Am. Chem. Soc. 1998, 120, 7020. 34) Tyryshkin, A. M . ; Dikanov, S. A. ; Evelo, R. G.; Hoff, A . J. J. Chem. Phys. 1992, 97, 42. 35) Zimmerman, J.-L.; Schneider, B.; Morlet, S.; Amano, T.; Sigalat, C. Spectrochim. Acta A 2000, 56, 285. 36) Song, R.; Zhong, Y . C ; Noble, C. J.; Pilbrow, J. R.; Hutton, D. R. Chem. Phys. Lett. 1995,237,86. 37) Fukui, K. ; Ohya-Nishiguchi, H. ; Nakai, M . ; Sakurai, H.; Kamada, H . FEBS Lett. 1995, 368, 31. 38) LoBrutto, R.; Hamstra, B. J.; Colpas, G. J.; Pecoraro, V . L.; Frasch, W. D. Am. Chem. Soc. 1998,120, 4410. 139 References begin on page 137 Chapter 3 39) Flanagan, H. L.; Gerfen, G. J.; Singel, D. J. J. Chem. Phys. 1988, 88, 20. 40) Reijerse, E. J.; Tyryshkin, A. M . ; Dikanov, S. A. J. Magn. Res. 1998, 131, 295. 41) Rossetti, L.; Laughlin, M . R. J. Clin. Invest. 1989, 84, 892. 42) Okumura, N . ; Shimazu, T. J. Biochem. (Tokyo) 1992,112, 107. 43) Strout, H. V. ; Vicario, P. P.; Biswis, C ; Saperstein, R.; Brady, E. J.; Pilch, P. F.; Berger, J. Endocrinology 1990,126, 2768. 44) Dikanov, S. A. ; Liboiron, B. D.; Thompson, K. FL; Vera, E.; Yuen, V . G.; McNeill , J. FL; Orvig, C. J. Am. Chem. Soc. 1999,121, 11004. 45) Chasteen, N . D.; Lord, E. M . ; Thompson, H. J. in Frontiers of Bioinorganic Chemistry; Xavier, A. , Ed.; V C H : Weinheim, Germany, 1985, p 133. 46) Melchior, M . Ph. D.. Dissertation 1999, University of British Columbia. 47) Thompson, K. H. ; Liboiron, B. D.; Sun, Y . ; Bellman, K. D. D.; Setyawati, I. A . ; Patrick, B. O.; Karunaratne, V. ; Rawji, G.; Wheeler, J.; Sutton, K.; Bhanot, S.; Cassidy, C ; McNeill , J. H. ; Yuen, V . G.; Orvig, C. J. Biol. Inorg. Chem. 2002, in press. 48) Hofer, P.; Grupp, A. ; Nebenfuhr, H. ; Mehring, M . M . Chem. Phys. Lett. 1986,132, 279. 49) Gemperle, C ; Aebli, G.; Schweiger, A. ; Ernst, R. R. J. Magn. Res. 1990, 88, 241. 50) Dikanov, S. A. ; Tyryshkin, A. M . ; Bowman, M . K. J. Magn. Reson. 2000, 144, 228. 51) Dikanov, S. A . in New Advances in Analytical Chemistry; Atta-ur-Rahman, Ed.; Gordon and Breach: Amsterdam, 2000, p 523. 52) Goldfine, A. B.; Patti, M.-E. ; Zuberi, L.; Goldstein, B. J.; LeBlanc, R.; Landaker, E. J.; Jiang, Z. Y . ; Willsky, G. R.; Kahn, R. C. Metabolism 2000, 49, 400. 53) Chasteen, N . D. Biol. Magn. Res. 1981, 3, 53. 54) Tyryshkin, A . M . ; Dikanov, S. A. ; Goldfarb, D. J. Magn. Reson. A 1993, 105, 271. 140 References begin on page 137 Chapter 3 55) Tipton, P. A. ; McCracken, J.; Cornellius, J. B.; Peisach, J. Biochemistry 1988, 28, 5720. 56) Tipton, P. A. ; Peisach, J. Biochemistry 1991, 30, 739. 57) Eaton, S. S.; Dubach, A. ; More, K. M . ; Eaton, G. R.; Thurman, G.; Ambruso, D. R. J. Biol. Chem. 1989, 264, 4776. 58) Fukui, K. ; Ohya-Nishiguchi, H. ; Kamada, H. J. Phys. Chem. 1993, 97, 11858. 59) Zhang, C.; Markham, G. D.; LoBrutto, R. Biochemistry 1993, 32, 9866. 60) Fukui, K. ; Ohya-Nishiguchi, FL; Kamada, H. Inorg. Chem. 1997, 36, 5518. 61) Dikanov, S. A. ; Samoilova, R. I.; Smieja, J. A. ; Bowman, M . K. J. Am. Chem. Soc. 1995, 117, 10579. 62) Atherton, N . M . ; Shackleton, J. F. Mol. Phys. 1980, 39, 1471. 63) Hauseman, A. ; Morgan, L.; LoBrutto, R.; Frash, W. D. Biochemistry 1994, 33, 4910. 64) Buy, C.; Matsui, T.; Rianambininstoa, S.; Sigalat, C.; Girault, G.; Zimmermann, J. L. Biochemistry 1996, 35, 14281. 65) Cosgrove-Larsen, S.; Singel, D. J. J. Phys. Chem. 1992, 96, 9007. 66) Fukui, K. ; Ohya-Nishiguchi, H.; Kamada, H. Inorg. Chem. 1998, 38, 2326. 67) Fukui, K. ; Fujisawa, Y . ; Ohya-Nishiguchi, FL; Kamada, FL; Sakurai, H . J. Inorg. Biochem. 1999, 77, 215. 68) Amano, R.; Enomoto, S.; Nobuta, M . ; Sakamoto, M . ; Tsujioka, R.; Ambe, F. J. Trace Elem. Med. Biol. 1996,10, 145. 69) Dikanov, S. A. ; Bowman, M . K. J. Magn. Reson. A 1995,116, 125. 70) Posner, A . S. Physiol. Rev. 1969, 49, 760. 71) Vega, E. D.; Pedregosa, J. C ; Narda, G. E. J. Phys. Chem. Solids 1999, 60, 759. 72) Mravcova, A. ; Jirova, D.; Janci, FL; Lener, J. Sci. Total Environ. Suppl 1993,1, 663. 141 References begin on page 137 Chapter 3 73) Poucheret, P.; Verma, S.; Grynpas, M . D.; McNeill, J. H. Mol. Cell. Biochem. 1998,188, 73. 74) Cullinan, G. J. 1996, U . S. Patent 5 547 685. 75) Cohen, N . ; Halberstam, M . ; Shlimovich, P.; Chang, C. J.; Shamoon, H.; Rossetti, L. J. Clin. Invest. 1995, 95, 2501. 76) Bhanot, S.; Girn, J.; Poucheret, P.; McNeill, J. H. Mol. Cell Biochem. 1999, 202, 131. 142 References begin on page 137 Chapter 4 Model Studies of In Vivo Vanadyl Coordination in Bone 4.1. Introduction The detection of three distinct 3 1P to vanadyl ion coupling constants in bone samples, described in the previous chapter, represented the first structural description of an important biological vanadium sink. It seems likely that a vanadyl-phosphate complex(es) serves as a long-term storage form for accumulated vanadium. It is conceivable that the action of vanadium pharmaceuticals extended beyond cessation of treatment may be a result of a steady-state release of accumulated vanadium bound in the bone mineral. The data afforded two possible structural interpretations. The first is that a distinct vanadyl-phosphate complex is present in bone mineral, with three magnetically-inequivalent 3 I P nuclei giving rise to the observed couplings. The second possibility is the formation of multiple vanadyl-phosphate complexes, with the E S E E M and HYSCORE spectra showing a superposition of all complexes present in significant concentration. Without additional study, it is impossible to discern between the two possibilities. Model studies are frequently used to provide analogous data applicable to an unknown system. Through comparison of coupling constants obtained for structurally-characterized species, parallels can be drawn to the unknown system. In the realm of E S E E M and HYSCORE spectroscopy, a great deal of effort is put towards the characterization of coupling constants arising from different donor types, i.e. between dissimilar atoms such as 1 4 N and 1 7 0 or between various donor types of the same element, such as imine versus amine nitrogen donors. Such work has resulted in the rapid assignment 143 References begin on page 181 Chapter 4 of coupling constants to specific donor groups in several bioinorganic systems. Such data can also be used to obtain structural insights as well, such as delineation between equatorial and axial nitrogen donation. Several E S E E M and HYSCORE studies have been reported to assist in the structural interpretation of hyperfine and quadrupolar coupling constants. 1~10 Model vanadyl complexes of amine and imine donors, in axial and equatorial positions relative to the V=0 bond, have been studied, resulting in a correlation between the magnitude of the isotropic coupling constant and the type and position of the donor l4N.2>7,10 In this chapter, the complete spectroscopic characterization of two model systems is presented. The interaction of vanadyl ions with phosphate donors is of particular biological importance. Triphosphate chains, most notably adenosine triphosphate (ATP) and other nucleotides, interact with divalent metal ions (typically Mg ) in numerous enzyme systems, but the complete description of the coordination environment around the spectroscopically-silent metal ion has remained elusive. A pulsed EPR study of the triphosphate-vanadyl system, therefore, has the potential to not only provide useful structural insights into the in vivo coordination of vanadyl ions in bone, but also to provide new information regarding the interactions of divalent metal ions with polyphosphate chains. The interaction of vanadium ions (vanadates and vanadyl) with hydroxyapatite has also been studied in some detail.^ 1-14 Concern over vanadium accumulation due to increased environmental exposure has led researchers to examine the effects of vanadium species on bone mineral. 15 Hydroxyapatite is a synthetic calcium phosphate polymer that is considered to be an excellent model for the inorganic phase of bone. 16,17 Due to its similarity to bone mineral, vanadyl binding to H A should prove a suitable reproduction of the in vivo coordination state. Ideally, the model systems should accurately reproduce the P 144 References begin on page 181 Chapter 4 coupling constants observed in the in vivo sample, so that through knowledge of the known system, insight into the in vivo coordination state can be obtained. 4.1.1. The Vanadyl-Triphosphate System Previous investigations into the coordination of triphosphate chains to the vanadyl cation have provided only incomplete descriptions of the metal ion coordination. Two previous characterizations using potentiometry and EPR in aqueous solution drew different conclusions regarding the formation of mono- and bis(triphosphate) species. 18,19 Triphosphate forms relatively strong complexes with vanadyl, protecting the cation from hydrolysis above pH 7.5. Cini et alA% found only monoligand complexes ([(VO)HL] and [(VO)L], where L = P 3 Oi 0 5 ") at 2.5 < pH < 6.5, while Kiss, Micera and co-workersl 9 described formation of two bisligand species ([(VO)(HL)L] and [(VO)(L)2]) at pH > 4, in addition to the two previously proposed mono complexes. The latter group also predicted tridentate chelation by the fully deprotonated triphosphate, based on the lowered pK a of the coordinated triphosphate molecule and on small differences in g and A tensors. The proposed solution structures of these complexes are shown in Figure 4.1. Based on these earlier studies, the pH of the VO-TPH solution was set at 5.0 where the concentration of 1:1 vanadyl:triphosphate species was maximized. 18,19 Additionally, at that particular pH, both previous works indicated that the largest concentration component was a fully deprotonated vanadyl-triphosphate species, which would allow for the possible detection of a tricoordinate triphosphate ligand (Figure 4.1(c)). 145 References begin on page 181 Chapter 4 (a) (b) (c) Figure 4.1: Proposed solution structures of the vanadyl-triphosphate system: (a) [VO(HL)] 2- and [VO(L)] 3" monoligand complexes 1 8> 1 9; (b) [VO(HL)(L)]7" and [VO(L) 2] 8" bisligand complexes19; (c) tridentate [VO(L)]3" species proposed by Buglyo et al.,^ based on potentiometric and EPR data. (L = P3O10 5 ") Much work has also been conducted on the related VO-ATP system.20-23 ATP possesses three possible coordinating moieties: the triphosphate chain, the C-3 and C-4 hydroxyls of the ribose moiety, and the nitrogens of the adenine base. Potentiometric results, in combination with EPR spectra, indicate that vanadyl is bound by the phosphate chain in acidic to neutral media, but coordinated to the ribose in alkaline solution.24 The 1:1 complex at pH < 6 is thought to involve bidentate coordination by the y- and P-phosphates of the triphosphate chain. 146 References begin on page 181 Chapter 4 4.1.2. Hydroxyapatite and Vanadyl Interactions T h e structural s imi la r i ty between hydroxyapati te and the minera l fraction o f bone has been k n o w n since powder X - r a y and neutron diffraction patterns o f H A were compared against those o f ground bone powder nearly a h a l f century ago.25,26 while the patterns shared some s imilar i t ies , a number o f key differences existed between the laboratory and natural products. Firs t ly , b io log i ca l H A is generally less crystal l ine than synthetic HA.27 T h i s observat ion is predominately due to i o n substitutions i n the lattice; several u n i - and divalent metal ions can substitute for C a 2 + (e.g. N a + and S r 2 + ) , w h i l e carbonate and halogen anions can replace phosphates. The second major difference is i n overa l l compos i t ion . B i o l o g i c a l samples studied b y Parfitt had a 1.3:1 C a 2 + : P 0 4 3 " ratio w h i l e for synthesized H A , the ratio was 1.67 w h i c h agrees w i t h the unit ce l l fo rmula o f C a i 0 ( P O 4 )6 ( O H ) 2 . '7 it is l i k e l y this i on def ic iency that a l lows for the h igh degree o f lattice substitution. B o n e and H A samples possess two monolayers o f water molecules , he ld i n p lace b y hydrogen bond ing to exposed hydroxides.28 The water layers impart a degree o f se lect iv i ty to i o n substi tution by l i m i t i n g access to the crystal lattice to ions o f a certain size. S o d i u m ions , for instance, can incorporate into the lattice, w h i l e potassium ions do not penetrate the hydrat ion she l l . 1 7 The interactions o f vanad ium w i t h H A are dictated b y the ox ida t ion state o f the metal i o n . V a n a d i u m ( V ) , not surpris ingly because o f its structural analogy to phosphate, incorporates into the lattice, although not without causing charge compensat ing vacancies and modi f ica t ions i n unit ce l l parameters.29 The interaction o f V O ^ is not c lear ly defined. Th rough a precipi ta t ion method, it appears that V 0 2 + ions can enter the crystal lattice for C a 2 + , 3 0 although this result is i n confl ict w i t h s imi la r V 0 2 + so lu t ion-so l id interactions w i t h 147 References begin on page 181 Chapter 4 alumina and anatase (both hydroxyl-bearing solid phases) which found surface-binding only.31,32 A more recent study reported only pH-dependent surface binding of vanadyl ions to H A through specific adsorption to surface hydroxides. 14 Through the study of these two model systems, the in vivo coordination chemistry of vanadyl ions in bone will be described. The structure of this complex is important because of the role of bone in the long-term storage of administered vanadium, and its likely role in the steady-state release of vanadium leading to an extended anti-diabetic effect. Some of the work presented in this chapter appears in a recent publication.33 4.2. Experimental Pentasodium triphosphate, calcium chloride and sodium hydrogen phosphate were obtained from Sigma-Aldrich and used without modification. Vanadium atomic absorption standard solution (in 5% HC1) was obtained from Aldrich and carefully neutralized to pH 1.0 with 10% NaOH solution. Vanadyl concentrations were confirmed by titration against potassium permanganate. Pulsed EPR studies were conducted in collaboration with Dr. Sergei Dikanov of the University of Illinois at Urbana-Champaign. 4.2.1. V0 2 +-triphosphate (VO-TPH) Pentasodium tripolyphosphate (0.0606 g, 0.127 mmol) and vanadyl stock solution (2.0 mL of 0.0181 M , 0.036 mmol) were combined and diluted to 50.0 mL with degassed, distilled, deionized water. The solution was kept under Ar and stirred for 20 minutes; the pH was adjusted to 5.0 by slow, drop-wise addition of 0.1 M NaOH. Glycerol (21 mL, 30% by volume) was added and the solution stirred for a further 30 minutes. An aliquot of this 148 References begin on page 181 Chapter 4 solution was then loaded into a 4 mm i.d. quartz cryotube for spectroscopic study. The triphosphate to V 0 2 + ratio was 3.1:1. 4.2.2. Vanadyl-Hydroxyapatite (VO-HA) CaCl 2 -2H 2 0 (2.20 g, 14.9 mmol) was dissolved in 100 mL dd H 2 0 , previously degassed by Ar sparge. In a separate solution, Na 2HP04 (1.42 g, 10.0 mmol) was dissolved in 100 mL degassed dd H 2 0 . The two solutions were mixed (under Ar); a white solid precipitated immediately. An aliquot of a vanadyl stock solution ( [V0 2 + ] = 0.0181 M , 11.0 mL, 0.199 mmol) was adding with vigorous stirring. The initial pH was ~ 5 by pH paper. The pH adjusted to ~8 with concentrated NaOH and readjusted to ~8 every hour for 4 hours; the mixture was then allowed to stir overnight. A pale blue solid was obtained by filtration, dried in vacuo at 60°C for 48 hours and then ground to a fine powder. 4.2.3. E P R Measurements The continuous wave (cw) EPR and ESEEM experiments were carried out using X -band Bruker ESP-380E and Bruker Elexsys E580 spectrometers equipped with Oxford CF 935 cryostats, at 25 K. Preliminary cw EPR study of VO-HA was carried out using a Bruker ECS-106 X-band spectrometer. Several types of ESEEM experiments with different pulse sequences were employed; among those used were the two-pulse and both ID- and 2D four-pulse sequences. The length of the nil pulse was 16 ns. A four step phase cycle, +(0,0,0,0), -(0,0,0,TI), +(0,0,7t,0), -(0,0,7t,7i:) was used to eliminate unwanted features from experimental echo envelopes.3 4 H Y S C O R E data were collected as 2D time-domain patterns containing 256x256 points with a step of 16 ns. Spectral processing of E S E E M patterns was performed 149 References begin on page 181 Chapter 4 using Bruker WIN-EPR software. Simulations were carried out using SimFonia simulation software. 4.3. Results and Discussion 4.3.1. Characterization of V O - H A The synthesis of V O - H A was based on established methods for the isolation of microcrystalline hydroxyapatite.28,35-38 Studies have shown that initially, precipitation of calcium phosphate results in formation of an amorphous phase (amorphous calcium phosphate, ACP), which converts to crystalline hydroxyapatite (HA) over time.35,36 The rate of this conversion is pH-dependent and is favoured at higher pH vs. lower. Therefore, the synthesis of H A in the presence of V 0 2 + was carried out at pH 8, at which value conversion to crystalline H A is relatively rapid, but not so high as to ensure hydrolysis of the vanadyl cations. Buffer solutions were avoided to prevent formation of vanadyl-buffer complexes, which could interfere with subsequent spectroscopic study. Characterization of an entirely inorganic substance can be problematic. Previous studies of H A have relied on the comparison of X-ray powder patterns to observe and quantify the conversion of ACP to H A . 3 5-3 8 Blumenthal and Posner have noted very specific infra-red (IR) bands from free - O H stretching at 3572 cm"1 and - O H librational modes at 630 cm"1 that are diagnostic for crystalline HA.28 m the absence of powder X-ray data, IR spectroscopy was used to confirm the conversion of ACP to H A for the V O - H A sample (Figure 4.2). Finally, a significant concentration of vanadyl ions in the sample was confirmed by EPR spectroscopy on the ground sample (vide infra). 150 References begin on page 181 Chapter 4 o H 1 1 1 1 i 1 1 1 3700 3650 3600 3550 3500 700 650 600 550 Wavenumber (cm"1) Figure 4.2: IR spectra of hydroxyapatite (HA) and amorphous calcium phosphate (ACP) (in KBr). 4.3.2. E P R / E S E E M C W EPR and field-sweep ESE spectra of the frozen VO-TPH solution showed a lineshape with axially symmetric g tensor anisotropy and 5 1 V hyperfine structure. The H A sample was predictably similar, but showed slight differences compared to VO-TPH. Spin Hamiltonian values, determined by simulation, are listed in Table 4.1. These values are typical for square pyramidal oxovanadium(IV) complexes with an O 4 coordination sphere.39,40 The spectral lineshape in all samples is characteristic of isolated V O * species with no oligometallic species detected. 151 References begin on page 181 Chapter 4 Table 4.1: Spin Hamiltonian values for V O model complexes and in vivo V 0 2 + in bone mineral (20 K, v = 9.7170 GHz). Sample gx(± 0.005) #,(±0.01) ^ ± ( ± 0 . 5 ) a ^ | ( ± 0 . 5 ) a V O - T P H b 1.983 1.92 202 83 V O - H A c 1.996 1.92 198 80 VO-boned 1.996 1.93 190 80 a in units of 10"4cm"'. b frozen aqueous solution sample, 33% glycerol by volume. c powder sample. d in vivo sample. Based on spin Hamiltonian values determined from cw and FS-ESE spectra, the most accurate model of VO-bone is VO-HA, while VO-TPH is substantially different from the in vivo sample. Detailed pulsed EPR studies of the two systems, however, give a better indication of the applicability of the VO-TPH and V O - H A systems as models of the in vivo coordination of vanadyl ions. Figure 4.3 shows the modulus FT two-pulse E S E E M spectrum of VO-TPH measured at the tri|V = -1/2 component of the hyperfine structure from 5 1 V . This particular peak was selected for study due to the coincidence of parallel and perpendicular components of the axial g and A tensors at a magnetic field value of 3477 G. The lineshape of this component resembles more of a single peak in contrast to other resonances that possess extended lineshapes typical for an axially symmetric hyperfine tensor with well-separated parallel and perpendicular singularities.41 Therefore, the excitation of the miV = -1/2 component is 152 References begin on page 181 Chapter 4 essentially non-selective, and hence E S E E M spectral analyses normally used for orientationally-disordered (i.e. powder) systems can be used. Unlike the two-pulse ESEEM spectra of Chapter 3, the spectrum depicted in Figure 4.3 exhibits discernable features that allow some preliminary quantitative conclusions to be made about the vanadyl coordination. This is likely because of the high degree of structural homogeneity of the vanadyl centers in the sample. A doublet of lines appear at 6.7 MHz and 5.3 MHz, centred relative to the Zeeman frequency of 3 1 P (suppressed at this x value), v P = 6.00 MHz. The combination peak for phosphorus appears at 12.05 MHz, on the shoulder of the ' H matrix peak at 14.84 MHz. The sum combination harmonic is split into two peaks, one at 30.35 MHz shifted to the higher frequencies relative to the proton double Zeeman frequency, 2v H = 29.62 MHz. Suggestions of other peaks appear below 3 MHz, similar to those observed for the bone sample. 0 5 10 15 20 25 30 [MHz] Figure 4.3: Modulus two-pulse E S E E M spectrum of VO-TPH (B = 3479 G, T = 30 K, v 9.713 GHz). 153 References begin on page 181 Chapter 4 The doublet of lines indicates 3 1 P nucleus(i) with A = —1-1.5 MHz either very near or directly bound to the V 0 2 + ion. Such splitting cannot result from interactions with randomly distributed, remote 3 1 P nuclei and therefore, the direct coordination of at least one phosphate to the metal ion can be inferred. The peaks in the low frequency region of the spectrum (with a maximum at ~1.5 MHz) may be a result of other 3 1 P nuclei, strongly coupled with the V 0 2 + unpaired electron, although the exact source is not obvious from the one dimensional spectrum. The presence of sum combinations with a shift of ~0.7 MHz at a field strength of -3300 G is typical of equatorial water coordination to vanadyl. 1'42 Thus, analysis of the two-pulse spectrum provided the first indication of mixed coordination of V 0 2 + by phosphate and water molecules in VO-TPH. The modulus two-pulse ESEEM spectrum (Figure 4.4) for V O - H A is not nearly as definitive as that obtained for VO-TPH. While the spectrum has the same basic appearance as the two-pulse spectrum in Figure 4.3, the peaks are broad and ill-defined. This result is probably due to inhomogeneity in the sample, or a shorter phase memory time for V O - H A . The 3 I P matrix peak is visible at 5.88 MHz, slightly lower than the predicted frequency of 5.97 M H z at 3460 G. A weaker signal appears at 11.03 MHz, which likely represents a nuclear transition, coupled to another signal in the broad peak centred at -1.5 MHz. The 2v P signal is not visible, as it is in Figure 4.3; perhaps it is buried beneath the strong ' H matrix peak at 14.70 MHz. Lastly, the sum combination peak for proton appears at 29.66 MHz, but it appears as a single peak with no observed splitting due to sum combinations of the basic nuclear frequencies. 154 References begin on page 181 Chapter 4 Figure 4.4: Modulus two-pulse E S E E M spectrum of V O - H A (B = 3460 G, T = 20 K, v = 9.717 GHz). The application of more sophisticated ID and 2D E S E E M methods led to more complete information about the phosphorus and proton environments of the vanadyl ions. 4.3.3. HYSCORE Contour HYSCORE spectra of VO-TPH are shown in Figure 4.5 and Figure 4.6. The spectrum in Figure 4.5(a), defined between -18 and +12 MHz in one dimension and 0 and 15 M H z in the second, contains three different pairs of cross-peaks, designated Pi , P 2 , and P3. Cross-peak P i , a low intensity pair appearing in the (+,-) quadrant, possess maxima at frequencies ~[+1.8; ±13.5] MHz, oriented approximately parallel to the diagonal and centered at 8 MHz. The coordinates of each peak are separated by close to double the Zeeman frequency o f 3 ! P (2vP). Two other cross features P 2 and P3are located in the (+,+) quadrant. Both appear as extended narrow ridges between [10.1; 2.2] M H z and [7.0; 5.0] MHz (for P 2), plus a pair of peaks at [6,5; 5.5] MHz (P3), centered symmetrically about the 155 References begin on page 181 Chapter 4 diagonal point of vp= 6.03 MHz. The cross-peaks are better visualized in a stacked plot, shown in Figure 4.5(b). The observation of the P3 cross-peaks was assisted by the suppression of the 3 I P matrix peak through adjustment of the x delay to 200 ns. From the location and properties of the observed cross-peaks, three estimated vanadyl-phosphorus couplings of-15, 9 and 1 MHz can be derived. These constants, however, do not take into account the individual contribution of the isotropic A and anisotropic T components. (a) ^ R 3 - 1 5 10 -5 0 ui[MHz] 1 0 u 2 [ M H z ] - 1 4 - 1 2 1 0 8 6 4 [2 0 1.25 2.50 3.75 5.00 6.25 7.50' 8.75 10!00 l l !25 ui[MHz] Figure 4.5: (a) H Y S C O R E spectrum of VO-TPH, in the region of cross-peaks P1-P3 (B = 3479 G, T = 30 K , x = 120 ns, v = 9.713 GHz), (b) Enlarged stacked plot presentation of the (+,+) quadrant to visualize the P2 and P3 cross peaks. 156 References begin on page 181 Chapter 4 Figure 4.6 shows the HYSCORE spectrum with both frequency domains centred at the proton matrix frequency of 14.8 MHz. Several groups of strong lines are grouped around this strong peak, in contrast to the in vivo sample where only one cross-peak set was observed. There is a pair of clearly resolved cross-peaks labelled Hi that extend between [18.8; 11.5] and [16.5; 13.9] MHz, with a maximum at [17.8; 12.3] MHz in this particular spectrum. The extended features of cross-peak Hi are distorted asymmetrically along the bottom ridge, possibly a result of the close proximity of the intense peaks of Hi with lower intensity H 2 lines. This interpretation is supported by an observation of proton sum combination lines in ID four-pulse ESEEM spectra (vide infra). The final pair of cross-peaks (H 3) is located close to the diagonal. One of the cross-features of H 3 was partially obscured by the proton matrix peak, despite careful selection of x to suppress strongly the matrix signal. The HYSCORE spectrum for V O - H A is presented in two parts, corresponding to the 3 1 P and ' H cross-peak regions (Figure 4.7 and Figure 4.8). For 3 1 P , three cross-peaks are visible, and they closely approximate the location of the cross-peaks of the VO-TPH spectrum. Cross-peak P A appears in the (+,-) quadrant, centred at [±13, +1.5]. Two pairs of cross-peaks (PB and Pc) appear in the opposite region. The second cross-peak P B is manifested by two extended ridges between [11.7; 0.6] to [9.0; 3.0] while Pc peaks are centred about the 3 1 P matrix peak in a similar fashion as P 3 i n the VO-TPH spectrum. 157 References begin on page 181 Chapter 4 w,[MHz] -18 .16 -14 -12 10 , . . . . I . . . . . . . . . . . . . . . . . . . ( . . . . ( . . . . . . . . . 10.00 1250 15.00 17.50 v, [KHz] Figure 4 .6 : HYSCORE spectrum of VO-TPH, in the region of cross-peaks H 1 - H 3 (spectral parameters as in Figure 4.5). - 1 5 - 1 0 - 5 0 5 1 0 1 5 v2[MHz] Figure 4 . 7 : HYSCORE spectrum of VO-HA, in the region of cross-peaks PA-Pc (B = 3465 G, T = 30 K, x = 128 ns, v = 9.710 GHz). 158 References begin on page 181 Chapter 4 Analysis of proton coupling in V O - H A was again assisted by the ID four-pulse experiment, which exhibited two sum combination peaks shifted to higher frequencies relative to 2v H (A\\ = -0.4 and -0.8). The HYSCORE spectrum of the proton region (v H = 14.84 MHz) is shown in Figure 4.8. The spectrum appears as an extended ridge spanning across perpendicular to the diagonal about the proton matrix peak. From the contours of the ridge and evidence of two components from the ID four-pulse spectrum, two cross-peaks can be identified ( H A and H B ) . Despite acquisition of several HYSCORE spectra with various x delay times to suppress the matrix proton signal, no third cross-peak about the diagonal peak could be discerned. Cross-peak H A extends from [18.7; 11.5] to [17.0; 12.9] while H B lies closer to the diagonal from [16.7; 13.1] to [15.7; 14.0]. Figure 4.8: HYSCORE spectrum of VO-HA, in the region of cross-peaks H A and H B (B = 3465 G, T = 30 K, x = 192 ns, v = 9.710 GHz). V,[MHz] 159 References begin on page 181 Chapter 4 4.3.4. Analysis of H Y S C O R E Spectra Contour lineshape analysis was applied to extract the isotropic and anisotropic coupling constants from both the 3 1 P and *H cross peaks, in a manner completely analogous to that detailed in Chapter 3. The experimental data plots of v a 2 versus v p 2 along with their linear regression lines for the HYSCORE spectra of VO-TPH are shown in Figure 4.9 and Figure 4.10. Slopes, intercepts and hyperfine coupling values derived from the contour lineshape analyses are listed in Table 4.2 for both model systems. 300 250 200 r I N S 150 8 > 100 50 0 0 10 20 30 40 50 v p 2 , M H z 2 Figure 4.9: Contour lineshape analysis plot for 3 1 P cross-peaks observed in H Y S C O R E spectra of VO-TPH (e.g. Figure 4.5) recorded at various x values. The larger point was arbitrarily chosen as v a and the smaller coordinate as vp. The solid lines depict the linear regressions of the experimental data points, while the dashed curve represents a plot of | v a + v p | = 2v P with v P = 6.00 MHz. 160 References begin on page 181 Chapter 4 Included with the experimental points, the curve | v a + vp | = 2v\ (for vi corresponding to v P or v H ) is plotted in Figure 4.9 and Figure 4.10 for 3 1 P and ' H , respectively. This plot assists in the assignment of the cross-peaks as arising from coupling to a specific nucleus, in this case, that o f 3 ' P . The points at which the curve crosses each extrapolated straight line correspond to the nuclear frequencies at canonical orientations. They define the range of possible frequencies in powder spectra for each nucleus at this particular value of vi- These canonical frequencies can also be used to determine the hyperfine parameters. For an axial hyperfine tensor, there are two possible assignments of the parallel or perpendicular orientations and consequently two sets of hyperfine tensors, one for each assignment. This approach gives hyperfine couplings identical to those determined from the slope and intercept. The placement of this curve on each of Figure 4.9 and Figure 4.10 provides additional evidence for the correct assignment of these peaks to protons and phosphorus. The fit of the calculated lines to the experimental points is excellent (minimum r = 0.947). The extrapolated regression lines cross the nuclear frequency curve at two points; these intercepts would not be observed i f the signals originated with other elements. The intersections of the regression lines with the nuclear frequency curve serves to affirm the treatment of the data with an axial approximation. Dikanov et al. showed that for strongly non-axial systems (i.e. highly rhombic), HYSCORE cross-peaks are expected to appear as three separate ridges.43 The regression lines of these contours, when plotted v a 2 versus vp2, form a triangle shape, the apices of which intersect with the | v a + vp | = 2\\ curve. The regression lines of the two figures, however, do not form such as shape, confirming the correct use of the axial approximation and assignment of the lines as arising from three distinct nuclei.43 161 References begin on page 181 Chapter 4 4 5 0 4 0 0 \ 3 5 0 X S! 3 0 0 2 5 0 2 0 0 150 \ v \ \ \ \ \ \ \ \ \ \ V \ \ \ \ \ \ \ \ 'A \ \ > NS \ 5 0 100 150 v R 2 , M H z 2 2 0 0 25( Figure 4.10: Contour lineshape analysis plot for H cross-peaks observed in H Y S C O R E spectra of VO-TPH of various x values. The larger point was arbitrarily chosen as v a and the smaller coordinate as vp. The solid lines depict the linear regressions of the experimental data points, while the dashed curve represents a plot of | v a + vp | = 2v H with v H = 14.81 MHz. There does appear to be a significant deviation from linearity for the points of the Hi cross-peak. This observation may indicate contributions from protons with different coupling constants. If the presentation of the Hi peaks was assumed to be a superposition of two lines, regression of either case leads to an assignment of the couplings with a spread of A = ±2 MHz and a highly stable T constant of about 4.3 MHz. 162 References begin on page 181 Chapter 4 Table 4.2: Parameters3 derived from contour lineshape analysis of HYSCORE spectra of VO-TPH and VO-HA. Cross-peakb Q« G a , M H z 2 A, MHz | T | , M H z Pi 9.09 (0.38) 145 (1.50) -15.9 (0.2)/ 14.0 (0.1) 1.94 (0.18) P 2 -5.11 (0.15) 128 (0.85) -8.82 (0.03)/7.33 (0.18) 1.48 (0.16) P 3 -1.27 (0.02) 82.4 (0.54) -2.01 (0.1)/0.88 (0.25) 1.14(0.3) H , -1.51 (0.02) 554 (2.36) -8.17 (0.16)/3.87 (0.2) 4.29 (0.05) H 2 -1.46 (0.01) 532(1.8) -6.9 (0.3)/4.15 (0.08) 2.79 (0.03) H 3 -1.22 (0.03) 485 (4.6) -3.66 (0.3)/2.39 (0.32) 1.22 (0.03) PA 7.74 (0.42) 150 (2.0) -16.6 (0.2)/ 14.6 (0.1) 2.00 (0.09) PB -4.01 (0.15) 124(1.2) -8.10 (0.06)/6.33 (0.12) 1.77 (0.12) Pc -0.852 (0.030) 69.8(1.4) -1.88 (0.10)/0.04 (0.04) 1.84 (0.08) H A -1.36(0.08) 516(14) -5.76 (0.2)/3.26 (0.1) 2.49 (0.03) H B -2.15 (0.07) 636(8.3) -13.1 (0.15)/8.61 (0.12) 4.45 (0.02) a Values in parentheses represent standard errors. b P # , H # = VO-TPH model; P x , H x = V O - H A model. The difference between these two cases is therefore only a minor effect. It is also possible, however, that other additional factors disturb the lineshape of cross-peaks and contribute to the observed deviations of the experimental points from the straight contour described by v a = Q a vp 2 + G a . These factors might be g and A tensor anisotropy, inaccuracy in the determination of the (v a, vp) coordinates from the discrete character of the 2D data acquisition, and spectral distortions associated with the data collection over a limited time interval. 163 References begin on page 181 Chapter 4 Contour lineshape analysis of several HYSCORE spectra of V O - H A formed a plot very similar in appearance to Figure 4.9, with slight differences in Q and G values leading to minor differences in determined A and T constants (Table 4.1). The three regression lines produced a very similar shape and all experimental points of the three linear sections were bounded by the | v a + vp | = 2vi curve. Cross-peak P A possessed the greatest isotropic coupling constant of any phosphorus detected in the model study. Depending on the relative signs of A and T, Pc has a near zero isotropic component. Generally, however, each V O - H A 3 1 P cross-peak can be roughly correlated to a cross-peak set observed in VO-TPH (P| ~ P A ; P2 ~ P B ; P 3 ~ Pc)- The anisotropic constants of the V O - H A showed very little variance, suggesting a relatively uniform distance of 3 I P atoms around the vanadyl ion. 4.3.5. Sum Combination Lines in E S E E M Spectra The extended ridges that comprise predominately the Hi cross-peak were unequivocally determined to be in fact two overlapping cross-peaks from the observation of sum combination harmonic peaks in the ID four-pulse E S E E M experiment. Previous E S E E M studies of [VO(H20) 5] 2 + in powder and glassy solution have shown that sum combination peaks, shifted relative to the matrix line at 2v H , appear from protons of ligand molecules.1 '42 while the two-pulse ESEEM spectra of VO-TPH show a splitting of the 2v H matrix peak, four-pulse E S E E M spectra provide superior resolution of sum combination peaks arising from anisotropically-coupled protons. The two-pulse experiment is limited by T 2 relaxation that generally leads to peak broadening which may obscure even the usually narrow sum combination peaks. The one-dimensional four-pulse sequence is superior for the observation of sum combination peaks because the echo decay time is limited by the much longer relaxation time T i that allows for improved resolution. The four-pulse E S E E M 164 References begin on page 181 Chapter 4 spectrum of VO-TPH (Figure 4.11) contains three well-resolved lines in the region of the proton 2v H . The matrix peak dominates the spectrum, but two other sum combination peaks are visible at A ~ 0.7 and 0.3 MHz. The higher A value is resolved in the two-pulse spectrum, but the four-pulse spectrum reveals a new weak peak with a smaller shift relative to 2v H . This experiment was also able to resolve similar combination peaks in the V O - H A sample. The peaks in this case, however, were not a clear as those observed in the VO-TPH model, possibly due to inhomogeneity of the paramagnetic centres in the sample. The two peaks appeared at ~ +0.85 and ~ +0.4 MHz from the proton matrix peak. This observation confirmed the treatment of the broad ridge observed in the proton coupling region of the H Y S C O R E spectrum is in fact composed of two overlapping cross-peaks. The frequency of the sum combination harmonic maximum (v+, in a powder or frozen solution spectrum) from an I = Vi nucleus with hyperfine couplings A and T is described by Equation 4.1.44 2 u H 27.5 28.0 2B.5 29.0 29.5 30 Q 30.5 31.0 31.5 32.0 [MHz] Figure 4.11: Proton sum combination peaks in the ID four-pulse E S E E M spectrum (magnetic field 3479 G, x = 120 ns, v = 9.713 GHz). 165 References begin on page 181 Chapter 4 Equation 4.1 ^ = 2v, 1 + 9T 2 16v2 - ( T + 2^)2 11 / 2 Using the ^ and T values obtained from Table 4.2, A shifts of 0.73, 0.32, 0.06 (for H|. 3), 0.86 and 0.26 (for H A -B ) can be calculated from Equation 4.1. The two largest calculated shifts agree with the shifts of the two lines resolved in the four-pulse spectrum shown in Figure 4.11. The third shift is far too small to be resolved, even in a four-pulse spectrum. Through the combination of HYSCORE, ID 4-pulse ESEEM and contour lineshape analysis, the complicated proton coupling in the VO-TPH sample was resolved into three distinct coupling constants from three different protons. The predicted shifts correlate reasonably well to those observed in the four-pulse experiment of VO-HA, although the larger shift correlation is particularly good. The structural interpretation of the relative magnitudes of the A and T values for the 3 1 P and 'H nuclei reported in Table 4.2 is detailed in the next section. 4.3.6. Phosphorus Couplings Two-dimensional HYSCORE spectra of VO-TPH and VO-HA each show the presence of three 3 1 P couplings with significantly distinct isotropic hyperfine constants. These constants can be compared against those obtained in analogous systems.45,46 The binding of adenosine diphosphate (ADP) and ATP was studied by EPR titration. Measurement of the peak-to peak intensity of one component of the spectrum indicated a 1:2 metal:ligand stoichiometry for ADP and ATP, with vanadyl chelation occurring via the phosphate groups.4^ The spectra of ADP were of a more symmetric system, however, as frozen solution spectra of [VO(ADP)2]" exhibited additional superhyperfine structure. Observation of such coupling is rare for V 0 2 + complexes. It was assigned to four equivalent 166 References begin on page 181 Chapter 4 phosphorus nuclei with a hyperfine coupling constant of 18.54 MHz.45 Therefore, in the ADP system, the equatorial coordination of the phosphate groups of two ADP molecules is geometrically equivalent with respect to the metal center. This symmetric binding mode was not conserved in the ATP system. Due to inequivalent hyperfine couplings with phosphorus, no regular superhyperfine structure was resolved in the cw EPR spectra, however, ENDOR spectra of these complexes determined a single 3 1 P coupling of 20.6 MHz.45 It is possible, however, that the binding of phosphate moieties to V 0 2 + demonstrates the low frequency limitations of ENDOR spectroscopy. HYSCORE spectroscopy, an E S E E M method, resolved two smaller 3 1 P coupling constants of 14.8 and 9.0 MHz in the VO-ATP system, assigned respectively to the P- and y-phosphorus atoms of ATP.46 Similar A coupling constants were also reported for V0 2 + used as a spin probe for the Mg 2 +binding site of wild type FI ATPase (TF1) .46 The equatorial coordination sites were filled by two nitrogen donors (lysine amine and histidine imine), also detected by HYSCORE. Comparison of the HYSCORE spectra (Figures 6 and 7 of ref 46) reported by Buy et al. to those in this thesis (Figure 4.5) reveals a similar extended cross-peak (labelled P 3 and Pc in this work) about the 3 1 P diagonal matrix peak. This cross-peak was not discussed by Buy et al, likely due to its relatively weak intensity at the x values used to acquire the spectra. The presence of these weak cross-peaks, however, parallels the results for the model systems discussed in this thesis, for which a third 3 1 P coupling of ~1 MHz was determined (from P3 and P c ) . ENDOR spectroscopy could be used to determine the relative signs of A and T from the two possibilities indicated in Table 4.2, at least for the strongest 3 1 P couplings. The large isotropic coupling of 18-20 MHz observed in the powder EPR and ENDOR spectra of V O -167 References begin on page 181 Chapter 4 A D P and VO-ATP most likely corresponds to the perpendicular component of the hyperfine tensors. The two sets of hyperfine constants derived from the analysis of the P i cross-peaks give A±=\A - T | = 17.8 and 12.1 MHz, respectively, while for P A yields 18.6 and 12.6 MHz. Only the larger of the two possible A values listed in Table 4.2, -15.9 (Pi) and -16.6 (PA), approach the reported EPR and ENDOR splittings (corresponding to opposite signs of A and T). 4.3.7. Proton Couplings A wide variety of studies have been conducted to analyze the coupling constants of protons adjacent to the vanadyl ion. 1,42,47-53 p r o ton hyperfine couplings for coordinated water and alcohol molecules in V 0 2 + complexes in single crystal,47 powder,1 frozen aqueous1,42,48,49 a n c j alcohol50-52 solutions, as well as in frozen protein solution^3 have been investigated in detail. These studies differ markedly in the reported data. The most complete and accurate single crystal ENDOR study of [VO(H20)5] 2 +doped in Tutton salt lists the principal components of hyperfine tensors for all ligand protons and orientation of their principal axes correlated with the location relative to the metal.47 ]n general, the magnitude of the isotropic coupling constant is dependent on the orientation of the proton to the equatorial plane of the vanadyl ion. The anisotropic coupling constant is also sensitive to the orientation of the proton to the molecular axis system; protons oriented in an axial position have much smaller T values, typically ~3 MHz. Protons in the equatorial plane possess T = 4 MHz or greater.47 ESEEM analysis of sum combinations 1,42,53 a i s o allows direct measurement of | T |. These studies reported variations for the equatorial proton anisotropic couplings within the same range. Thus, the magnitude of the isotropic and 168 References begin on page 181 Chapter 4 anisotropic coupling constants can be used to gain some insight into the overall geometric arrangement of the coupling protons. ENDOR studies of frozen solutions^,50,51 h a v e limited bearing on the constants reported in this thesis. ENDOR usually provides the splitting values along directions parallel and perpendicular to the V=0 axis through orientationally-selective measurement at the parallel or perpendicular hyperfine features of the cw EPR spectrum. The hyperfine coupling constants determined by this method are therefore equal to the projection of the principal values of the hyperfine tensors onto the g tensor coordinate system, since in general the axes of the g and proton hyperfine tensors are not coincident with each other. As a result, the couplings reported for equatorial protons in these studies cannot be directly compared with the values obtained from single-crystal ENDOR and E S E E M data. Based on the most applicable studies, 1,42,47,53 therefore, the value of | T | = 4.29 MHz found for cross-peaks Hi (and confirmed by the corresponding shift of sum combination peaks) belongs to protons from equatorially coordinated water(s). Atherton and Shackleton47 have stated that the A and T values for such protons typically have opposite signs with the anisotropic constant being negative, leading to the choice of A = 8.8 MHz and T = -4.3 MHz for the protons producing cross-peaks Hi and H B . Due to the potential coincidence of two lines, the magnitude of A for the Hi cross-peaks has a range of ± ~2 MHz. 169 References begin on page 181 Chapter 4 Table 4.3: Isotropic (| A |) and anisotropic (| T |) ' l l coupling constants of various V O complexes compared to the proton couplings of VO-TPH and VO-HA. \A\, MHz | T |, MHz Assignment Reference ~ 0 0 - 9 3.3 4.2-5.0 H 2 0 (axial) H 2 0 (equatorial) 47 0 0.65-8.6 3.4 4.23-4.69 H 2 0 (axial) H 2 0 (equatorial) 1,42 — 4.34 ' H cis to V=0 53 — 4.15-4.90 ' H ( ± t o g 2 ) 54 8.17 or 3.87 4.29 H | cross-peak ! H (equatorial) 6.9 or 4.15 2.79 H 2 cross-peak Remote *H, perhaps from -O-P-OH 3 3.66 or 2.39 1.22 H 3 cross-peak V = 0 " ' H - O a This work 5.76 or 3.26 2.49 H A cross-peak Remote ' H , perhaps from -O-P-OH 3 13.1 or 8.61 4.45 H B cross-peak *H (equatorial) 3 see text for assignment, p. 171. 170 References begin on page 181 Chapter 4 According to several studies, protons from axially coordinated water molecules typically possess T = 3.1-3.4 MHz and isotropic constants close to zero. Signals from this type of proton have been reported in the HYSCORE spectra of [VO(ImH) 4] 2 + (ImH = imidazole) in aqueous solution.55 HYSCORE spectra of this complex did not show cross-peaks with characteristics similar to Hi or HDbecause all four equatorial positions were occupied by imidazoles and only the axial position was filled with a water molecule. Cross-peaks H 2 , H 3 and H A possess anisotropic couplings that are much smaller than those determined for protons of water ligands. The protons producing them, therefore, must arise from different sources. Previous ENDOR spectra of VO 2 * in frozen aqueous and alcohol solutions show the presence of couplings assigned to the proton involved in a hydrogen bond to the vanadyl oxo (V=O---H).50 Based on the combination of isotropic and anisotropic components, cross-peaks H 3 can be assigned to this type of proton. Splittings of 4^1| = 4.34 MHz andv4± = 1.40 MHz were reported by Mustafi and Makinen;50 calculated values of A \\ = \A + 2T | = 4.8 MHz and AL= \A-T \ =1.2 MHz for proton H 3 (based on A = 2.39 M H z and T = 1.22 MHz) compare favourably to these values. Therefore H 3 coupling likely arises from a hydrogen-bound proton attached to the vanadyl oxygen. This assignment appears more likely i f one invokes only small deviations of the principal directions of the H 3 hyperfine tensor from the principal directions of g tensor (as for protons of axial ligands). The ID four-pulse E S E E M spectra also provide support for this assignment. Through Equation 4.1, an estimate of the sum combination peak shift of H 3 = -0.06 M H z is obtained, which is not detectable in the experimental spectra. Such a miniscule shift has been shown to be undetectable by theoretical analyses of the peak intensities of axial and equatorial protons in ID four-pulse E S E E M spectra. 1 '42 171 References begin on page 181 Chapter 4 The hyperfine parameters obtained for cross-peaks H 2 or H A do not correlate with any values previously reported for the protons of ligands in the first coordination sphere. The anisotropic components IT | for cross peaks H 2 , H 3 and H A are smaller than typical values obtained for both equatorial and axial ligands and so it is likely that these peaks for H 2 and H A arise from one or more proton(s) bound as hydroxyls on the triphosphate ligand. The most significant observation of the HYSCORE studies of the 3 I P and *H coupling constants is that no axially-bound protons are detected. Coupled with the detection of three distinct 3 1 P couplings, it is consistent with experimental data to invoke a solution structure with some other ligand (i.e. phosphate) coordinating in the axial position. 4.3.8. Comparison Between the Models The VO-TPH and V O - H A model systems yield similar coupling constants by E S E E M and HYSCORE spectroscopies, including number, type and relative magnitude. Significant differences are the failure to detect any equivalent H 3 cross-peak in V O - H A , suggesting that no water molecules sit above the V=0 bond. This observation is consistent with the state of each model; VO-TPH was a frozen solution sample and V O - H A was a solid state sample that had been previously dried to remove coordinated water. The range in T values for the 3 1 P cross-peaks is also much greater for the VO-TPH sample. This perhaps reflects less restriction on the orientations of the phosphates relative to the orbitals of the vanadium centre. In VO-HA, phosphates are not tethered together and therefore can be oriented to energetically favourable positions that likely allow greater interaction with the 3p orbitals of the phosphorus atoms (i.e. dipolar coupling), and hence indicate shorter V-O-P distances. Magnitudes of isotropic couplings also show some degree of variation between VO-TPH and VO-HA, but considering the size of these couplings they likely do not 172 References begin on page 181 Chapter 4 represent major differences in compound structure. Such differences are likely a result of the differences between a frozen solution and a true powder sample. 4.3.9. Structure of the Complex Determination of three unique 3 1 P and ! H couplings indicate a mixed coordination state of the V 0 2 + ions in both model samples. The magnitude of these couplings allows for some structural conclusions to be made about the complexes in solution and the solid state. Strong Hi cross-peaks and a readily visible shifted peak in the ID four-pulse experiment indicates the presence of at least one water molecule, coordinated equatorially. The H Y S C O R E and E S E E M results are entirely consistent with equatorial water coordination to V 0 2 + previously reported.' ,23,42,47,53 The appearance of the four-pulse spectrum can also yield a rough estimate on the number of coordinated water molecules. Relative to the matrix peak v H , the intensity of the shifted lines in the VO-TPH sample is about half of those observed for [V0(H 20)5] 2 + ,1'42 suggesting the coordination of two water molecules in the equatorial plane. A more accurate determination of equatorially bound water molecules could be obtained through study of the EPR linewidth of the VO-TPH complex in H 2 0 and D 2 0 . 5 6 From the data presented, a solution structure can be proposed, which includes a facial, tridentate triphosphate coordinated to the vanadyl ion, along with two water molecules located near to the equatorial plane (Figure 4.12). The difference in ^( 3 I P) constants can be explained by differences in V-O-P bond angles relative to the vanadyl oxo bond. Isotropic 3 1 P coupling constants, in a manner similar to ' H , would be expected to vary relative to the degree of equatoriality, that is, the degree of 173 References begin on page 181 Chapter 4 orbital overlap between the metal center and the orbitals of the phosphorus atoms. 4 7 Less orbital overlap leads to a decreased contact interaction and spin density on the 3 1 P atoms. Unit spin density for the 3s orbital of phosphorus leads to a calculated isotropic coupling constant of 13306 M H z . 5 7 O(H) o | / 0 H 20 / / , .N. ,x \0 — H 2 0 ^ | O o \ / / / \ \ (H)01/P 0 Figure 4.12: Proposed solution structure of VO-TPH at pH 5.0, based on E S E E M and H Y S C O R E data. In contrast, anisotropic coupling is determined by two factors, dipole-dipole coupling and indirect spin transfer. The distance between the spin bearing nucleus and the 3 1 P atom determines both factors. If the V-P distance of 3.44 A, determined by Mustafi et alf1^ is used as an estimate of the vanadium-phosphorus distance in the VO-TPH sample, a dipolar coupling of 0.79 MHz is obtained through the point dipole approximation. Indirect spin transfer to the 3p orbital of 3 I P leads to a computed value of 367 MHz, approximately 36 times less than that determined for the 3s orbital.57 Thus, it would be expected that the isotropic coupling would be much more sensitive to changes in coordination state (i.e. axial versus equatorial) than the anisotropic coupling constant. This prediction is indeed what is 174 References begin on page 181 Chapter 4 observed in the coupling values determined for cross-peaks P1-P3, which exhibit a large variation in isotropic coupling from ~ 1 to 15 MHz, while the T component is relatively stable between 1.14 and 1.94 MHz. The anisotropic coupling constants for V O - H A are even less variant. Similar behaviour was observed by Atherton and Shackleton with proton anisotropic coupling constants for the [VO(H 2 0) 5 ] 2 + complex, in which the anisotropic variation did not exceed 1.5 times.47 The largest isotropic coupling (P| and P A ) likely corresponds to the O-P fragment closest to true equatorial coordination. Axial phosphate coordination to vanadyl has never been spectroscopically characterized, and has some implications for the related coordination of divalent metal ions in ATP-utilizing enzymes. Special attention, therefore, will be paid to this component of the proposed structure. The low isotropic coupling and relatively invariant anisotropic coupling (compared to the other two cross-peaks) of the P 3 and Pc cross-peaks therefore support the presence of an axial phosphate coordinated to the vanadyl ion. Recent quantum mechanical calculations^ confirmed the weak influence of an axial ligand on the characteristics of the EPR spectrum resulting from low spin density transfer to this ligand. Density functional (DFT) calculations of [VO(H 2 0) 5 ] 2 + , [VO(H 2 0) 4 ] 2 + and [VO(NH 3 ) 4 (H 2 0)] 2 + complexes yielded anisotropic hyperfine coupling values of | T | ~4 MHz for equatorial protons and | T | ~3 M H z for axial protons with isotropic constants of ~9-10 MHz and ~0 MHz, respectively.58 These results correlate very well with experimental results reported in Table 4.2 and provide further support for the assignment of the weakest 3 1 P coupling constants as an axially-ligated phosphate residue. The observed A coupling constant of-2.01 or 0.88 M H z likely arises from a more complex mechanism of spin density distribution which does not take place in aqua ligands. Likely contributions to the isotropic constant include the 175 References begin on page 181 Chapter 4 transfer of spin density to the 3 1 P of the axially coordinated phosphate from the metal center through the coordinated oxygen as well as through the chain of equatorial 3 1 P - 0 donors. The latter mechanism is likely the major contributor to the observed 3 1 P isotropic constant of the axial phosphate. Further DFT calculations would be required to analyze these possibilities. The ratio of anisotropic constants of equatorial over axial 3 1 P donors approximates previous results for coupled protons. The VO-TPH model possesses a | T e q | /1 T a x | ratio of 1.3-1.7; for V O - H A it is 1.1 - 1.2 compared the vanadyl-aqua complex at ~ 1.5. These values are also consistent with tridentate TPH coordination. One other structural interpretation involves transfer of spin density through the equatorial 0 - 3 l P donors only, that is, to an uncoordinated third phosphate moiety. This model, however, is inconsistent with the observed couplings. In the case of bidentate equatorial coordination, the dipole-dipole contribution to | T | from the 3 1 P of the noncoordinating phosphate decreases proportionally to r"3. An increase of the V-P distance, even to 4.5 - 5 A, compared to 3.44 A for a coordinated phosphate,45 would yield | T | = 0.25 - 0.35 MHz instead of | T | = 0.79 MHz from the point dipole approximation. If the contribution to | T | resulting from indirect spin density on the equatorial phosphates is estimated to be ~ 1 MHz, it can be suggested that the concomitant decrease of the anisotropic coupling constant of the noncoordinating phosphate would be proportional to the decrease of the isotropic coupling found to be 5-7 times. Therefore, a roughly estimated value of | T | for the third noncoordinating phosphate would not exceed -0.5 MHz (compared to | T | = 1.14 for the coordinated model), thus precluding only a bidentate coordination mode for the TPH ligand. The VO-TPH model and its coupling constants are quite similar to those reported in the VO-TF1 -ATP system. 4 6 HYSCORE spectra of both VO-ATP (at pH 6.3) and the V O -176 References begin on page 181 Chapter 4 TF1-ATP protein system demonstrate three distinct 3 I P couplings (of which two were reported) similar to those observed in the VO-TPH system. In the protein system, the two other equatorial positions were occupied by nitrogen donors while in the VO-TPH system likely two water molecules occupy these sites. Therefore, coordination of phosphate moieties to the two remaining equatorial positions can be inferred. Unfortunately, Buy and co-workers did not analyze the proton environment of the V 0 2 + ions, which could have indirectly indicated the number of equatorial donors (either proton or phosphorus containing), or the presence/absence of axial protons.46 Comparison of the H Y S C O R E spectra between the two studies however reveals a conserved extended ridge about the phosphorus matrix peak, which in the VO-TPH system gives rise to the third set of coupling constants (P3). The similarity of these results obtained from HYSCORE spectra of three different complexes indicates a similar binding motif of one triphosphate chain to the vanadyl ion. The other high resolution paramagnetic resonance study of the coordination of a triphosphate moiety to V 0 2 + was the ' H ENDOR study of VO-nucleotide systems reported by Makinen and co-workers.45 For the VO-ATP system at pH 6.5, only a [VO(H 20)(ATP)2] species, with the coordinated water in the axial position, was detected; no monoligand, tridentate ATP (i.e. analogous to the VO-TPH model) complex was found. These results are therefore inconsistent with the solution structure proposed in Figure 4.12. The differences in experimental conditions, however, are particularly important. The V O -TPH model was studied at pH 5.0, where the concentration of monoligand species was maximized (according to potentiometric and EPR studies). 18,19 The ENDOR study was conducted at pH 6.5 and, additionally, in a 50:50 mixture of water and methanol. Higher pH 177 References begin on page 181 Chapter 4 would favour the formation of bisligand species. There are, however, no detailed 31 measurements of the magnitude and number of different hyperfine couplings from P nuclei, except some discussion of features assigned to the coupling at ~20 MHz. The presence of an axially-coordinated water molecule follows from the observation of a 3 MHz coupling in the spectrum recorded at a perpendicular component of the hyperfine structure. The proposed structure, fully characterized by high resolution E S E E M and H Y S C O R E techniques, represents the first observed tridentate triphosphate coordination to the vanadyl ion, either for triphosphate or for a triphosphate-containing ligand such as ATP. While previous studies of these systems have been reported, the axial phosphate interaction, consistent with the experimental data reported in this chapter, has not been previously detected. A l l other previous work with vanadyl has detected only equatorial phosphate coordination, corresponding to the P- and y-phosphates of ATP. The proposed structure is consistent, however, with several X-ray crystal structures of other metal complexes with triphosphate (two containing Co(III), 5 9> 6 0 and one Cd(II), 6 1) or ATP (including Mn(II) 6 2 , Mg(II), Ca(II) 6 3 , Ni(II) , 6 4 Cu(II) and Zn(II) 6 5) as tridentate ligands. An example of one such complex is shown in Figure 4.13. Evidence also exists for the formation of a tridentate triphosphate Rh(III) complex from N M R spectroscopy and H P L C . 6 6 The triphosphate moiety is coordinated in a facial fashion, analogous to the VO-TPH model structure. 178 References begin on page 181 Chapter 4 O Figure 4.13: Structure of [(tacn)CoP30ioH2], based on the X-ray crystal structure reported by Haight, Jr. et a/. 59 (tacn = 1,4,7-triazacyclononane). 4.3.10. Comparison to In Vivo Coordination in Bone. Table 4.4 presents the hyperfine coupling constants for the in vivo bone sample and VO-TPH together for comparison. It is clear that both the isotropic and anisotropic coupling constants for the two systems correlate very well, indicating a high degree of structural similarity between the vanadyl complexes. One major difference is the absence of abundant proton coupling in the bone sample while VO-TPH HYSCORE spectra contained 3 sets of cross-peaks. The bone sample, however, did possess an A = 3 MHz coupling constant (see Chapter 3), corresponding to a water molecule. An isotropic coupling of this magnitude is mostly likely due to water coordination in the equatorial plane. It is likely that the H 2 and H 3 protons found in the VO-TPH sample are either not present in the bone sample, or such proton coupling is not homogeneous throughout the sample and hence the cross-peak signals for these protons are not observed. The in vivo coordination structure of vanadyl ions in the bone mineral^? i s modeled well by the VO-TPH sample, and the vanadyl ions are likely found in a similar tridentate, triphosphate coordination state. 179 References begin on page 181 Chapter 4 Table 4.4: Comparison of hyperfine coupling constants obtained by pulsed EPR study of VO-TPH, V O - H A and vanadyl-treated bone. Cross-peaka A, MHz I T | , MHz Pi -15.9/14.0 1.94 P 2 -8.82/7.33 1.48 P3 -2.01/0.88 1.14 PA -16.6/14.6 2.00 PB -8.10/6.33 1.77 PC -1.88/0.04 1.84 Bone P, -14.25/ 12.38 1.87 Bone P 2 -9.19/7.72 1.47 Bone P 3 -3.78/2.40 1.39 a P # = VO-TPH model; P x = V O - H A model. The weak in vivo coupling of 3 MHz could indicate a solid state axial phosphate interaction with the paramagnetic centres, while the two stronger couplings correlate very well with the four equatorial phosphate coupling constants (P2-3, PA-B)- The VO-TPH sample demonstrated that it is feasible for a single triphosphate moiety to generate three distinct vanadyl-phosphate coupling constants in a 1:1 solution complex between triphosphate and V 0 2 + . Hydroxyapatite has also demonstrated to be a suitable model for bone mineral and its coordination to vanadyl ions, although the inhomogeneity leads to somewhat broader peaks in the pulsed EPR spectra. The similarities between the systems presents the possibility that in vivo vanadyl coordination in bone is relatively uniform. The two largest 3 1 P couplings are 180 References begin on page 181 Chapter 4 conserved in VO-TPH, VO-HA, bone, VO-ATP and V O - T F 1 - A T P 4 6 , with only slight variation in the P3 and Pc coupling constants, indicating a common tridentate binding motif for triphosphate moieties to the vanadyl ion. Lastly, the presence of at least one water molecule correlates well with a previous study of vanadyl interactions with the bone mineral model compound hydroxyapatite. 14 Vanadyl ions did not incorporate into the apatitic lattice and could therefore be expected to interact strongly with the surrounding aqueous bilayer.68 In the solid state, however, it appears that some water molecules remain within the H A sample, even after vigourous drying of the sample at elevated temperatures in vacuo. AA. References 1) Tyryshkin, A . M . ; Dikanov, S. A. ; Evelo, R. G.; Hoff, A . J. J. Chem. Phys. 1992, 97, 42. 2) Kofman, V. ; Dikanov, S. A. ; Haran, A. ; Libman, J.; Shanzer, A. ; Goldfarb, D. J. Am. Chem. Soc. 1995, 7/7,383. 3) Dikanov, S. A. ; Bowman, M . K. J. Magn. Reson. Ser. A. 1995,116, 125. 4) Tyryshkin, A . M . ; Dikanov, S. A. ; Reijerse, E. J. J. Magn. Reson. A 1995,116, 10. 5) Poppl, A. ; Kevan, L. J. Phys. Chem. 1996, 100, 3387. 6) Hamstra, B. J.; Houseman, A. L. P.; Colpas, G. J.; Kampf, J. W.; LoBrutto, R.; Frasch, W. D.; Pecoraro, V . L. Inorg. Chem. 1997, 36, 4866. 7) Fukui, K. ; Ohya-Nishiguchi, H. ; Kamada, H. Inorg. Chem. 1997, 36, 5518. 8) Lee, H. C ; Scheuring, E.; Peisach, J.; Chance, M . R. J. Am. Chem. Soc. 1997,119, 12201. 9) Reijerse, E. J.; Tyryshkin, A . M . ; Dikanov, S. A. J. Magn. Reson. 1998, 131, 295. 181 References begin on page 181 Chapter 4 10) LoBrutto, R.; Hamstra, B. J.; Colpas, G. J.; Pecoraro, V . L.; Frasch, W. D. J. Am. Chem. Soc. 1998, 120, 4410. 11) Amano, R.; Enomoto, S.; Nobuta, M . ; Sakamoto, M . ; Tsujioka, R.; Ambe, F. J. Trace Elem. Med. Biol. 1996, 10, 145. 12) Etcheverry, S. B.; Apella, M . C.; Baran, E. J. J. Inorg. Biochem. 1984, 20, 269. 13) Narda, G. E.; Vega, E. D.; Pedregosa, J. C ; Etcheverry, S. B.; Baran, E. J. J. Naturforsch. 1992, 47b, 395. 14) Vega, E. D.; Pedregosa, J. C.; Narda, G. E. J. Phys. Chem. Solids 1999, 60, 759. 15) Nriagu, J. O. in Vanadium in the Environment. Parti: Chemistry and Biochemistry; Nriagu, J. O., Ed.; John Wiley and Sons, Inc.: New York, 1998, p 1. 16) Posner, A. S. Physiol. Rev. 1969, 49, 760. 17) Parfitt, A . M . in Physiology and Pharmacology of Bone; Mundy, G. R. and Martin, T. J., Eds.; Springer-Verlag: New York, 1993, p 1. 18) Cini, R.; Giorgi, G.; Laschi, F.; Sabat, M . ; Sabatini, A. ; Alberto, V . J. Chem. Soc, Dalton Trans. 1989, 575. 19) Buglyo, P.; Kiss, T.; Alberico, E.; Micera, G.; Dewaele, D. Coord. Chem. 1995, 36, 105. 20) Etcheverry, S. B.; Ferrer, E. G.; Baran, E. J. Z. Naturforsch. 1989, 44b, 1355. 21) Sakurai, Ft.; Goda, T.; Shimomura, S. Biochem. Biophys. Res. Commun. 1982,108, 474. 22) Urretavizcaya, G.; Baran, E. J. Z. Naturforsch. 1987, 42b, 1537. 23) Williams, P. A . M . ; Baran, E. J. J. Inorg. Biochem. 1992, 48, 15. 24) Alberico, E.; Dewaele, D.; Kiss, T.; Micera, G. J. Chem. Soc, Dalton Trans. 1995, 425. 25) Posner, A. S.; Perloff, A. ; Diorio, A . F. Acta. Cryst. 1958,11, 308. 182 References begin on page 181 Chapter 4 26) Kay, M . I.; Young, R. A . ; Posner, A. S. Nature 1964, 204, 1050. 27) Posner, A . S. Clin. Orthop. 1985, 200, 87. 28) Blumenthal, N . C ; Posner, A. S. Calc. Tiss. Res. 1973,13, 235. 29) Baran, E. J.; Apella, M . C. J. Mol. Struct. 1980, 61, 203. 30) Oniki, T.; Doi, Y . Calcif. Tissue Int. 1983, 35, 538. 31) Wehrli, B.; Stumm, W. Langmuir 1988, 4, 753. 32) Wehrli, B.; Sulzberger, B.; Stumm, W. Chem. Geol. 1989, 78, 167. 33) Dikanov, S. A. ; Liboiron, B. D.; Orvig, C. J. Am. Chem. Soc. 2002, 124, 2969. 34) Gemperle, C ; Aebli, G.; Schweiger, A. ; Ernst, R. R. J. Magn. Reson. 1990, 88, 241. 35) Boskey, A . L.; Posner, A . S. J. Phys. Chem. 1973, 77, 2313. 36) Boskey, A . L.; Posner, A. S. J. Phys. Chem. 1976, 80, 40. 37) Bonar, L. C.; Grynpas, M . D.; Roberts, J. E.; Griffin, R. G.; Glimcher, M . J. in The Chemistry and Biology of Mineralized Tissues; Slavkin, H. and Price, P., Eds.; Elsevier Science Publishers: Amsterdam, 1992, p 226. 38) Addadi, L.; Moradian-Oldak, J.; Furedi-Mihofer, H. ; Weiner, S.; Veis, A . in The Chemistry and Biology of Mineralized Tissues; Slavkin, H. and Price, P., Eds.; Elsevier Science Publishers: Amsterdam, 1992. 39) Chasteen, N . D. in Vanadyl(IV) EPR Spin Probes: Inorganic and Biochemical Aspects; Berliner, L. J. and Reuben, J., Eds.; Plenum Press: New York, 1981; Vol . 3, p 53-119. 40) Cornman, C. R.; Geiser-Bush, K. M . ; Rowley, S. P.; Boyle, P. D. Inorg. Chem. 1997, 36, 6401. 41) Grant, C. V. ; Cope, W.; Ball, J. A. ; Maresch, G.; Gaffney, B. J.; Fink, W.; Britt, R. D. J. Phys. Chem. B 1999,103, 10627. 183 References begin on page 181 Chapter 4 42) Tyryshkin, A . M . ; Dikanov, S. A. ; Goldfarb, D. J. Magn. Reson. Ser. A 1993,105, 271. 43) Dikanov, S. A. ; Tyryshkin, A. M . ; Bowman, M . K. J. Magn. Reson. 2000,144, 228. 44) Reijerse, E. J.; Dikanov, S. A. J. Chem. Phys. 1991, 95, 836. 45) Mustafi, D.; Telser, J.; Makinen, M . W. J. Am. Chem. Soc. 1992,114, 6219. 46) Buy, C ; Matsui, T.; Andrianambininstoa, S.; Sigalat, C ; Girault, G.; Zimmerman, J.-L. Biochemistry 1996, 35, 14281. 47) Atherton, N . M . ; Shackleton, J. F. Mol. Phys. 1980, 39, 1471. 48) van Willigen, H. J. Magn. Reson. 1980, 39, 37. 49) Mulks, C. F.; Kirste, B.; van Willigen, H. J. Am. Chem. Soc. 1982,104, 5906. 50) Mustafi, D.; Makinen, M . W. Inorg. Chem. 1988, 27, 3360. 51) Kirste, B. FL; van Willigen, H. J. Phys. Chem. 1983, 87, 781. 52) Dikanov, S. A. ; Evelo, R. G.; Hoff, A . J.; Tyryshkin, A. M . Chem. Phys. Lett. 1989,154, 34. 53) Gerfen, G. J.; Hanna, P. M . ; Chasteen, N . D.; Singel, D. J. J. Am. Chem. Soc. 1991,113, 9513. 54) Goldfarb, D.; Bernardo, M . ; Thomann, FL; Kroneck, P. M . FL; Ullrich, V . J. Am. Chem. Soc. 1996,118, 2686. 55) Dikanov, S. A . ; Samoilova, R. I.; Smieja, J. A. ; Bowman, M . K. J. Am. Chem. Soc. 1995,117, 10579. 56) Albanese, N . F.; Chasteen, N . D. J. Phys. Chem. 1978, 82, 910. 57) Morton, J. R.; Preston, K. F. J. Magn. Reson. 1978, 30, 577. 58) Carl, P. J.; Isley, S. L.; Larsen, S. C. J. Phys. Chem. A 2001,105, 4563. 184 References begin on page 181 Chapter 4 59) Haight Jr., G. P.; Hambley, T. W.; Hendry, P.; Lawrence, G. A. ; Sargeson, A . M . J. Chem. Soc., Chem. Commun. 1985, 488. 60) Merritt, E. A. ; Sundaralingam, M . Acta. Cryst. 1981, B37, 1505. 61) Lutsko, V. ; Johansson, G. Acta. Chem. Scand. A 1984, 38, 415. 62) Sabat, M . ; Cini, R.; Haromy, T.; Sundaralingam, M . Biochemistry 1985, 24, 7827 and references therein. 63) Cini, R.; Burla, M . C ; Nunzi, A. ; Polidori, G. P.; Zanazzi, P. F. J. Chem. Soc, Dalton Trans. 1984, 2467. 64) Butenhof, K. J.; Cochenour, D.; Banyasz, J. L.; Stuehr, J. E. Inorg. Chem. 1986, 25, 691. 65) Cini, R.; Cinquantini, A. ; Burla, M . C ; Nunzi, A. ; Polidori, G.; Zanazzi, P. F. Chim. Ind. (Milan) 1982, 64, 826. 66) Lin, I.; Knight, W. B.; Ting, S.-J.; Dunaway-Mariano, D. Inorg. Chem. 1984, 23, 988. 67) Dikanov, S. A . ; Liboiron, B. D.; Thompson, K. H.; Vera, E.; Yuen, V . G.; McNeill , J. H. ; Orvig, C. Am. Chem. Soc. 1999, 121, 11004. 68) Posner, A . S. Clin. Orthop. Rel. Res. 1985, 200, 88. 185 References begin on page 181 Chapter 5 Conclusions and Future Work 5.1. Protein Transport of Insulin-Enhancing Vanadium Complexes The abstraction of the vanadyl ion from the B M O V complex by endogenous serum proteins correlates well to a recent report from our laboratories that studied the biodistribution of 14C-labeled B E O V . 1 Through the use of a radiolabel on the ligand (as opposed to radiolabeled 4 8 V in an earlier study2), Thompson et al. were able to track the biodistribution of the two components (i.e. the ligand and the metal ion) of the vanadium complex. The distribution and pharmacokinetics of l-1 4C-ethylmaltol and vanadium were completely different. Ethylmaltol was rapidly removed from the bloodstream and predominately excreted, while the vanadium was shuttled to its primary sites of accumulation. 1 This result strongly suggests a rapid decomposition of the complex, resulting in free ligand release, at some point post-administration. This thesis has shown that two serum proteins, apo-transferrin (apo-Tf) and albumin (HSA), are capable of removing at least one of the maltolato ligands and therefore effecting the dissociation of the metal ion from its original chelating agents. It is not known definitely, however, whether serum proteins are responsible for the delivery of vanadium to tissues, or represent the active species themselves. It is conceivable that upon production of (VO)2-transferrin, the metal-protein complex could react with cellular receptors to effect the intracellular release of vanadyl ions in a mechanism similar to that used for incorporation of Fe(III) into cells. 3 The relative binding strengths of HSA versus apo-Tf to chelated vanadyl sources have not been elucidated. Studies detailed in this thesis suggest that by presenting a chelated 186 References begin on page 195 Chapter 5 source of vanadyl ions to HSA, the protein's binding ability is actually increased through the formation of a ternary complex (the three components being the V=0 unit, maltol and the protein). Such a complex likely possesses a thermodynamic stability several orders of magnitude higher than the basic interaction of free vanadyl ions with HSA, previously defined by Chasteen and Francavilla4 and Purcell et a/. 5 Without invoking the formation of a mixed ligand species, speciation calculations predict no formation of a vanadyl-HSA complex, a result that is inconsistent with observed EPR spectra. Chasteen and co-workers reported competition studies between HSA and apo-Tf for V O 2 1 - by equilibrium dialysis. 6 They found that most of the available vanadyl ions were bound by apo-Tf. Due to the likely formation of a ternary complex with HSA, the distribution of chelated vanadyl ions would be different. The presence of a strong co-chelating agent in solution may allow HSA to compete with apo-Tf for binding of vanadyl ions. The definitive detection of a ternary complex, however, could not be achieved through the use of conventional cw EPR or difference U V spectroscopies. Preliminary mass spectrometry experiments (using matrix-assisted laser desorption ionization, MALDI) were attempted, however, the resolution of such methods was insufficient to provide differentiation of a VO-HSA species with an additional maltol moiety attached. The maltol anion difference amounts to only 125 mass units, which compared to the 66 kDa protein is only 0.19%. HSA is also notoriously heterogeneous.7 Due to its role as a serum transporter for a wide variety of substrates, it is difficult to obtain a relatively uniform sample of HSA for mass spectrometric study, free from bound fatty acids and other metabolites. This difficulty results in rather broad peaks in the mass spectrum, and a high degree of uncertainty on the parent peak mass. Therefore, at least by M A L D I , ternary versus binary (ma)-VO-187 References begin on page 195 Chapter 5 HSA complexes could not be differentiated. Perhaps through a combination of chromatographic procedures with a higher resolution method (liquid chromatography mass spectrometry, LC-MS), such as electrospray ionization (ESI) coupled with time-of-flight (TOF) detection, the proposed ternary complex could be observed. High resolution EPR spectroscopies such as *H ENDOR or E S E E M could detect the number and type of coupled nuclei in the vicinity of the bound vanadyl ion and hence some insight into the donor types. HSA possesses a single histidine residue at the strong binding site; previous studies have shown that the imidazole functionality plays a role in binding the V 0 2 + ion.4 A study of the interaction of 1 -methylimidazole (ImMe, a model for histidine) with B M O V demonstrated that the formation of a ternary BMOV-ImMe adduct is a thermodynamically-favourable reaction at pH 7.4. This reaction could explain the inconsistent result when 1:1 V 0 2 + : H S A was titrated with maltol. During the titration, no formation of B M O V was detected by EPR. Early in the titration, however, a movement of the bound vanadyl ions from the weak to the strong site was observed. These observations suggest a reorganization of the bound vanadyl ions to the strong site only upon the addition of bidentate chelate. The likely mechanism for such a process maybe that the provision of a chelating agent increases the stability of strong site bound V 0 2 + ions which increases the difference in binding strengths between the strong and weak sites to such a degree that the weak non-specific sites can no longer effectively compete for the vanadyl ions. This conclusion is therefore consistent with formation of a ternary complex, either ma-VO-HSA or B M O V - H S A , at the strong site of HSA only. Ternary interactions of VO(acac)2 were recently examined by Makinen and Brady by ' H ENDOR. 8 They concluded that VO(acac)2 188 References begin on page 195 Chapter 5 reacted with HSA as an intact compound to form a ternary complex of (acach-VO-HSA, interacting with the protein via the vacant axial position. E S E E M could also be used to detect formation of a ternary complex with HSA. The former would be useful in the detection and quantification of ligands, particularly water, in the first coordination sphere. Of particular interest is the detection of equatorial water molecules via strong isotropic and anisotropic 'H coupling constants. If a ternary complex is formed, it is likely that one or two maltolato ligands occupy the equatorial positions. Since all oxygen isotopes either possess 1=0 spin states ( 1 6 0, l s O), or are of insufficient concentration to be accurately detected ( 1 70,1 = 5/2, natural abundance = 0.04%), maltolato oxygen donors are ENDOR and ESEEM silent. Further, the presence of maltolato ligands precludes strong proton coupling via two positions in the first coordination sphere. If, however, water molecules do coordinate to the V 0 2 + ions in either equatorial or axial positions, the protons of these ligands would be easily detected. The relative magnitude of the coupling constants could be used to differentiate between equatorial or axial positions. Alternatively, orientation selection could be used, which in combination with ! H ENDOR and 2 D isotopic substitution, could reveal accurate binding distances and geometries. ENDOR is generally limited to the detection of relatively strong coupling constants (compared to ESEEM) and so i f no water is present in the coordination sphere, ENDOR may yield ambiguous results.9> 10 Perhaps a better approach to these high resolution studies would be the detection and quantification of 1 4 N coupling, for which ESEEM is superior to ENDOR. 1 1 Since imidazole and amine donors have distinct .4 coupling constants to vanadyl ions,12-15 m e number and type of 1 4 N donors could be determined. These studies would, therefore, directly detect the 189 References begin on page 195 Chapter 5 protein coordination to the complex, as opposed to inferring protein binding indirectly by the presence of coordinated water molecules. Coordination of a single nitrogen donor, particularly axially, would provide complementary evidence to the intact complex interaction suggested by Makinen and Brady.8 If more than one nitrogen donor is detected, it is highly likely that HSA causes the dissociation of a single maltolate from the B M O V complex; more than two nitrogen donors infers that the complete decomposition of the B M O V complex has taken place, although such a result would be inconsistent with the EPR studies reported in this thesis. The number of coordinating nitrogens is easily determined from both the number and appearance of peaks in the ID E S E E M spectrum.9 Lastly, such studies could differentiate between axial and equatorial coordination of histidine to the B M O V complex. 1 5 5.1.1. The Ligand Taxi Mechanism The main conclusion from the study of BMOV-protein interactions is that, once introduced into the bloodstream, vanadyl complexes with stabilities on the order of B M O V will likely be delivered to the site(s) of action as identical metabolites. Due to the strong binding capabilities of apo-Tf and HSA, abstraction of the metal ion from vanadium complexes represents a major biotransformation pathway of chelated vanadyl sources. Thus, complexes such as BEOV, VO(acac)2 and its derivatives 1 6 as well as VO(pic)2 and its derivatives 1 7 are all likely delivered to the target sites as the same vanadyl-protein complex (where acac = acetylacetonate; pic = picolinate). This process is referred to as the ligand taxi mechanism, referring to the role served by the ligand set as a method to deliver higher levels of vanadium into the bloodstream from the GI tract. Once the complex is absorbed into the bloodstream, endogenous chelating agents such as HSA and apo-Tf possess sufficient 190 References begin on page 195 Chapter 5 affinities for vanadyl ions so as to abstract them from the administered complex. A recent V biodistribution study from our groupz highlights the similar tissue distributions of 4 8 V O S 0 4 and 4 8 V - B M O V in the organs of the body; administration of 4 8 V - B M O V merely results in higher concentrations being found in the main sites of accumulation. Interactions with apo-Tf, because of its requirement of a synergistic binding anion, almost certainly result in a transformation of this type. The possibility of ternary complex formation with HSA, however, does provide a potential pathway by which an intact vanadium compound could be delivered to the target tissue. The final determination of the existence of such a complex, perhaps by methods suggested above, would be important in order to support or refute this possibility. Further, competition studies between HSA and apo-Tf for binding of chelated sources would examine whether ternary complex formation with HSA is even a factor in vivo, or i f apo-Tf binding dominates the transport mechanism of vanadyl ions, regardless of chelation state. If apo-Tf does in fact dominate the in vivo distribution of vanadyl ions in the bloodstream, much of the current effort to improve the pharmacological effectiveness of insulin-enhancing vanadium complexes is merely improving upon the absorption of these compound from the GI tract only. It would not be altering, as several researchers claim, 16,18-20 m e action at the target site. Future studies of the whole body action of vanadium pharmaceuticals should take into account the strong metal binding capabilities of serum proteins. Results reported herein highlight the dangers of discussing mechanistic aspects of drug action without considering in vivo transport and biotransformation functions. A study to test this hypothesis could involve the administration of a neutral vanadyl complex with a high thermodynamic stability and/or low kinetic lability, such that serum 191 References begin on page 195 Chapter 5 proteins would have difficulty in abstracting the metal ion. To prevent axial interactions, the ligand should also be tetra- or pentadentate, to limit interaction with the sixth coordination position. In vivo speciation studies could be conducted to confirm that the complex survives challenges from endogenous chelating agents, and the anti-diabetic effects could be measured and compared against those of "releasing" complexes such as B M O V . The effects of strong chelation on the anti-diabetic activity of vanadyl complexes could indicate a pharmacological role for serum proteins in the action of insulin-enhancing vanadium compounds. 5.2. Tissue Accumulation of Vanadyl Species The pulsed EPR studies of bone, liver and kidney samples described in Chapter 3 are consistent with the ligand taxi mechanism. Comparison of previous E S E E M studies of liver and kidney from rats treated with VOSO4 versus those treated with BEOV indicates that vanadyl ions from either source are ligated in a similar manner. Isotropic coupling constants from both samples were very similar and are of a magnitude typically found for a coordinated amine. 13,15 Evidence also exists for coordinated water as well.21,22 Bone samples show a complete decomposition of the BEOV complex in order to accommodate at least one water ligand and three different 0- 3 l P donors.23 Studies of the interactions of B M O V with serum proteins suggest this decomposition takes place in the bloodstream, although it is possible such reactions could take place in the tissues themselves. Further, bone represents the long-term storage site for accumulated vanadium and is likely responsible for the sustained glucose-lowering activity after cessation of administration.2 Vanadium subsequently released from bone mineral is not in the form of the administered complex and is likely transported to the site of action by serum proteins. 192 References begin on page 195 Chapter 5 Model studies of in vivo bone coordination have demonstrated that vanadyl ions are likely coordinated in a stable mixed ligand species, consisting of three phosphate donors and up to two water molecules.24 The presence of water in the coordination spheres of both the bone and synthetic hydroxyapatite (HA) samples strongly suggests a surface interaction only, with no incorporation of vanadyl ions into the crystal lattice. The study of VO-TPH confirmed a facial tridentate triphosphate coordination state, previously suggested by Buglyo et al.,25 but never spectroscopically characterized. Until these studies, many researchers described only bidentate coordination of triphosphate moieties to V0 2 + .26,27 The superiority of E S E E M methods for the resolution of small and overlapping coupling constants was clearly demonstrated. Further study of vanadyl-phosphate systems could involve the spectroscopic characterization of a simple solution of vanadyl ions and phosphate at physiological pH. With phosphate anions in large excess, such a solution would represent the most energetically favourable arrangement of phosphates around a vanadyl ion and hence yield 3 1 P coupling constants that are free from the effects of bonded phosphates (as in triphosphate) or distortions present in the solid state (as in hydroxyapatite). Despite the structural limitations imposed by the triphosphate ligand, the VO-TPH model is superior to that of V O - H A for the modeling of the three 3 1 P coupling constants determined in the in vivo sample. Such a conclusion is more striking when one considers that the synthesis of the hydroxyapatite model even parallels the likely in vivo coordination mechanism of V 0 2 + , where vanadyl ions in solution associate with the solid calcium phosphate mineral surface. The VO-TPH model is entirely a solution phase model complex. A key step in the synthesis of V O - H A was the addition of vanadyl ions while the calcium phosphate solution was likely in an amorphous state. The crystal state of the calcium 193 References begin on page 195 Chapter 5 phosphate polymer could influence the coordination of vanadyl ions to the surface, or may perhaps allow incorporation of vanadyl ions into the lattice itself. The study of the effects on the 3 1 P coupling constants by varying the vanadyl administration time could demonstrate the slow change of the amorphous phase to the crystalline H A lattice. Conversely, perfectly crystalline H A would not be expected to be a better model, due to the poor crystallinity and numerous ion vacancies of bone mineral. This thesis has examined the transport and bioaccumulation of both free and complexed vanadyl ions. In general, complexation to ligands of moderate binding strength does not alter the transport or bioaccumulation mechanisms compared to free vanadyl sources such as VOSO4. Significantly, interactions with serum proteins are likely to cause delivery of identical vanadium metabolites to target tissues. Alteration of complex structure likely brings about pharmacokinetic changes that affect the overall amount and rate of vanadium distribution in body tissues after uptake from the gastrointestinal tract. Differences between the absorption characteristics of chelated vanadyl sources vs. salts could be clarified through pulsed EPR studies. The Caco-2 cell line has seen increasing use as a model system for intestinal endothelial cells.28,29 ESEEM and HYSCORE spectroscopies could be applied to in vitro cell samples to determine the chemical form of absorbed vanadium for each compound. In conjunction with accurate vanadium determinations at selected time points, the suspected absorption differences between vanadyl complexes (such as B M O V ) and salts (such as VOSO4) could be examined. Administration of vanadyl compounds, either as vanadyl salts or neutral complexes, results in similar coordination structures in liver and kidney, and formation of a stable vanadyl-phosphate species in bone mineral. Whether this accumulated vanadium represents 194 References begin on page 195 Chapter 5 the percentage of drug metabolites in an active or inactive antidiabetic state remains to be elucidated. Suffice to say, the results reported in this thesis suggest that the absorption difference between chelated and free vanadyl sources accounts for most of the variations observed in the insulin-enhancing action of vanadyl compounds. 5.3. References 1) Thompson, K. H.; Liboiron, B. D.; Sun, Y . ; Bellman, K. D. D.; Setyawati, I. A . ; Patrick, B. O.; Karunaratne, V. ; Rawji, G.; Wheeler, J.; Sutton, K. ; Bhanot, S.; Cassidy, C.; McNeill , J. H. ; Yuen, V . G.; Orvig, C. J. Biol. Inorg. Chem. 2002, in press. 2) Setyawati, I. A . ; Thompson, K. FL; Sun, Y . ; Lyster, D. M . ; Vo, C.; Yuen, V . G.; Battell, M . ; McNeill , J. FL; Ruth, T. J . ; Zeisler, S.; Orvig, C. J. Appl. Physiol. 1998, 84, 569. 3) Lippard, S. J . ; Berg, J . M . Principles of Bioinorganic Chemistry; University Science Books: M i l l Valley, CA, 1994. 4) Chasteen, N . D.; Francavilla, J. J. Phys. Chem. 1976, 80, 867. 5) Purcell, M . ; Neault, J. F.; Malonga, FL; Arakawa, FL; Tajmir-Riahi, H . A . Can. J. Chem. 2001, 79, 1415. 6) Chasteen, N . D.; Grady, J. K.; Holloway, C. E. Inorg. Chem. 1986, 25, 2754. 7) Harford, C ; Sarkar, B. in Handbook of Metal-Ligand Interactions in Biological Fluids: Bioinorganic Chemistry; Berthon, B., Ed.; Marcel Dekker: New York, 1995; Vol . 1, p 405. 8) Makinen, M . W.; Brady, M . J. Biol. Chem. 2002, 277, 12215. 9) Deligiannakis, Y . ; Louloudi, M . ; Hadjiliadis, N . Coord. Chem. Rev. 2000, 204, 1. 10) Dikanov, S. A. ; Tsvetkov, Y . D. Electron Spin-Echo Envelope Modulation (ESEEM) Spectroscopy; CRC Press: Boca Raton, Fl . , 1992. 195 References begin on page 195 Chapter 5 11) Garner, C. D.; Collison, D.; Mabbs, F. E. Metal Ions Biol. Syst. 1995, 31, 617. 12) Fukui, K. ; Ohya-Nishiguchi, H.; Nakai, M . ; Sakurai, H. ; Kamada, H. FEBSLett. 1995, 368, 3\. 13) Fukui, K. ; Ohya-Nishiguchi, H.; Kamada, H. Inorg. Chem. 1997, 36, 5518. 14) Dikanov, S. A . ; Samoilova, R. I.; Smieja, J. A. ; Bowman, M . K . J. Am. Chem. Soc. 1995,117, 10579. 15) LoBrutto, R.; Hamstra, B. J.; Colpas, G. J.; Pecoraro, V . L.; Frasch, W. D. J. Am. Chem. Soc. 1998,120, 4410. 16) Reul, B. A. ; Amin, S. S.; Buchet, J.-P.; Ongemba, L. N . ; Crans, D. C ; Brichard, S. M . Br. J. Pharmacol. 1999, 126, 467. 17) Sakurai, FL; Fujii, K.; Watanabe, FL; Tamura, H. Biophys. Res. Commun. 1995, 214, 1095. 18) Willsky, G. R.; Goldfme, A . B.; Kostyniak, P. J.; McNeill , J. FL; Yang, L. Q.; Khan, H . R.; Crans, D. C. J. Inorg. Biochem. 2001, 85, 33. 19) Sakurai, FL; Kojima, Y . ; Yoshikawa, Y. ; Kawabe, K.; Yasui, H . Coord. Chem. Rev. 2002,226, 187. 20) Rehder, D.; Pessoa, J. C ; Geraldes, C. F. G. C ; Castro, M . ; A . , M . C.;-Kabanos, T.; Kiss, T.; Meier, B.; Micera, G.; Pettersson, L.; Rangel, M . ; Salifoglou, A. ; Turel, I.; Wang, D. J. Biol. Inorg. Chem. 2002, 7, 384. 21) Atherton, N . M . ; Shackleton, J. F. Mol. Phys. 1980, 39, 1471. 22) Tyryshkin, A . M . ; Dikanov, S. A. ; Evelo, R. G.; Hoff, A. J. J. Chem. Phys. 1992, 97, 42. 23) Dikanov, S. A. ; Liboiron, B. D.; Thompson, K. FL; Vera, E.; Yuen, V . G.; McNeill , J. FL; Orvig, C. J. Am. Chem. Soc. 1999,121, 11004. 196 References begin on page 195 Chapter 5 24) Dikanov, S. A. ; Liboiron, B. D.; Orvig, C. Am. Chem. Soc. 2002, 124, 2969. 25) Buglyo, P.; Kiss, T.; Alberico, E.; Micera, G.; Dewaele, D. J. Coord. Chem. 1995, 36, 105. 26) Buy, C.; Matsui, T.; Andrianambininstoa, S.; Sigalat, C.; Girault, G.; Zimmerman, J.-L. Biochemistry 1996, 35, 14281. 27) Mustafi, D.; Telser, J.; Makinen, M . W. J. Am. Chem. Soc. 1992,114, 6219. 28) Thompson, K. H.; Tsukada, Y . ; Xu, Z.; Battell, M . ; McNeill , J. H. ; Orvig, C. Biol. Trace Elem. Res. 2002, 85, 31. 29) Hidalgo, I. J.; Raub, T. J.; Borchardt, R. T. Gastroenterology 1989, 96, 736. 197 References begin on page 195 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061319/manifest

Comment

Related Items