Chapter 1. A brief history of the use of ionizing radiation for cancer therapy is presented, as the motivation for modern clinical investigations on alpha therapy and the use and development of bifunctional chelators in targeted medicine. Recent studies on targeted therapy with alpha-emitters, such as 223Ra and 227Th, are described. The chemistry of Th(IV) is presented with a survey of important coordination compounds, with special emphasis on those developed by the Raymond group.

Chapter 2. The rational design of a ligand as a suitable bifunctional chelator for 227Th therapy is described in the context of ligands developed by the Raymond group for actinide sequestration. A novel Φ ligand topology incorporating macrocyclic and pendant terephthalamide binding groups aims to address the kinetic and thermodynamic requirements of targeted alpha therapy. The syntheses of the model ligands Φ(2,2)moeTAM and Φ(3,3)moeTAM are detailed, from which a derivative of Φ(2,2)moeTAM is developed. Functionalization of a pendant terephthalamide of the model ligand gives the asymmetric bifunctional chelate Φ(2,2)NBuTAM via the strategic use of protecting groups.

Chapter 3. The evaluation of the solution thermodynamic behavior of Φ(2,2)moeTAM is extensively detailed. The protonation constants of the ligand and its affinity for Th(IV) is measured by spectrophotometric titration. The thorium complex was found to have a remarkably high thermodynamic stability, with a formation constant of 1054, surpassing the stabilities of previously measured complexes. Efforts to elucidate the effects of the Φ topology on the coordination of Th(IV) using thermodynamic studies of the linear analog 3,4,3-LiMeTAM as well as the shape analyses of the crystal structure of Th[Φ(2,2)moeTAM] are described. The surprising geometry adopted by Φ(2,2)moeTAM in the thorium complex relative to its structure as the free ligand is further investigated with DFT studies.

Chapter 4. While the measurement of the thermodynamic parameters of the thorium complex provides an assessment of its inertness, its formation kinetics are also of paramount importance due to the inherent time sensitivity of radiotherapeutic agents. Difficulties encountered in preliminary kinetic experiments with existing octadentate ligands and Φ(2,2)moeTAM lead to the use of indirect kinetics. Dye-displacement kinetic studies with the Φ ligands, 3,4,3-LiMeTAM, and the aminocarboxylic acid ligands provide a useful qualitative comparison of the rates of complexation of these ligands with Th(IV). Φ(2,2)moeTAM is observed to be a much more rapid chelator than the prevailing DTPA and DOTA, and its association (a second-order reaction with a rate constant of k2 = 1.8(1) x 104 M-1s-1) can be directly measured using stopped-flow kinetics.

Chapter 5. The application of Φ(2,2)moeTAM toward the coordination of the lanthanide analogs cerium and praseodymium is presented. The crystal structure of the Ce(IV) complex with Φ(2,2)moeTAM is isomorphous to that of the Th(IV) complex, prompting solution thermodynamic and electrochemical studies to investigate the stability of the Ce(IV) complex. Spectrophotometric titration indicates an unprecedented formation constant on the order of 1061, and a Nernstian shift of the Ce(III)-(IV) couple of 2.0 V upon complexation is observed by cyclic voltammetry. The strong preference of Φ(2,2)moeTAM for the +4 oxidation state relative to the +3 is further highlighted by the solution thermodynamics of Pr(III).

Abstract

Siderophore-mediated Competition for Iron Between Humans and Pathogenic Bacteria

by

Allyson Kawile Sia

Doctor of Philosophy in Chemistry

University of California, Berkeley

Professor Kenneth N. Raymond, Chair

Iron is the most abundant and essential transition metal used by many living organisms. To understand the means by which organisms acquire, transport and use iron expands the understanding we humans have of life. Organisms evolved multiple strategies to solubilize iron hydroxides from the environment and to increase iron bioavailability. Most bacteria, fungi and plants solubilize iron from the environment by synthesis and secretion of siderophores, or small molecules that have a high affinity for ferric iron. Alternative iron acquisition strategies are often employed simultaneously which include piracy of iron stores from other organisms and exploitative uptake of nonnative siderophores and hemes.

Generally this work is meant to provide detailed information for mechanisms and strategies involved in iron competition between pathogenic bacteria and human hosts. Detailed structural and kinetic investigations of iron acquisition pathways in the Bacillus cereus group of bacteria are discussed in Chapters 1-4. The B. cereus group of bacteria includes threatening pathogens such as the anthrax pathogen (B. anthracis) and enterotoxic bacteria (B. cereus). Iron complexes of molecules as small as citrate and as large as a polymeric bacterial capsule are probed for structural and kinetic data related to iron acquisition and transport. Studies of the iron binding ability and subsequent transport of synthetic analogs of the B. anthracis stealth siderophore petrobactin are also included. Taken together, these studies shed light on multiple mechanisms and structure-function relationships necessary for the pathogenic bacteria to obtain an adequate iron supply and regulate virulence.

Bacillus bacteria can infect humans and compete with the host iron supply. The consequences of iron piracy by Bacillus anthracis are especially likely to be fatal. Crucial to the human defense against bacterial infections is the immune protein siderocalin. Siderocalin is antibacterial and limits bacterial growth by intercepting bacterial siderophores. Chapter 5 highlights siderocalin and its recently established alternative role as an iron transport protein. For siderocalin to transport iron, an endogenous siderophore equivalent is required. A putative endogenous, or `mammalian' siderophore, is evaluated by solution thermodynamic analyses, inorganic spectroscopy and biochemical assays. The multipurpose role of siderocalin could have important clinical impact, and it is important that the foundation of these studies are clear and consistent at the chemical level. The elucidated roles of siderocalin in humans convey key iron-restricting strategies by humans to compete with invading pathogens for iron. The collection of chapters in this work portrays the competition for iron as dynamic, aggressive and evolving by both humans and invading bacterial pathogens and is meant to provide a snapshot of the competition in current times.

Bacterial pathogens acquire the iron they need for survival and growth in a host by using siderophores. The structures of siderophores are specialized for binding ferric ion with high affinity. Siderophore structures are also specialized to specifically interact with the proteins that mediate siderophore function. These proteins include the bacterial proteins involved in siderophore uptake and utilization, as well as host proteins that inhibit bacterial iron acquisition by intercepting siderophores.

The interactions between siderophores and iron underlie biological function. The fundamental coordination chemistry of catecholate and hydroxamate siderophores affects protein interactions during siderophore uptake and host defense. Chapter 1 reviews previous studies on siderophore coordination chemistry and the effects it has on protein interactions and biology with emphasis on research from the Raymond laboratory.

The human protein siderocalin defends against siderophore-mediated iron acquisition. Human siderocalin recognizes the metal center of catecholate siderophores, including enterobactin, with high affinity. Several pathogens modify the catecholate metal binding units to make stealth siderophores that are not recognized by human siderocalin. As presented in Chapter 2, the pathogens Vibrio fluvialis and Vibrio cholerae use the siderophores fluvibactin and vibriobactin, respectively, which have catechol-oxazoline metal-binding units. The catechol-oxazoline had been proposed to be a stealth mechanism, but the results herein presented clearly demonstrate that it is not a stealth mechanism. Catechol-oxazoline coordinates iron in either a catecholate mode or a phenolate-oxazoline mode. The phenolate-oxazoline mode is not recognized by siderocalin while the catecholate mode is. Siderocalin stabilizes the catecholate mode sufficiently to cause a shift from the phenolate-oxazoline mode at physiological pH. The high affinity recognition of the ferric triscatecholate metal center allows siderocalin to defend against iron acquisition by a large number of bacterial siderophores.

Siderocalin defense has been identified in hosts other than humans. Ex-FABP is a protein found in chickens that has high structural homology to human siderocalin. Chapter 3 reports that Ex-FABP binds many of the same siderophores that are bound by human siderocalin. Unlike human siderocalin, the binding pocket of Ex-FABP is expanded to allow it to bind glucosylated enterobactin. Many of the pathogens specific to chickens use glucosylated enterobactin siderophores known as salmochelins. Salmochelins are stealth siderophores in humans. The siderophore recognition of Ex-FABP demonstrates that siderophore binding proteins may be a general host defense mechanism, and that siderocalins have adapted to the pathogens most frequently encountered by the host. Siderocalin defense and stealth siderophores are at the edge of the arms race for iron.

Siderophores carry iron into a bacterial cell through specific transport systems. Once inside the cell, the iron must be removed from the siderophore. Bacteria that use ferric enterobactin remove the iron by hydrolyzing the backbone with an esterase followed by reduction of the ferric ion. Hydrolysis is necessary because the high stability of intact ferric enterobactin prevents biological reduction and iron release. V. cholerae had been reported to use ferric enterobactin, but it does not have an esterase to hydrolyze the backbone. Chapter 4 reports that V. cholerae does not use intact ferric enterobactin, but that it most likely uses ferric complexes of enterobactin hydrolysis products.

Uptake of ferric siderophores relies on specific cell membrane receptors. Many siderophore receptors recognize the apo-siderophores as well as the ferric complexes. Binding apo-siderophores does not directly deliver iron to the bacteria, but it plays a role in the uptake mechanism. Chapter 5 describes YxeB, the ferrichrome/ferrioxamine receptor of Bacillus cereus. YxeB transports the siderophores using a Gram-positive siderophore-shuttle in which metal exchange between a ferric siderophore and the bound apo-siderophore is facilitated by the receptor. Metal exchange is not required for uptake, but the siderophore-shuttle is faster than transport without metal exchange.

Metal exchange, iron release, and the sterics and electronics of the metal center are coordination chemistry principles that influence the interactions between siderophores and proteins. Proteins usher siderophores through the biological functions of removing iron from the host, passing through the bacterial membrane, and releasing iron to the cell. Therefore, siderophores act as the intermediaries between ferric ion and the biology of bacterial iron uptake.

Magnetic resonance imaging (MRI) is one of the most powerful diagnostic techniques at the disposal of the medical community. Its success in the clinic, with 75 to 90 million scans performed worldwide annually, can be attributed in part to the use of injectable contrast agents to improve signal differentiation between healthy and pathological tissue. These contrast agents primarily use Gd(III) as the paramagnetic metal ion to induce contrast. With seven unpaired electrons, Gd(III) has the most paramagnetic character of any nonradioactive element. Aqueous Gd(III), however, is highly toxic; hence contrast agents use chelators to encapsulate the Gd(III) ion, which protects the patient from from the Gd(III) ion.

While these chelators are necessary, they greatly decrease the relaxivity of the current commercial contrast agents. Commercial contrast agents are similar in that they are heteroatom chelators (N, O) and octadentate coordination, leaving only one open site for water coordination. Additionally, given their steric bulk, the water exchange mechanism with bulk solvent is a laboriously hindered dissociative mechanism. These factors contribute to the low efficiency of these Gd(III) complexes, as measured in relaxivity. Diagnostic scans typically inject 8-10 g of these complexes to achieve sufficient signal.

Hydroxypyridinone (HOPO) chelators have emerged as a superior alternative to current commercial compounds. Using a tris(2-aminoethyl)amine (TREN) capping moiety, three bidentate HOPO chelators form a hexadentate ligand. These TREN-tris-HOPO ligands leave multiple open sites for water coordination and exhibit rapid water exchange with bulk solvent, due to their reduced steric bulk and associative exchange mechanism. These ligands use all-oxygen-donor chelators, capitalizing on the oxophilicity of Gd(III) to form highly stable complexes.

From this superior family of chelators, a variety of approaches can be used to develop the next generation of MRI contrast agents. Increasing molecular weight and tumbling time has been a strategy for increasing relaxivity and efficiency of MRI contrast agents. Through macromolecular conjugation, relaxivity is readily increased; simultaneously, these macromolecules provide the potential for building multimodal and multifunctional diagnostic and therapeutic agents. The potential applications for this class of materials are further increased with the addition of targeting functionality. These agents must have the ability to be fully and rapidly excreted and have facile and uniform large-scale syntheses to be candidates for the clinic.

The esteramide (EA) dendrimer is one such macromolecular platform. With eight sites for contrast agent conjugation, the esteramide dendrimer readily loads many distinct HOPO ligands with multiple lanthanides for multimodal imaging. With close to 40 kDa of polyethylene glycol units, the Gd-HOPO-EA macromolecular architecture is highly soluble and biocompatible. Furthermore, the ester core of the dendrimer is degradable under in vivo conditions, easing renal clearance with four smaller moieties. The superior properties of this system inspired investigation into a variety of other macromolecular systems.

Porous silica mesoparticles provide a rigid architecture that is much larger than other macromolecules evaluated and can hold greater than 108 small molecule MRI contrast complexes. The surfaces of these particles are readily functionalized and suitable for conjugation with most small molecule MRI contrast agents. These structures use a nontoxic silica infrastructure and are excreted renally despite their large size, making them viable candidates for further in vitro and in vivo study.

Gold nanoparticles (AuNP) as a solid-support system have the most potential for use as multifunctional diagnostic and therapeutic compounds. AuNP have been long used for enhancing computed tomography (CT) imaging and have recently emerged as a cancer therapeutic when their structure is irradiated. These compounds are readily synthesized in large scales and have loading sites that are close together to hold multi-tethered Gadolinium-HOPO systems for multifunctional imaging.

Using a variety of macromolecules to capitalize on the structural relationship between relaxivity and size, per-Gd and per-macromolecule-Gd relaxivity have been increased dramatically at clinically and physiologically relevant conditions. These improvements show that the combination of carefully designed macromolecules with excellent HOPO chelators produces an ideal MRI contrast agent for the clinic of the future.

The work described in this thesis examines the noncovalent interactions, dynamics and structure of a self-assembled supramolecular host-guest system. The highly-charged, water soluble supramolecular assembly is able to selectively bind cationic molecules to the host interior and exterior and mediate the physical properties and chemical reactivity of bound guests. Herein, physical organic chemistry in the context of this supramolecular system is used to elucidate some of the fundamental host-guest interactions that underlie guest binding and reactivity.

Chapter 1. First, an overview of different noncovalent interactions and their importance to biological systems is presented. Selected examples of synthetic supramolecular host-guest systems are then reviewed, focusing on examples that illustrate how chemists have used noncovalent interactions to construct complex host architectures and affect guest chemical reactivity. Finally, the [Ga4L6]12- supramolecular assembly, which is the topic of Chapters 2 - 5, is introduced and previous work examining the host-guest chemistry of this system is briefly reviewed.

Chapter 2. Ortho-substituted benzylphosphonium guest molecules are used to quantitatively probe the steric effects of confinement within the [Ga4L6]12- host. Encapsulated guest bond rotational barriers and tumbling rates are measured in different solvents and at elevated external pressures. These studies reveal that despite the flexibility of the host assembly, guest molecules experience significant steric confinement on the host interior and their bond rotational barriers are increased by up to 6 kcal/mol. Significant solvent and pressure effects on the bond rotational rates are also observed; these suggest that the host cavity is smaller or less flexible in organic than in aqueous solution, and that internal solvent pressure is responsible for the observed changes in ligand framework flexibility. The apparently smaller cavity volumes in organic solvents are further supported by NOE distance measurements, which show shorter average host-guest distances in organic, as compared to aqueous, solutions.

Chapter 3. Building on the solvent effects observed in Chapter 2, some additional, qualitative solvent effects are first presented; these show that guest binding and exchange is also sensitive to bulk solvent and that N,N-dimethylformamide can act as weakly bound guest. The second part of this chapter presents a quantitative study of the solvent effects on the thermodynamics and kinetics of guest binding and exchange in the [Ga4L6]12- host. No correlation between the guest binding or exchange parameters across different solvents are observed, illustrating the complexity of host and guest solvation in these highly-charged supramolecular systems.

Chapter 4. A brief literature review of isotope effects on guest binding and exchange in supramolecular host-guest systems is first presented. The measurement of kinetic and equilibrium isotope effects on guest binding and exchange in the [Ga4L6]12- assembly are then described. Protiated guests are found to be more strongly bound to both the host interior and exterior, with significantly larger equilibrium isotope effects observed for C-H/D bonds that can participate in cation-π interactions. DFT-level computations reveal that the equilibrium isotope effects arise from changes in low frequency C-H/D vibrational motions upon guest association. Kinetic isotope effects during the guest ejection process are also observed. A general model to explain both equilibrium and kinetic isotope effects, based only on changes in C-H/D vibrational force constants and zero-point energies is presented. The observation of significant isotope effects on both guest binding and exchange demonstrates the remarkable sensitivity of the [Ga4L6]12- host assembly to even the most subtle changes in guest architecture.

Chapter 5. Guest molecules encapsulated in [Ga4L6]12- experience dramatic changes in their 1H nuclear magnetic resonance (NMR) chemical shifts due to the unique magnetic environment of the host interior. Gauge-independent atomic orbital NMR chemical shift calculations are carried out and compared with experiment to provide information about host-guest conformation and structure.