WO2000056300A2

WO2000056300A2 - Synthesis and structure of metallocene compounds and their interactions with lipid membranes

Info

Publication number: WO2000056300A2
Application number: PCT/US2000/007067
Authority: WO
Inventors: Phalguni Ghosh; Ann T. Kotchevar; Fatih M. Uckun
Original assignee: Parker Hughes Institute
Priority date: 1999-03-19
Filing date: 2000-03-17
Publication date: 2000-09-28
Also published as: WO2000056300A3; AU3756000A

Abstract

Metallocene compounds of Formula (I) useful for modulating the permeability of a lipid membrane, wherein Formula I is where M is a metal ion, Cp is unsubstituted or substituted cyclopentadienyl, and R1 and R2 are together a bidentate ligand are described. Some suitable compounds have a unique structural requirement, particularly the hydrophobicity, planarity, and rigidity of the coordinated ancillary ligands, which alter the membrane of intercalation. Some suitable compounds can modulate the permeability of a lipid membrane through oxidization of lipids without generating hydroxyl radicals. Additionally, methods of using such compounds and pharmaceutical compositions including such compounds.

Description

SYNTHESIS AND STRUCTURE OF METALLOCENE COMPOUNDS AND THEIR INTERACTIONS

WITH LIPID MEMBRANES

This application claims the benefit of Provisional Application Serial No. 60/125,144, filed 19 March 1999.

Field of the Invention

The invention relates to metallocene complexes, and pharmaceutically acceptable salts thereof, pharmaceutical compositions including such metallocene complexes and methods of use thereof. More particularly, the invention relates to metallocene complexes for use in modulating the permeability of a membrane.

Background

Biological membranes are of fundamental importance to living cells by serving as selective barriers for transport and boundaries for energy and information (Gennis, R. B. Biomembranes, Molecular Structure and Function; Springer-Nerlag: New York, 1989; p 85). Observations of simple model membrane systems, like those composed of vesicles or liposomes, have proven experimentally useful in offering insights into the fundamental mechanisms of biological membrane functions (Fendler, J. H. Membrane Mimetic Chemistry; Wiley: New York, 1982; pp 3-4).

Permeation of liposomal membranes has been effected by peptides, polyether compounds, and surfactants. Mechanisms proposed for the permeation include pore formation and development of localized inverted micelle structures within the lipid bilayer. Polypeptides (Langs, D. A. Science 1988 241, 188; and O'Connell, et al. Science 1990, 250, 1256; Galvez, et al. J Bacteriol. 1991, 886; and Matsuzaki, et al. Biochemistry 1977, 36, 9799), macrocyclic ionophores (Hartsel, et al. Biochemistry 1988, 27, 2656; Fuhrhop, et al. J. Am. Chem. Soc. 1988 110, 6840; Lindloy, et al. J. Am. Chem. Soc. 1990, 112, 3659; Bolard, et al. Biochemistry 1991, 30, 5707; and Tsukube, et al. Inorg. Chem. 1994, 33, 2984.), and polymeric crown ethers (Fyles, et al. J Am. Chem. Soc. 1993, 115, 12315; Carmichael, et al. J Am. Chem. Soc. 1989, 111, 767; Voyer, et al. J Am. Chem. Soc. 1995, 117, 6599; Kragten, et al. J. Chem. Soc. Chem. Commun. 1985, 1275; Neevel, et al. Tetrahedron Lett. 1984, 21 2263; and Pregel, et al. J. Chem. Soc, Per kin Trans. 2 1995, 417) form artificial channels by spanning the bilayer. To form such a channel, it is optimal for the compound to have hydrophilic end groups, lipophilic portions in the channel, and appropriate infrastructural size to span the bilayer, which is approximately 40N (Fyles, et al. J. Am. Chem. Soc. 1993, 115, 12315; and Fuhrhopet al. J Am. Chem Soc. 1984, 106, 4643.). Low levels of surfactants and polymers have been found to increase the permeability of liposomes, without destroying the membrane (Hunt, G. R. A. FEBS Lett. 1980, 119, 132; Jayasuriya, et al. J Am. Chem Soc. 1990, 112, 5844; Jayasuriya, et al. J. Am. Chem Soc. 1990, 112, 5851; Νagawa, et al. J. Am. Chem. Soc. 1992 114, 1668; Νaka, et al. J. Am. Chem. Soc. 1993, 115, 2278; and Scrimin, et al. J. Am. Chem. Soc. 1998, 120, 1179). Triton X-100 is postulated to form inverted micelle structures within the bilayer, promoting both permeability and membrane fusion (Hunt, G. R. A. FEBS Lett. 1980, 119, 132.). The model proposed by Νagawa et al. (J. Am. Chem. Soc. 1992 114, 1668) for surfactants, bolaphiles, and polymers implies that the leakage of dye from the vesicle is due to aggregates of surfactant causing membrane rupture. In addition, surfactants designed with rigid, wedge-shaped hydrophobic units show increased ability to release the dye encapsulated in osmotically stressed vesicles (Νaka, et al. J. Am. Chem. Soc. 1993, 115, 2278). Recently, Scrimin et al. (J Am. Chem. Soc. 1998, 120, 1179) reported that the addition of lipophilic amines affects the permeability of vesicular membranes by forming "leaky patches" in the membrane. Addition ofthe long-chain amines promotes the concentration-dependent leakage of a fluoroescent dye without the destruction ofthe vesicles.

Several metal salts also have been shown to perturb the bilayer structures of liposomal membranes. Cu²⁺ and Al³⁺ induce changes in the permeability of membranes and cause damage due to Fe -initiated lipid peroxidation (Ohsumi, et al. J. Bacteriol. , 1998, 170, 2676; Zhang, et al. Acad. Med. Okayama 1994, 48, 131; Gutteridge, et al. Biochim. Biophys. Acta 1985, 835, 441; Quinlan, et al. Biochim. Biophys. Ada 1988, 962, 196; and Deleers, et al. Biochim. Biophys. Acta 1985, 813, 195). Sc³⁺, Ga³⁺, In³⁺, Y³⁺, and Be²⁺ demonstrate the ability to promote aggregation, fusion, permeabilization, and membrane rigidification (Verstraeten, et al. Arch. Biochem. Biophys. 1995, 322, 284; and Verstraeten et al. Arch. Biochem. Biophys. 1997, 555, 121). These effects correlate with their capacity to induce Fe ⁺-initiated lipid peroxidation, prompting the hypothesis that the metal ions alter the membrane by creating rigid clusters where the hydrocarbon chains are closer together, thus increasing the susceptibility ofthe lipids to peroxidation (Verstraeten et al. Arch. Biochem. Biophys. 1997, 338, 121). Clearly, many compounds exist that can modulate membrane permeability through many different mechanisms of action. However, because biological membranes are of fundamental importance to living cells, the development and use of new compounds for modulating membrane permeability are needed.

Summary of the Invention

Applicants have developed metallocene complexes useful for modulating the permeability of a lipid membrane. In one aspect, the invention relates to novel metallocene complexes, and pharmaceutically acceptable salts or esters thereof. In another aspect, the invention relates to pharmaceutical compositions comprising one or more of such metallocene complexes, and pharmaceutically acceptable salts thereof. In another aspect, the invention relates to a method of modulating the permeability of a lipid membrane by contacting the membrane with the one or more metallocene complex thereby inserting the one or more metallocene complex in the membrane. In another aspect, the invention relates to a method of modulating the permeability of a lipid membrane by oxidizing membrane lipids in the presence ofthe one or more metallocene complex, without the generation of hydroxy radicals. In another aspect, the invention relates to a method of modulating the permeability of a lipid membrane by oxidizing membrane lipids through the generation of free radicals in the presence ofthe one or more metallocene complex. In another aspect, the invention relates to increasing the uptake or transport of a therapeutic agent across a lipid membrane by co-administering the therapeutic agent with a metallocene complex that modulates the permeability of a membrane. In some embodiments, the lipid membrane is a lipid bilayer. Preferably, the lipid membrane is a liposome or artificial skin.

Suitable metallocene complexes, for example, include metallocene complexes of Formula I, or a pharmaceutically acceptable salt or ester thereof.

Formula I

M is a transition metal atom or ion, preferably V or Ti. Cp is unsubstituted cyclopentadiene or cyclopentadiene substituted with one or more substituents that can be the same or different, and are preferably selected from Cι-₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably, Cp is unsubstituted cyclopentadiene.

Ri and R together are a bidentate ligand.

Suitable bidentate ligands include N,N'; O,O'; N,O; and S,S' bidentate ligands. Examples of suitable N, N' bidentate ligands include diamines and other such known suitable N, N' bidentate ligands. Examples of diamines include bipyridyl, derivatives of bipyridyl, bridged bipyridyl, such as phenanthroline, derivatives of phenanthroline, and other such compounds. Examples of suitable N, O bidentate ligands include amino acids and hydroxylamino type groups. Examples of suitable O, O' bidentate ligands include dicarboxylate, 2-hydroxyacetophenone, acetylacetone and catechol type groups. Examples of suitable S, S' bidentate ligands include diethyldithiocarbamate.

Preferred examples of bidentate ligands include: phenanthroline; bipyridyl; bridged bipyridyl; N-phenyl benzohydroxamato; N, N-diethyldithiocarbamato; acetylacetonato; catacholato; and acetophenone. Each of these bidentate ligands can be unsubstituted or substituted with one or more substituents selected from Cι-₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine.

Specific examples of suitable metallocene complexes include chelated vanadocene complexes or chelated titanocene complexes. Specific examples of chelated vanadocene complexes and chelated titanocene complexes include Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂V(cat), Cp₂V(Et₂ (dtc)), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3-metm), Cp₂V(3-mmm), Cp₂V(3-mcm), Cp₂V(3-mnm), pharmaceutically acceptable salts thereof, and mixtures thereof.

In some embodiments, the metallocene complexes provide for rapid, but temporary modulation ofthe permeability of a lipid membrane.

Brief Description of the Drawings

Figure 1 shows the ability of various doses of metallocenes to cause permeation of CF out of liposomes. Figure 1A shows the ability of individual vanadocene complexes, Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), and Cp₂V(cat), to cause permeation of CF out of PC liposomes. Figure IB shows the ability of individual vanadocene complexes, Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp V(acac), and Cp₂V(cat), to cause permeation of CF out of PG liposomes. Figure IC shows the ability of individual vanadocene, Cp₂V(phen) and Cp₂V(bpy), and titanocene, Cp₂Ti(phen) and Cp₂Ti(bpy), complexes to cause permeation of CF out of PC liposomes.

Figure 2 shows the ability of various doses of Cp₂V(bpy) to cause permeation of CF out of liposomes over time.

Figure 3 shows the absorbance at 300 nm of PC liposomes incubated with various concentrations of individual vanadocene compounds, Cp₂V(phen), Cp₂V(bpy), and Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), and Cp₂V(cat).

Figure 4 shows the polarization of a fluorescent probe diphenyl hexatriene in PC liposomes incubated with various concentrations of individual vanadocene compounds, Cp₂V(phen), Cp₂V(bpy), Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), and Cp₂V(cat).

Figure 5 shows the ability of various doses of metallocenes to peroxidize lipids of liposomes. Figure 5 A shows the ability of individual vanadocene complexes, Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), and Cp₂V(cat), to peroxidize lipids of PC liposomes. Figure 5B shows the ability of individual vanadocene complexes, Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), and Cp₂V(cat), to peroxidize lipids of PG liposomes. Figure 5C shows the ability of individual vanadocene, Cp V(phen) and Cp₂V(bpy), and titanocene, Cp₂Ti(phen) and Cp Ti(bpy), complexes peroxidize lipids of PC liposomes.

Figure 6 is a ORTEP drawing of Cp₂V(phen) with 30% probability anisotropic displacement parameters at room temperature.

Figure 7 is a ORTEP drawing of Cp₂V(bpy) with 30% probability anisotropic displacement parameters at room temperature.

Figure 8 shows the UV-Vis spectrum of Cp₂V(phen) (2.15 x 10^~5 M) in acetonitrile; inset shows the spectrum in the visible region (5.88 x 10 M); arrows indicate the positions of weak shoulders. Figure 9 shows the X-band EPR spectrum of 1 xl0^~3 M Cp₂V(phen) in PBS; temp = 298 K; pH = 7.2; t = 5 min; v = 9.5597 GHz; modulation amplitude = 3.984 G at 100 kHz; and receiver gain = 5 x 10³.

Figure 10 shows the cyclic voltammograms of Cp₂V(bpy) in acetonitrile (0.1 M TABP); scan rate = 200 mV/s; referenced to Cp₂Fe^+/0 in acetonitrile.

Figure 11 shows the chemical structures ofthe following examples of metallocene complexes: Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(Et₂(dtc)), Cp₂V(acac), Cp₂V(cat), Cp₂Ti(phen), and Cp₂Ti(bpy).

Figure 12 shows the chemical structures ofthe following examples of metallocene complexes: Cp₂V(acac), Cp V(3-metm), Cp₂V(P3-mmm), Cp₂V(3-mcm), Cp₂V(3- mnm), and Cp₂V(Cl₂).

Figure 13 shows a cyclic voltammogram overlay of the V^IV/V^V redox couple in acetonitrile of [Cp₂V(acac)][OTfJ compounds substituted in the 3 -position (1-4) at a scan rate = 0.2 V/s, where potentials are referenced vs. Cp₂Fe ⁺ in acetonitrile.

Figure 14 shows a time course ofthe lipid peroxidation reaction initiated by 400 μM 1 in the presence (•) and absence (■) of oxygen.

Figure 15 shows an EPR spectrum of Cp₂V(acac): Figure 15(a) shows a spectrum of Cp₂V(acac) taken in the liposome solution within 90 s; Figure 15(b) shows a spectrum of Cp₂V(acac) taken in the liposome solution after 90 min incubation at 37 °C exposed to oxygen; and Figure 15(c) shows a spectrum of Cp₂V(acac) taken in the liposome solution after 90 min incubation at 37 °C in the absence of oxygen.

Figure 16 shows the effect of pH on the extent of lipid peroxidation after 90 min incubation at 37 °C when initiated by 200 μM 1 (•) and 6 (■).

Figure 17 shows the effect ofthe addition of hydrogen peroxide on the extent of lipid peroxidation after 90 min incubation at 37 °C when initiated by 200 μM 1 (•) and 6

(^■)•

Figure 18 shows the EPR spin-trapping experiments with Cp₂V(acac) and Cp₂V(Cl₂): Figure 18(a) shows the EPR spectra of Cp₂V(acac) taken in the liposome solution in the presence ofthe spin-trapping agent, POBN, after incubation at 37 °C for 90 min; and Figure 18(b) shows the EPR spectra of Cp₂V(Cl₂) taken in the liposome solution in the presence ofthe spin-trapping agent, POBN, after incubation at 37 °C for 90 min. Figure 19 shows the correlation ofthe V /V redox potential and lipid peroxidation rates with the Hammett constant (σ) for the 3-substituted [Cp₂V(acac)][OTfj complexes, Cp₂V(acac), Cp₂V(3-metm), Cp₂V(P3-mmm), Cp₂V(3- mcm), and Cp₂V(3-mnm).

Figure 20 shows the proposed pathway for the activation of a vanadocene superoxo compound from Cp₂V(acac).

Detailed Description of the Invention

The present invention concerns the finding that certain metallocenes modulate the permeability of a membrane. Thus, as compounds having membrane permeability modulating properties, these metallocene complexes may be formulated by known methods into compositions, including pharmaceutical compositions, for use as agents to improve the uptake or transport of compounds, including therapeutic compounds, across membranes such as, but not limited to, skin, eyes, or lung tissue.

Terms and Definitions:

As used herein, the following terms have the following meaning unless clearly indicated otherwise:

"Lipid membrane" means the assembly of lipids in aqueous media in a bilayer formation whereby the hydrocarbon tails associate leaving the polar head groups in contact with an aqueous environment, and includes biological membranes such as cell plasma membranes and membranes surrounding organelles such as the nucleus, mitochondria, golgi bodies, liposomes, and the like; lipid bilayers; phospholipid bilayers; liposomes, artificial skin, and the like.

"Modulation of membrane permeability" means the ability of a compound or a chemical species to either gain access to volume defined within the interior confines of the membrane from outside the membrane or to exit the volume defined within interior confines ofthe membrane to outside the membrane is altered.

"Pseudotetrahedral" means a geometric arrangement where the ligands approach the metal from four ofthe eight corners of a cube with a slight deviation. "Bent sandwich" means a complex with at least two cyclic organic ligands containing delocalized π systems bound to the metal ion in which the two cyclic ligands are not parallel to each other.

"Organometallic compound" is an organic compound comprised of a metal attached directly to carbon (R-M).

"Coordination compound" or "Complexed compound" is a compound formed by the union of a central metal atom or ion with a nonmetal atom, ion or molecule called a ligand or complexing agent.

"Ligand" or a "complexing agent" is a molecule, ion or atom that is attached to the central metal atom or ion of a coordination compound.

"Chelate" or "chelated compound" a type of coordination compound in which a central metal ion is attached by coordinate links to two or more non-metal atoms in the same molecule, called ligands. One or more heterocyclic rings are formed with the central metal atom as part of each ring.

"Monodentate ligand" is a ligand having a single donor atom coordinated to the central metal atom or ion.

"Bidentate ligand" is a ligand having two donor atoms coordinated to the same central metal atom or ion.

"Metallocene" is an organometallic coordination compound obtained as a cyclopentadienyl derivative of a transition metal or transition metal halide.

"Transition metals" is any of a number of elements in which the filling ofthe outermost shell to eight electrons within a period is interrupted to bring the penultimate shell from 8 to 18 or 32 electrons. Transition metals include elements 21 through 29, 39 through 47, 57 through 79, and from 89 on.

"Aryl" refers to monovalent unsaturated aromatic carbocycle having a single ring, such as phenyl, or multiple condensed rings, such as naphthyl or anthryl, which can be optionally substituted by substituents such as halogen, alkyl, arylalkyl, alkoxy, arylkoxy, and the like.

"Alkyl" is straight chained or branched chained alkyl of 1-4 carbons, and includes substituted alkyl. Alkyl can be substituted with halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. "Alkoxy" is straight chained or branched chained alkoxy or 1-4 carbons, and includes substituted alkoxy. Alkoxyl can be substituted with halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine.

"Halo" is Br, Cl, F, or I.

The following is a list of abbreviations and their corresponding meanings which are used throughout the specification:

"Cp" is cyclopentadienyl or cyclopentadiene;

"bpy" is 2,2' bipyridine;

"phen" is phenanthroline;

"acac" is acetyl acetonate;

"cat" is catecholate;

"Et₂.(dtc)" is NN-diethyldithiocarbamate;

"PH" is N-phenylbenzohydroxamate;

"OTf ' is O₃SCF₃;

"PBS" is phosphate buffered saline;

"THF" is Tetrahydrofuran;

"POBΝ" is α-(4-pyridyl-l-oxide)-Ν-tert-butylnitrone;

"PC" is phosphatidylcholine;

"N-NBD-PE" is N-(7-nitrobenz-2oxa-l,3-diazol-4-yl)dipalmitoyl-L-α- phosphatidylethanolamine;

"N-Rh-PE" is N-(lissamine rhodamine B sulfonyl)- dipalmitoyl-L-α- phosphatidylethanolamine;

"3-mcm" is 3-chloro acetylacetonate;

"3-mmm" is 3-methyl acetylacetonate; "3-metm" is 3-ethyl acetylacetonate;

"3-mnm" is 3-nitro acetylacetonate;

"DLPC" is dilinoleoyl phosphatidyl choline.

Table 1 below provides a glossary of compound number, abbreviated name and compound name of example metallocenes used throughout the specification. Table 1: Glossary of metallocenes used throughout the specification.

Compounds

The present invention provides metallocene complexes useful for modulating the permeability of a lipid membrane, compositions comprising such metallocenes, and methods of using such metallocenes to modulate the permeability of a lipid membrane. Metallocenes typically have a so-called "bent-sandwich" structure where cyclopentadienyl moieties in a tetrahedral symmetry are positioned in a bent conformation with respect to the metal center.

Formula I

M is a transition metal atom or ion, preferably V or Ti.

Cp is unsubstituted cyclopentadiene or cyclopentadiene substituted with one or more substituents that can be the same or different, and are preferably selected from from Cι-₄ alkyl, aryl, Cn alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably, Cp is unsubstituted cyclopentadiene.

Ri and R₂ together are a bidentate ligand.

Suitable bidentate ligands include N,N'; O,O'; N,O; and S,S' bidentate ligands. Examples of suitable N, N' bidentate ligands include diamines and other such known suitable N, N' bidentate ligands. Examples of diamines include bipyridyl, derivatives of bipyridal, bridged bipyridal, such as phenanthroline, derivatives of phenanthroline, and other such compounds. Examples of suitable N, O bidentate ligands include amino acids and Schiff base type groups. Examples of suitable O, O' bidentate ligands include dicarboxylate, 2-hydroxyacetophenone, acetylacetone type and catechol type groups. Examples of suitable S, S' bidentate ligands include diethyldithiocarbamate.

Preferred examples of bidentate ligands include: phenanthroline; bipyridyl; bridged bipyridyl; N-phenyl benzohydroxamato; N, N-diethyldifhiocarbamato; acetylacetonato; catacholato; and acetophenone. Each of these bidentate ligands can be unsubstituted or substituted with one or more substituents selected from from Cι-₄ alkyl, aryl, C,-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine.

Specific examples of suitable metallocene complexes include chelated vanadocene complexes or chelated titanocene complexes. Specific examples of chelated vanadocene complexes and chelated titanocene complexes include Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂V(cat), Cp₂V(Et₂ (dtc)), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3-metm), Cp₂V(3-mmm), Cp₂V(3-mcm), Cp₂V(3-mnm), pharmaceutically acceptable salts or esters thereof, and mixtures thereof.

Modulation of Membrane Permeability

The present invention provides compounds, methods and compositions useful for modulating the permeability of a lipid membrane. Modulating the permeability of a membrane can be useful for to increase the uptake of agents, including therapeutic agents, across the membrane, which can be useful for formulations to cross membranes such as, but not limited to, skin in topical applications, eyes in opthamalogical formulations, or lung tissue in the formulation of inhalants and aerosols. Accordingly, the invention provides compounds, compositions and methods useful for improving drug delivery in a subject in need thereof. When using compounds, compositions or methods ofthe invention to modulate the permeability of a lipid membrane to enhance the uptake or transport of therapeutic agents, it is preferred that the metallocene complex not catalyze sufficient lipid peroxidization to cause damage to the membrane or cell death.

The concentration of metallocene complex in media surrounding a membrane effective to modulate the permeability ofthe membrane will vary according to metallocene complex used. Any concentration of metallocene complex effective for modulating membrane permeability can be useful. Preferably, the concentration of metallocene complex in media surrounding a membrane is between 50 μm and 400 μm. Insertion into Membrane

In one embodiment, the invention provides compounds, methods and compositions useful for modulating the permeability of a lipid membrane by inserting one or more metallocene complexes of formula I into the membrane.

Formula I

M is a transition metal atom or ion, preferably V or Ti.

Cp is unsubstituted cyclopentadiene or cyclopentadiene substituted with one or more substituents that can be the same or different, and are preferably selected from -₄ alkyl, aryl, C,_₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably, Cp is unsubstituted cyclopentadiene.

Ri and R₂ together are a bidentate ligand.

The degree of modulation of membrane permeability can be associated with the size, hydrophobicity, and planarity ofthe metallocene complex, particularly the size, hydrophobicity, and planarity ofthe bidentate ligand. Preferably, the bidentate ligand comprises one or more aromatic rings. More preferably, the bidentate ligand comprises two or more aromatic rings.

Preferred examples of bidentate ligands include: phenanthroline; bipyridyl; bridged bipyridyl; N-phenyl benzohydroxamato; N, N-diethyldithiocarbamato; acetylacetonato; catacholato; and acetophenone. Each of these bidentate ligands can be unsubstituted or substituted with one or more substituents selected from Cι-₄ alkyl, aryl, Ci-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably, the bidentate ligand is unsubstituted. Most preferably the bidentate ligand is unsubstituted phenanthroline or unsubstituted bipyridyl.

It is envisioned that any metallocene complex having the appropriate size, hydrophobicity, and planarity would be suitable for modulating membrane permeability, regardless of central metal ion. Any central metal ion, regardless of charge, is suitable, provided that the central metal ion adapts a "bent sandwich" shape when the metallocene is of Formula I.

Specific examples of suitable metallocene complexes include chelated vanadocene complexes or chelated titanocene complexes. Specific examples of chelated vanadocene complexes and chelated titanocene complexes include Cp₂V(phen), Cp V(bpy), Cp₂V(PH), Cp₂Ti(phen), Cp₂Ti(bpy), pharmaceutically acceptable salts or esters thereof, and mixtures thereof. In some embodiments, the metallocene complexes provide for rapid, but temporary modulation ofthe permeability of a lipid membrane.

While not intending to limit the scope ofthe invention, it is believed that it is the bidentate ligand portion ofthe metallocene complexes useful for the invention that inserts into a lipid membrane to modulate the permeability ofthe membrane.

Lipid Oxidation without 'OH Radical Production

In another embodiment, the invention provides compounds, methods and compositions useful for modulating the permeability of a lipid membrane by oxidizing membrane lipids in the presence of a metallocene complex of formula I without generating hydroxy radical (OH) intermediates.

Formula I

M is a transition metal atom or ion, preferably V or Ti.

Cp is unsubstituted cyclopentadiene or cyclopentadiene substituted with one or more substituents that can be the same or different, and are preferably selected from Cι-₄ alkyl, aryl, C,-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably, Cp is unsubstituted cyclopentadiene.

Ri and R₂ together are a bidentate ligand.

Oxidization of lipid membranes in the presence of a metallocene can be catalyzed by metal complexes where metal-centered oxidants are directly responsible for C-H bond cleavage. Lipid peroxidization can alter the physical properties of a membrane to an extent effective to alter the permeation of a membrane.

Preferred examples of bidentate ligands include: phenanthroline; bipyridyl; bridged bipyridyl; N-phenyl benzohydroxamato; N, N-diethyldithiocarbamato; acetylacetonato; catacholato; and acetophenone. Each of these bidentate ligands can be unsubstituted or substituted with one or more substituents selected from Cι-₄ alkyl, aryl, Ci-4 alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine. Preferably the bidentate ligand is substituted or unsubstituted acetylacetonato, phenanthroline or bipyridyl. More preferably the bidentate ligand is substituted or unsubstituted acetylacetonato. If substituted, preferably acetylacetonato is 3-substituted acetylacetonato.

Selected preferred compounds include Cp₂V(phen), Cp₂V(bpy), Cp₂V(acac), Cp₂V(3-metm), Cp₂V(3-mmm), Cp₂V(3-mcm), and Cp₂V(3-mnm).

The presence of substituents on the bidentate ligand can alter the ability ofthe molecule to oxidize lipids. The presence of electron-withdrawing groups, such as , halo, NO₂, CN, OCN, SeCN, SCN, N₃, and the like, can decrease the rate of lipid oxidization, whereas the presence of electron-donating groups, such as unsubstituted alkyl, aryl, and the like can increase the rate of lipid oxidization. Accordingly, if an increased rate of lipid peroxidization is preferred, then the bidentate ligand is preferably substituted with an electron-donating substituent. Conversely, if a decreased rate of lipid peroxidization is preferred, then the bidentate ligand is preferably substituted with an electron- withdrawing substituent.

Selected preferred compounds for increased rates of lipid oxidization include Cp₂V(acac), Cp₂V(3-metm), and Cp₂V(3-mmm).

Lipid Oxidation with 'OH Radical Production

In one embodiment, the invention provides compounds, methods and compositions useful for modulating the permeability of a lipid membrane by oxidizing membrane lipids in the presence ofthe metallocene complex through the generation of hydroxy radical (^*OH) intermediates. Oxidization of lipid membranes in the presence of a metallocene can be catalyzed by metal complexes where the catalyst generates metal-free radicals (HO", RO^*, ROO\ etc.) that cause C-H bond cleavage. Lipid peroxidation is generally shown to proceed through hydroxy radical-initiated pathways such as the classic Fenton-type reaction (Konings, A.W.T. In Liposome Technology, 1992, pp, 139— 161; and Spiteller, G. Chem. Phys. Lipids, 1998, 95:105). Lipid peroxidization resulting from metal-free radical generation can alter the physical properties of a membrane to an extent effective to alter the permeation of a membrane. However, metal-free radical generation can have indiscriminate negative effects and can lead to cell death.

Any metallocene compound that can catalyze the oxidization of membrane lipids through the generation of free hydroxyl radicals can be useful in compositions useful according to the methods ofthe invention. An example of a such a metallocene compound is vanadocene dichloride. Such compounds, because of their generation of hydroxy radicals, are least preferred when attempting to minimize cell death.

Uptake of Agents across a Membrane

The compositions and methods ofthe invention can be useful for increasing the uptake of agents across a membrane.

In one embodiment, the compositions ofthe invention can be used to increase the uptake of therapeutic agents across membranes such as skin in topical applications, eyes in opthamalogical formulations, or lung tissue in the formulation of inhalants and aerosols for the treatmetns of diseases such as asthma, cystic fibrosis, and pulmonary fibrosis. If it is desired to minimize cell damage and/or cell death, it is preferably that little or no lipid oxidation occur. Thus, the invention provides compounds, compositions, and methods for modulating the permeability of a membrane by inserting inserting one or more metallocene complexes into the membrane without the one or more metallocene catalyzing sufficient lipid peroxidization to cause cell death. Preferably, metallocene complexes not capable of catalyzing lipid oxidization can be used. Such metallocene complexes preferably have titanium as the central metal ion. However, it will be understood that the concentration in media surrounding a membrane of a metallocene complex capable of catalyzing lipid oxidization can be altered to allow for effective permeability of the membrane without causing sufficient lipid peroxidization to cause cell death. It will also be understood that combinations of two or more metallocene complexes can be used to obtained the desired effect. Accordingly, metalocene complexes that permeabilize membranes through lipid oxidization can be used in this embodiment, provided that the permeability ofthe membrane can be effectively modulated without causing cell death. Preferably, metallocene complexes that permeabilize membranes through lipid oxidization are metallocene complexes that do not generate metal-free hydroxy radicals.

The compositions ofthe invention can also be used to increase the permeability of a non-biological membrane such as an artificial skin; e.g., to increase the permeability of water or a therapeutic agent through the artificial skin.

Pharmaceutical Compositions

The invention provides compositions, including pharmaceutical compositions, comprising metallocene compounds useful for modulation ofthe permeability of a lipid membrane. The membrane permeabilizing compositions ofthe present invention are suitable for use in mammals. As used herein, the term "mammals" means any class of higher vertebrates that nourish their young with milk secreted by mammary glands, e.g., humans, rabbits and monkeys.

The membrane permeabilizing compositions ofthe present invention comprise one or more metallocene complexes. The total amount of metallocene complex thereof will typically range from about 1 to 25 weight percent based on the weight ofthe membrane permeabilizing composition. Preferably, the amount of metallocene complex employed will be that amount necessary to achieve the desired membrane permeabilizing effects. Appropriate amounts can be determined by those skilled in the art. Preferably, the amount ofthe metallocene complex employed, an amount effective to modulate the permeability of a lipid membrane, will comprise from about 1 to 25 weight percent, based on the weight ofthe composition.

When used in vivo to modulate the permeability of a membrane, the administered dose is that effective to have the desired effect, e.g., sufficient to increase permeation or transport of an agent, including a therapeutic agent, through a membrane. The appropriate dose can be extrapolated using known methods and relationships. A useful dose will vary with the desired effect, the mode of administration, and the composition administered.

The compositions ofthe invention contain not only the metallocene complex but also necessary pharmaceutically acceptable carriers, diluents or vehicles, i.e., one that appropriately delivers metallocene complexes to a site for insertion into a lipid membrane and modulates lipid membrane permeability.

The compounds can be administered by known techniques, such as orally, intranasally, parentally (including subcutaneous injection, intravenous, intramuscular, intrasternal or infusion techniques), by inhalation spray, dermally, transdermally, intrathecal, intracerebroventricular, buccal, sublingual, topically, by absorption through a mucous membrane, or rectally, in dosage unit formulations containing conventional non- toxic pharmaceutically acceptable carriers, adjuvants or vehicles. Pharmaceutical compositions ofthe invention can be in the form of suspensions or tablets suitable for oral administration, nasal sprays, eye drops, nose drops, creams, sterile injectable preparations, such as sterile injectable aqueous or oleagenous suspensions or suppositories.

For oral administration as a suspension, the compositions can be prepared according to techniques well-known in the art of pharmaceutical formulation. The compositions can contain microcrystalline cellulose for imparting bulk, alginic acid or sodium alginate as a suspending agent, methylcellulose as a viscosity enhancer, and sweeteners or flavoring agents. As immediate release tablets, the compositions can contain microcrystalline cellulose, starch, magnesium stearate and lactose or other excipients, binders, extenders, disintegrants, diluents and lubricants known in the art.

For administration by inhalation or aerosol, the compositions can be prepared according to techniques well-known in the art of pharmaceutical formulation. The compositions can be prepared as solutions in saline, using benzyl alcohol or other suitable preservatives, absorption promoters to enhance bioavailability, fluorocarbons or other solubilizing or dispersing agents known in the art.

For administration as injectable solutions or suspensions, the compositions can be formulated according to techniques well-known in the art, using suitable dispersing or wetting and suspending agents, such as sterile oils, including synthetic mono- or diglycerides, and fatty acids, including oleic acid. For rectal administration as suppositories, the compositions can be prepared by mixing with a suitable non-irritating excipient, such as cocoa butter, synthetic glyceride esters or polyethylene glycols, which are solid at ambient temperatures, but liquefy or dissolve in the rectal cavity to release the drug.

Preferred administration routes include orally, parenterally, as well as intravenous, intramuscular or subcutaneous routes.

Solutions or suspensions ofthe compounds can be prepared according to methods known in the art. Under ordinary conditions of storage and use, these preparations may contain a preservative to prevent the growth of microorganisms.

The pharmaceutical dosage form suitable for injection or infusion use can include sterile, aqueous solutions or dispersions or sterile powders including an active ingredient which are adapted for the extemporaneous preparation of sterile injectable or infusible solutions or dispersions. In all cases, the ultimate dosage form should be sterile, fluid and stable under the conditions of manufacture and storage. The liquid carrier or vehicle can be a solvent or liquid dispersion medium including, for example, water, ethanol, a polyol such as glycerol, propylene glycol, or liquid polyethylene glycols and the like, vegetable oils, nontoxic glyceryl esters, and suitable mixtures thereof. The proper fluidity can be maintained, for example, by the formation of liposomes, by the maintenance ofthe required particle size, in the case of dispersion, or by the use of nontoxic surfactants. The prevention ofthe action of microorganisms can be accomplished by various antibacterial and antifungal agents, for example, parabens, chlorobutanol, phenol, sorbic acid, thimerosal, and the like. In many cases, it will be desirable to include isotonic agents, for example, sugars, buffers, or sodium chloride. Prolonged absorption ofthe injectable compositions can be brought about by the inclusion in the composition of agents delaying absorption — for example, aluminum monosterate hydrogels and gelatin.

Sterile injectable solutions are prepared by incorporating the compounds in the required amount in the appropriate solvent with various other ingredients as enumerated above and, as required, followed by filter sterilization. In the case of sterile powders for the preparation of sterile injectable solutions, the preferred methods of preparation are vacuum drying and freeze-drying techniques, which yield a powder ofthe active ingredient plus any additional desired ingredient present in the previously sterile-filtered solutions. It will be understood that pharmaceutical compositions and preparations ofthe compositions can be prepared therapeutic agents.

The invention may be further clarified by reference to the following Examples, which serve to exemplify some ofthe preferred embodiments, and not to limit the invention.

EXAMPLES

Example 1. Synthesis and characterization of metallocene complexes-

Methods and Materials

Syntheses ofthe metallocene complexes were performed under an inert atmosphere using standard Schlenk techniques, unless otherwise noted.

Egg yolk phosphatidylcholine (PC) and egg yolk phosphatidylglycerol (PG) was purchased from Avanti Polar Lipids (Alabaster, AL). Cholesterol, 1,6-diphenyl- 1,3,5- hexatriene, and Triton X-100 were obtained from Aldrich Chemical Co. (Milwaukee, WI). 5(6)-Carboxyfluorescein, N-(7-nitrobenz-2-oxa-l ,3-diazol-4-yl)dipalmitoyl-L- α-phosphatidylethanol-amine (Ν-ΝBD-PE) and N-(lissamine rhodamine B sulfonyl)- dipalmitoyl-L-α-phosphatidylethanolamine (Ν-Rh-PE) were from Molecular Probes, Inc. (Eugene, OR). Nil other reagents were purchased from Aldrich Chemical Co. (Milwaukee, WI) and used without modification unless stated otherwise. Acetonitrile and methylene chloride were dried over CaH₂, then distilled prior to use. THF was dried over sodium and distilled prior to use.

IR. Infrared spectra were obtained on a FT-Νicollet Protege 460 spectrometer as a KBr pellet or nujol mull. IR spectra are reported in cm^"1.

UV/vis. A UV/vis polarizer attachment was used for the polarization experiments. UV-vis spectra were recorded in a quartz cell or cuvette on a Beckman Model DU 7400 spectrophotometer and the spectral bands were registered in the 250 - 800 nm range.

1H ΝMR. Η ΝMR spectra were recorded on a Varian XL-300 spectrometer operating at 300.110 MHz. Chemical shifts were referenced to residual protio solvent peaks in the sample. Magnetic moments were determined using Evan's Method at 20°C (Evans, D. F. J. Chem. Soc. 1959, 2003).

Mass spectra. Mass spectra were recorded on a HP G2025 A MALDI-TOF mass spectrometer using α-cyano-4-hydroxycinnamic acid as the supporting matrix. Spectra were averaged over 50 shots.

EPR. Electron paramagnetic resonance (EPR) spectra were recorded in PBS (0.015 M NaHPO₄, 0.10 M NaCl, 0.02M Kcl, pH 7.2) or acetonitrile on a Bruker ESP 300 EPR spectrometer (9.64 GHz). The g values were calibrated with a Varian strong pitch (0.1%> in KC1) standard (g value of 2.0028). The samples for the EPR spectral analysis were studied in a Willmad WG-814 standard TE102 aqueous cell cavity (0.3- mm inner path length) to minimize the dielectric loss.

Fluorescence. All fluorescence measurements were made using a Shimadzu spectrofluorophotometer (Model RF-5301PC).

Electrochemical measurements. Electrochemical measurements were performed on a Bioanalytical Systems B/W 100b electrochemical analyzer with internal resistance (IR) compensation. The cyclic voltammograms taken in acetonitrile were obtained in a 0.1 M BmNPFe (TABP) electrolyte solution with a 0.1 M Ag/AgNO₃ reference electrode, a Platinum Disc working electrode, and a platinum wire auxilliary electrode. Solutions were purged with nitrogen and scanned at 200 mV/s. Aqueous cyclic voltammograms were taken in a standard PBS solution, using a Ag/AgCl reference electrode, a glassy carbon working electrode, and a platinum wire auxilliary electrode. Solutions were purged with nitrogen and scanned at 200 mV/s. All potentials were referenced to the ferrocene-ferrocinium couple

V vs Ag/AgCl in PBS).

Elemental analyses. Elemental analyses were performed by Atlantic Microlab,

Inc. (Norcross, GA).

Synthesis of Compounds

[Cp₂V(phen)][OTf]₂ (l)

The synthesis was based on a modification ofthe synthetic procedure reported by Thewalt et al. (J. OrganomeL Chem. 1986, 302, 193) for [Cp₂Ti(phen)][OTf]₂. Cp₂VCl₂ (253 mg, 1.0 mmol) and AgOTf (521 mg, 2.1 mmol) were placed in a 25 mL round bottomed flask. THF (20 mL) was added and the solution was stirred for 1.5 hr. The solution was quickly filtered in air through celite, and the solution was placed again under argon. A methylene chloride solution (15 mL) of 1 , 10-phenanthroline (271 mg, 1.5 mmol, 1.5 eq.) was added to the stirring solution, causing an immediate color change from green to muddy brown. The solution was stirred vigorously for 13 h, opened to the air, then filtered. The gummy brown residue was washed with THF and hexanes. The resultant brown powder was then washed with acetone (3 x 5 mL), yielding a fine brown powder. The powder was recrystallized from a minimum of acetonitrile layered with ether. 224 mg of Cp V(phen) (1) was obtained as large dark brown crystals (34%> yield; 0.34 mmol). IR: 3134 (w), 3101 (m), 2951 (m), 1630 (w), 1606 (m), 1587 (m), 1448 (w), 1271 (s), 1260 (s), 1215 (m), 1152 (s), 1030 (s), 872 (w), 854 (m), 724 (m), 637 (s), 572 (w), 517 (m). UV-vis (CH₃CN, nm (ε)): 558 (92), 448 (sh), 355 (sh), 270 (29,300), 229 (38,800). μ_eff(CDCl₃) = 1.68 (12) BM. MALDI-TOF MS: m/z = 361 (m - 2 OTf+). Anal. Calcd. for C₂₄Hι₈F₆N₂O₆S₂V: C, 43.72; H, 2.75; N, 4.25. Found: C, 43.93; H, 2.87; N, 4.08.

[Cp₂Ti(phen)][OTf]₂ (2)

This compound was synthesized according to the procedure reported by Thewalt et al (J. OrganomeL Chem. 1986, 302, 193). Unlike the reaction to prepare the vanadium complex Cp₂V(phen) (1) which required more than 12 hours for completion, the reaction to prepare the titanium complex Cp Ti(phen) (2) was instantaneous.

[Cp₂N(bpy)][OTf]₂ (3)

This synthesis was based on a modification ofthe synthetic procedure reported by Thewalt et al. (J OrganomeL Chem. 1986, 302, 193) for [Cp₂Ti(bpy)][OTf]₂. Cp₂VCl₂ (252 mg, 1.0 mmol) and AgOTf (523 mg, 2.1 mmol) were placed in a 25 mL round bottomed flask. THF (20 mL) was added and the solution was stirred for 1.5 h. The solution was quickly filtered in air through celite, and the solution was placed again under argon. A methylene chloride solution (15 mL) of 2,2'-bipyridine (325 mg, 2.08 mmol, 2.08 eq.) was slowly added to the solution, causing an immediate color change from deep green to a dark brown. The solution was stirred for a few minutes and allowed to stand for 14 h. The flask was opened to the air and the solution was then filtered. The fine brown precipitate was washed with excess THF (40 mL) and hexanes (2 x 10 mL). The resultant brown powder was then washed with acetone three times, giving a fine brown powder. The powder was recrystallized from a minimum of acetonitrile layered with hexanes. 267 mg of small reddish brown crystals of Cp₂V(bpy) (3) were recovered (42%> yield; 0.42 mmol). IR: 3121 (w), 3099 (m), 2951 (m), 1604 (m), 1474 (m), 1450 (w), 1437 (m), 1404 (m), 1285 (w), 1266 (vs), 1257 (s), 1225 (m), 1156 (s), 1028 (s), 864 (m), 772 (m), 637 (s), 573 (w), 516 (m). UV-Vis (CH₃CN, nm (ε)): 598 (sh, 16,722), 556 (120), 444 (22,522), 319 (9,900), 242 (25,800). μ_efϊ(CDCl₃) = 1.66 (12) BM. MALDI- TOF MS: m/z = 337 (m - 2 OTf+). Anal. Calcd. for C₂₂Hι₈F₆N₂O₆S₂V: C, 41.59; H, 2.90; N, 4.41. Found: C, 41.81 ; H, 2.98; N, 4.36.

[Cp₂Ti(bpy)][OTfj₂ (4)

The compound was synthesized according to the procedure reported by Thewalt et al (J. OrganomeL Chem. 1986, 302, 193). Unlike the reaction to prepare the vanadium complex Cp₂V(bpy) (3) which required 12 hours for completion, the reaction to prepare the titanium complex Cp₂Ti(bpy) (4) was instantaneous.

[Cp₂V(PH)][OTf] (5)

The reaction mixture composed of VCp₂Cl (0.2 g, 0.79 mmol) and AgCF₃SO₃ (0.43 g, 1.7 mmol) in lOmL of H₂O was stirred for 2 h and then filtered through a fine glass frit with Celite in air. A solution of N-phenylbenzohydroxamic acid (0.17 g, 0.79 mmol) in 5 mL of ethanol was added to the filtrate with stirring, and the resulting solution was kept for 4 h to complete the precipitation of a dark compound. The precipitate was collected by filtration, thoroughly washed with hexane, and dried overnight under vacuum to give 265 mg (62%>) ofthe title compound, mp 160°C. Anal. Calcd for VC₂₄H₂₀ΝF₃O₅S (542.429): C, 53.10; H, 3.69; N, 2.58; S, 5.90. Found: C, 52.48; H, 3.72; N, 2.51; S, 5.73. UV/vis (CH₂C1₂) λ_max: 680, 501 (d-d), 377 (LMCT), 314 (π-π* of hydroxamic moiety), 261, 233 (π-π* of Cp ring) nm. IR: 3345 (sb), 3117 (s), 1651 (mb), 1600 (m), 1539 (vs), 1495 (m), 1450 (m), 1300 (m), 1281 (s), 1244 (vs), 1173 (s), 999 (m), 758 (m), 694 (m), 638 (s), 578 (w), 515 (m) cm^"1. [Cp₂V(Et₂(dtc))][OTf] (6)

The compound was prepared following literature procedures of Doyle, et al., Inorg. Chem. 1968, 7, 2479; and Casey et al. Aust. J. Chem. 1974, 27, 757.

[Cp₂V(acac)][OTf] (7)

The compound was prepared following literature procedures of Doyle, et al., Inorg. Chem. 1968, 7, 2479; and Casey et al, Aust. J. Chem. 1974, 27, 757.

[Cp₂V(cat)] (8)

The compound was prepared as in Gosh, et al. Clin. Can. Res., 2000, in press. Briefly,one hundred twenty-six mg of Cp VCl₂ (0.50 mmol) was placed in a 250-ml flask and dissolved in 100 ml of THF. In another flask, sodium cat was prepared by the addition of NaH (25 mg, 1.0 mmol) to 55.5 mg of catechol (0.50 mmol) in 15 ml of THF. The solution was stirred for 2 h, resulting in a deep blue solution. The cat solution was cannulated into the vanadium solution and stirred for 4 h. The reaction mixture was opened to the air and quickly flash-chromatographed under nitrogen on alumina (neutral; acetonitrile mobile phase). The solvent ofthe deep blue solution was then removed under vacuum, and the product was collected. Yield: 26%; m.p., decomposition starts at 95°C.

Crystallographic Structure Determination of [Cp₂V(phen)][OTf]₂ and [Cp₂V(bpy)][OTf]2: Rectangular shaped crystals of Cp₂V(phen) were grown by vapor diffusion of hexane into an acetone solution of Cp V(phen) at room temperature. A single crystal of dimensions 0.5 x 0.3 x 0.1 mm was attached to a glass fiber using epoxy and was used for cell constant determination and data collection. Dark green, prism shaped crystals of Cp₂V(bpy) were grown by vapor diffusion of hexane into an acetonitrile solution of Cp₂V(bpy) at room temperature. A single crystal of dimensions 0.4 x 0.2 x 0.2 mm was attached to a glass fiber using epoxy and was used for cell constant determination and data collection. Data were collected using a Bruker SMART platform CCD (SMART Software Reference Manual. Version 5.0, Bruker Analytical X-ray Instruments Inc., Madison, Wisconsin, USA), with graphite monochromatic MoKα radiation (λ = 0.71073 A ) at room temperature (293 K). The SMART software (Bruker, 1998, SMART Software Reference Manual. Version 5.0, Bruker Analytical X-ray Instruments Inc., Madison, Wisconsin, USA) was used for data collection, and the SAINT program (Bruker, 1996, SAINT Software Reference Manual. Version 4.0, Bruker Analytical X-ray. Instruments Inc., Madison, Wisconsin, USA) was used for data reduction. An empirical absorption correction (Bruker, 1997, SHELXTL- Reference Manual. Version 5.10, Bruker Analytical X-ray Instruments Inc., Madison, Wisconsin, USA) was applied and the structure was solved by direct methods using the SHELXTL V 5.10 suite of programs (Bruker, 1997, SHELXTL- Reference Manual. Version 5.10, Bruker Analytical X-ray Instruments Inc., Madison, Wisconsin, USA). All non- hydrogen atoms were refined anisotropically and the hydrogen atoms were placed in ideal positions and refined as riding atoms with individual isotropic thermal parameters. Atomic scattering factors were taken from the International Tables for X-Ray Crystallography (" International Tables for X-ray Crystallography", Volume C, 1991). All calculations were performed on a Pentium computer. Crystallographic data are summarized in Table 3. Selected bond lengths and angles are presented in Table 4, and ORTEP drawings are shown in Figures 6 and 7.

EPR Spin Trapping Experiments: The following is a general procedure for the spin trapping experiments: a lxl 0^~3 M solution ofthe metallocene compound was prepared in water. POBN (5 eq.) was added to the solution, followed by the addition of excess H₂O₂ (10 eq.). The solution was transferred to a 50 μl capillary tube, mounted in the cavity ofthe EPR spectrometer, and the spectrum was taken. The spectra were taken within 90 seconds ofthe addition of H₂O₂ unless otherwise noted, and each spectrum was averaged from four independent scans. The EPR spectra ofthe POBN -"OH adducts were identified by their doublet of triplet EPR signals ofthe hydroxyl radical at g = 2.008 with hyperfme splitting ofthe nitrosyl nitrogen, <A_N> = 15xl0^-4 cm^-1, and the β- hydrogen atom, <A_H> = 2.81xl0^-4 cm^-1 using VOSO₄ as a standard (Setaka, et al. J. Catal. 1969, 15, 209). Control solutions ofthe organometallic vanadium compounds and POBN indicated no interaction between the compounds and the spin trap. Control solutions of POBN and H₂O also indicated no interaction between just the spin trap and H₂O₂. Results Synthesis: Reaction of in situ generated Cp₂V(OTf)₂ in dry THF with 2,2'- bipyridine or 1 , 10-phenanthroline produced [Cp₂V(phen)][OTf]₂ (1) and [Cp₂V(bpy)][OTfJ₂ (3) respectively (eq. 2). The compounds were characterized by elemental analysis and MALDI-TOF mass spectrometry. The magnetic moments, as measured by Evan's method, gave a μ_efr = 1.68 (12) BM for Cp₂V(phen) (1) and μ_eff = 1.66 (12) BM for Cp V(bpy) (3). Both are close to the expected value for a d' system (μ_eff = 1.72 BM). Most chelated complexes of vanadocene (IV) are synthesized via an oxidative pathway

(1 ) with vanadocene (II) and the appropriate organic substrates (Gambarotta, et al. Inorg. Chem. 1984, 23, 1739; Stephan, D. Inorg. Chem. 1992, 31, 4218; Fochi, et al. C. J. Chem. Soc. Dalton Trans. 1983, 1515; Gambarotta, et al. Inorg. Chem. 1984, 23, 3532; Gambarotta, et al. Organometallics 1986, 5, 2425; and Muller, et al. J. OrganomeL Chem. 1976, 111, 73), but a few have been prepared through the direct substitution ofthe ancillary positions ofthe Cp₂V^lv ion (Casey, et al. Aust. J. Chem. 1972, 25. 2085; ibid. 1974, 27, 757; Doyal, et al. Inorg. Chem. 1968, 7, 2479; and Doyal, et al. Inorg. Chem. 1968, 7, 2484). Compounds Cp₂V(phen) (1) and Cp₂V(bpy) (3) were prepared using a method similar to that described for the titanium analogues, [Cp₂Ti(phen)][OTfj₂ (2) and [Cp₂Ti(bpy)][OTfj₂ (4) (Thewalt et al. OrganomeL Chem. 1986, 302, 193). Although the reaction to prepare the titanium complexes (Cp₂Ti(phen) (2) and Cp Ti(bpy) (4)) was instantaneous, the reactions to obtain Cp₂V(phen) (1) and Cp V(bpy) (3) required more than 12 h for completion.

UV-Vis and EPR: The UV-Vis absorption spectra of Cp₂V(phen) (1) and Cp₂V(bpy) (3) in acetonitrile have absorption bands at ~ 17,900 cm^-1 (ε = 90-120 mol^-1 cm^-1) along with a shoulder at -16,600 cm^-1 (Figure 8). The two weak absorption bands can be assigned to the first two Laporte forbidden d-d transitions expected from a low symmetry C_2v type molecule. These two transitions are very close in energies, comparable to the well resolved bands detected for [Cp₂V(DeDtc)]⁺ (DeDtc = diethyl di hiocarbamate) which are at 16,100 cm^-1 and 18,700 cnf '( Ghosh, et al. J Inorg. Biochem. 1998, 72(1-2), 89). This feature is different than the Cp₂VCl₂ spectrum which has a single intense band at 11,800 cm^-1, assigned by Stewart et al. based on molecular orbital calculations (Stewart, et al. J Chem Soc. Dalton. Trans. 1973, 722). A third d-d transition also predicted from the Huckel LCAO calculations is observed for Cp₂V(phen) (1) and Cp₂V(bpy) (3) as a weak shoulder at ~ 22,400 cm^-1 (ε « 90 mof¹ cm^-1) (Stewart, et al. J. Chem Soc. Dalton. Trans. 1973, 722; and Petersen, et al. J. Am. Chem. Soc. 1975, 97, 6422). There is a significant difference (hypsochromic shift) in the energies ofthe d- d transitions compared to the Cp VCl₂ spectral bands. This large blue shift is probably due to the difference in the ligand field effects ofthe two types of donor ligands (Lever, A. P. B. Inorganic Electronic Spectroscopy, Elsevier : Amsterdam, 1984; and Figgis, B. N. Introduction to Ligand Fields: Wiley-Interscience: New York, 1966). The characteristic intense band for a typical Cp-ring to V(IV) metal ion charge transfer transition (LMCT) (Stewart, et al. J. Chem Soc. Dalton. Trans. 1973, 722) is also detected at approximately 28,200 cm^-1 (ε « 2,500 mol^-1 cm^-1) for both complexes. This LMCT band is very pronounced for Cp₂V(phen) (1) but appears as a shoulder for Cp₂V(bpy) (3). In the latter case, the π-π* transition ofthe coordinated bipyridine ligand is shifted about 50 nm more to the low energy region than Cp₂V(phen) (1). This is expected from the higher degree of π - electron delocalization in the expanded planar ring system of phen compared to bpy (Seddon, et al. 77ze Chemistry of Ruthenium; Elsevier, Amsterdam, 1984, pp. 1193-1214 and references therein). This shift is also observed in the UV-Vis spectrum ofthe titanocene (IV) analogues, Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4). The UV-Vis spectral data is not reported in the literature. UV-Vis for Cp₂Ti(phen) (2) (CH₃CN, nm (ε)): 406 (2,400), 347 (sh, 3000), 272 (45,000), 223 (50,000). UV-Vis for Cp₂Ti(bpy) (4) (CH₃CN, nm (ε)): 383 (sh, 1,400), 319 (15,000), 241 (23,000), 215 (32,000).

The room temperature EPR spectra of Cp₂V(phen) (1) and Cp V(bpy) (3) both in acetonitrile and PBS solution show a typical isotropic eight-line resonance ranging from g = 1.98 - 1.999 with a ⁵¹V (I = 7/2) isotropic hyperfme coupling constant <A_Α > of 62.66 x 10^~4 cm^-1 and 61.06 x 10^~4 cm^""1 respectively (Figure 9 and Table 2). Due to second order perturbation effects (Rogers, et al. J. Chem. Phys. 1960, 33, 1107) a non-uniform line spacing is observed for both complexes. The <A_av> values are comparable to those of Cp₂V(trop)⁺ (trop = 2-hydroxy-2,4,6-cycloheptatriene-l-onate) (<A_m> - 61 x 10^"4 cm^" ') (Doyal, et al. Inorg. Chem. 1968, 7, 2479) but are significantly lower than Cp VX systems, where X= Cl^", N₃ ^", SCNT, and SeCN^" (<A_av> ~ 68 x 10^~4 cm^"1 ) (Stewart, et al. J. Chem Soc. Dalton. Trans. 1973, 722), and considerably higher than the ethane dithiolato (<A_av> = 51 xlO^-4 cm^"1) or mercaptocatecholato (<A_m> = 42 x 10^"4 cm^"1) chelated complexes (Gambarotta, et al. Inorg. Chem. 1984, 23, 1739). The observed difference in the anisotropic hyperfme coupling constant for Cp₂V(phen) (1) and Cp₂V(bpy) (3), compared to the vanadocene (IV) halides or pseudohalides, is due to the dπ-pπ^* back-bonding effects of a d system in the presence of coordinated phen or bpy type π-acid ligands (Taube, H. Pure Appl. Chem. 1979, 51, 901). The significant difference in the coupling constant parameter between η²-iminoacyl or η²-acetylene vs. dihalide or dipseudohalide coordinated complexes cannot be due to the reduced L-V-L angles as suggested by Petersen, et al. (Inorg. Chem. 1980, 19, 1852). In fact, this contradicts the results obtained for Cp₂V(phen) (1) and Cp₂V(bpy) (3) when compared to the sulfur coordinated complexes of vanadocene (IV). The S-V-S angles ofthe 1,2- benzenedithiolato or 1 ,3-propane dithiolato complexes are considerably higher than the N-V-N angles for Cp₂V(phen) (1) and Cp V(bpy) (3) (vide infra), yet a substantial decrease in the <A_Α > value for the sulfide coordinated species compared to Cp₂V(phen) (1) and Cp₂V(bpy) (3) was seen. These discrepancies may be rationalized from an increased covalent character in the vanadium-sulfur bonding (Casey, et al. Aust. J. Chem. 1972, 25, 2085; and ibid. 1974, 27, 757) over a V-N coordinated bond. This higher covalancy in V-S bonding would provide greater π-electron delocalization compared to the bond caused by the dπ-pπ^* back-bonding ofthe phen or bpy type π-acceptor ligands in Cp₂V(phen) (1) or Cp V(bpy) (3). Presumably, a similar back-bonding effect also occurs for Cp V(trop)⁺ as the tropolone ligand also has a π-delocalized ring in the vicinity of oxygen donors and could be responsible for the comparable <A_ay> value. Table 2: Electrochemical and EPR Data for [Cp₂V(phen)][Otf]₂ (1) and [Cp₂V(bpy)][Otf]₂ (3)

C. Electrochemistry. Although extensive heterogeneous electron transfer properties have been documented for Cp₂VCl₂ (Holloway, et al. J. Am. Chem. Soc. 1979, 101, 2038; Mugnier, et al. New J. Chem. 1982, 6, 197; and Kuzharenko, et al. Izv. Akad. NaukSSSR, Ser. Khim. 1987, 297) and its related analogues (Dorer, et al. J OrganomeL Chem. 1992, 427, 245), there is very little information available on the chelated vanadocene (IV) complexes (Ghosh, et al. J. Inorg. Biochem. 1998, 72(1-2), 89; and Bond, et al. Inorg. Chem. 1973, 4, 887). Cyclic voltammetric studies of Cp₂V(phen) (1) and Cp₂V(bpy) (3) in acetonitrile show that both complexes undergo identical electron transfer processes over the potential range +2.1 to -2.5 V (vs. Cp₂Fe^+/0 couple). As illustrated in Figure 10 and Table 2, each complex displays one reversible couple, centered at -0.62 V, attributable to the V^{(II1) (IV)} redox process. This is followed by a le^" irreversible reduction wave at -1.73 V and another quasireversible redox process with an Eι_/2 of about - 2.0 V. The v^{(III) (IV)} couple is significantly more positive (-400 mV) than the corresponding dithiocarbamato complexes (Eι ₂ for [Cp₂V(DeDtc)]^0/+ = -1.027 V vs. Cp₂Fe^{0 +} in CH₃CN) (Ghosh, et al. J. Inorg. Biochem. 1998, 72(1-2), 89). This may be due to the destabilization ofthe e_g orbitals ofthe degenerate d¹ system in the presence of σ-donor and π-acceptor ligands like phen or bpy compared to the dithiocarbamate ligand which is believed to be σ - and π - donor in character (Coucouvanis, D. Prog. Inog. Chem. 1970, 77, 233-371; and Steggerda, et al. Reα.e/7 1981, 100(2), 41). The dπ -pπ^* back-bonding effect could make the e_g levels of Cp₂V(phen) and Cp V(bpy) more susceptible to the accommodation of an extra electron, whereas the highly stabilized vanadium d-orbitals occurring through V-S bonding require higher electrode potential. The third reductive couple located at — 1.98 V represents considerably less than a full le^" process, probably resulting from an electron transfer induced chemical reaction (EC mechanism) after the system undergoes a second redox reaction at -1.73 V. It has been noted that on the cyclic voltammetric time scale, vanadium (III) complexes can release their ligands from the ancillary position after a one or two electron reduction (Taube, H. Pure Appl. Chem. 1979, 51, 901; and Petersen, et al. Inorg. Chem. 1980, 19, 1852). The origin of this third redox couple is not the Cp₂V^(I)/(n) couple, as this is reported to be at - 3.3 V (Dorer, et al. J OrganomeL Chem. 1992, 427, 245), nor is it the reduction ofthe phen or bpy coordinated ligands (Saji, et al. J. Electroanal. Chem. Interfac. Electrochem. 1975, 60, 1 ; Guadualupe, et al. J. Am. Chem. Soc. 1988, 110, 3462; and Saji, et al. J. Electroanal. Chem. Interfac. Electrochem. 1975, 63, 31). It may be a metal centered reduction from an unknown vanadium (II) complex. To our surprise, unlike Cp₂VCl₂ or Cp₂V(acac)⁺ (acac = acetyl acetonate), there is no detectable V^(IV)/(V) couple observed in the available solvent window in acetonitrile (0.0 to +2.2 V). The extra stabilization ofthe d¹ electron through back-bonding effects could impart such results. The complete electrode stoichiometry is represented by eq. 3.

+ e" + e~ + e" unknown

[Cp₂V^lv(phen)]²⁺=*- [Cp₂V^lll(phen)]⁺^^- [Cp₂v"(phen)] *- "™ <™T ^=^ reduction (eq. 3)

- e' _ - e- proαucis _ _e- _products

In physiological buffer these two complexes display only one v^{(III) (IV)} reversible redox couple centered at — 0.72 V over a potential range of -0.9 to +0.9 V indicating no alteration ofthe stereochemistry around the metal center on the cyclic voltammetric time scale upon reduction. Both the V(IV) complexes show fair stability in aqueous solution. The peak current height (I_p) of the v^{(III) (IV)} couple was reduced approximately 15% after 3 days of standing at room temperature. Similar results were also observed by EPR spectroscopy.

D. Structural Studies. ORTEP diagrams of 1 and 3 are depicted in Figures 6 and 7, respectively, along with the atom numbering scheme. The crystallographic data for [Cp₂V(phen)][OTf] (1) and [Cp₂V(bpy)][OTfj (3) are presented in Tables 3 A and 3B and selected bond distances and angles are given in Tables 4A and 4B.

Table 3 A. Crystal data and structure refinement for [Cp₂V(phen)][OTf] (1).

Empirical formula C^ Hjg Fβ ^ Oβ Sz V,

Formula weight 659.46

Temperature 293(2) K

Wavelength 0.71073 A

Crystal system, space group Monoclinic, P2(l)/n

Unit cell dimensions a = 10.2763(5) A alpha = 90 b = 18.1646(9) A beta = 99.4150(10)° c = 13.5741(7) A gamma = 90,

Volume 2499.7(2) A^

Z 4

Calculated density 1.752 g/cm³

Absorption coefficient 0.655 mm^-1

F(000) 1332

Crystal size 0.50 x 0.30 x 0.10 mm

Theta range for data collection 1.89 to 28.28°

Limiting indices -12<=Λ<=13, -22<= <=23, -17<=/<=17

Reflections collected / unique 15058 / 5723 [R(int) = 0.0466]

Completeness to theta = 28.28° 92.2 %

Absorption correction Empirical

Max. and min. transmission 0..9374 and 0.6222

Refinement method Full-matrix least-squares on F

Data / restraints / parameters 5723 / 0 / 371

Goodness-of-fit on F 1.042

Final R indices [I>2sigma(I)] RI = 0.0400, wR2 = 0.0965

R indices (all data) Rl = 0.0503, wR2 = 0.1048

Extinction coefficient 0.0009(3)

Largest diff. peak and hole 0.441 and -0.369 e. A^"3 Table 3B. Crystal data and structure refinement for [Cp V(bpy)][OTf] (3).

Empirical Formula C₂₂H₁₈F₆N₂O₆s₂Vι

Formula weight 635.44

Temperature 293(2) K

Wavelength 0.71073 A

Crystal system Monoclinic

Space group P2ι/c

Unit cell dimensions α = 10.6451(6) A alpha = 90° b = 18.3863(10)A beta = 98.6220(10)° c = 12.6993(7) A gamma = 90°

Volume, Z 2457.5(2) A³ , 4

Density (calculated) 1.718 g/cm³ Absorption coefficient 0.663 mm^"1 F (000) 1284 Crystal size 0.40 x 0.20 x 0.20 mm θ range for data collection 1.93 to 28.30° Limiting indices -14 < h < 9, -24 < k < 24, -16 < 7 < 16 Reflections collected 14839 Independent reflections 5626 (R_int = 0.0435) Absorption correction Empirical Max. and min. transmission 0.8789 and 0.7775 Refinement method Full-matrix least-squares on F Data / restraints / parameters 5620 / 0 / 353 Goodness-of-fit on F² 1.085 Final R indices [I>2σ(I)] Rl = 0.0460, wR2 = 0.1006 R indices (all data) RI = 0.0646, wR2 = 0.1128 Extinction coefficient 0.0010(3) Largest diff peak and hole 0.408 and -0.376 eA^~3

Table 4A. Bond lengths [A] and angles [°] for [Cp₂V(phen)][OTf] (1).

V-N(l) 2.1352(18) V-C(l) 2.268(3)

V-N(2) 2.1388(18) V-C(2) 2.297(3)

S(l)-O(l) 1.440(2) V-C(3) 2.296(3)

S(l)-O(2) 1.4360(19) V-C(4) 2.262(3)

S(l)-O(3) 1.430(2) V-C(5) 2.276(2)

S(2)-O(4) 1.428(2) V-C(6) 2.263(2)

S(2)-O(5) 1.431(2) V-C(7) 2.298(2)

S(2)-O(6) 1.4289(19) V-C(8) 2.312(2)

S(l)-C(23) 1.826(3) V-C(9) 2.309(2)

S(2)-C(24) 1.825(3) V-C(IO) 2.285(2)

F(l)-C(23) 1.327(3) C(l)-C(2) 1.432(5)

F(2)-C(23) 1.317(3) C(2)-C(3) 1.375(6)

F(3)-C(23) 1.331(3) C(3)-C(4) 1.342(5)

F(4)-C(24) 1.321(3) C(4)-C(5) 1.350(5) F(5)-C(24) 1.326(3) C(l)-C(5) 1.390(5)

F(6)-C(24) 1.327(3) C(6)-C(7) 1.406(4)

N(l)-C(l l) 1.332(3) C(7)-C(8) 1.400(4)

N(l)-C(15) 1.361(3) C(8)-C(9) 1.416(4)

N(2)-C(22) 1.335(3) C(9)-C(10) 1.388(3)

N(2)-C(16) 1.365(3) C(6)-C(10) 1.415(3)

C(l l)-C(12) 1.394(4) C(12)-C(13) 1.363(4)

C(13)-C(14) 1.402(4 C(14)-C(15) 1.407(3)

C(14)-C(19) 1.435(3) C(15)-C(16) 1.421(3)

C(16)-C(17) 1.402(3 C(17)-C(20) 1.403(4)

C(17)-C(18) 1.436(4) C(18)-C(19) 1.339(4)

C(20)-C(21) 1.364(4) C(21)-C(22) 1.397(4)

N(l)-V-N(2) 76.66(7) N(2)-V-C(4) 90.95(11)

N(l)-V-C(4) 131.79(11) N(2)-V-C(6) 81.87(8)

N(l)-V-C(6) 84.86(8) N(2)-V-C(l) 134.89(11)

N(l)-V-C(l) 97.77(11) N(2)-V-C(5) 125.56(10)

N(l)-V-C(5) 132.70(9) N(2)-V-C(10) 82.29(8)

N(l)-V-C(10) 119.83(8) N(2)-V-C(3) 77.07(10)

N(l)-V-C(3) 97.77(13) N(2)-V-C(2) 100.01(13)

N(l)-V-C(2) 77.96(9) C(4)-V-C(6) 139.97(11)

C(4)-V-C(l) 59.29(12) C(6)-V-C(l) 142.90(11)

C(4)-V-C(5) 34.61(12) C(6)-V-C(5) 134.34(9)

C(l)-V-C(5) 35.62(12) C(4)-V-C(10) 103.85(11)

C(6)-V-C(10) 36.25(9) C(l)-V-C(10) 134.18(12)

C(5)-V-C(10) 105.33(9) C(4)-V-C(3) 34.25(13)

C(6)-V-C(3) 157.51(12) C(l)-V-C(3) 59.13(12)

C(5)-V-C(3) 57.19(12) C(10)-V-C(3) 131.14(14)

C(4)-V-C(2) 58.32(13) C(6)-V-C(2) 161.71(11)

C(l)-V-C(2) 36.57(13) C(5)-V-C(2) 58.65(10)

C(10)-V-C(2) 161.89(11) C(3)-V-C(2) 34.83(14)

N(l)-V-C(7) 79.79(8) N(2)-V-C(7) 114.81(8)

C(4)-V-C(7) 144.74(12) C(6)-V-C(7) 35.91(9)

C(l)-V-C(7) 107.86(11) C(5)-V-C(7) 115.05(11)

C(10)-V-C(7) 59.59(9) C(3)-V-C(7) 166.53(12)

C(2)-V-C(7) 132.52(14) N(l)-V-C(9) 138.63(8)

N(2)-V-C(9) 114.50(8) C(4)-V-C(9) 89.05(12)

C(6)-V-C(9) 59.51(8) C(l)-V-C(9) 99.21(12)

C(5)-V-C(9) 75.20(9) C(10)-V-C(9) 35.17(8)

C(3)-V-C(9) 123.29(13) C(2)-V-C(9) 133.04(10)

C(7)-V-C(9) 59.13(9) N(l)-V-C(8) 109.66(9)

N(2)-V-C(8) 139.31(8) C(4)-V-C(8) 109.63(12)

C(6)-V-C(8) 59.59(8) C(l)-V-C(8) 85.17(12)

C(5)-V-C(8) 80.72(10) C(10)-V-C(8) 59.27(8)

C(3)-V-C(8) 137.75(11) C(2)-V-C(8) 120.68(14)

C(7)-V-C(8) 35.37(9) C(9)-V-C(8) 35.68(9)

O(3)-S(l)-O(2) 115.37(14) O(3)-S(l)-O(l) 115.07(15)

O(2)-S(l)-O(l) 114.82(15) O(3)-S(l)-C(23) 102.87(14)

O(2)-S(l)-C(23) 103.51(13) O(l)-S(l)-C(23) 102.65(13) O(4)-S(2)-O(6) 114.92(14) O(4)-S(2)-O(5) 1 15.61(16)

O(6)-S(2)-O(5) 114.96(14) O(4)-S(2)-C(24) 102.70(14)

O(6)-S(2)-C(24) 103.36(13) O(5)-S(2)-C(24) 102.66(13)

C(l l)-N(l)-C(15) 118.25(19) C(l l)-N(l)-V 126.79(16)

C(15)-N(l)-V 114.94(14) C(22)-N(2)-C(16) 117.94(19)

C(22)-N(2)-V 127.45(16) C(16)-N(2)-V 114.59(13)

C(5)-C(l)-C(2) 105.1(3) C(5)-C(l)-V 72.53(15)

C(2)-C(l)-V 72.83(16) C(3)-C(2)-C(l) 106.7(3)

C(3)-C(2)-V 72.55(16) C(l)-C(2)-V 70.60(15)

C(4)-C(3)-C(2) 109.7(3) C(4)-C(3)-V 71.47(17)

C(2)-C(3)-V 72.61(18) C(3)-C(4)-C(5) 108.7(3)

C(3)-C(4)-V 74.29(19) C(5)-C(4)-V 73.28(16)

C(4)-C(5)-C(l) 109.7(3) C(4)-C(5)-V 72.10(16)

C(l)-C(5)-V 71.85(15) C(7)-C(6)-C(10) 107.6(2)

C(7)-C(6)-V 73.37(13) C(10)-C(6)-V 72.71(12)

C(8)-C(7)-C(6) 108.2(2) C(8)-C(7)-V 72.89(13)

C(6)-C(7)-V 70.72(13) C(7)-C(8)-C(9) 107.6(2)

C(7)-C(8)-V 71.74(13) C(9)-C(8)-V 72.04(13)

C(10)-C(9)-C(8) 108.4(2) C(10)-C(9)-V 71.46(12)

C(8)-C(9)-V 72.28(13) C(9)-C(10)-C(6) 108.1(2)

C(9)-C(10)-V 73.37(13) C(6)-C(10)-V 71.04(12)

C(13)-C(12)-C(l l) 119.8(2) C(12)-C(13)-C(14) 119.8(2)

C(13)-C(14)-C(15) 117.1(2) C(13)-C(14)-C(19) 125.0(2)

C(15)-C(14)-C(19) 117.9(2) N(l)-C(15)-C(14) 122.7(2)

N(l)-C(15)-C(16) 116.79(19) C(14)-C(15)-C(16) 120.5(2)

N(2)-C(16)-C(17) 123.0(2) N(2)-C(16)-C(15) 116.82(18)

C(17)-C(16)-C(15) 120.2(2) C(16)-C(17)-C(20) 117.4(2)

C(16)-C(17)-C(18) 118.4(2) C(20)-C(17)-C(18) 124.2(2)

C(19)-C(18)-C(17) 121.2(2) C(18)-C(19)-C(14) 121.7(2)

C(21)-C(20)-C(17) 1 19.3(2) C(20)-C(21)-C(22) 120.2(2)

N(2)-C(22)-C(21) 122.1(2) F(2)-C(23)-F(l) 107.2(2)

F(2)-C(23)-F(3) 107.3(2) F(l)-C(23)-F(3) 106.2(2)

F(2)-C(23)-S(l) 113.0(2) F(l)-C(23)-S(l) 1 11.7(2)

F(3)-C(23)-S(l) 111.07(19) F(4)-C(24)-F(5) 107.9(2)

F(4)-C(24)-F(6) 106.5(2) F(5)-C(24)-F(6) 106.1(2)

F(4)-C(24)-S(2) 111.74(19) F(5)-C(24)-S(2) 112.0(2)

F(6)-C(24)-S(2) 112.24(19)

Table 4B. Bond lengths [A] and angles [°] for [Cp₂V(bpy)] [OTf] (3).

V-N(l) 2.129(2) SO)-O(l) 1.420(2) V-N(2) 2.128(2) S(l)-O(2) 1.428(2) V-C(l) 2.299(3) S(l)-O(3) 1.429(3) V-C(2) 2.266(3) S(2)-O(4) 1.430(2) V-C(3) 2.283(3) S(2)-O(5) 1.428(2) V-C(4) 2.302(3) S(2)-O(6) 1.429(2) V-C(5) 2.304(3) S(l)-C(21) 1.810(4) V-C(6) 2.315(3) S(2)-C(22) 1.821(3)

V-C(7) 2.293(3) F(l)-C(21) 1.296(5)

V-C(8) 2.280(3) F(2)-C(21) 1.327(4)

V-C(9) 2.298(3) F(3)-C(21) 1.314(5)

V-C(IO) 2.294(3) F(4)-C(22) 1.314(4)

N(l)-C(l l) 1.349(3) F(5)-C(22) 1.321(4)

N(l)-C(15) 1.346(3) F(6)-C(22) 1.318(4)

N(2)-C(16) 1.358(3) C(l)-C(2) 1.405(4)

N(2)-C(20) 1.346(3) C(2)-C(3) 1.410(4)

C(l l)-C(12) 1.379(4) C(3)-C(4) 1.383(4)

C(12)-C(13) 1.372(5) C(4)-C(5) 1.413(4)

C(13)-C(14) 1.375(4) C(l)-C(5) 1.391(4)

C(14)-C(15) 1.392(4) C(6)-C(7) 1.391(4)

C(15)-C(16) 1.465(4) C(7)-C(8) 1.415(4)

C(16)-C(17) 1.385(4) C(8)-C(9) 1.412(5)

C(17)-C(18) 1.367(4) C(9)-C(10) 1.407(5)

C(18)-C(19) 1.379(5) C(6)-C(10) 1.392(4)

C(19)-C(20) 1.371(4)

N(2)-V-N(l) 75.69(8) N(2)-V-C(2) 86.46(10)

N(l)-V-C(2) 80.75(10) N(2)-V-C(8) 89.02(10)

N(l)-V-C(8) 128.30(10) C(2)-V-C(8) 148.25(12)

N(2)-V-C(3) 121.57(9) N(l)-V-C(3) 83.19(9)

C(2)-V-C(3) 36.10(11) C(8)-V-C(3) 142.68(11)

N(2)-V-C(7) 78.50(10) N(l)-V-C(7) 92.24(10)

C(2)-V-C(7) 164.60(11) C(8)-V-C(7) 36.06(11)

C(3)-V-C(7) 157.02(11) N(2)-V-C(10) 137.49(10)

N(l)-V-C(10) 100.46(11) C(2)-V-C(10) 135.42(11)

C(8)-V-C(10) 59.75(12) C(3)-V-C(10) 99.39(11)

C(7)-V-C(10) 59.18(11) N(2)-V-C(9) 124.66(11)

N(l)-V-C(9) 134.72(11) C(2)-V-C(9) 134.17(11)

C(8)-V-C(9) 35.93(11) C(3)-V-C(9) 108.90(11)

C(7)-V-C(9) 59.50(11) C(10)-V-C(9) 35.69(12)

N(2)-V-C(l) 79.84(9) N(l)-V-C(l) 112.78(10)

C(2)-V-C(l) 35.83(11) C(8)-V-C(l) 112.49(12)

C(3)-V-C(l) 59.49(11) C(7)-V-C(l) 141.37(12)

C(10)-V-C(l) 136.37(12) C(9)-V-C(l) 110.69(12)

N(2)-V-C(4) 138.70(9) N(l)-V-C(4) 116.13(10)

C(2)-V-C(4) 59.24(10) C(8)-V-C(4) 107.85(11)

C(3)-V-C(4) 35.11(10) C(7)-V-C(4) 135.76(11)

C(10)-V-C(4) 81.69(11) C(9)-V-C(4) 76.84(11)

C(l)-V-C(4) 58.93(10) N(2)-V-C(5) 108.62(10)

N(l)-V-C(5) 138.98(10) C(2)-V-C(5) 59.41(11)

C(8)-V-C(5) 92.72(11) C(3)-V-C(5) 59.31(10)

C(7)-V-C(5) 128.78(11) C(10)-V-C(5) 101.47(12)

C(9)-V-C(5) 77.69(11) C(l)-V-C(5) 35.19(11)

C(4)-V-C(5) 35.73(10) N(2)-V-C(6) 105.43(10)

N(l)-V-C(6) 77.60(10) C(2)-V-C(6) 151.68(12)

C(8)-V-C(6) 59.18(12) C(3)-V-C(6) 122.27(11) C(7)-V-C(6) 35.14(11) C(10)-V-C(6) 35.15(11)

C(9)-V-C(6) 58.78(11) C(l)-V-C(6) 169.45(12)

C(4)-V-C(6) 115.65(11) C(5)-V-C(6) 134.99(11)

O(l)-S(l)-O(2) 115.72(17) O(l)-S(l)-O(3) 115.11(18)

O(2)-S(l)-O(3) 114.31(15) O(l)-S(l)-C(21) 102.20(17)

O(2)-S(l)-C(21) 103.79(17) O(3)-S(l)-C(21) 103.2(2)

O(5)-S(2)-O(6) 115.10(16) O(5)-S(2)-O(4) 115.32(16)

O(6)-S(2)-O(4) 114.16(16) O(5)-S(2)-C(22) 103.38(15)

O(6)-S(2)-C(22) 103.55(15) O(4)-S(2)-C(22) 103.03(15)

C(15)-N(l)-C(l l) 119.2(2) C(15)-N(l)-V 117.10(17)

C(l l)-N(l)-V 123.69(19) C(20)-N(2)-C(16) 118.5(2)

C(20)-N(2)-V 124.54(18) C(16)-N(2)-V 117.01(17)

C(5)-C(l)-C(2) 108.2(3) C(5)-C(l)-V 72.59(16)

C(2)-C(l)-V 70.83(15) C(l)-C(2)-C(3) 107.7(3)

C(l)-C(2)-V 73.34(16) C(3)-C(2)-V 72.59(15)

C(4)-C(3)-C(2) 107.9(3) C(4)-C(3)-V 73.18(15)

C(2)-C(3)-V 71.31(15) C(3)-C(4)-C(5) 108.5(3)

C(3)-C(4)-V 71.71(15) C(5)-C(4)-V 72.21(15)

C(l)-C(5)-C(4) 107.6(3) C(l)-C(5)-V 72.22(16)

C(4)-C(5)-V 72.06(15) C(7)-C(6)-C(10) 109.0(3)

C(7)-C(6)-V 71.58(16) C(10)-C(6)-V 71.60(16)

C(6)-C(7)-C(8) 107.9(3) C(6)-C(7)-V 73.28(16)

C(8)-C(7)-V 71.45(16) C(9)-C(8)-C(7) 107.4(3)

C(9)-C(8)-V 72.75(16) C(7)-C(8)-V 72.49(16)

C(10)-C(9)-C(8) 107.8(3) C(10)-C(9)-V 71.99(16)

C(8)-C(9)-V 71.32(16) C(6)-C(10)-C(9) 108.0(3)

C(6)-C(10)-V 73.25(16) C(9)-C(10)-V 72.32(16)

N(l)-C(l l)-C(12) 121.7(3) C(13)-C(12)-C(l l) 119.5(3)

C(12)-C(13)-C(14) 119.2(3) C(13)-C(14)-C(15) 119.6(3)

N(l)-C(15)-C(14) 120.9(3) N(l)-C(15)-C(16) 115.4(2)

C(14)-C(15)-C(16) 123.7(3) N(2)-C(16)-C(17) 121.0(2)

N(2)-C(16)-C(15) 114.8(2) C(17)-C(16)-C(15) 124.2(2)

C(18)-C(17)-C(16) 119.9(3) C(17)-C(18)-C(19) 119.1(3)

C(20)-C(19)-C(18) 119.1(3) N(2)-C(20)-C(19) 122.5(3)

F(l)-C(21)-F(3) 106.2(4) F(l)-C(21)-F(2) 107.8(3)

F(3)-C(21)-F(2) 106.7(4) F(l)-C(21)-S(l) 113.2(3)

F(3)-C(21)-S(l) 111.8(3) (2)-C(21)-S(l) 110.8(3)

F(4)-C(22)-F(6) 107.8(3) F(4)-C(22)-F(5) 107.7(3)

F(6)-C(22)-F(5) 106.7(3) F(4)-C(22)-S(2) 112.0(2)

F(6)-C(22)-S(2) 111.6(2) F(5)-C(22)-S(2) 110.7(2)

The two π-bonded Cp rings and the two nitrogen atoms of phenanthroline (1) or bipyridine (3) formally occupy the pseudotetrahedral coordination sites around the vanadium (IV) center. The cyclopentadienyl ring A (C1-C5) and ring B(C6-C10) are planar in both complexes. The dihedral angle between the two cyclopentadienyl rings is 46.07° for Cp₂V(phen) (1) and 46.17° for Cp₂V(bpy) (3). The bisecting angles defined by the VN₂ plane with respect to the two Cp rings are 26.97 (6)° (ring A) and 19.10 (7)° (ring B) for Cp₂V(phen) (1), and 21.24 (6)°(ring A) and 24.93 (13)° (ring B) for Cp₂V(bpy) (3). The chelated ring is inclined closer to one of its neighboring Cp rings and it is not clear to us at present why the heterocyclic chelated ring is shifted toward one of the Cp rings over the other. The degree of inclination between the VN₂ plane with respect to the plane ofthe remaining carbon atoms is more pronounced for phen (3.89 (5)°) than bpy (1.69 (11)°). This could be attributed to the difference between the heterocyclic ligands themselves. A similar effect is also observed for titanium analogues (Thewalt, et al. J. OrganomeL Chem. 1986, 302, 193). The vanadium atom projects 1.96 A below the center of ring A and 1.96 A above the mean plane of ring B for the complexes. The angles between the ring centroid-V-ring centroid vectors for Cp₂V(phen) (1) and Cp₂V(bpy) (3) are 133.63° and 133.24°, respectively, which are slightly lower than Cp₂V(DeDtc)⁺ (134.6°) (Ghosh, et al. J. Inorg. Biochem. 1998, 72(1-2), 89) but comparable to those of the titanocene (IV) analogues Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4) (Table 5).

Table 5 Comparison of Selected Bond Lengths and Angles for Compounds 1-4

While not intended to limit the invention, this could reflect the higher steric demands ofthe five membered chelated phen or bpy unit compared to the four membered DeDtc ligand. The N1-V-N2 angles are 76.66 (8)° and 75.69 (8)° in Cp₂V(phen) (1) and Cp₂V(bpy) (3), respectively. These values are comparable to the reported five membered chelated oxametalacycles of Cp₂V^(IV) (Gambarotta, et al. Organometallics 1986, 5, 2425), as well as to the titanocene (IV) analogues (Table 5), but are significantly smaller when compared with the Cp₂V(l, 2-benzenedithiolate) complex where the bite angle of the five membered ring is 79.9°( Stephan, D. Inorg. Chem. 1992, 57, 4218). The relatively larger S-V-S angle compared to the N-V-N angle is probably due to the greater flexibility offered by the thiolate chelated ring vs. the rigid phen or bpy ligand where the donor nitrogen centers are fixed in a six membered heterocycle. The V-N distances are 2.135 (1) A and 2.129 (1) A for Cp₂V(phen) (1) and Cp₂V(bpy) (3) which are slightly shorter than the Ti-N bond distances reported for Cp₂Ti(phen) (2) or Cp₂Ti(bpy) (4) (Table 5). There is no literature precedence for a single bonded V-N distance coordinated to a Cp₂V(IV) unit except for η²-iminoacyl coordinated vanadocene(IV) (Carrier, et al. Organometallics 1987, 6, 454; and Rettig, et al. Inorg. Chem. 1969, 8, 2685). Crystal structure data revealed that with the three membered chelation, the V-N distance is 2.054 (4) A which is significantly shorter than observed for Cp₂V(phen) (1) or Cp₂V(bpy) (3). In the three cases, ligand moieties have the vacant π^* orbital available for back-bonding and the difference in bond lengths could be rationalized in terms of their differences in dπ-pπ* back-bonding which is more pronounced for the three membered η² chelated ring. This is supported by the comparison ofthe V-N (2.054 (4) A) bond distance ofthe η²-iminoacyl coordinated complex to its d° Ti(IV) analogue (2.149 (4) A) (Van Bolhuis, et al. J. OrganomeL Chem. 1979, 770, 299). Without the involvement of a single d¹ electron in bonding, the M-N distances in such systems would be expected to be very close to each other. This is indeed the case for the titanium compounds when their Ti-N distance is compared with the η -iminoacyl vs. bpy coordinated complexes (Thewalt, et al. J. OrganomeL Chem. 1986, 302, 193; and Van Bolhuis, et al. J OrganomeL Chem. 1979, 770, 299). The small observed difference between vanadium- (Cp2V(phen) (1) or Cp₂V(bpy) (3)) or titanium- (Cp₂Ti(phen) (2) or Cp₂Ti(bpy) (4)) nitrogen bond distances arises from a delocalization imparted by a d¹ electron via dπ-pπ^* back-bonding. Example 2. Interaction of Metallocenes with Membrane

Methods

A. Carboxyfluorescein Leakage Experiments. Liposomes with PC:cholesterol (3:1) or PG-cholesterol (3:1) were formed in 0.1 M phosphate buffer, pH 7.2, using the ethanol injection method (Betageri, et al. Liposome Drug Delivery Systems; Technomic: Lancaster, PA 1993; pp 13-14) to a 1 mM total lipid concentration with carboxyfluorescein encapsulated under self-quenching conditions (0.1 M carboxyfluorescein). The liposomes encapsulated with carboxyfluorescein were separated from free carboxyfluorescein by passage over a Sephadex 25 column (Pharmacia Biotech) to give a final lipid concentration of 0.7 mM. Compounds Cp₂V(phen) (1), Cp₂V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), and Cp₂V(acac) (7) were each dissolved in methanol and added to the liposome solutions to a final metal complex concentration of 50 - 400 μM. (Control experiments demonstrated that the small amount of methanol added did not effect the permeabilization ofthe liposome.) Compounds Cp2Ti(phen) (2) and Cp₂Ti(bpy) (4) were less stable in methanol over long periods of time and so were dissolved in acetonitrile and added to the liposome solutions in the same fashion. Compound Cp₂V(cat) (8) was unstable in methanol and so was dissolved in distilled water and added to the liposome solutions in the same way. The increase in fluorescence was monitored at λ_exc=550 nm and λ_em=490 nm for 6 min at 20°C (Deleers, M.; Servais, J.-P.; Wulfert, E. Biochim. Biophys. Acta 1985, 575, 195; Verstraeten, S. J.; Oteiza, P. I. Arch. Biochem. Biophys. 1995, 322, 284; and Bramhalll, J.; Hofmann, J.; DeGuzman, R.; Montestruque, S.; Schell, R. Biochemistry 1987, 26, 6330). Complete liposome disruption was achieved by the addition of excess Triton X- 100 (lOμL of a 10% aqueous solution).

Carboxyfluorescein release was calculated as shown in eq. 1 where CF = carboxyfluorescein release, Fo = fluorescence intensity ofthe intact liposome, F = fluorescence intensity at time = 6 min, and F_t = fluorescence intensity with Triton X.

CF (%) = [(F - Fo)/(F_t - F₀)] 100 (eq. 1)

B. Lipid Peroxidation. Aliquots of 0.5 mL of liposomes (0.7 mM total lipid) were incubated with the metal complexes (50-400 μM) at 37°C for 90 min. The incubation was stopped by the addition of 0.1 mL 4% butylated hydroxy-toluene in EtOH. Sodium dodecyl sulfate (3%; 0.25 mL) was added to destroy the liposomes followed by the addition of 0.5 mL of 1%> 2-thiobarbituric acid in 0.05 M NaOH and 0.5 mL of 25%) HC1. The samples were mixed and heated in boiling water for 15 min. The 2-thiobarbituric acid reactive substances were extracted into 3 mL of 1 -butanol, and the fluorescence ofthe butanol layer was measured at λ_exc=515 nm and λ_em=555nm (Ohsumi, Y.; Kitamoto, K.; Anraku, Y. J. Bacteriol, 1998, 770, 2676;Gutteridge, J. M. C; Quinlan, G. J.; Clark, I.; Halliwell, B. Biochim. Biophys. Acta 1985, 835, 441; Quinlan, G. J.; Halliwell, B.; Moorhouse, C. P.; Gutteridge, J. M. C. Biochim. Biophys. Acta 1988, 962, 196; Deleers, M.; Servais, J.-P.; Wulfert, E. Biochim. Biophys. Acta 1985, S75, 195; Verstraeten, S. J.; Oteiza, P. I. Arch. Biochem. Biophys. 1995, 322, 284; and Verstraeten, S. V.; Nogueira, L. V.; Schreier, S.; Oteiza, P. I. Arch. Biochem. Biophys. 1997, 338, 121).

The 2-thiobarbituric acid reactive substances are reported as malonaldehyde equivalents (Gutteridge, J. M. C. Anal. Biochem. 1975, 69, 518).

C. Lipid Packing Order. Liposomes (0.7 mM) containing 1 mol % of 1,6- diphenyl-l,3,5-hexatriene were prepared as described (Bramhall, J., Hofmann, J., DeGuzman, R., Montestruque, S., Schell, R. Biochemistry 1987, 26, 6330). Only liposomes with zwitterionic lipids could be used because the presence of diphenylhexatriene caused the negative liposomes to precipitate out immediately. The metal complexes were added to the liposome solutions in increments of 50μM. The solutions were incubated for 5 min at 20°C prior to measurement of fluorescence intensities at λ_exc=360 nm and λ_em ^:=450 nm. The extent of polarization was calculated according to the method described by Jahnig (Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 6361).

D. Liposome Aggregation. The metal complexes were added to 0.5 mL liposome solution (0.7 mM total lipid) to yield a final concentration of 50-400 μM vanadium or titanium. The aggregation of liposomes was measured as the increase in absorbance at 300 nm in a UV-vis spectrophotometer over 15 min at 20°C (Verstraeten, et al. Arch. Biochem. Biophys. 1995, 322, 284). E. Liposomes Fusion. A 1 :4 ratio of liposomes (0.7 mM total lipid) containing 2% N-NBD-PE, N-Rh-PE, and liposomes with no labeled lipids was prepared in phosphate buffer, pH 7.2. Fusion was measured after the addition ofthe vanadocene- chelated complexes by changes in the fluorescence intensity at λ_exc^SO nm and λ_em=470 nm at 20°C (Verstraeten, S. J.; Oteiza, P. I. Arch. Biochem. Biophys. 1995, 322, 284; and Struck, D. K.; Hoekstra, D.; Pagano, R. E. Biochemistry 1981, 20, 4093). Complete liposome disruption was achieved by the addition of excess Triton X-100 (lOμL of a 10%) solution). When liposomes containing N-NBD-PE and N-Rh-PE fuse with liposomes not containing probes, the surface density decreases, resulting in a decreased efficiency of resonance energy transfer from N-NBD-PE to N-Rh-PE (Struck, D. K.; Hoekstra, D.; Pagano, R. E. Biochemistry 1981, 20, 4093).

Results A. Effects of Metallocene-Chelated Complexes on Liposome Leakage. The concentration dependence for metal-complex-induced permeability of liposomes, as measured by the release of 5(6)-carboxyfluorescein, is shown in Figure 1 A for zwitterionic (PC) liposomes and Figures IB and 2 for negative (PG) liposomes. The metal complexes themselves do not interfere with the fluorescence ofthe carboxyfluorescein probe. At a concentration of 400 μM, the phenanthroline chelated complexes, Cp₂V(phen) (1) and Cp V(bpy) (3), induce a release of 51 %> and 34 %, respectively, ofthe entrapped carboxyfluorescein, while the bipyridine chelated complexes, Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4), release 54% and 33%>, respectively, of the entrapped carboxyfluorescein (Figure IC). The kinetics of Cp₂V(bpy) (3)-induced liposome leakage shows that the process has reached a plateau after 2 min of incubation (Figure 2). Similar kinetics are observed for Cp₂V(PH) (5). The vanadocene derivatives Cp₂V(bpy) (3) and Cp₂V(PH) (5) cause liposome permeability in a concentration- dependent manner. Cp2V(PH) (5) induces the release of approximately 20% ofthe entrapped carboxyfluorescein at a concentration of 400μM, while vanadocene-chelated complexes, Cp2V(Et2(dtc)) (6), Cp2V(acac) (7), and Cp2V(cat) (8), have no effect on carboxyfluorescein leakage under the conditions tested. Control experiments with all of the chelating ligands confirmed that the free ligands had no effect on the permeation. B. Effects of Metallocene-Chelated Complexes on Liposomal Aggregation and Fusion. The membrane permeability effects ofthe metallocene complexes are not associated with aggregation of liposomes indicative of membrane disruption, as shown by the lack of an increase in turbidity with addition of any ofthe vanadocene complexes at concentrations up to a 400 μM as measured by an increase in absorbance at 300 nm for PC or PG liposomes (Figure 3). Similarly, no fusion ofthe liposomes is seen with the addition of any ofthe vanadocene-chelated complexes, Cp₂V(phen) (1), Cp₂V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), Cp₂V(acac) (7), and Cp₂V(cat) (8), as indicated by a constant efficiency of energy transfer between fluorescent liposomes.

C. Effects of Vanadocene-Chelated Complexes on Lipid Packing Order.

Since the PC: cholesterol liposome is already in the liquid phase, an increase in fluorescence polarization would indicate membrane rigidification. No such increase occurs after the addition ofthe vanadocene-chelated complexes, Cp₂V(phen) (1), Cp₂V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), Cp₂V(acac) (7), and Cp₂V(cat) (8), at 20°C (Figure 4).

D. Effects of Vanadocene-Chelated Complexes on Lipid Peroxidation. The amount of lipid peroxidation caused by the incubation of PC liposomes with the vanadocene-chelated complexes, Cp₂V(phen) (1), Cp V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), Cρ₂V(acac) (7), and Cp₂V(cat) (8), is shown in Figure 5A, and amount of lipid peroxidation caused by the incubation of PG liposomes with the vanadocene-chelated complexes, Cp₂V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), Cp₂V(acac) (7), and Cp₂V(cat) (8), is shown in Figure 5B. A comparison between the amount of lipid peroxidation caused by vanodecene complexes Cp₂V(phen) (1) and Cp₂V(bpy) (3) and their corresponding metallocene complexes Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4) is shown in Figure 5C. The vanadocene complexes Cp₂V(phen) (1), Cp₂V(acac) (7) and Cp₂V(bpy) (3) induce lipid peroxidation, as measured by the production of 2-thiobarbituric acid reactive substances. However, the other tested complexes, Cp₂Ti(phen) (2), Cp₂Ti(bpy) (4), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), and Cp₂V(cat) (8), have no effect on lipid peroxidation. E. Cyclic Voltammetry. The V(IV)/V(V) redox potential was not observed for any ofthe complexes over the potential range of +1.0 to 1.0 V with a glassy carbon electrode, although the V(IV)/V(III) couple was readily observed (Table 6). Table 6 also shows the membrane permeability and lipid peroxidation effects ofthe vanadocene- chelated complexes in the presence of liposomes. The data reveal that the V(IV)/V(III) redox couple does not appear to correlate with membrane permeability or lipid peroxidation.

Table 6: Redox couple, membrane permeability, and lipid peroxidation by vanadocene complexes.

Discussion

Vanadocene complexes Cp₂V(bpy) (3) and Cp₂V(PH) (5) cause the leakage of carboxyfluorescein from both zwitterionic and negatively charged liposomes as seen in Figure 2. The vanadocene-chelated complexes, Cp₂V(Et₂(dtc)) (6), Cp₂V(acac) (7), and Cp₂V(cat) (8), on the other hand, have little or no effect on the permeability of liposomes under the conditions tested, indicating special effects imparted by Cp₂V(bpy) (3) and Cp₂V(PH) (5). None ofthe vanadocene-chelated complexes cause appreciable liposome aggregation or fusion in either type of liposome. If the complexes were acting in such a way as to create localized areas of inverted micelles, fusion and aggregation would have been expected, as areas of hydrophobic patches were exposed to the aqueous solution (Hunt, G. R. A. FEBS Lett. 1980, 779, 132.). Experimental evidence herein indicates that this is not the case. In addition, based on the cyclic voltammetric measurements as well as from the characteristic sharp eight line spectral features in the EPR ofthe complexes, all ofthe vanadocene-chelated complexes studied retain their pseudotetrahedral-like structure in the liposomal matrices with respect to the central metal V(IV) ion.

The type of permeation that we observe caused by the two vanadocene-chelated complexes Cp₂V(bpy) (3) and Cp2V(PH) (5) is similar to what is seen with the addition of lipophilic amines. The amine additives caused concentration-dependent permeation of probe molecules without altering the gross features ofthe vesicles. Low concentrations of surfactants also cause an increase in membrane permeability; however, what changes occur in the liposome or vesicle morphology are not known. Metallocenes, however, are not surfactants nor do they contain long hydrocarbon chains. In addition, the metallocene complexes are much shorter than the 30-40 A required to span the bilayer and form channels (Fuhrhop, J.-H.; Liman, U. J. Am. Chem Soc. 1984, 106, 4643). Therefore, the mechanism of activity of these organometallic complexes to cause such effects is likely to be quite different from that ofthe surfactants or ionophores.

The permeation effect caused by Cp₂V(bpy) (3) and Cp₂V(PH) (5) could be explained in terms ofthe difference in their overall structural configurations compared to the three other chelated vanadocene derivatives. Both Cp₂V(bpy) (3) and Cp₂V(PH) (5) possess two aromatic rings besides their core VCp₂ unit in the inner coordination sphere ofthe V(IV) ion. In the former complex, the two heterocyclic aromatic rings are fused in a plane constituting the V, N, N atoms ofthe ancillary positions of a pseudotetrahedral geometry, while in the latter complex, the two planar phenyl rings are covalently linked to the carbon and nitrogen atoms ofthe hydroxamate moiety, thus conferring flexibility along the C-C and C-N axis, respectively. While not intended to limit the invention, it is reasonable that both configurations are more able to insert inside the hydrophobic portion ofthe membrane compared to the other three chelated complexes.

The vanadocene-chelated complexes described here are tetrahedral in geometry in which the two Cp rings are positioned in a bent sandwich conformation with respect to the V(IV) central metal ion. Presumably, this geometry is necessary for leakage to occur because the same effect could not be detected for the free ligand in control experiments using 2,2 -bipyridine or phenylbenzohydroxamic acid instead of their respective vanadocene-chelated complexes. The effect ofthe vanadocene-chelated complexes wedged in localized patches ofthe membrane may render it temporarily "leaky" without affecting the overall integrity ofthe liposome.

The difference in the level of permeation, viz. 35% vs 20%), between Cp₂V(bpy) (3) and Cp₂V(PH) (5) at 400μM concentration could be attributed to their relative difference in perturbation once intercalated. The two fused heterocyclic rings in the bipyridine ligand, by virtue of its rigidity, may cause relatively wider patches of perturbation in the packing ofthe hydrocarbon chains than the two separate rotatable phenyl rings present in the phenylbenzohydroxamate ligand. The vanadocene derivative Cp₂V(cat) (8) also contains a planar, aromatic ring as a chelating ligand; however, it does not modulate membrane permeability. Although not intended to limit the invention, it is possible that in this case one ring may not be large enough to create the required patch in the liposome through which the carboxyfluorescein can permeate. These results lead to our conclusion that it is the preferential configuration of Cp₂V(bpy) (3) and Cp₂V(PH) (5) that is responsible for interacting with the membrane to cause permeation of encapsulated dye molecules.

Lipid peroxidation does not seem to be correlated with membrane permeabilization. For example, compound Cp₂V(acac) (7) causes significant lipid peroxidation, but does not cause appreciable permeation ofthe membrane under the conditions tested. Also the titanocene complexes Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4) do not cause lipid peroxidization, but cause significant permeation ofthe membrane. The present results lead us to believe that peroxidation is not the cause ofthe formation of leaky patches since Cp₂V(acac) (7) exhibits the strongest peroxidation ofthe lipids but does not appreciably increase the permeability of a dye molecule through the membrane under the conditions tested.

Furthermore, since rigidification of membranes correlates with metal-ion- stimulated propagation of lipid peroxidation, changes in the membrane fluidity may facilitate the lipid peroxidation process. However, since none ofthe vanadocene- chelated complexes, Cp₂V(bpy) (3), Cp₂V(PH) (5), Cp₂V(Et₂(dtc)) (6), Cp₂V(acac) (7), and Cp V(cat) (8), have an effect on the overall packing order ofthe hydrophobic chains ofthe zwitterionic liposomes at the temperature the experiments were conducted, it is highly unlikely that the differences in lipid peroxidation can be explained by rigidification ofthe hydrocarbon portion ofthe liposome. It is also not likely that charge effects play an important role with respect to the vanadocene-chelated complex-liposome interaction because there is very little difference between zwitterionic and negatively charged liposomes for any ofthe properties studied. No correlation has been found between the charge on the vanadocene derivative and its ability to cause permeation or peroxidation effects. While not intended to limit the invention, the hydrophobic interactions between the chelated ligands and the hydrocarbon portion ofthe liposome appear to be more important than electrostatic interactions between the complexes and the charged headgroups.

Aggregation, measured by turbidity with UV/vis spectroscopy, fusion properties, by fluorescence energy resonance transfer measurements, and rigidification studies, through fluorescence polarization, clearly reveal the fact that none ofthe added complexes change the structural integrity ofthe liposomes. These results closely resemble the observations made by D'Cruz et al. from biological assays, which indicated that the spermicidal vanadocene-chelated complexes do not disrupt the sperm plasma membrane (D'Cruz, et al. Biol. Reprod. 1998, 58, 1515; and D'Cruz, et al. Mol. Hum. Reprod. 1998, 4, 683).

While not intending to limit the invention, in comparing vanadocene complexes to titanocene complexes, it appears that the magnitude ofthe effect may be modulated by the degree of hydrophobicity and planarity ofthe ancillary ligand ofthe complex since there was no correlation between permeation and the charge on the metallocenes and their ability to initiate lipid peroxidation. For example, changing the ligand from bipyridine (complexes Cp V(bpy) (3) and Cp₂Ti(bpy) (4)) to phenanthroline (complexes Cp₂V(phen) (1) and Cp₂Ti(phen) (2)) increases the hydrophobicity and rigidity ofthe molecule, and the result is an increase in the permeability ofthe liposomal membranes from approximately 35%> to approximately 50%>, regardless ofthe cenral metal ion (Figure IC). Furthermore, X-ray data reveals that the structural features are nearly identical in both the vanadocene complexes (Cp₂V(phen) (1) and Cp₂V(bpy) (3)) and their titanium analogues (Cp₂Ti(phen) (2) and Cp₂Ti(bpy) (4)) in terms of bond lengths and conformational geometry (Table 5). This shows that the observed trend may indeed be due to the structural details ofthe complexes. Most likely, the metal complexes are wedged in localized patches ofthe membrane, rendering it temporarily leaky without affecting the overall integrity ofthe liposome. While not intended to limit the invention, the better binding ofthe phenanthroline complex, as well as its greater size and rigidity. as compared to the bipyridine coordinated complex, could be creating larger areas of leaky patches and hence causing a greater percentage ofthe dye molecule to leak out.

Conclusions

Applicants have found that some metallocene complexes cause the liposomes to become permeable, and this effect does not appear to be related to the extent of peroxidation ofthe lipids by the metal-chelated complexes or to a tendency to cause aggregation or fusion or alter the membrane packing order. While not intended to limit the invention, applicants believe that in order to observe such properties, these compounds have a unique structural requirement, particularly the hydrophobicity, planarity, and rigidity ofthe coordinated ancillary ligands, which could alter the membrane by intercalation.

Example 3. Synthesis and characterization of metallocene complexes.

Methods and Materials

Dilinoleoyl phosphatidyl choline (DLPC) was purchased from Avanti Polar Lipids (Alabaster, AL). 3-Nitro-2,4-pentanedione was prepared by literature methods (Yoshida, et al. Tetrahedron, 1970, 26, 5691). All other reagents were obtained from Aldrich Chemical Co. (Milwaukee, WI) and used without modification unless otherwise stated.

Synthesis. The vanadocene chelated complexes used in this study are shown in Figure 12. Syntheses were performed under an inert atmosphere using standard Schlenk techniques, unless otherwise noted. [Cp₂V(acac)][OTf] (7) was prepared according to literature procedures (Doyle, et al. Ingor. Chem. 1968, 7, 2479). [Cp₂VCl₂][OTf] (13) was purchased from Aldrich Chemical Co. (Milwaukee, WI) and purified by slow recrystalhzation at -30 °C in a dry, saturated solution of hydrochloric acid in methylene chloride. The bis(cyclopentadienyl)-vanadium(IV) (acetylacetonate) complexes substituted in the 3-position were synthesized as follows. Cp₂VCl₂ (1.0 mmol) and AgOTf (2.1 mmol) were dissolved in water. The color turned to the dark blue of Cp₂V(OTf)₂ and AgCl precipitated out of solution. After stirring for 15 min the solution was filtered through celite. A THF solution ofthe carbonyl, 3-ethyl-2,4-pentadione (2.2 eq.) for 9, 3-methyl-2,4-pentadione (2.2 eq.) for 10, 3-chloromalonyl (2.8 eq.) for 11, and 3-nitro-2,4-pentadione (3.1 eq.) for 12, was added dropwise. The amount of time the reaction was allowed to proceed was critical to the synthesis ofthe compound. 9 and 10 were stirred for 4 h, 11 was allowed to stand for 16 h, and 12 was stirred for 5 min. The brown solution was then extracted with methylene chloride. The solvent ofthe organic layer was removed under vacuum. The brown residue was dissolved in 10 mL of methylene chloride, layered with pentane and allowed to stand overnight. A brown precipitate crystallized out of solution and was washed with copious amounts of hexanes.

[Cp₂V(3-metm)][OTf] (2). 41% yield. IR: 3119 (m), 3095 (w), 2969 (w), 1567 (s), 1521 (w), 1460 (w), 1449 (w), 1373 (w), 1356 (w), 1321 (w), 1274 (vs), 1258 (s), 1223 (m), 1159 (s), 1141 (m), 1029 (s), 838 (m), 636 (s), 571 (w), 517 (m). μ_eff(CDCl₃) = 1.69 (12) BM. MALDI-TOF MS: m/z = 309 (m+1 - OTf). Anal. Calcd. for C₁₈H₂₁F₃O₅SV: C, 47.27; H, 4.63. Found: C, 47.51; H, 4.73.

[Cp₂V(3-mmm)][OTf] (10). 68% yield. IR: 3122 (m), 3096 (w), 2969 (w), 1571 (s), 1555 (w), 1463 (w), 1451 (w), 1369 (w), 1323 (w), 1301 (w), 1272 (vs), 1259 (s), 1223 (m), 1160 (s), 1141 (m), 1029 (s), 837 (m), 636 (s), 571 (w), 517 (m). μ_efϊ(CDCl₃) = 1.68 (12) BM. MALDI-TOF MS: m/z = 295 (m+1 - OTf). Anal. Calcd. for C_I7H₁₉F₃O₅SV: C, 46.06; H, 4.32. Found: C, 46.01; H, 4.57.

[Cp₂V(3-mcm)][OTf] (11). 13% yield. IR: 3119 (m), 3093 (w), 2958 (w), 2926 (w), 1565 (s), 1458 (m), 1436 (w), 1449 (w), 1369 (w), 1341 (w), 1298 (w), 1274 (vs), 1257 (s), 1222 (m), 1164 (s), 1153 (m), 1049 (s), 1030 (s), 904 (w), 841 (m), 637 (s), 516 (m). μ_eft{CDCl₃) = 1.68 (12) BM. MALDI-TOF MS: m/z = 314 (m - OTf). Anal. Calcd. for C,₆H_l6ClF₃O₅SV: C, 41.43; H, 3.48. Found: C, 41.81; H, 3.98.

[Cp₂N(3-mnm)][OTf] (12). 36% yield. IR: 3115 (m), 1577 (s), 1527 (s), 1464 (w), 1452 (w), 1374 (w), 1339 (m), 1271 (m), 1256 (s), 1222 (w), 1159 (s), 1121 (w), 1028 (s), 848 (m), 827 (m), 754 (w), 638 (s), 572 (w), 516 (m). μ_efϊ(CDCl₃) = 1.68 (12) BM. MALDI-TOF MS: m/z = 326 (m+1 - OTf). Anal. Calcd. for C₁₆H₁₆ΝF₃O₇SV: C, 40.52; H, 3.40; N, 2.95. Found: C, 40.29; H, 3.37; N, 2.87.

Instrumentation. Infrared spectra were obtained on a FT-Nicollet Protege 460 spectrometer as a KBr pellet or nujol mull. IR spectra are reported in cm- . UV-vis spectra were recorded in a quartz cell or cuvette on a Beckman Model DU 7400 spectrophotometer and the spectral bands were registered in the 250 - 800 nm range. Fluorescence measurements were done using a Shimadzu spectrofluorophotometer (model RF-5301PC). H NMR spectra were recorded on a Varian XL-300 spectrometer operating at 300.110 MHz. Chemical shifts were referenced to residual protio solvent peaks in the sample. Magnetic moments were determined using Evan's Method at 20 °C (Evans, D. F. J. Chem. Soc. 1959, 2003). Mass spectra were recorded on a HP G2025A MALDI-TOF mass spectrometer using dithranol as the supporting matrix. Spectra were averaged over 50 shots. Electrochemical measurements were performed on a Bioanalytical Systems B/W 100b electrochemical analyzer with IR compensation. The cyclic voltammograms taken in acetonitrile were obtained in a 0.1 M Bu NPF₆ (TABP) electrolyte solution with a 0.1 M Ag/AgNO₃ reference electrode, a Platinum Disc working electrode, and a platinum wire auxiliary electrode. Solutions were purged with nitrogen and scanned at 0.2 V/s. Aqueous cyclic voltammograms were taken in a standard phosphate buffer saline solution, using a Ag/AgCl reference electrode, a glassy carbon working electrode, and a platinum wire auxiliary electrode. Solutions were purged with nitrogen and scanned at 0.2 V/s. All potentials were referenced to the ferrocene- ferrocinium couple. Elemental analyses were performed by Atlantic Microlab, Inc. (Norcross, GA). Electron paramagnetic resonance (EPR) spectra were recorded in PBS (phosphate buffer saline, 0.015 M NaHPO₄, 0.10 M NaCl, pH 7.2) on a Bruker ESP 300 EPR spectrometer (9.64 GHz). The g values were calibrated with a Varian strong pitch (0.1% in KC1) standard (g value 2.0028). The samples for EPR spectral analysis were studied in Willmad WG-814 standard TE102 aqueous cell cavity (0.3 mm inner path length) to minimize the dielectric loss.

Lipid Peroxidation. Liposomes of dilinoleoyl phosphatidyl choline lipids were formed in 0.1 M phosphate buffer (pH 7.2) using the ethanol injection method (Betageri, et al. Liposome Drug Delivery Systems; Technomic Publishing Co., Inc.: Lancaster, 1993; pp 13-14). Aliquots of 0.5 mL liposomes (2 mM lipid) were incubated with the vanadocene chelated complexes (50-400 μM) at 37 °C. The incubation was stopped at various time points by the addition of 0.1 mL 4%> butylated hydroxyl toluene in EtOH. Sodium dodecyl sulfate (3%>, 0.25 mL) was added to destroy the liposomes followed by the addition of 0.5 mL 1% 2-thiobarbituric acid in 0.05 M NaOH and 0.5 mL 25% HC1. The samples were mixed and heated in boiling water for 15 min. The 2-thiobarbituric acid-reactive substances were extracted into 3 mL 1 -butanol and the fluorescence ofthe butanol layer was measured at λ_exc = 530 nm and λ_em = 550 nm (Verstraeten, et al. Arch. Biochem. Biophys. 1997, 338, 121; Gutteridge, et al. Biochim, Biophys, Acta 1988, 835, 441; and Deleers, M.; et al Biochim. Biophys. Acta 1985, 813, 195). The 2-thiobarbituric acid-reactive substances are reported as malonaldehyde equivalents (Gutteridge, J. M. C. Anal. Biochem. 1975, 69, 518).

Kinetic analysis ofthe reaction was determined from the linear slopes ofthe conversion versus time plots at the beginning ofthe reaction to give the conversion of mol malonaldehyde per mol vanadocene complex per min. The values are an average of three measurements and the standard deviation was no more than 5%>.

The experiments measuring the effect of pH and presence of hydrogen peroxide on the reaction were conducted with liposome solutions containing 1 mM dilinoleoyl phosphatidyl choline (DLPC) in 0.1 M phosphate buffer adjusted to the appropriate pH or containing the appropriate amount of hydrogen peroxide. The reactions were initiated by the addition of 200 μM 7 or 13 and incubated for 90 min at 37 °C, then quenched and worked up as previously described. Control experiments confirmed that neither the variation in the pH range or the presence of hydrogen peroxide affected the thiobarbituric acid assay.

EPR Spin-Trapping Experiments. The following is a general procedure for the spin trapping experiments: a 5x10 M solution ofthe vanadocene compound (80 μL) was added to 1 mL of a 2 mM liposome solution containing α-(4-pyridyl-l-oxide)-N- tert-butylnitrone (POBN) (10 eq.) and the mixture was incubated at 37 °C for 0 - 90 min. The solution was transferred to a 0.9 mm ID fused quartz capillary tube, mounted in the cavity ofthe EPR spectrometer, and the spectrum was taken. Each spectrum was averaged from four independent scans. The EPR spectra ofthe POBN - OH adducts were identified by their doublet of triplet EPR signals ofthe hydroxyl radical at g = 2.008 with hyperfme splitting of the nitrosyl nitrogen, <A > = 15x10^" cm^-1, and the β- hydrogen atom, <A > = 2.81x10 cm using VOSO as a standard (Setaka, et al J Catal. 1969, 15, 209). solutions ofthe organometallic vanadium compounds and POBN indicated no interaction between the compounds and the spin trap. Control solutions of POBN and liposomes also indicated no interaction between the spin trap and liposomes.

Results Effects of Substitution to [Cp₂V(acac)] [OTf] on Lipid Peroxidation. The rates for vanadocene-chelated induction of lipid peroxidation in DLPC liposomes, as measured by the production of 2-thiobarbituric acid reactive substances, are given in Table 6. The kinetic profiles indicate that the reaction has an initial lag time of about 15 min before proceeding linearly. Substitution at the 3-position ofthe acetylacetonate ligand altered the rate ofthe reaction in the following order:

CH₂CH₃ (9) > CH₃ (10) > H (7) > Cl (ll) > NO₂ (12) Thus, electron-donating substituents increased and electron-withdrawing substituents decreased the rate of lipid peroxidation.

Table 6. Rates ofthe lipid peroxidation reaction initiated by the vanadocene complexes.

Compound k (malonaldehyde eq/min)^a

[Cp₂V(3-metm)][OTf] (9) 0.74

[Cp₂V(3-mmm)][OTf] (10) 0.55

[Cp₂V(acac)][OTf] (7) 0.50

[Cp₂V(3-mcm)][OTf] (11) 0.45

[Cp₂V(3-mnm)][OTf] (12) 027 ^aEach rate is the average of at least three independent measurements and the standard deviation was no greater than 5%.

Electrochemistry. As illustrated in Figure 13 and Table 7, 7 and 9-11 display one reversible couple in acetonitrile attributable to the V /V redox process. 11 has the highest couple at 1.14 V followed by the unsubstituted acetylacetonate, 7, with 9 and 10 having the lowest values at 0.95 V and 0.97 V, respectively. 12 has an irreversible couple of 1.59 V. Observation ofthe V /V couple was not possible in aqueous or liposome solutions due to the constraints ofthe available solvent window. However, although the absolute value ofthe redox potential would be different, the trends in the inductive effect are expected to remain the same regardless ofthe solvent.

Table 7. Cyclic voltammetry and EPR results ofthe vanadocene complexes.

Compound V^1V/V^V redox couple, E_1/2 (V)^a EPR g [A„ (10^"V)]

[Cp₂V(3-metm)][OTf] (9) +0.97 1.994 [67.8]

[Cp₂V(3-mmm)][OTf] (10) +0.95 1.994 [68.4]

[Cp₂V(acac)][OTf] (7) +1.09 1.995 [69.6]

[Cp₂V(3-mcm)][OTf] (11) +1.14 1.992 [64.5]

[Cp₂V(3-mnm)][OTf] (12) +1.59⁶ 1.979 [67.4]

Potentials listed vs. Cρ₂Fe⁺ in acetonitrile. Scan rate = 0.2 V/s. E_pc value.

Effects of Oxygen on Lipid Peroxidation Initiated by 7. The time dependence for lipid peroxidation of DLPC liposomes inititated by 7 in the presence and absence of oxygen is shown in Figure 14. There is no production of malonaldehyde. the breakdown product of lipid peroxidation, in the absence of oxygen. The EPR spectrum ofthe reaction mixture was taken at 0 min and 90 min in the presence and absence of oxygen. Figure 15a shows the spectrum of 7 in the liposome solution seconds after addition. After a 90 min incubation at 37 °C in the presence of oxygen, there is significant loss ofthe intensity of the EPR spectrum (Figure 15b) indicating that roughly 70 % ofthe complex has been converted to an EPR-silent species. However, the remaining 7 retains its eight line characteristic spectrum. In Figure 15c the same experiment has been carried out in the absence of oxygen and clearly shows the emergence of two additional EPR-active species. The formation ofthe additional V(IV) species was also obtained with 7 in buffer alone and is time-dependent, reaching an equilibrium around 3 h at 37 °C. Comparison of the pH Effect on Lipid Peroxidation Initiated by 7 and 13.

The effect of pH on the extent of lipid peroxidation in 90 min when initiated with 7 and 13 is shown in Figure 16. Reaction with 13 was chosen to compare with the acetylacetonate complexes because of its known ability to form hydroxyl radicals. The extent of lipid peroxidation induced by 13 is sharply increased at pH 4 and non-existent at pH 10, whereas the lipid peroxidation initiated by 7 only slightly increases with increasing acidity.

Comparison of the Effect of the Addition of Hydrogen Peroxide on Lipid Peroxidation Initiated by 7 and 13. The effect of hydrogen peroxide on the extent of lipid peroxidation in 90 min when initiated with 7 and 13 is shown in Figure 17. The addition of hydrogen peroxide to the reaction initiated by 13 shuts down the reaction whereas the lipid peroxidation reaction initiated by 7 is unaffected. With hydrogen peroxide, in the absence of oxygen, no lipid peroxidation occurred with either 7 or 13 (results not shown), indicating that hydrogen peroxide cannot substitute for oxygen in the reaction.

EPR Spin-Trapping Experiments. EPR experiments with POBN as the spin- trap were used to examine formation of hydroxyl radicals during the lipid peroxidation reaction. When the spin trap was added prior to incubation of 7 with the liposome solution for 90 min at 37 °C, the typical eight-line EPR spectrum of 7 (Figure 18a) was seen. When 13 was used in place of 7 the typical triplet spectrum indicating the presence of POBN- OH appeared as shown in Figure 18b.

Discussion Bis(cyclopentadienyl)-vanadium(IV) (acetylacetonate) complexes substituted in the 3 -position ofthe acetylacetonate ligand induce peroxidation of dilinoleoyl phosphatidyl choline in the presence of oxygen as determined by the generation of malonaldehyde. Greater electron donating capacities ofthe substituent lead to a faster rate of peroxidation such that an ethyl-substituted complex (9) is 50% faster than the unsubstituted species and nitro substitution (12) decreases the rate by almost one half compared to the unsubstituted. The correlation ofthe Hammett constant (Taft, R. W., Jr, In Steric Effects in Organic Chemistry; Newman, M.S. Ed.; John Wiley & Sons: New York, 1956; and Pross, A. Theoretical and Physical Pricniples of Organic Reactivity; John Wiley & Sons: New York, 1995). with the rate ofthe oxidation reaction (eq.4) is log(k_x/k_H) = ρσ_x (eq. 4) shown in Figure 19 (right axis). The reactivity order parallels that of increasing electron availability at the vanadium center. Based on the observed kinetic profiles of lipid peroxidation, it is clear that the vanadium metal center plays an important role and it is likely that the mechanistic pathway involves the oxidation ofthe complex to initiate the process. Electron-withdrawing substituents shift the metal d orbital energy to a lower level, increasing the reorganization energy and the overall activation energy required for generating the active oxidant.

The Hammett substituent constant, σ, and the ΔEι_/ value, as expected, correlate linearly with equation 5 (Figure 19, left axis) (Zuman, P. Substituent Effects in Organic Polargraphy, Plenum Press: New York, 1967; pp 1-41). The reaction constant, p, which in this case

ΔE_1/2 = E_1/2(X) - E_1/2(H) = pσ_x (eq. 5) measures the susceptibility ofthe electron-transfer process to polar effects, is 0.65 V. This is somewhat greater than that reported for substituted ferrocene complexes (0.48 V) (Silva, et al. J. Organometallilc Chem. 1991, 421, 75). or substituted peroxovanadium salicylaldehyde complexes (0.35 V) (Nakajima, et al. J. Bull. Chem. Soc. Jpn. 1990, 63, 2620) , reflecting a higher degree of transmission ofthe electron density from ligand to metal center.

The inclusion of electron withdrawing or donating groups in organometallic complexes has been observed to have an effect on the oxidation of unsaturated substrates. The catalytic activity of a series of iron(II) and oxovanadium(IV) salen complexes increased with increasing Fe /Fe or V^V/V^IV reduction potential for the aerobic epoxidation of unsaturated olefins using an alkyl hydroperoxide as a promoter (Bottcher, et al. J. Mol. Cat. A. 1997, 117, 229; and Chang, C. j. Et al. Inorg. Chem. 1997, 36, 5927). From the results ofthe product distribution ofthe unsaturated hydrocarbon oxidation reaction it was proposed that the key factor is the rate of reduction ofthe metal center by the hydroperoxide in a metal-catalyzed Haber- Weiss cycle (Chang, C. J. et al. Inorg. Chem. 1997, 36, 5927). Greater electron withdrawing groups in the salen ligand increased the reduction potential ofthe metal which in turn accelerated the rate of decomposition of alkyl peroxides. In our study however, the trend for the rate of peroxidation of unsaturated lipid with 3 -substituted acetylacetonate ligands coordinated to the vanadocene is reversed. In addition, the reaction is not catalytic nor is a promoter needed for the initiation of lipid peroxidation. In fact, the use of even 100 fold excess hydrogen peroxide did not affect the overall rate of lipid peroxidation.

Oxygen apparently plays a critical role since in its absence lipid peroxidation does not occur. In the present series of vanadocene acetylacetonate complexes, outer sphere electron transfer from vanadium to molecular oxygen (O₂ + e = O₂ , E_1/2 = -0.16 V) (Sawyer, D. T. In Oxygen Complexes and Oxygen Activation by Transition Metals, Martell, A. E.; Sawyer, D. T. Eds.; Plenum Press: New York, 1988; p 131.) is thermodynamically unfavorable even with the ethyl-substituted acetylacetonate complex (9) which displays the lowest V /V redox potential (E^ = 0.97 V) ofthe series examined. This huge kinetic barrier for oxygen activation can only be overcome if coordination with V(IV) takes place. However, there is no valence molecular orbital available in the Cp₂V(IV) system for σ or π-bonding with oxygen if the acetylacetonate ligand remains chelated (Lauher, et al. J. Am. Chem. Soc. 1976, 98, 1729). The first clue of ancillary ligand dissociation from the V(IV) coordination sphere comes from an observed lag time of ~15 min in the initiation of lipid peroxidation which is independent of substitution in the 3-position ofthe acetylacetonate ligand. When we monitored the EPR spectrum of 7 in both buffer and liposome solutions in the presence of oxygen, the spectrum retained its typical eight line characteristics although there is approximately a 70%) decrease in intensity after 90 min at 37 °C (Figure 15b). In the absence of oxygen, however, the features of an additional species emerge as shown in Figure 15c. The formation of this species is time dependent reaching an equilibrium after 3 h at 37 °C. These results lead us to propose that the dissociation of one arm ofthe acetylacetonate ligand from the coordination sphere ofthe V(IV)-acetylacetonate complex occurs. The dissociation ofthe Cp ring is not feasible as it has been shown by several investigators that under aqueous conditions the Cp₂V(IV) unit is very robust (Toney, et al. J Am. Chem. Soc. 1985, 107, 947) whereas monodentate acetylacetonate metal complexes are well known in the literature (Garnovskii, et al. Coordin. Chem. Rev. 1998, 173, 31). Attachment ofthe acetylacetonate ligand in a non-chelated fashion not only allows an open coordination sphere but also lowers the redox potential for the V /V couple while influencing the observed rates of lipid peroxidation by maintaining the same redox trend. The binding of dioxygen then occurs when the reaction is carried out under aerobic conditions. The resulting decrease in electrode potential and the association of a dioxygen molecule with the vacant paramagnetic vanadium center helps to overcome the kinetic barrier to dioxygen reduction (22.5 kcal/mol) (Taube, H. J Gen. Physiol. 1965, 49, 29; and Hamilton, G .A. In Molecular Mechanisms of Oxygen Activation, Hayaishi, O.; Ed.;

Academic Press: New York, 1974; p 405) and thus overcomes the spin restriction ( O₂

(triplet) -> O₂ (doublet)) by an inner sphere one electron transfer step to form the very

reactive V(V)-OO species. The fate of this oxidized species has not be determined from our experiments, but further reaction is possible to form a variety of vanadium oxo compounds. If this reactive species abstracts a hydrogen atom from the lipid, the resulting

V(V)-OOH has the possibility of undergoing homolytic cleavage to form OH , which is a powerful oxidant.

Lipid peroxidation is generally shown to proceed through hydroxyl radical- initiated autooxidation pathways such as the classical Fenton-type reaction (Konings, A. W. T. In Liposome Technology; Gregoriadis, G. Ed.; CRC: Boca Raton, Fl, 1992; Vol. 1, pp 139-161; and Spiteller, G. Chem. Phys. Lipids 1998, 95, 105). However, when 7 was used to initiate the reaction in the presence ofthe spin trapping agent, POBN, no signals of spin-trapped radicals were observed in EPR spectroscopy experiments (Figure 18a) indicating that the peroxidation reaction does not involve a hydroxyl radical-mediated pathway. Mechanistic proposals for hydrocarbon oxidation catalyzed by metal complexes include mechanisms which involve metal-centered oxidants that are directly responsible for C-H bond cleavage and those in which the role ofthe catalyst is to generate metal- free species (HO , RO , ROO , etc.) that cause C-H bond cleavage (Bottcher, et al. J Mol. Cat. A. 1997, 117, 229; Chang, C. j. et al. Inorg. Chen. 1997, 36, 5927; Talsi, et al. J. Mol. Catal. 1993, 81, 235; and Mimoun, et al. J. Am. Chem. Soc. 1986, 108, 3711). Since no spin-trapped products were observed during the peroxidation reaction of DLPC by 7 and given that 7 is incapable of causing hydrogen peroxide decomposition (Ghosh, et al. J. Inorg. Biochem. 1999, 75, 135) we concluded that lipid peroxidation induced by the acetylacetonate vanadium compounds (7 and 9-12) likely involves a vanadium- centered oxidant. The postulated scenario involves the acetylacetonate ligand of 7 becoming unidentate and opening a coordination site (Figure 20). With the binding of oxygen, a V(V)-OO species could form through an inner sphere electron transfer which then reacts directly with the lipid. The step in which oxygen is bound and activated is, at least partially, rate limiting in this reaction and modulated by the substitution on the acetylacetonate ligand. Our data cannot distinguish between the direct insertion of peroxide into the double bond ofthe lipid hydrocarbon chain or the abstraction of a hydrogen atom from the allylic position ofthe hydrocarbon. What is clear is that the bound peroxide is never free in solution to generate hydrogen peroxide by the abstraction of protons from the solvent. A similar situation is seen in oxygen-carrying metalloproteins whereby the metal-dioxygen moiety remains intact rather than dissociating (Bertini, et al. New. J. Chem. 1996, 20, 187). Lipoxygenases and prostaglandin H synthases, two families of nonheme iron enzymes which oxygenate polyunsaturated fatty acids, catalyze the abstraction of a hydrogen atom from the unsaturated hydrocarbon chain and insertion of oxygen through the formation of ferryl oxygen intermediates without the generation of hydroxyl or peroxyl free radicals (Su, et al. J. Biol. Chem. 1998, 273, 20744; Elliott, et al. J Biol. Chem. 1986, 261, 6719; and Oliw, et al. Arch. Biochem. Biohys. 1993, 305, 288). The proposed mechanism for the vanadocene acetyacetonate complexes differs from the enzymatic reaction which involves the formation of a protein radical to initiate a new catalytic cycle because there is no possibility of further oxidation ofthe vanadocene.

Since 13 has been shown to generate hydroxyl radicals by the decomposition of hydrogen peroxide, (Ghosh, et al. J. Inorg. Biochem. 1999, 75, 135) we also followed the lipid peroxidation reaction with 13 under identical experimental conditions. The amount of lipid peroxidation induced by 13 in the presence of oxygen is found to be similar to that observed for complex 7. In contrast to the acetylacetonate complexes, the addition of 13 to a lipid solution produces a spin trapped POBN- OH product (Figure 18b). As no hydroxyl radicals are formed in the absence of lipid, they are presumably formed during the lipid peroxidation reaction. Hydrogen peroxide is not necessary for the formation of hydroxyl radicals but a reactant, such as lipid, is necessary. In fact, the lipid peroxidation reaction initiated by 13 is adversely affected by the addition of hydrogen peroxide indicating that hydrogen peroxide is capable of oxidizing 13 to form an inactive V(V) species which is no longer capable of binding and/or activating dioxygen. In aqueous media, kinetic studies have shown that the release ofthe first Cl^" from 13 is too fast to be measured and the rate of hydrolysis ofthe second chloride is reported to be on the order of 15 min (Toney, et al. J. Am. Chem. Soc. 1985, 107, 947; and Kuo, et al. T. J. J. Am. Chem. Soc. 1991, 113, 9027). The subsequent diaquo product, Cp₂V(H₂O)₂ ⁺, is very susceptible towards protonolysis at pH 7.2 (Kopf-Maier, et al. Naturwiss., 1981, 68, 272). During the association and dissociation ofthe ancillary coordinated water molecules, association of dioxygen to the V(IV) center could take place with subsequent formation ofthe vanadium (V) superoxide species occurring after transfer of an electron from V(IV) to the bound dioxygen. This reactive species is capable of abstracting an allylic hydrogen atom from the lipid to form a V(V)-OOH species. It could also undergo decomposition via homolytic cleavage to form the hydroxyl radicals that we observe by the EPR spin-trapping experiments and are known to initiate lipid peroxidation. The activated vanadocene acetylacetonate complex, on the other hand, does not undergo this type of decomposition possibly due to a stabilizing influence ofthe acetylacetonate ligand.

To obtain further insight into the mechanistic differences between the modes of action of compounds 7 and 13, we have also followed the lipid peroxidation reaction in solutions with varied pH (Figure 16). Unlike the 13-induced lipid peroxidation, the lipid peroxidation initiated with compound 7 is minimally affected by pH. As expected, the chelated acetylacetonate ligand provides more stability to 7 at higher pH compared to the labile unidentate chlorides in 13. The differences in the pH-profile are consistent with the different aquo forms of 13 seen at different pH values (Kuo, et al. T. J. J. Am Chem. Soc. 1991, 113, 9027). Above pH 9, the formation ofthe neutral and inert Cp₂V(OH)₂ species from 13 takes precedence. Due to the formation of the strong σ-bonded OH ligands. the system is considerably less favorable towards dissociation ofthe ligands compared to its aquo-ligated counterpart which hinders the binding of neutral dioxygen to an open coordination sphere of vanadium. Low pH accelerates the formation ofthe more labile diaquo adducts and thus leads to greater oxygen binding which, in turn, increases the extent of lipid peroxidation.

Our proposed mechanism for 7-initiated lipid peroxidation is also consistent with the results ofthe peroxidation in the presence of added hydrogen peroxide (Figure 17). Additional hydrogen peroxide acts as an excellent oxidant towards 13, as seen in previous experiments where a five-fold excess of hydrogen peroxide effectively oxidized all of the V(IV) to V(V) within one minute at room temperature (Ghosh, et al. J. Inorg. Biochem. 1999, 75, 135). Oxidation of 13 by hydrogen peroxide shuts down the activation of oxygen and subsequent lipid peroxidation. The lipid peroxidation induced by 7 is insensitive to the addition of hydrogen peroxide because it is much less susceptible to oxidation than 13 (An E_pc value of 0.65 V was measured for 6 vs. C ₂Fe^+/0 in acetonitrile under the same experimental conditions used for the acetylacetonate complexes). Addition of hydrogen peroxide to an oxygen-free reaction did not result in any lipid peroxidation with 7 or 13 confirming that hydrogen peroxide cannot compete with oxygen for binding in the open coordination sphere.

Thus 7 and 13 induce lipid peroxidation in DLPC liposomes by two different pathways. Both complexes form a vanadium(V)-superoxo complex as the active oxidizing species but the oxidized complex formed from 7 must act directly to initiate lipid peroxidation. The activated complex from 13, on the other hand, also has the ability of decomposing to form hydroxyl radicals which are known initiators of lipid peroxidation.

Conclusion

Investigation of a series of vanadocene acetylacetonate complexes with substitution in the 3-position (7 and 9-12) demonstrated their ability to initiate lipid peroxidation in the presence of oxygen. Substitution with electron-donating groups increased the rate of reaction by facilitating the oxidation as measured by the decrease in the V^IV/V^V redox potential. Experimental evidence suggested that a secondary vanadocene species is first formed which becomes the active oxidizing agent after binding dioxygen. Lipid peroxidation induced by the vanadocene acetylacetonate complex is not initiated by the formation of hydroxyl radicals as seen by vanadyl salts or vanadocene complexes such as 13. To our knowledge, this is the first example of non- enzymatic metal-initiated lipid peroxidation which does not proceed via the classic Fenton-type mechanism. All publications, patents, and patent documents described herein are incorporated by reference as if fully set forth. The invention described herein may be modified to include alternative embodiments. All such obvious alternatives are within the spirit and scope ofthe invention, as claimed below.

Claims

ClaimsWhat is claimed is:

1. A method of modulating the permeability of a lipid membrane, the method comprising: contacting a metallocene complex with the lipid membrane.

2. The method of claim 1 , wherein the lipid membrane is a lipid bilayer.

3. The method of claim 1, wherein the lipid membrane is a liposome or artificial skin.

4. The method of claim 1, wherein the lipid membrane is a biological membrane.

5. The method of claim 1, wherein the metallocene complex comprises a vanadocene complex or a titanocene complex.

6. The method of claim 1 , wherein the permeability of a lipid membrane is modulated temporarily.

7. The method of claim 1, wherein the metallocene complex comprises a compound having the formula:

Cp Rl

\ N M

Cp R2

where M is a transition metal ion or atom, Cp is unsubstituted cyclopentadienyl or cyclopentadienyl substituted with one or more substituents that can be the same or different, and are selected from Cι- alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO?, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine, and Ri and R₂ are together a bidentate ligand; or a pharmaceutically acceptable salt or ester thereof.

8. The method of claim 7, wherein the bidentate ligand comprises one or more aromatic rings.

9. The method of claim 8, wherein the bidentate ligand comprises two or more aromatic rings.

10. The method of claim 8, wherein the bidentate ligand is selected from phenanthroline; bipyridyl; bridged bipyridyl; N-phenyl benzohydroxamato; N, N- diethyldithiocarbamato; acetylacetonato; catacholato; and acetophenone. wherein each of these bidentate ligands is unsubstituted or substituted with one or more substituents selected from Cι-₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine.

11. The method of claim 1 , wherein the metallocene complex is selected from Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂V(cat), Cp₂V(Et₂ (dtc)), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3-metm), Cp₂V(3-mmm), Cp₂V(3-mcm), Cp₂V(3-mnm), and mixtures thereof.

12. The method of claim 11 , wherein the metallocene complex is selected from Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3- metm), Cp₂V(3-mmm), Cp₂V(3-mcm), and Cp₂V(3-mnm).

13. A composition for modulating the permeability of a lipid membrane, the composition comprising a metallocene complex having the formula:

Cp Rl

M

Cp R2 where M is a transition metal ion or atom, Cp is unsubstituted cyclopentadienyl or cyclopentadienyl substituted with one or more substituents that can be the same or different, and are selected from d-₄ alkyl, aryl, Ci-4 alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine, and Ri and R₂ are together a bidentate ligand; or a pharmaceutically acceptable salt or ester thereof; and a pharmaceutically acceptable carrier, diluent or vehicle.

14. The composition of claim 13, wherein the bidentate ligand comprises one or more aromatic rings.

15. The composition of claim 14, wherein the bidentate ligand comprises two or more aromatic rings.

16. The composition of claim 14, wherein the bidentate ligand is an N,N'; O,O'; N,O; S,S'; N,S; or O,S bidentate ligand.

17. The composition of claim 13, wherein the bidentate ligand is selected from phenanthroline; bipyridyl; bridged bipyridyl; N— phenyl benzohydroxamato; N, N- di ethyl dithiocarbamato; acetylacetonato; catacholato; and acetophenone, wherein each of these bidentate ligands is unsubstituted or substituted with one or more substituents selected from Cι-₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine.

18. The composition of claim 13, wherein the metallocene complex is selected from Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂V(cat), Cp₂V(Et₂ (dtc)), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3-metm), Cp₂V(3-mmm), Cp₂V(3-mcm), Cp₂V(3-mnm), and mixtures thereof.

19. The composition of claim 13, wherein the metallocene complex is selected from Cp₂V(phen), Cp₂V(bpy), Cp₂V(PH), Cp₂V(acac), Cp₂Ti(phen), Cp₂Ti(bpy), Cp₂V(3- metm), Cp₂V(3-mmm), Cp₂V(3-mcm), and Cp₂V(3-mnm).

20. A compound having the formula:

where M is a transition metal ion or atom, Cp is unsubstituted cyclopentadienyl or cyclopentadienyl substituted with one or more substituents that can be the same or different, and are selected from Cι-₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF , NO₂, CN, OCN, SeCN, SCN, N₃, OH, SH, and amine, and Ri and R₂ are together a bidentate ligand.

21. A compound of claim 20, wherein the compound is Cp₂V(phen).

22 A compound of claim 20, wherein the compound is Cp₂V(bpy).

23. A compound of claim 20, wherein the compound is Cp₂V(phen).

24. A compound of claim 20, wherein the compound is Cp₂V(cat).

25. A compound of claim 20, wherein the compound is Cp₂V(3-metm).

26. A compound of claim 20, wherein the compound is Cp₂V(3-mmm).

27. A compound of claim 20, wherein the compound is Cp₂V(3-mcm).

28. A compound of claim 20, wherein the compound is Cp₂V(ph3-mnm).

29. A method for enhancing the uptake or transport of a therapeutic agent across a lipid membrane, the method comprising co-administering the therapeutic agent with a compound ofthe formula:

where M is a transition metal ion or atom, Cp is unsubstituted cyclopentadienyl or cyclopentadienyl substituted with one or more substituents that can be the same or different, and are selected from Cι_₄ alkyl, aryl, Cι-₄ alkoxy, carboxylate, halo, CF₃, NO , CN, OCN, SeCN, SCN, N₃, OH, SH, and amine, and Ri and R₂ are together a bidentate ligand.