CA2240183A1 - Cobalt schiff base compounds - Google Patents

Abstract

The invention relates to cobalt compounds in which divalent or trivalent cobalt is complexed with water soluble tetradentate Schiff's bases. The tetradentate Schiff's bases preferably contain two nitrogen atoms and two oxygen atoms as coordinating atoms. The compounds can contain polypeptide or nucleic acid targeting moities and can be used to inhibit enzymes such as thrombin and to inhibit zinc finger proteins.

Description

COBALT SCHIFF BASE COMPOUNDS

FIELD OF THE INVENTION

The invention relates to cobalt compounds, and methods of using such compounds to reduce the biological activity of proteins.

BACKGROUND OF THE INVENTION

The use of metals in medicine has grown impressively in recent years as the result of a greatly advancing underst~n-ling of the structures of biologically active metal complexes and metal-cont~ining proteins.

Currently, a class of cobalt-cont:~ining complexes, where the cobalt is Co(III),has been shown to have antiviral, antiturnor and antimicrobial activities~ as well as showing use in the treatment of infl~mm:~tion and burns (see U.S. Patent Nos. 4,866,054, 4,866,053, 5,049,557, 5,106,841, 5,142,076, and 5,210,096, and Wooley et al., Agents and Actions 35:274 (1992)).

W O 97/21431 PCT/~S96/19900 These complexes have a basic core structure shown below:
X' RC ~ ~ RD
>~=N~ ~N=~
RB~ ~ RE

RA RF

These complexes are hypothesized to be active-oxygen or superoxide antagonists, thus suppressing medical conditions associated with free radicals such as infl~mm~tion.

Additionally, a Co(II) complex of isopropyl salicylic acid has been made and reported to be cytotoxic. (Ranford et al., J. Chem. Soc. Dalton Trans. (1993) 3393).

Finally, the oxidation of certain Co(II~) complexes cont~ining coordinated nitrogen mustards causes the release of activated aliphatic mustards which can act as diffusible cytotoxins. (Ware et al., J. ~led. Chem. 36:183g (1993)).

SUMMARY OF THE INVENTION

It is an object of the invention to provide novel cobalt compounds, including cobalt compounds cont~ining targeting moities. It is a further object to providemethods for the inhibition of proteins, such as enzymes, using these cobalt 1 5 complexes.

In accordance with these objects, compositions are provided comprising water soluble tetradentate Schiff's base complexes of Co(II).

W O 97/~1431 PCT/~S96/19900 Further provided are compounds having the structure comprising Formula 1:
Formula 1 R3 N, ~N==~<R6 R2~\ Co ~R7 ~ 0~

In Formula 1, Co is either Co(II) or Co(III), and each of Rl, R2, R3, R4, R5, R6, R7 and R8 is a hydrogen, an alkyl group, an aryl group, a hydrophobic organic acid, an alkyl alcohol, an alcohol, an alkyl amine group, an amine group, or a targeting moiety.

Also provided are compounds of Formula I wherein R~, R7, R3 and R4 are each hydrogen, alkyl, or aryl.

Also provided are protein-cobalt compound complexes comprising a protein attached to the cobalt compound of Formula 1.

Further provided are methods of inhibiting a selected protein comprising contacting the selected protein with the compound of Formula 1.

Additionally provided are methods of inhibiting zinc finger proteins comprising contacting a zinc finger protein with the cobalt compounds described herein.
.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure I depicts the inhibition of thrombin. 3.07 X 10-9 M thrombin at 25C was assayed using Spectrozyme TH (American Diagnostics), and the react;on followed at 406 nm using a Hewlett Packard HP8452A diode array spectrophotometer with temperature control. All assays were performed in 10 mM Tris, 10 mM HEPES, 0.1% polyethylene glycol, 0.5 M NaC1, pH 7.8. The Co(III) carboxypropyl(NH3)2 (labeled as Co(carboxypropyl)) was coupled to the active site directed peptide NH2-GGGdFPR-CO-NH~ (labeled as peptide dFPR). The observed inhibition greatly exceeded Co(carboxypropyl)(NH3)2, the peptide and Co(III)acacen(N~3)2 Cl (labeled as Co(acacen)). This demonstrates ehe principle that coupling known inhibitors to the cobalt compound can greatly increase the potency of enzyme inhibition compared with the inhibitory activity of the uncoupled components.

Figure 2 depicts the structure of the Co(III)(acacen-GGGFPR)(NH3),.

E~igures 3A and 3B shows the temperature dependence of the enzyme inhibition rate correlates to the ligand exchange rate. (A): temperature dependence (see Example 3). (B): ligand exchange ratc (see Example 3).

Figure 4 shows the inhibition of thrombin by ~CoIII(acacen)~NH3)21Cl.
Thrombin ~ as incubated for 12 hours at room temperature with (A) 0 M
Co(III) and B) 2.5 mM Co(III). Spectra were taken every 30 seconds for 30 minutcs to monitor the release of p-nitroaniline by enzymatic hydrolysis of a commercial substrate.

Figurc 5 sho~vs that inhibition of thrombin is dependent on the concentration ofthe inhibitor~ the length of incubation, and the temperature of incubation.

~igures 6A, 6B, 6C, 6D, 6E, 6F, 6G, 6H, and 6I depict the structures of cobalt compounds of the invention that have been made. Unless noted, all of the compounds are Co(III).
.

DETAILED D:~SCRlPTION OF THE INVENTION

S As is described below, the present invention is directed to cobalt compounds that can exchange or bind functional moieties such as histidine on a protein's surface resulting in the inactivation of a biological activity of the protein due to the complexing of the functional moiety to the cobalt compound.

The cobalt compounds of the invention utilize either Co(II) (also depicted herein as Co+2) or Co(III) (also depicted herein as Co+3). Generally, Co(II) compounds have up to four coordination atoms, and may contain a first axial ligand, although it is possible that water molecules may be weakly associated in one or both a~ial ligand positions. Similarly, Co(IlI) compounds have up to six coordination atoms, of which two are defined herein as axial ligand positions. By "axial ligand" herein is meant a ligand L, or L2 located at eitherthe fifth or sixth coordination sites, generally depicted in the structure below:
R4 L1 R~;

~N~ <\N==< 6 R2~ Co ~R7 Without being bound by theory, the cobalt compounds of the invention derive v their biological activity by the substitution or addition of ligands in the axial positions. The biological activity of the cobalt compounds of the invention W O 97/21431 PCT/~S96/19900 results from the binding of a new axial ligand, most preferably the nitrogen atom of imidazole of the side chain of histidine. Presumably the amino acid serving as a new axial ligand of the cobalt compound is required by the target protein for its biological activity. Thus, as is more fully described below, S proteins such as enzymes that utilize a histidine in the active site, or proteins that use histidine, for example, to bind essential metal ions, can be inactivated by the binding of the histidine in an axial ligand position of the cobalt compound, thus preventing the histidine from participating in its normal biological function.

When the cobalt is Co(III), the Co(III) complex is synthesized or formulated with two particular axial ligands, and then when the complex is added to a protein, for example, the original axial ligand or ligands are replaced by one or more ligands from a protein. This will occur either when the affinity of the protein axial ligand is higher for the cobalt compound as compared to the l S original axial ligand, or when the new axial ligand is present in elevated concentrations such that the equilibrium of axial ligand binding favors the binding of the new axial ligand from the protein. Thus, Co(III) complexes are made with axial ligands that can be substituted with other ligands.

~,Vithout being bound by theory, when the cobalt is Co(lI), such complexes may, under certain circumstances, have a first axial ligand. The Co(II) compounds of the invention are preferably synthesi7~d with no axial ligands.
Upon incubation with a protein, certain moieties, such as the nitrogen atom of the imida~ole of the side chain of histidine, within the protein can become an axial ligand, resulting in a tightly-bound protein-cobalt compound complex.
This occurs when the Co(II) compound, with its four coor~in~ting atoms from the Schiff s base, binds an imidazole moiety, for example, and is oxidized to a Co(III) compound. In one sense, this may be considered a redox reaction~ since 02,0~,a8 FRI l0:4a FA~ 212 s18a47a ~r~r~Y
CA 02240183 1998-06-10 ~ ~ U Us 9 ~ ~ 1 9 ~Ul~ S ~ ~12 ~AN 1998 the Co(lI) c~Tnpo~ 1 is oxiti:zed to a Co(III) compound upon binding to the protein. Thus, tbe imioazole axial ligand serves as a fif[h cvvldilLsLing atom, and is tightly bound.

In a ~r~f.,.-cd emho~ . t, the ni~ogen atom of an imidazole site chain of the 5 amino acid residue hic1i~in~A, c- ,..~ d within a ta~get protein, is the new axial ligand. WhiletheeYU~pl and~ b~,lva~ herei~parhcularlydesc~bethis hicitirlinP embo.ii~.,uL, any "cobalt-reactive amillo acid" may serve as the newaxial ligand. A "cobalt-reactive arnino acid" is one which is capable of bindingto the cobalt cu~ uu~lds of the invention as an axial ligand. Thus, wile the 10 nitrogen of ~e ;miri~71 IG sidc chain of histidine is particularly ~erc..~d.
alL~ ive Pml~u~ utili7e the r itrogen atom of the arornatic indole side chain of t~tupll_l, the sulfilr ato_s of the side chai~s of cysteine and ~ h:.~..;..r., the amino groups of ar~inine, Iysine, -, ~ - or ~l~t~in~ as the moieties which may become axial ligands as outlined above. The 15 availabili1~ of these moieties may depcnd o~ the pH of the solution c~ ; . . ;. .g the protein or er~zyme, since in the L,.uLu~tcd state ~any of these moieties arenot good electron donors suitable ~s axiàl ligands.

The p~esent invention provides cobalt cull~oullda that ma~ be c~mrl~Y-~ with a protein c- ~-.~ a suitable new ~xial ligan~L

;!O In one e~ho~ , ~e present hl~ L;0ll provides water-soluble t~re~ nt~t_ SchifPs base c-~,.ru-~ of Co(II).

By thc term "1~ *" hcrein is Tneant that the Schif~s base colll~u~ld, which is a ligand fo~ the Co(Il), has four coordinating atorns. In a ~leI~
~~ho.l;.,....t, there are t~o nitrogen atoms a~d two oxygen atoms which serve 25 as the cuv~ At;~l~ atoms.

RNIEN~E3 SH~r By the term "Schiffis base" herein is meant a substituted imine. The substituentgroups are outlined below. Schiffs bases are generally the condensation products of amines and aliphatic aldehydes forming azomethines substituted on the nitrogen atom.

By the term "cobalt compound" herein is meant a tetradentate Schiffs base with a bound cobalt atom. The Schiff's base may be substituted or unsubstituted, and the cobalt may be Co(II) or Co(lII).

In a preferred embodiment, the cobalt compounds have the structure depicted in Forrnula 1:
Formula 1 ~= ". ."" ~
R2 ~ Co ~ R7 In this embodiment, Co is either Co(II) or Co(III~. Each of R" R2, R3, R4, R
R6, R7 and R8 is a hydrogen, an alkyl group, an aryl group, a hydrophobic organic acid, an alkyl alcohol, an alcohol, an alkyl amine group, an amine group, or a targeting moiety. When Co is Co(II), at least one of Rl through R8 is hydrophilic such that the compound is soluble in aqueous solution. When Co is Co(III), at least one of R, through R8 is a targeting moiety.

By "alkyl" or "alkyl group" or grammatical equivalents herein is meant a straight or branched chain alkyl group, with straight chain alkyl groups being preferred. If branched, it may be branched at one or more positions, and unless specified, at any position. Also included within the definition of an all; l group are cycloalkyl groups such as C5 and C6 rings. In some cases, two R groups may be part of a ring structure, that is, they may be linked to form a cyclic structure.
.

The alkyl group may range from about 1 to 20 carbon atoms (C1 - C20), with a preferred embodiment ~ltili7ing from about 1 to about lû carbon atoms (C1 -C10), with about C1 through about C5 being preferred. However, in some embodiments, the alkyl group may be larger, particularly if it is a straight chain alkyl. Particularly preferred is methyl.

By "aryl" or "aryl group" herein is meant aromatic rings including phenyl, benzyl, and naphthyl, heterocyclic aromatic rings such as pyridine, furan, thiophene, pyrrole, indole and purine, and heterocyclic rings with nitrogen, oxygen, sulfur or phosphorus.

The alkyl and aryl groups may be substituted, for example, a phenyl group may be a substituted phenyl group. Suitable substitution groups include, but are notlimited to, alkyl and aryl groups, halogens such as chlorine, bromine and fluorine, amines, carboxylic acids, and nitro groups.

By "hydrophilic organic acid" or grammatical equivalents herein is meant an alkyl group con~ining one or more carboxyl groups, -COOH, i.e. a carboxylic acid. As defined above, the alkyl group may be substituted or unsubstituted.
C1 - C20 alkyl groups may be used with at least one carboxyl group attached to any one of thc alkyl carbons, with C1 - C5 being preferred. In a preferred embodiment, the carboxyl group is attached to the terminal carbon of the alkyl group. Other ~ler~ d hydrophilic organic acids include phosphonates and sulfonates. A preferred hydrophilic organic acid is propionic acid.

W O 97/21431 PCT~US96/19900 In a pl' r~lL~d embodiment, only one of the R groups is a hydrophobic organic acid, since, in the case of Co(III~, this may result in a compound that is neutrally charged, and thus may cross the blood-brain barrier. Particularly preferred is the structure depicted in Formula 2:
Formula 2 NH
3 >=N~¦~ =<CH3 H~_ Co(li~ H

~ NH3 H3 Q=~
O-In addition, the length of the alkyl group shown in ~ormula 2 may be altered, either to encourage or prevent the carboxylic acid from "swinging around" to become an axial ligand.

By the terrn "amine" herein is meant an -NRR' group. In this embodiment, R
and R'may be the same or different, and may be hydrogen, alkyl or aryl. A
preferred-NRR'group is-NH7.

By the term "alkyl amine group'~ herein is meant an alkyl group, as defined above, with a -NRR' group, as defined above. As defined above, the alkyl group may be substituted or unsubstituted. Preferred alkyl amine groups are n-propylamine and n-butylamine.

By the term "alkyl alcohol" herein is meant an alkyl group with an -O~I group.
As defined above, the alkyl group may be substituted or unsubstituted. The all;yl alcohol may be primary, secondary or tcrtiary, depending on the alkyl group. In a preferred embodiment, the alkyl alcohol is a straight chain primary alkyl alcohol, generally contS~;ning at least 3 carbon atoms. Preferred alkyl alcohols include, but are not limited to, n-propyl alcohol, n-butyl alcohol, n-pentyl alcohol, n-heptyl alcohol, or n-octyl alcohol.

By the term "alcohol" herein is meant an -OH group.

In a preferred embodiment, at least one of R,-R8 is a targeting moiety. It is preferred that only one of the R groups be a targeting moiety. In an alternativeembodiment, more than one of the R groups may be a targeting moiety. ~Then the Co of ~ormula 1 is Co(IlI), at least one of R~, R2, R3, R4, R5, R6, R7 and R8 is a targeting moiety.

By the term "targeting moiety" herein is meant a functional group that ~
specifically interact with the target protein, and thus is used to target the cobalt compound to a particular target protein. That is, the cobalt compound is covalently linked to a targeting moiety that will specifically bind or associatewith a target protein. For examplc, the cobalt compounds of the invention may include a polypeptide inhibitor that is kno~,vn to inhibit a protease, thus effectively increasing the local concentration of the cobalt compound at a functional site on the target protein. Suitable targeting moieties include. but are not limited to, polypeptides, nucleic acids, carbohydrates, lipids, hormones including proteinaceous and steroid hormones, growth factors, receptor ligands, antigens and antibodies, and the lilce.

In a pre~erred embodiment, the cobalt compound cont~ining a targeting moiety as one of the R groups inhibits a protein, which may or may not be an en~me.
By "inhibition of a protein" herein is meant that a biological activity of the ~ protein is decreased or elimin~t~ ~ upon binding of the inhibitor. In the case of enzymes, inhibition results in a decrease or loss of enzymatic activity. For CA 02240183 1998-06-lO

example, polypeptides comprising protease substrates or inhibitors are used as an ~ group on the cobalt compounds, to form cobalt compounds that will selectively inhibit the protease. Similarly, a cobalt compound cont~ining an R
group comprising a nucleic acid that specifically binds to a particular nucleic acid binding protein such as a transcription factor is used to selectively inhibit the transcription factor. These targeted cobalt compounds preferentially bind tothe target site on the protein, favoring that site over non-specific binding to other sites or other proteins. This makes the resulting compound more effective at lower concentrations since fewer molecules interact at other sites and minimi7e~ the side-effects due to inhibition of other proteins. Secondary interactions also increase the time spent at thc target, giving more opportunityfor ligand exchange.

In designing a cobalt compound for a particular protein, it is to be understood that the high affinity of the cobalt compound for an imidazole moiety, or the other possible reactive axial ligand moieties, is such that the cobalt compound need not be a perfect fit in the acti~e site. Rather, what is important is that the cobalt compound be able to approach the target axial ligand moiety. ~or targeting active site residues of enzymes, for example, the cobalt compounds should generally not be larger than typical enzyme substrates or inhibitors. The2~ gross structure and surface properties of the cobalt compound reagent willdetermine its outer sphere interaction with the desired biological active site.
Specificity in outer sphere interactions is optimized by variations in size, charge, flexibility, stereochemistry, and surface properties of the cobalt compound reagent. Thus, in designing an appropriate inhibitor, the characteristics of the protein or enzyme target are exploited. In addition, as is sho~,vn in the Examples, increasing the local concentration of the cobalt compound at or near the active site of the protein is sufficient to increase thebinding of the cobalt compound and thus the inhibition of the biological -~3-activity of the protein, effectively decreasing the K~, or K; values, in the case of enzymatic inhibition.

When the target protein is known to have a histidine or other cobalt reactive amino acid that is functionally important, either Co(II~ or Co(III) may be used,with Co(III) being preferred. When the functional mech~ni~m ofthe target protein is unknown, either Co(II) or Co(III) may be used, with Co(II) being preferred. In a preferred embodiment, the cobalt reactive amino acid is also close to the binding pocket or site of the targeting moiety.

By the term "polypeptide" herein is meant a compound ranging from about 2 to about 15 amino acid residues covalently linked by peptide bonds. Preferred embodiments utilize polypeptides from about 2 to about 8 amino acids, with about 4 to about 6 being the most preferred. Preferably, the amino acids are naturally occurring amino acids in the L-configuration, although amino acid analogs are also useful, as outlined below. Under certain circumstances, the l S polypeptide may be only a single amino acid residue. Additionally, in some embodiments, the polypeptide may be larger, and may even be a protein, although this is not preferred. In one embodiment, the polypeptide is glycosylated.

Also included within the definition of polypeptide are peptidomimetic structures or amino acid analogs. Thus, for example, non-naturally occurring side chains or linkages may be used, for example to prevent or retard i~l vivo degradations. Alternatively, the amino acid side chains may be in the ~R) or D-configuration. Additionally, the amino acids, normally linked via a peptide bond or linkage, i.e. a peptidic carbamoyl group, i.e. -CONH-, may be linked via peptidomimetic bonds. These peptidomimetic bonds include CH,-NH-, CO-CI-I;, azapeptide and retroinversion bonds.

CA 02240183 1998-06-lO
W O 97/21431 PCT/US96/~9900 As used herein, "nucleic acid" may refer to either DNA or RNA, or molecules which contain both deoxy- and ribonucleotides. Generally, the nucleic acid is an oligonucleotide, ranging from about 3 nucleotides to about 50 nucleotides, with from about 12 to about 36 being particularly preferred, and at least 21 nucleotides being especially ~l~r~ ,d. When the nucleic acid is used solely to confer solubility, the nucleic acid may be smaller, and in some embodiments may be a single nucleotide. The nucleotides may be naturally occurring nucleotides, or synthetic nucleotides, and may be any combination of natural and synthetic nucleotides, although uracil, adenine, thymine, cytosine, guanine, and inosine are preferred. As is more fully described below, the nucleic acids include genomic DNA, cDNA and oligonucleotides including sense and anti-sense nucleic acids. The nucleic acid may be double stranded, single stranded, or contain portions of both double stranded or single stranded sequence. In a preferred embodiment, for example when the nucleic acid is used to target a zinc finger transcription factor, the nucleic acid is double stranded, as zinc fingers bind to the major groove of double stranded nucleic acids.

A nucleic acid will generally contain phosphodiester bonds, although in some cases, as outlincd below, a nucleic acid may ha~e an analogous backbone, comprising, for example, phosphoramide (Beaucage et al., Tetrahedron 49(10):1925 (1993) and references therein; Letsinger, J. Org. Chem. 35:3800 (1970); Sprinzl et al., ~ur. J. Biochem. 81 :579 (1977); Letsinger et al., Nucl.Acids Res. 14:3487 (1986); Sawai et al, Chem. Lett. 805 (1984), Letsinger et al., J. Am. Chem. Soc. 110:4470 (1988); and Pau~vels et al., Chemica Scripta 26:141 91986~), phosphorothioate, phosphorodithioate, O-methylphosphoroamidite linl~ages (see Eckstein, Oligonucleotides and Analogues: A Practical Approach, Oxford Uni~ersity Press), or peptide nucleic acid lin~;ages (see Egholm, J. Am. Chem. Soc. 114:1895 (1992); Meier et al., .

CA 02240183 1998-06-lO
WO 97/21431 PCT~US96/19900 . -15-Chem. Int. Ed. Engl. 31:1008 (1992); Nielsen, Nature7 365:566 (1~93)). These ~ modifications of the ribose phosphate backbone may be done to facilitate the addition of cobalt compounds, as outlined below, or to increase the stability and half-life of such molecules in physiological environments.

By "carbohydrate" herein is meant a compound with the general formula Cx(H2O)y~ Monosaccharides, disaccharides, and oligo- or polysaccharides are all included within the definition and comprise polymers of various sugar molecules linked via glycosidic linkages. Particularly preferred carbohydrates are those that comprise all or part of the carbohydrate component of glycosylated proteins, including monomers and oligomers of galactose, mannose, fucose, galactosamine, (particularly N-acetylglucosamine), glucosamine, glucose and sialic acid, and in particular the glycosylation component that allows binding to certain receptors such as cell surface receptors. Other carbohydrates comprise monomers and polymers of glucose, ribose, lactose, raf~mose, fructose, and other biologically significant carbohydrates.

"Lipid" as used herein includes fats, fatty oils, waxes, phospholipids, glycolipids, terpenes, fatty acids, and glycerides, particularly the triglycerides.
Also included within the definition of lipids are the eicosanoids, steroids and sterols, some of which are also hormones, such as pros~gl~n~lin~, opiates, and cholesterol. Hormones include ~oth steroid hormones and proteinaceous hormones, including, b-ut not li~lited t~, epineph~ine, ~.hyroxir.e, oxytoGin, insulin, thyroid-stimulating hormone, calcitonin, chorionic gonadotropin, - cortictropin, follicle-slim~ ing hormone, glucagon, leuteinizing hormone, lipotropin, melanocyte-stimutating hormone, norepinephrine, parathryroid hormone, vasopressin, enkephalins, seratonin, estradiol, progesterone, testosterone, cortisone, and glucocorticoids. Receptor ligands include ligands _ CA 02240183 1998-06-lO
W O 97/21431 PCT~US96/19900 that bind to receptors such as cell surface receptors, which include hormones, lipids, proteins, glycoproteins, signal transducers, growth factors, cytokines, and others.

In one embodiment, the targeting moiety is chosen just to confer solubility on S the Co(II) or cobalt compound. Thus, for example, the actual sequence of amino acid residues or nucleotides is not critical. Alternatively, as outlined above, the targeting moiety is chosen to target a particular protein or enzyme, and thus, when the targeting moiety is a polypeptide for example, the actual sequence of amino acids is important.

In a preferred embodiment, at least one of Rl-R8 of ~ormula I is a polypeptide.
In this embodiment, the polypeptide is chosen on the basis of the target proteinor enzyme to be inhibited.

For example, when the target enzyme is a protease, the polypeptide will mimic or comprise an enzyme substrate or the reactive site of an inhibitor. When the polypeptide comprises an inhibitor, the inhibitor may be either a reversible or irreversible inhibitor. The sequence of the polypeptide is chosen to allow the binding of the polypeptide to the active site of the protease.

The polypeptide and the site of attachment of the polypeptide to the cobalt compound, will be chosen to maximize the interaction of the cobalt with the active site histidine. That is, as is explained below, the polypeptide may be attached to the cobalt compound at the N-terminal end, the C-terminal end, or internally, via one or more amino acid side chains.

As is well l~nown in the art, the active site histidine of many enzymes is closeto the S 1 -S 1 ' position of the enzyme's substrate (or inhibitor) binding site.

CA 02240183 1998-06-lO
W O 97/21431 PCT~JS96/19900 Examples include the serine and cysteine proteases. Thus, in a preferred embodiment, the polypeptide is chosen to allow optimum interaction of the cobalt compound with the active site histidine. For example, the polypeptide may comprise roughly the P4 through P 1 residues of a substrate or inhibitor S ~which occupy the S4 to S1 positions of the enzyme's binding site), and beattached at the ~-t~rmin~l end (P 1 ) to the cobalt compound, to maximize the steric interaction of the cobalt compound with the active site of the enzyme, and particularly the active site histidine. Alternatively, the polypeptide may comprise the P 1 ' through P4' residues (corresponding to the S 1 ' to S4' positions), with att~hment at the N-terminus (Pl'). In a further embodiment, the polypeptide spans the P 1 -P 1 ' site, but has an internal attachment at or near the Pl or Pl' residues, to similarly maximize the interaction of the cobalt compound with the active site histidine. These types of attachments are described below. However, as notcd above, the interaction need not be perfect to allow inhibition, since it appears that increasing the local concentration ofthe cobalt compound near the active site is sufficient.

Thus, the present invention allows a known enzymatic substrate to be used as an inhibitor, as well as increasing the efficiency of known inhibitors, for example via decreasing the Kl. A wide variety of enzyme substrates and inhibitors for a variety of proteases containing either an active site histidine or an essential metal ion coordinated by a histidine are known in the art. In addition, the morphological properties of enzymes for which the crystal structures are known are used to design appropriate cobalt compounds.
Alternative embodiments utilize known characteristics about surface charge and hydrophobicity, and substrate and inhibitor specificity.

In a preferred embodiment, the K~ of the polypeptide inhibitor is decreased as aresult of attachment to the cobalt compound. That is, the inhibitor becomes a better inhibitor as a result of the ~tt:~hment of the cobalt compound. Thus, thecobalt compound is effective at lower concentrations since fewer molecules are wasted at other sites.

In a preferred embodiment, at least one of the R groups is a nucleic acid used to target the cobalt compound to a particular protein or enzyme. For exarnple, the target protein can be a nucleic acid binding protein that has at least one histidine that is important in biological activity, such as a zinc finger protein.

As is known for zinc finger proteins that bind nucleic acids, it appears that each zinc finger interacts or binds to three base pairs of nucleic acid (see Berg, supra). Thus, the actual sequence of the nucleic acid used to target a nucleic acid binding protein will vary depending on the target protein. Nucleic acid sequences and their target binding proteins are known in the art.

As with the polypeptides, the cobalt compound can be attached to the nucleic acid in a variety of ways in a variety of positions; the actual methods are described below. The attachment site is chosen to maximize the interaction of a cobalt-reactive amino acid such as histidine that is essential for metal ion binding (or an active site histidine) with the cobalt compound. In a preferred embodiment, the backbone of the nucleic acid is modified to contain a functional group that can be used for attachment to the cobalt compound. This functional group may be added to either the 5' or 3' end of the nucleic acid, orto an internal nucleotide. ~or example, the nucleic acid may be synthesized to contain amino-modified nucleotides using techniques well known in the art (see for example Imazawa et al., J. Org. Chem. 44:2039-2041 (1979); Miller et al., Nucleosides, Nucleotides 12:785-792 (1993); and WO95/15971, and references cited therein). In this embodiment~ amine groups are added to the ribophosphate backbone at the 2' or 3' position, thus allowing the attachment of CA 02240183 1998-06-lO
W O 9~/21431 PCT/US96/19900 the nucleic acid to the cobalt at either the S' or 3' end, or to an internal nucleotide. These amine groups are then used to couple the nucleic acid to the cobalt compound. Alternatively, nucleotide dimers, c~-nt~ining phosphoramide, phosphorothioate, phosphorodithioate, or O-S methylphosphoroamidite linkages may be made, and added to the nucleic acid at any position during synthesis, and the nitrogen or sulfur atom used for attachment using well known techniques, as outlined below. Additionally, the phosphorus atom of the backbone may be used, or linkers, as is known in the art (see for example Thuong et al., Angew. Chem. Intl. Ed. Engl. 32:666-690 (1993); and Mergny et al., Nucleic Acid Res. 22:920-92~ (1994)).

When Co is Co(II) in Formula 1, at least one of R~, R~, R3, R4, R5, R6, R~ and R8 is hydrophilic such that the Co(II) compound is soluble in aqueous solution. In one embodiment, only one of the R groups is hydrophilic and the other R
groups are chosen such that the single hydrophilic R group is sufficient to confer water solubility to the Co(II) compound. In a preferred embodiment, R, is hydrophilic, for example, n-propyl alcohol. In altemative embodiments, two, three, four, five, six, seven or eight R groups are hydrophilic. In a preferred embodiment, the hydrophilic group is a targeting moiety, and preferably either a polypeptide or a nucleic acid.

By "soluble in aqueous solution" herein is meant that the Co(II) compound has appreciable solubility in aqueous solution and other physiological buffers and solutions. Solubility may be measured in a variety of ways. In one embodiment, solubility is measured using the United States Pharmacopeia solubility classifications, with the Co(II) compound being either very soluble (requiring less than one part of solvent for 1 part of solute), freely soluble (requiring one to ten parts solvent per 1 part solute), soluble (requiring ten to thirty parts solvent per 1 part solute), sparingly soluble (requiring 3Q to 100 parts solvent per I part solute), or slightly soluble (requiring 100 -1000 partssolvent per 1 part solute). Alternatively, since cobalt cont~ining compounds are generally colored, the appearance of a color upon addition to a colorless aqueous solution indicates an acceptable level of solubility. For example, many Co(II) Schiff's base complexes have a yellow or orange eolor.

Testing whether a particular Co(II) eompound is soluble in aqueous solution is routine, as will be appreciated by those in the art. For example, as noted above, the appearance of a Schiff's base Co(II) complex color upon addition to a colorless aqueous solution indieates solubility. Alternatively, the parts of solvent required to solubilize a single part of Co(II) eompound may be measured, or solubility in gm/ml may be determined.

In a preferred embodiment, the cobalt compounds depicted in Formula 1 have a regiospecific hydrophilieity. That iS7 R" R2, R3, and R4, are either hydrogen, alkyl or aryl, and are therefore hydrophobie, and at least one of R5, R6, R7 andR8 is hydrophilic. ~Iowever, other combinations resulting in amphiphathic characteristics are also possible, as will be appreciated by those in the art. This is particularly desirable since this regiospecific hydrophilicity allows bctter positioning of the eobalt eompound in or near the active site or on the surface of a protein or enzyme, as is discussed below. Without being bound by theory, it appears that this regiospecific hydrophilicity/hydrophob;city allows the cobalt compound to more efficiently interact with the protein or enzyme, which generally displays both hydrophobic and hydrophilic regions.

Particularly preferred embodiments of the present invention include the structure depicted in Formula 4, wherein R~ is n-propyl alcohol, R~ is hydrogen, R3 is methyl, R6 is methyl. R7 is hydrogen, R8 is methyl, and R4 and R5 are hydrogen:

o ~ Ub, 8~ FRI lO:~la FA~ 212 81~179 CA 02240i-83~l998 06 10 ~ 9 ~005 c~s~c~orY ~S12 JA~N~9~8~

Formula 3 >= "~N=<
H~ Co ~H

Or H3 In this ~ ....~..l;.... ~ t, ~e Co can be either Co(II) or Co~II).

The ~ ureg depicted in Formulas 4 and 5 a~e also preferred:
Formula 4 H3C ~\ CH3 >=N" ~N=<
H~ Co ~H

O=~OH
For~ula 5 H3C>=N~s ,~N=<CH3 H~ Co ~H

> H3 In a ~.cr~.~c,d Pmho~irnrnt, the cobalt cr~rnrlPx~.5 may have groups that alter the retox pot~-nti~l, oxidation stability, or ability of the compound to exchange anaxial ligand. Fo} example, many of the Ca(II) ~ of the ir~vention are ~Mr ~ _ r ~) ~ 0~ ~8 ~RI lo: 50 F~ 212 ~18a~7~ Ui;"~ A.K oo CA 0224~l83 1998 - 06 ~ 10 PCT/US 9 6 / 1 9 9 ~ ~
c P I~YUS 12 Jhl'~ 1998 -~-ser~citive to air nYi~lAAtinn. Ihat is, in the prese~ce of ~1Tnt~phrri~ oxygen, they may be nYi~i7~1 Thus, in a p~erc~ er~bo~ , the Co(II) complexes ~ re ~r .1~ . d and utilized in tho abseuce of air.

Thus, in a filIther ernl-u.l;...~ ~.L, the Co c .~1 are AA~l~itinnAAlly mn~lifiecl to 5 make them ~r stabk cv~l.vu~ld . For e~ 1 of a methyl R
group in a complex of the inventiorl ~:vit~ a L illuu~u~_Lh,~l group results in a positive shift of the metal ~~Y;~ n F ~ "-1, ctAl~jli7.ir~ the metal complex with respect to air oxida~on. For exarnple, 1.1.1-trifluoro-2,4-p- . ,1 A ~ Ir. I ione is L~islly available, and may be used to ~llLIl~ ~ hinuu~v~,L~rl 10 d~L;vtlLi-_s of the Co(II) ~ ~--F1~ - '1te~ hcrein. Further well known Tnn~lifi~AAtinn~ such ~s ~hlor~ nn of the Srhil~s base - v~ .le also &reatly enhance the stability of these ~ JIL ~ ~ with respect to air nYi~ inn Thus, the usc of Li~luolv~nethyl groups alone or in ~l~ju l~,L on wLth rhlnrinAtinn Ofthe macrocycle will result iD a soluble air stable Co(rl) Il.~.v-,y-,le cnrnpl~Y.

15 These types of d_liY.-h~_~ may also be made to adjust thc redox potential of rnpl-~Yr~ to mnd7l1~te~ their .~ tiYiLy with other ~

Similarly, the addition of electron donating or electron w~L~ ing groups may ~cAect the activity of the cobalt c~ with respect to the ability to 20 ~ e an axial ligand. As shown in the i , ' the addition of Llinuvlv.l~ll~l R g,roups at the Rl ~And R~ positions basically ~ the ~.~d~iVitj of the Co(III) cvlll~uund towards rlew axial li~ands. Electron wil~ lg or donating groups are preferably added at the R~ and/or R~
p.~c;~ nc as this is easiest for ~ ais~ as wcll as the pl~fe.l~,l position for 2~ electronic wupli~ The R~ aT~d/oT R, positions are also ~.r.,.l.,d. ~t is ~Isopossible to put electrvn donating ~roups at the R3 ~nd R~ pn~iti~nc, but if R3 and/or R6 contain an electron witLd~w;l.g group t~erl tbe cvll~vu~ld may be I~lMFNDFn C~ T

CA 02240183 1998-06-lO
WO 97/21431 PCTiUS96/19900 difficult to synthesize using the schemes depicted herein. Suitable electron withdrawing groups include, but are not limited to, halides (F, Cl, Br, I, in decreasing order of electron withdrawing strength), phenyl and substituted ~ phenyl groups such as nitro-phenyl, amines and quaternary ~rnin~s~ thiols, nitro groups, carboxy groups, nitrile, alkynes and alkanes, sulfonyls, and others known in the art. Suitable electron donating groups include, but are not limitedto, -OCH3, methyl, carboxylate, and ether.

Once the R groups are chosen, the preparation of the cobalt compounds of the invention proceeds as outlined below.

Generally, the cobalt compounds of the invention are synthesized as generically disclosed below in Scheme I, using the general methods of Costa et al., L
Or~anometal. Chem., 6:181 C1966), which describes the preparation of derivatives of the components used to make the ligands used in the invention, such as acetylacetone ethylenediamine (acacen).

W O 97/21431 PCT~US96/19900 Scheme I

2N NH2 ~~ ~N NH2 Co / R6 O ~

R3 ~ R6 ~= N, ~N=<
R2~ Co ~R7 R1 !; R8 Compounds 1 (ethylene-1iAmine, "en"), 2 and 4 are generally made using techniques well known in the art. Compounds 2 and 4 are aliphatic ,B-diketones, and compound 2 is an aliphatic amine. It will be understood b~
those skilled in the art that compounds 2, 3 and 4 are thc resonance structures of compounds 6, 7, and 8 shown below in Scheme II. Compounds 6 and 8 are acetylacetone derivatives ("acac"), and compound 7 is the "monoacacen"
product.

Scheme II

R3 R4 ,R5 H2N NH2 +2~o H ~ R2~o,H N 72 H' O ~R7 ~

>=N N=<'R6 Co ~ ". ~'N==~<R6 R2~ ~H Hb~R7 ~ R2~ CO ~R7 R, 9 R8 R1 5 R8 When the Co is Co(III), the axial ligands are usually added in the last step.

Of particular use for att~chment of targeting moieties and particularly polypeptide and nucleic acid R groups are cobalt compounds with carboxy and amino groups. Cobalt compounds utilizing carboxylic acids are synthesized as depicted below in Scheme III:

W O 97/21431 ~CTAUS96/19900 Scheme III

X + R2~CH2C12 _~N\ NH3+

O~ 10 ~ 11 OH O

(CH3)3N, MeO~
/ H~o~R7 12 R4 R~; ~ R3 R R
~= N=<R5 Co >=N" ~N=<,Rs R2~ ~ b~ ~' R2~ - Co ~R

13 8 ~ 14 Rs o~ o~
OH OH

The am;no-derivative of the core cobalt compound is synthesized as follows:

CA 02240183 1998-06-lO
WO 97/21431 PCT~US96/19900 Scheme IV

R , H R ~< 5 ~0 1 6 NHAc <

/ H~o~R7 12 R R5 ~/ R8 R4 R~
R2 =--~R7 I R2~N ~N~R

17 R8 CH2C12 ~? '18 R8 NHAc NCS

The NCS group may then be used for coupling, as is known in the art.

In the case where the R group is a targeting moiety, the cobalt compounds are generally constructed in three phases. First, the corc cobalt compound is S synthesized with a functional moiety that can be used to couple the targeting moiety. For example, the core cobalt compound is made with an amine, a carboxy or a sulfhydryl group. Next, the R group, comprising a targeting moiety, is made, which also contains a functional moiety that can be used for attachment. In some instances, other reactive groups of thc targeting moiety CA 02240183 1998-06-lO
W O 97/21431 PCT~US96/199OO

are protected to prevent them from reacting with the functional group of the core cobalt compound. For example, amino acid side chains cont~ining amino groups, such as arginine, may need to be protected to prevent the side chain from reacting, although in some embodiments the z~tt~hment is done via a functional moiety of an amino acid side chain. Protecting groups and techniques are well known in the art. Once the core cobalt compound and the R group are made, they can be attached by reacting the functional groups. In some instances, the linkage is direct; for example a cobalt compound cont~ining a carboxy R group may be directly linked to an amino terminus of a polypeptide, as is depicted in the Examples. C-terminal attachment may be done using a cobalt compound with a amino functional moiety. As is kno-vn in the art, this direct linkage may be done in organic solvents or alternatively using coupling reagents such as 1-(3-dimethylaminopropyl)-3-ethylcarboiimide (EDC~ (sce generally, March, ~dvanced O~ganic Chen2istry, 3rd Ed. Kiley &
Sons, Inc. (1985); see also the 1994 Pierce Chemical Company catalog, technical section on cross-linkers, pages 155-200, incorporated herein by reference).

In a preferred embodiment, the linkage between the t~vo functional moieties may utilize a linker, also well known in the art. For example, two amino 2-0 groups may be linked via a stable bifunctional groups as are well known in the art, including homobifunctional and heterobifunctional linkers (see Pierce ~atalog and Handbook, pages 155-200). In an additional embodiment, carboxy groups (either from the polymer or from the cell targeting moiety) may be derivatized using well known linkers (see the Pierce catalog). For example, carbodiimides activate carboxy groups for attacL~ by good nucleophiles such as amines (see Torchilin et al., Critical Rev. Therapeutic Drug ~arrier Systems, 7(4):275-308 (1991), expressly incorporated herein). Sulfhydryl groups may be added to amines or carboxy groups with heterobifunctional linkers (see the Pierce catalog).

~ It should be understood that the attachrnent may be done in a variety of ways, including those listed above. What is important is that manner of attachment does not significantly alter the functionality of the targeting moiety; that is,they are still able to bind to the target protein. As will be appreciated by those in the art, this is easily verified.

As will be appreciated in the art, a number of functional groups of the targeting moiety may be used for covalent coupling, such as alcohols, amino groups, and carboxy groups. Alternatively, the targeting moiety may be derivati zed to contain a functional moiety, such as tllrough the addition of a linker cont~ininp;
a functional moiety. When a polypeptide is to be used as an R group, a preferred embodiment utilizes an amino group of the polypeptide. The N-terminal amino group may bc used, or alternatively, an amino group of an amino acid side chain, such as the amine groups of arginine, asparagine, gl~ mine, Iysine, histidine and tryptophan. Similarly, the linkage may be accomplished using the sulfur atoms of the side chains of methionine or cysteine. The carboxy groups of the side chains of glutamic acid and aspartic acid may also be used.

When the R group is a nucleic acid, a variety of positions may be used as the site of covalent attachment to the cobalt compound. In a preferred embodiment, the ribophosphate backbone of the nucleic acid is modified to contain a functional moiety (see for example Meade et al., Angewandte Chemie, F.nglic.l1 Edition, 34(3):352-354 (1995), and references cited therein;
Imazawa et al., supra, Miller et al., supra). For example, in a preferred embodiment, an amino group is added at the 2' or 3' position of the sugar using CA 02240l83 l998-06-lO

techniques well known in the art. In one embodiment, this is done by adding additional nucleotides that have an added amino group to the nucleic acid; that is, as shown in the Examples, one or more "extra" nucleotides is added to the targeting nucleic acid. Alternatively, the phosphodiester linkage between two nucleotides may be altered to form phosphoramide, phosphorothioate, phosphorodithioate, or O-methylphosphoro~mi~lit~ linkages, as is known in the art. The nitrogen or sulfur atoms are then used as functional moieties.
The nucleotide dimer, cont~ining the altered linkage, may be added to the nucleotide at any position. Functional groups on the nucleotide bases themselves may also be used, such as the amino groups on adenosine and cytosine, or modified bases such as is lcnown for thymine (see for example Telser et al., J. Amer. Chem. Soc. 111:7221-7226 (1991); Unglisch et al., Angew. Chem 103:629-646(1991); Angew. Chem. Int. Ed. Engl. 30:613-629 (1991); Goodchild, Bioconjugate Chem. 1:165-187 (1990); and Brun et al., J. Amer. Chem. Soc. 113:8153-8159 (1991)). Then the nucleic acid cont~ining the functional group may be added to the cobalt compound either directly or via a linker, as is outlined above for polypeptides.

Similarly, other targeting moieties such as carbohydrate, lipid, and hormone targeting moieties may be altered to contain functional groups for linkages, as will be appreciated in the art, or derivatized with linkers cont:~ining functional groups. As discussed above, the functional group for coupling should not prevent the binding of the targeting moiety to the target protein, and preferably does not affect the binding. Generally, these targeting moieties cont~ining suitable functional groups are made using well known techniques.

CA 02240l83 l998-06-lO
W O 97/21431 PCTnJS96/19900 Once synthesized, the cobalt compounds of the invention find use in a number of applications. At the broadest level, the Co(II) compounds are useful as reducing agents in aqueous solution.

In one embodiment, the cobalt compounds of the invention are useful as general bacteriostatic or bactericidal agents, antimicrobial agents and/or antiviral agents, for both topical and other therapeutic applications. For example, topical antimicrobial agents may be useful in cleaning and disinfectant compositions, as will be appreciated in the art. Therapeutic uses of antimicrobial and antiviral agents are also well known.

The compounds are assayed for antiviral, antimicrobial and antibacterial activity using techniques well known in the art; for example, bactericidal activity may be measured using the techniques outlined in example VI of U.S.
Patent 5,049,557. Both in vitro and in vivo antiviral activity may be measured using the techniques outlined in U.S. Patent No. 5,210,096.

The cobalt compounds of the invention can also be used to label proteins. The Co(II) compounds of the invention are preferably made with no axial ligands, and the Co(III) compounds are generally made with two axial ligands. Upon incubation with a protein, certain moieties on the protein will become axial n~l~7 resulting in a tightly bound protein-cobalt compound complex. Since cobalt-cont~;ning compounds may be detected spectrophotometrically, the result is a labeled protein. The pl~fe"~d axial ligand from a protein is the imidazole side chain of histidine. Thus, a protein with one or more histidine residues either at the surface of the protein or otherwise accessible to the solvent can be labeled using the cobalt compounds of the invention.
-CA 02240183 1998-06-lO
W O 97/21431 PCTrUS96/19900 In this embodiment, the cobalt compounds of the invention are added or contacted with the target protein. The excess cobalt compound may be separated, and the labeled protein, with the attached Co(III) compound, is detected spectrophotometrically. The Co(lII) compounds are generally S (letecfecl at 280,338, and 451 nm, although a broad range from 2~0 to 500 nm may be useful.

The stoichiometry of the bound cobalt compound to protein will vary depending on the number of potential axial ligands in or at the active site or on the surface of the protein, and may be ~lett~.rmined spectrophotometrically, as is understood in the art. Thus, for example, a protein which has four accessible histidines will generally bind four cobalt compounds, etc.

Thus, the cobalt compounds of the present invention are also useful in probing the surface characteristics of a protein.

When used to bind or label proteins, the cobalt compounds can be coupled, using standard tcchnology, to affinity chromatography columns. These columns may then be used to separate proteins from a sample. For example, depending on the specificity of the cobalt complex, proteins may be removed from a sample, or specific proteins, such as those conf:~inin~ histidines at or near the active site may be separated from other components of the sample.

In a ~-ef~ d embodimcnt, the cobalt compounds are useful as enzyme inhibitors. The mechanism of inactivation is similar to the mechanism of protein labeling. In this embodiment, an enzyme has one or more moieties capable of binding in an axial position in the cobalt compounds of the invention. One or more of such moieties are also functionally important for enzymatic activity, and are inactivated upon contact with the cobalt compounds of the invention.

For example, enzymes which have histidine as an active site catalytic residue orhave histidines which are functionally important for enzymatic activity are particularly preferred. Enzymes such as the serine proteases (trypsin, subtilisin, chymotrypsin, elastase, thrombin, factor Xa, lysozyme, and others known in the art), cysteine proteases such as the cathepsins and interleukin converting enzyme; RNAse H, thermolysin and lactate dehydrogenase all have active site histidines and thus may be inhibited with the compounds of the present I 0 invention.

In this embodiment, a cobalt compound is contacted with the target enzyme.
The imidazole side chain of an active site histidine binds to the cobalt compound as an axial ligand. In the case of Co(II), this occurs with a simultaneous or rapid oxidation of the Co(II) compound to form an enzyme-Co(III) compound complex. This is termed "redox coupling".

The binding (and oxidation, in the case of the Co(lI) compound) results in the inhibition of the enzyme. The exact mechanism of the inactivation is unknown; however, several possibilities exist. Thc bound cobalt compound, which after binding and oxidation is a Co~III) compound, may sterically interfere with catalytic activity, i.e. it may be bound in or near the catalyticactive site. Alternatively, the bound cobalt compound may interfere with the catalytic mech~nism, i.e. by binding to a catalytic histidine. Additionally, in the case of Co~II), it is also possiblc that a functionally important moiety at the active site is reduced by the Co(II) compound, and thus the enzyme is ~ 25 inactivated.

W O 97/21431 PCT~US96/19900 In a preferred embodiment, the inactivation of the enzyme by the cobaIt compound inhibitor is effectively h-t;v~ ible.

In alternative embodiments, the reactive axial ligand from the enzyme is the indole side chain of tryptophan or the side chains of cysteine, methionine, arginine, Iysine, asparagine, gl~ mine, aspartate or gl~lt~m~te. As outlined above, the availability of these moieties may depend on the pH of the solution cont~ining the protein or enzyme, since in the protonated state these moieties are not good electron donors suitable as axial ligands. Thus, enzymes with these groups within the active site, or enzymes which have functionally important tryptophans, cysteines, or methionines may be inactivated by the cobalt compounds of the present invention, as outlined above.

In an additional embodiment, metalloproteins are inactivated with the cobalt compounds of the present invention. Generally, the metals of metalloproteins have ligands such as histidine, cysteine and methionine. If one or more of these residues are inactivated using these cobalt compounds, the binding of the metal atom may be decreased or elimin~ted, thus reducing or elimin~ting biological activity. Particular metalloproteins include, but are not limited to,nucleic acid binding proteins such as "zinc finger" proteins and hemerythrin.
Zinc finger proteins utilize histidine and cysteine to bind zinc ions (see Berg,Ann. Rev. Biophys. Biophys. Chem 19:405-421 (1990), Berg, Science 232:485 (1986), and Berg, Prog. Inorg. Chem. 37:143 (1989), hereby expressly incorporated by refcrence3. Zinc finger proteins have been shown to bind nucleic acids and thus play a role in a variety of gene regulatory processes.
Zinc finger proteins include transcription factors and other nucleic acid-binding and gene-regulatory proteins (see Berg, Science, supra), and are found in eul;aryotes, prol~aryotes, and viruses. Other zinc finger proteins suitable for inactivation by the compounds of the present invention include the nucleic acid o ~ o~. as FRI lo: So F~ 212 slsa l7a UI~ K 007 CA 02240183 1998 - 06 - 10 P~TIUS 9 6 / 1 9 9 ~ (~
12 ~ AN l9g8 bindirlg doma~n of steroid and thyroid hor~none ~ a and the human . .,g, ... product GLI (see Pavletch et al., Science 261:1701 (1993)), Kin~er et al.. Natu~e 332:371 ~1988), that contains five zi~c finger domains. In a ,...c;r~ mhv.1 ~.-- .t one or ~ore of the zinc finger domai~}s utilizes at least5 one hiatitine to bind z~nc, with t_e proteins that utilize two hi~ inrg being .eI:_..ed. In some cases the metal is bount exclusively by L.yatUill~

When the Jnrt~ ptotein is a 'lrl~n7yme~ ~ a of the active site metal by the cobalt complex may mn~ Atr enzy~ne activity. Such mrts~ l,y~L~c3 in~Alude. but a~c not lilr~ited to, the c~1.uAy~ Lidases, 10 carbollic d~y~h~c, thermolysin, collsL;~n~-, hist;~in~l dehydrogenase, lc~ol-;_,.e A4 l.~Lol~c, ~ d rl- l~ ., a~u.~dc .1;~.1...~ . ~, alcohol oge.-~sc~ lactate dcLydlu~,la3cl stromalycin, allunoacyclase, Lly~Lu~hc~llyl-tRNA a~l~lilL,tbs~" and others known in the art.

In a ~-~f...-.,d e~nho~limPnt, serine ant cysteine ~ are inhibjtlq-l 15 In a ~.ef~ mholl~ , the qme to be inhibited is carbonic anhydrase.
Carbonic ~ ~ has been; ~ t- 'i in tiabetes, ocul~r disease such as Ll~'l""~s, and seizures and ~ll'~/~lb;~JllS. A-,-,u~ ly, inhibitors of carbonic , such as the Co(II) ~ ' of ~e present ' ' ' ~ ~hull, are useful in ~e t~ ...c.lt of ~ese c~-n.~

20 Thus, in one e~nho~linnl~nt~ the Co(l~) c~ leY~ ~ are useiul in ~e treat~ent of elevated intr~lel~ presSure and ~]~ Carbonic ~Ihy~se has been il.llll;. ~t d in elevated intr~c--lAr pressure, ~nd carbonic anhydrase inhibitors have been sho~n to be efficacious in dc~..e&,illg this pressure in anirnals and humans Isee Sharir et al., F~-- ~;Il-- .lls~l F~vc Rcs. 58(1):107-116 (1994);
Rassa n et al., Eve 1(Pt 5):697-102 (1993); Gunn~ng et al, Graefes Arrhivc for Clinir-AI a.1.1 F~U~ .1L-~I ODhth~lnnolnAv 231(7):384 ~1993)).

W O 97/21~31 PCT~US96/19900 Rassarn et al., Eve 7(Pt 5):697-702 (1993), Gunning et al., Graefes Archive for CliniGal and ~xperimental Ophthalmolo~y 231(7):384 (1993)).

In an additional embodiment, the Co(II) compounds are useful in the tre~tment of seizures and conwlsions. Carbonic anhydrase II deficient mice have been shown to have increased resistance to chemically induced seizures, and ~,eL.e~l",ent with carbonic anhydrase inhibitors has been shown to increase the resistance of normal mice to chemically induced seizures. See Velisek et al., EpilepsyRes. 14(2):115-121 (1993).

In a further embodiment, the Co(II) compounds are useful in the treatment of diabetes and abnormal renal function. Elevated levels of carbonic anhydrase have been associated with metabolic diseases like diabetes mellitus and hypertension, and carbonic anhydrase inhibitors have been suggested for treatment. See Parui et al., Biochem. International 26(5):809-820 (1992); Parui et al, Biochem. International 23(4):779-89 (1991); Dodgson et al., ~
Biochem. Biophvs. 277(2):410-4 (1990); Hannedouche et al., Clinical Sci.
81(4):457-64 (1991).

In a preferred embodiment, the cobalt compounds find use in the inhibition of proteins and enzymes of tumor cells. As outlined above, Co(III) "acacen"
compounds can exchange an axial ligand for a different one by a dissociative mechanism with the slow loss of one axial ligand to form a five coordinate intermediate, followed by binding to another suitable ligand. For most cobalt complexes, ligand exchange is a slow process because there is a large loss of ligand field stabilization energy when a ligand is removed from an octahedral d6 complex (see Huheey et al., Inorganic ChenZistry: Principles of St~ ucture and Reactivity, 4th Ed. ~arperCollins, N.Y., chapter 13). Generally, the exchange is slow; for example, [Co(III)(acacen)(NH3)2]Cl in water with excess s ~ 3 . ~

CA 02240183 1998-06-lO

imidazole exchanges ammonia for imidazole with a half-life under an hour at 25 ~C, with the rate of exchange increasing with temperature. However, reduction to cobalt(II) puts an electron into the antibonding dz~ orbital, labilizing the axial ligands. Typical one-electron reduction potentials with S irreversible loss of an axial ligand are around -360 mV vs NHE (Darbieu et al., Transition Met. Chem., 7:149 (1982)~. This property may be exploited as a "redox switch" to conkol the activity of the cobalt compound. For example, certain regions within tumors are often oxygen-starved due to high metabolic demands and inadeguate blood supply; therefore~ reductive reactions might be more favorable in such an environment than in a healthy cell (see A. C.
Sartorelli, Cancer Research, (1988), 48, 77S; Brown et al., J. Nat. Cancer Inst.
~3:178(1991))-Raising the reduction potential of a cobalt acacen compound with substituents such as halides may place it high enough for reduction to occur readily in tumor cells, but not in healthy cells. Ware and coworkers use a similar approach to attempt selective release of cobalt-bound cytotoxins in cancer cells(see Ware et al., supra).

Testing the efficacy of the cobalt compounds as inhibitors is routine, as ~vill be appreciated in the art. When the target protein is an enzyme, testing is similarto testing any enzyme inhibitor, as is known in the art. Generally, the enzyme is assayed in the presence and absence of the putative inhibitor, and kinetic parameters are calculated as is known in the art.

The amount of cobalt compound inhibitor needed to inhibit a given enzyme will vary depending on the number of other reactive axial ligands on the surface of the enzyme, as is outlined above for protein labeling. For example, an enzyme with an acti~e site histidine and two other "surface" llistidines willgenerally require at least a 3:1 ratio of cobalt compound inhibitor:enzyme. The CA 02240183 1998-06-lO

total amount bound to the enzyme may be determined spectrophotometrically, as outlined above.

In a preferred embodiment, the Co(II) compound inhibitors are generated i~2 situ by reducing the corresponding Co(III) compound. By "corresponding S Co(III) compound" herein is meant a Co(III) compound which has the identical R groups as the Co(II~ compound. Generally, the Co(III) compounds are synthesized with axial lig~ncls, such as, but not limited to~ amines, 2-methyl imidazole, and water.

In this embodiment, the Co(III) compound is synthesized, and then added to the l 0 enzyme under conditions which can result in the reduction of the Co(III) to Co(II). This may be done in several ways. For example, the in situ environment of the enzyme, whether it be in vit~o or i~ vivo, may be a reducing environment for the Co(III) compound, such that the Co(III) is reduced to Co(II). Alternatively, the Co(III) compound may contain an electron acceptor l 5 group as one of the R groups, such that in a given i71 situ environment, the electron acceptor group will pick up an electron and donate the electron to the Co(III), thus reducing it to the Co(lI) form. Suitable electron acceptor groups include, but are not limited to, cations such as methyl violgen (N-N-dimethyl 4,4' bipyridine), or ethyl or propyl violgen, as is understood in the art.
Additionally, the reduction potential of the compound may be tailored such that introducing the compound into a particu}ar environment causes reduction; for example, by glutathione in physiological systems. The resulting Co(II) compound then reacts with the reactive axial ligands of the enzyme to inhibit the enzyme as outlined above. Thus, the Co(II) compound is generated i~1 situ;
that is, a Co(III) compound is added to an enzyme, is reduced to the Co(II) form, which in turn inhibits the enzyme. In this embodiment, the Co(lII) compounds mav be inert with respect to a selected enzymatic target in a given CA 02240l83 l998-06-lO
WO 97/21431 PCT~US96/19900 oxidation state, yet inactivate the enzyme target in a second oxidation state.
This mechanism allows the in situ addition of a cobalt compound, whether i~
vzfro or in vivo, in an inactive form, with activation to the Co(II) compound form in a particular reducing environment.

The compounds of the present invention may be forrnulated into pharrnaceutical compositions, and ~(lmini~tered in therapeutically effective dosages. By "therapeutically effective dose" herein is meant a dose that produces the effects for which it is ~lministered. The exact dose will depend on the disorder to be treated and the protein to be inhibited, and will be ascertainable by one skilled in the art using known techniques. In a preferred embodiment, the pharmaceutical compositions of the invention are in a water soluble form, and contain a pharmaceutically acceptable carrier in addition to the cobalt compound. The pharmaceutical compositions may be ~-lministered in a variety of ways, inc}uding, but not limited to, orally, subcutaneously, l S intravenously, intraperitoneally, or topically.

Also provided are methods for inhibiting a selected protein or enzyme with the cobalt compounds of the invention. In this embodiment, the target protein is contacted or exposed to a cobalt compound. In a pler~ d embodiment, the cobalt compound has the structure depicted in Formula 1. Thc cobalt compound can be targetted to a particular protein by the addition of a targetingmoiety, such as a polypeptide or a nucleic acid.

Also provided are methods for inhibiting a zinc finger protein, comprising contacting a zinc finger protein with a cobalt compound. By "inhibiting a zinc finger protein" herein is meant that the biological activity of the zinc finger protein is decreased or elimin~ted upon exposure to the cobalt compound.
Generally, when the zinc finger protein is a nucleic acid binding protein. this 02:0~;a8 FRI iO:50 F.~ ~12 818a~7a CA 02240183 1998-06 lo P~TIUS 9~ I 1 9 9 ~
~ ~.r.~ ~nP~ l~E14/US 12 JAN 1998 ~o mean_ that tho zirlc fingcr will ao longer bind the nucleic acid to a q~ 9e~t degrec. Various prior art Co(~ll) cc ~u~u -ds are well known in the art, (see ~.S. Pate~t~os. 4,866,0~4, 4,866,053, 5,049,55~, 5,106,841, 5,142,076, aad 5,210,096). These c- ~I n~ , depicted below ia Forrnula 6, as well as the ~ )u~da ~rnhor~ 1 in Formula I, have utility ~n the LU~ inn of zi~c finger proteins. Accordingly, irl this ~ h~ t ~ when Co is Co(r[I), there is no rc~luu~ l. that at least one of R~ to R~ is a targeting moiety such as a ,Lde or a nucleic acid, althou~h this is prrf~ Lilcewise, when Co is CoaI), there ig no l~u;h .l,~lt that at least one of the R g~oups is hydrophilic, although this is preforred, Forcnula 6 RC>=N, ~N
R~ CO~ 3 $RE
RA RF
~ Forcnula 6, RA and RF are the same or di~f~rent and each is en all~yl group, aphenyl group or a ,~ t. i d~ ~;v..Li~_ of a phenyl group. RB and Re are the same or tifferent Ant each is L~ c~ b. ' ~ 1 alkyl gro~lp, a halide or 15 a group having the structure:
RG--C
Il o wherein R~; is hydrogen, an aL~co~cite group, an al~yl group, or OH. RC ~nd RD
are the sarne or different ant each is hy~O~, or an all~l group.

The followirlg ~ , ' ~ serve to more fully describe the ITlanner of using the above-~s~ibed iU'.~ iU~, as well as to set forth the best modes c~nt rnp1s~t ~q ~
, CA 02240183 1998-06-lO
W O 97/21431 PCT~US96/19900 for carrying out various aspects of the invention. It is understood that these examples in no way serve to limit the true scope of this invention, but rather are presented for illustrative purposes. All references cited herein are incorporated by reference.

EXAMPLES

Example 1 Synthesis of Cobalt Compounds A sample of ~Col"(acacen)(NH3)2~CI was obtained as a gift from Zvi Dori.
Acetylacetone, benzoylacetone, ethylene~ mine, and triethylamine (TEA) ~ere obtained from Aldrich (Milwaukee, WI).
Tris(hydroxymethyl)aminomethane (Tris, Trizma Base), polyethylene glycol (PEG g0û0) and 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide (EDC) were obtained from Sigma (St. Louis, MO).
N-(2-hydroxyethyl)piperazine-N'-2-ethanesulfonic acid (HEPES) was from J.
T. Baker (Phillipsburg, NJ). Cobaltous acetate tetrahydrate was obtained from EM Science (Gibbstown, NJ). Human a-thrombin and the assay agent Spectrozyme TH (H-D-hexahydrotyrosyl-L -alanyl-L-arginine-p-nitroanilide diacetate) were purchased from American Diagnostica (Greenwich, CT).
Antithrombotic peptides were manufactured as amides by the Beckman Institute Biopolymer Synthesis group at Caltech using solid phase methods.
Weak cation exchange resin Sephadex G-25 was from Pharmacia (Uppsala7 Sweden). Enzyme reactions were followed spectrophotometrically using a photodiode array spectrophotometer. Ultrafiltration materials were from Amicon (Beverly MA). HPLC used Vydac reverse phase columns. 'H NMR
~ 25 were obtained on a 300 MHz FT-NMR spectrometer. Solvents used include EM Omnisolve MeOH, Omnisolve C~l2Cl2 passed over basic alumina to _ remove residual acid, Fluka (Buchs, Switzerland) puriss. MeOH and dioxane, and Quantum Chemical (Tuscola, IL) absolute EtOH. Distilled water was prepared by a Barnstead Nanopure system. All other solvents were reagent grade.

Svnthesis of hvdroxypropyl acacen To 200 mL of deoxygenated CH~Cl~ was added l 0 mL of acetyl acetone (acac, 0.0974 mol) and cannulated into a 250 mL addition funnel, which was attached to a 500 mL 3-neck roundbottom flask containing 100 mL of deoxygenated CH2C12 and 32.6 mL ethylene~ mine (en, 0.488 mol). The solution cont~inin~
the acac was added dropwise to the en solution. The reaction mixture was extracted with two 50 mL portions of 0.Z NaPi, pH 5.5. The organic layer was separated and placed in a -20~C freezer overnight. The resulting solution was f1ltered through fluted filter paper and the solvent was removed in l~acuo. The compound was further purified using ~lash silica gel chromatography using 95:5:0.5 (v:v:v) CH,Cl~:MeOH:Et3N as the eluant. The resulting monoacacen was characterized by NMR.

Monoacacen (0.5 g, 3.5 X 10-3 mol) ~vas dissolved in 5 mL of ethanol and 7-hydroxy-2,4-heptanedione (0.51 g, 3.5 X 10-3 mol) was added. The dione was synthesized as described previously (Detty, M.R. J. Org. Chem., 44:2073-2077 (1979)). The reaction was allowed to proceed for 4 hours and the solvent was removed i~ acuo. The sample was purified using flash silica gel chromatography using 93:7 (v:v) CH~CI~:MeOH as the eluant. The resulting hydroxypropyl acacen was characterized by NMR.

Svnthesis of Co~ hydroxvpro~YI acacen Hydroxypropyl acacen (0.25 g, 9.4 X l o-4 mol) u~as dissolved in 2 mL of deoxygenated methanol in an inçrt atmosphere glove box. To this solution was added Co~II)(CH3COO-)2(H2O)4 (0.2338 g, 9.4 X 10-4 mol). The mixture was allowed to stir for an additional thirty minute~ The reaction vessel was sealed and the solvent was removed in vacuo. The compound was used w;thout further purification.

S Synthesis of [Co~ hvdroxypropvlacacen(NH3~J~3COO
Hydroxypropyl acacen was reacted with Co(acetate) as described earlier.
However after the reaction vessel was sealed, anhydrous ammonia gas was bubbled through the reaction mixture and subsequently exposed to air. The solvent was removed in vacuo, and the product was purified using an alumina column with neat methanol as the eluant. The sample was characterized by NMR.

Svnthesis of Acacen To 20 mL of ethanol was added 20 mL of acac (0.0973 mol). To this solution was added 6.5 mL of ethylene~ mine (0.0973 mol) using an addition funnel.
The solution was placed in a refrigerator at 4~C overnight, and the crystals were triturated three times with anhydrous diethylether (MP = 1 10.1-1 1 1.1).

Svnthesis of rCo(III)acacen(N~3~21~1 249.08 g of cobalt acetate, 6 H2O, (lmol) was dissolved in 1.750 L methanol and the solution was filtered through Whatman paper No. 1. Acacen (1 mol) was suspended in 150 mL methanol. Nitrogen dried by passage through a silica gel dessicant column was bubbled over the reagents for 15 minutes. The cobalt acetate solution was added dropwise (1/2 hour) and the orange-brown solution ~,vas left to react at room temperature under nitrogen for 2 hours. Theflask was opened to air and NH3 gas was bubbled into the solution; the mixture was concentrated on a rotary evaporator. An equivalent of sodium chloride dissolved in a minimum amount of water was added, poured into a wide vessel, and left to crystallize slowly. The brown crystalline powder was filtered, washed with methanol and dried.

Fl]rther synthesis of assymetrical or "mixed" ligands "Acacen" (compound 9 in Scheme I~): 1 equivalent ethylenerli~mine in anhydrous EtOH was added to 2 equivalents acetylacetone in EtOH with stirring. After 30 minutes, the mixture was put in the freezer to precipitate a white crystalline solid. The product was collected by vacuum filtration over a coarse glass frit and rinsed with diethyl ether. It can be recrystallized from benzene to desired purity. Purified crystals melted at 1 11 ~C.

"Monoacacen" (compound 7 in Scheme II): The 1:1 condensation product of acac and en was prepared according to literature procedures (Cros et al., C.R.Acad. Sc., Ser. II 294:173 (1982)) substituting CH2Cl2 for chloroform.
The resulting yellow oil often contained some acacen (about } 0%), which could be removed by flash chromatography on silica using 97 CH2CI2/ 3 MeOH/ 0.5 TEA either now or after the addition of another diketone.

"Bzacacacen"(the Formula 1 compound with Rl, R3 and R6 as methyl, R2 and R7 as hydrogen, and R8 as phenyl): 1 equiv. benzoylacetone in CH2Cl, was added to a solution of monoacacen in CH2CI2. Removal of solvent gave a white powder cont~ining some acacen impurity. Purification was accomplished by nash chromatography on silica using 97 CH2Cl2/ 3 MeOH/ 0.5 TEA.

"Aciden" (compound 11 in Scheme III): A solution of 1 equiv.
4,6-dioxoheptanoic acid in CH2C12 was added to 1 equiv. ethylenediamine in CH2Cl2 and the insoluble l: l condensation product immediately precipitated.
The product was collected over a frit and dried in vacuo. The melting point was 1 40~C, with decomposition. A direct reaction of 4,6-dioxoheptanoic acid with CA 02240183 1998-06-lO

monoacacen did not work, despite repeated attempts. l~vidently, the acid group was effecting decomposition, even under anhydrous conditions. Nor did using excess triethylamine to neutralize the diketoacid give satisfactory results.
"Acacaciden" ~compound 13 in Scheme III): 1 equiv. aciden was powdered and S slurried in Fluka puriss. MeOH. 1 equiv~ triethylamine and 2-2.5 equivalents acac were added and the mixture was allowed to stir overnight to give a yellow solution. It was evaporated to dryness to obtain the crude product as an orange oil. Further purification by flash chromatography over silica using a 5% to 25%
gradient of MeOH in CH2Cl2 with 0.5% TEA to guard against hydrolysis of imine bonds. Evaporation of solvent followed by recrystallization from EtOH
gave a beige solid. Mt was 282, as expected.

LCo(III)(acacen)(NH3)2~Cl: Procedure obtained from Zvi Dori (The Technion, Haifa, Israel). 1 equiv. of acacen was degassed in vact{o and placed under argon. Dry, degassed methanol was transferred into the flask via cannula. I
equiv. cobaltous acetate was treated in same manner and the resulting purple solution added via cannula to the clear solution of the ligand. An immediate color change from purple to orange was observed as the rcaction was stirred under argon for two hours. Ammonia gas was bubbled into the solution and the flask opened to air. Reaction was stirred with ammonia for 4 hours, evaporating solvent replenished as necessary. The reddish solution was filtered over a frit and concentrated on a hot plate. Addition of saturated aqueous NaCl precipitates the brown product. It can be recrystallized from ethanol to give a tan powder.

[Co(III)(acacaciden)(NH3)2~: The above metallation conditions were used, but with the acacaciden ligand. Crude reaction mixture did not afford precipitate, but purification over cation exchange resin using aqueous ammonium acetate W O 97/21431 PCT~US96/19900 followed by removal of the volatile buffer gave a light brown powder. M~ was 373, as expected.

Peptide synthesis of GGGdFPRarnide: The peptides were synthesized by the Beckman Institute Biopolymer synthesis group (Caltech). This was accomplished on p-methylbenzhydrylamine (MBHA) resin using N-tert-butyloxycarbonyl (Boc) arnino acid derivatives for Merrifield solid-phase synthesis on an ABI Model 430A peptide synthesizer. The terrninal Boc protecting group was removed with kifluoroacetic acid (TFA). Side chain protecting groups and the peptide-resin bond were cleaved under HF conditions (90% H~, 5% p-cresol, 5% p-thiocresol). After removal of HF under vacuum, the peptide/resin mixture was washed on a fritted funnel with ether. The peptide was then dissolved in 10% aqueous acetic acid and filtered through, leaving the resin behind. The crude peptide solution was subjected to gel filtration on anion exchange resin AG l-X2 to remove the scavengers. The peptide can bc further purified by reversed-phase HPLC on a Vydac C8 column using a 30-min. linear gradient of 6-26% acetonitrile/water/0.1% TFA with a 2.0 mL/min. flow rate.

Coupled product [Co(III)(acacen--GGFPR~(NH3),; shown in Figure 23: One potential difficulty in coupling the cobalt complex to this peptide is that the peptide's arginine side chain is more reactive than its N-terminus if the arginine is not protected or protonated (Bodzansky, Peptide Chemistry: ,4 Practical Textbook Springer-Verlag, Berlin, 1988). Since arginine has a pK~, around 12.0, it is easily protonated, but this renders the hydrophilic peptide insoluble in the organic solvents, such as dioxane, used for most coupling reactions. ~or this reason, we used the water-soluble coupling reagent 1 -(3-dimethylamino-propyl) -3-ethylcarbodiimide (EDC). At least a 1 0-fold excess of EDC is needed to compensate for its hydrolysis over the course of the CA 02240183 1998-06-lO
W O 97/21431 PCT~US96/19900 reaction. The large quantity of urea byproduct generated can be reduced by passing the solution through an Amicon YC05 filter or by extracting the crude oil with an organic solvent.

In addition, HPLC purification of a basic, hydrophilic peptide usually calls fora small amount of an organic acid in the eluting solvent to aid retention. Sincesuch an acid would attack the imine bonds of the free ligand, purification was attempted without it, but neither reverse phase C8 and C18 nor normal phase cyano columns were effective in resolving the mixture. Later, some progress was made using basic ammonium acetate buffer and acetonitrile on reversed phase, but there still was some decomposition of the product due to hydrolysis of the imine bonds. In order to prevent this, the imines were protected by inserting the metal into the ligand before ~tt~chin~ the peptide.

Synthesis was as follows. [CoIII(acacaciden)(NH3)~1 was dissolved in 0.1 M
HEPES buffer, pH 8 at 5~C. 1 equiv. peptide dissolved in the same buffer was added. 5 equiv. EDC were added directly. 4 hours later, another 5 equiv. were added. The reaction was stirred overnight at 5~C, then lyophilized to give a reddish brown product. The crude material was purified over cation exchange resin (Pharmacia G-25), eluting with ammonium acetate. Two products were collected, both of which contained the cobalt, based on reddish color, and the peptide, based on the presence of the phenylalanine signals in the 'H NMR
spectra (multiplets at 7.28 and 7.20 ppm). Mass spectrometry of both materials suggested that the first of the two to elute had lost an axial ligand, possibly replaced by coordination of the metal to arginine. This would likely deprotonate the arginine, lowering the overall charge to +1, causing it to eluteearlier. Mass spectrometry also suggests that the second band contains the desired product. as the first had M~ = 932, the second had M+ = 1066. The calculated mass of the desired product as a diacetate salt is 1062.

W O 97/21431 PCT~US96/19900 Example 2 Inhibition of Carbonic Anhydrase Tnhibition of Co(III) compound Bovine carbonic anhydrase (20 mg, 6.7 X 10-7 mol, Calbiochem) was dissolved S in 0.5 mL Tris buffer (pH = 8, 0.05 M). To this solution was added 30 mg of Co(III)hydroxypropyl acacen (30 mg, 7.6 X 10-5 mol) dissolved in 0.5 mL of H~O. This solution was incubated for 48 hours. The excess cobalt complex was separated from the protein using a PD-10 gel filtration column equilibrated with Tris buffer (pH 8, 0.05 M). This enzyme, which was incubated with the Co(III) complex, retained 100% of its activity.

Tnhibition with Co(lI) compound Two samples of bovine carbonic anhydrase (30 mg, 1 X 1 o-6 mol) were dissolved in 3 mL of degassed Tris buffer (pH 8, 0.05 M). To one of the sarnples was added Co(lI)hydroxypropyl acacen (30 mg, 9 X 10-5 mol). The other sample of bovine carbonic anhydrase served as the control. The solutions were incubated under inert atmosphere (glove box) and a 1 mL aliquot was removcd from each sample after 48 and 96 hours of incubation.
The protein was then exposed to air and the excess Co(II)hydroxypropyl acacen separated from the protein using a PD-I 0 gel filtration column (Pharmacia) equilibrated with Tris buffer (pH 8, 0.05 M). The enzymatic activity of thc protein was assayed using p-nitrophenylacetate as the substrate (Pocker et al., Biochem. 6:668-678 (1967)). The results are shown below:

Inhibition of Carbonic Anhydrase (CA) with Co(II)hydroxypropyl acacen Time of incubation % inhibition of CA
48 hours 33.8%

02, 0li,a8 F~I 10 50 ~.~ 212 8l8a~7a CA 02240183 l998-06- l~ ~ q 6 1 ~ 9 9 r~
~ 12 JAN l~gQ

96 hours 43.2%

Example 3 r~hihihnn of Thorrnolysin Therrnolys1n (2,500,000 units, C~1hiorhrn~) was dissolved in 20 mL of Tris S buffer (pH 7.Z, O.l M, 2.5 M NaBr, O.Ol M CaCl2) und stored at 4~C; enzyme r. ~ w4s d~fPrmiTI~d by using E1~2~0 5 17.65 ~nd a ~ weight of 34,600. This solution was filrther purified using gel filtration chrom~ ph~r on an FPLC using a ~;uperdex 75 colu~ (P~
eqni1ihn~t~ with O.l M Tris, O.l M NaBr, O.Ol CaCIl, pH 7.2. This stock 10 solution was stored at 4C. N-~3-(2-f~ryl)acryloyl~glycyl L l, --- 8~
(FAGLA) was obtained from Sigma as the thermolysin s1~hstrQt~ A stock solution of FAGLA (4.0 rnM) was p~eparcd by dissolving the substrate in ;Lyl r~.. ~.. ;~i (D~F) and diluting it with buffer to a final c~nr- .t~ ir n of 0.1 M Tris, 0.l NaBr and lO mM CaCI~, pH 7.0 (final w..~ ;on of DMF was 15 Z.5%; see Feder et al., Biochem. 9:2784-2791 (1970)). For all assays, the v- ~ ion of enzyme and substrate was 50 n~ and 2.0 mM ..~ ly. 'rhe pepti-l~sP acti~il~ of ~ermolysin was ~h ~ by follov~ing the deo~asc in -~sr~Tptifm at 346 due to the er~:ymatic hydrolysis of FAGLA. Initial velocitieswere ~3- ~ .. d for 5lO% of thc reaction.

20 Th~ ai.l (2 X 10 5 mol) was i1 - ~ b -t' .i with [Co(m)acacen~3)2~Cl (5 rnM) in HEPES buffer (pH 1.0, O.Ol M, 0.005 M CaC12). The concentration of ~errnolysin was 5 ~ lO-' M, while the ~ t~ ;011 of the cobalt inhibitor was l .25 X 10~ M. The results of this sb~dy are shown below and in Figure 3:
Inhibition of The~olysin Time of inruh~hnn % inhibition o ~ o~ ~8 ~RI ro~ F~L ~12 8l8a~7~ CA 02240183 1998 06- lO PCUUS 9 6 / 1 9 9 0 ~
- ~TB~ - ~PE~VlJS 12 jhl~l 1998 so 45 minutes 46.2%
190 minutes 63.9%
322 mirlutes 77.7%
Stoc~ solution of the~nolysin was mixed wit~ the cobalt c ~ u~ dissolved in 5 0.1 M Tris, 0.1 M NaBr, 0.01 CaCl2, pH ~.2 (run buffer) to yield a final enz;yme cu.~ c,lion of 10 mM and a cobalt ~ of 2.5 mM. These ~nhltion.c were inrl7hqtPt at 25C or 37C for several hours along with a control IEIcking cobalt CU~IJUUI~d~ pPrio~ ly S ml aliquots of these solutions were assqyed for residualenzs~mc activity by their atdition to a cuvettc contain 495 mL of run buffer and10 500 ml of FAGLA stock solution, and follow~ the absorption decrcase at 346 nmas described above. All enzyme assays were prrff~ at 25C. The results are shown in Figure 3.

Ligant ~ L~ are shown in Figure 3B. ~his c l ~ models the binding of the cobalt c~ v~ k to histidine residues on the~molysul by 15 r- .- .. I It~ g the bind~ng of in~ 7nle to thc bisaII~ine product (CoaII)acacen(NH3)2 1.35 mM of Co(lII~acacen~NH3)/ was ;..- ~Ih t~ -l with 0.1 M imida~701e in mn buffer. The risc in al.su.L,~.~ at 420 due to the . ~ h " ~&c of NH3 with irnidazole was Il~U~ u~.i with time at 25C ant 37C. Thc similarity ~n the t~ - c d r .-1< -~ ~ of ellzyme inhihlti~n ant ligand ~ - ~ h~ in thc modcl cobalt 20 complex suggests that ligant r ~ - lh~ with a histidine residue is the rate limiti~g step of enzyme inhibition The binding of the cobalt ~4.~.l.u~ 1 to the active site histidine was cu~ d as follows. Thermolysin (10 ~M in running buffer (0.1 M Tris, 0.1 M NaBr, 0.01 M
CaCl2, pH 7.2) was il....l..lt~ at 25C with Co(llI)(acacen)~NH3)2CI (5 mM~ in the 25 presence and absence of the inhibitor rho,~.h~ nn (~-~Y-L-LLa.---lv~ lv"ypLo~hvO-L-leucyl-L-tryFtoFhan, 50 ~M), whiGhhas a W O 97/2t431 PCTfUS96/19900 reported K, of 32 nM at pH 7.5 (Kitagishi et al., J. Biochem. 95:529-534 (1984). Phosphoramidon binds to thermolysin at the active site, and this enzyme-inhibitor complex has by crystallographically characterized (Weaver et al., J. Mol. Bio. 114:119-132 (1977)). After incubation with the cobalt compound overnight, the inhibitor was separated from the enzyme using gel filtration chromatography on an FP~C using a Superdex 75 column (Pharmacia) equilibrated with 0.1 M Tris, 5 mM CaCl2, pH 9. The resulting solution was transferred into the Tris running buffer using a PD-10 column (Pharmacia), and was characterized. There was no detectable loss of enzyme activity due to irreversible inactivation by the cobalt complex after removal ofthe inhibitor. Spectrophotometric characterization of this active enzyme revealed the binding of two cobalt complexes to the enzyme. Characterization of thermolysin, completely inactivated by the cobalt complex, showed the binding of three equivalents of the cobalt compound for each enzyme molecule.
Since protection of the active site prevents inhibitor of themolysin, and it prevents the binding of one equivalent of cobalt compound to the enzyme, the inhibition of the enzyme is a consequence of the binding of one cobalt compound at the enzyme active site.

Example 4 Inhibition of Thrombin Thrombin ~vas chosen as the first target enzyme.Several crystal structures of thrombin are available from the Protein Data Bank (e.g. file lPPB; see Bode et al., EMBO J. 8:3467 (~989)). Thrombin is a 34 kD serine protease with a well defined mechanism of action involving a histidine residue. It is vital to the coagulation cascade, but an unwanted clot is a severe, life-threatening condition. Thrombin was chosen for this investigation because its structure and mech~ni~m are well understood, there is a simple activity assay, and antithrombotic drugs are useful in the tre~tment of strokes.

Thrombin inhibition assay: As with any purified blood product, proper care was taken to avoid the tr~n~mi~ion of blood-borne pathogens. Thrombin was S taken as received (about 1 mL at 30 ,uM in 0.75 M sodium chloride storage solution) and divided into 100 ,uL aliquots. Each ali~uot was diluted to 10 mL
using clean, filtered (2 ,um) aqueous 0.75 M NaCl and divided into 1 mL
samples. Each sample was stored frozen at -80~C until ready for use. The protein should not be stored at -20~C as this is too close to the eutectic pointfor the thrombin-salt mixture and freeze-thaw cycling may damage the protein.

An assay buffer cont~ining 10 mM Tris, 10 mM HEPES, 0.1% polyethylene glycol (PEG 8000) and S00 mM sodium chloride was prepared to pH 8.
Following manufacturer's instructions, 5 ,umole Spectrozyme TH was dissolved in 1.000 mL filtered nanopure water. A series of ~Co"'(acacen)(NH3~z]Cl solutions was pL~a~ed by dissolving 16.6 mg ofthe compound in 0.75M NaCI and then diluting aliquots to a range of cobalt concentrations from 4.7 mM to 4.7 nM. Both of the purified materials were assayed.

Thrombin was preincubated with inhibitor in the kinetics buffer (total volume 2D of thrombin, buffer, and inhibitor of 992mL) for the times specified in Figure 1.
After incubation, the substrate, spectrozyme TH was added (8mL of 5mM
spectrozyme), and the thrombin-catalyzed hydrolysis rates were monitored at 406 nm. The final concentration of thrombin in the expcriments was 3 nM, and the concentration of substrate was 40 mM. The inhibitor concentrations are outlined in Figure 1. Thc rates of hydrolysis were determined from the linear portion of the saturation-kinetics plots. The percent activity is determined by dividing the rate of spectrozyme hydrolysis with inhibitor by the rate of hydrolysis without inhibitor and multiplying by 100.

A vial of thrombin prepared as above was thawed in warm water. 100 ~L
aliquots were added to 100 ~LL samples of the cobalt acacen solutions. One 100 ,uL aliquot of thrombin solution was diluted with 100 ,uL 0.75M NaCI to be used as a control. Samples were incubated as needed before assay. 980 ,uL of the assay buffer was placed in a 1.0 mL, 1 cm quartz cuvette and allowed to equilibrate to 25 ~C (Hewett-Packard Peltier constant temperature cell holder).
10 ,uL of a sample was mixed thoroughly into the cell's contents. The spectrophotometer was set for a 30 second delay during which 10 ,uL
Spectrozyme TH solution was added and mixed into the cell's contents by inverting the capped cell a few times before replacing it in the spectrophotometer. Scans from 250 to 500 nrn were taken every 30 seconds for 10 minutes, although a single-wavelength scan at 406 nm would suffice. After the runs, the control was allowed to hydrolyze to completion before determining the end point.

The absorbance values at 406 nm were extracted and used to find the pseudo-first order rate constant according to the formula:
- ln[(A.-A,~/(Aw-AO)] = kt where A is the absorbance at completion, Ao is the initial absorbance (approximately the first data point) and A, is the absorbance at each time, t. The slope of the linear fit (typically ~2 > 0.99) yields k. Comparison to an uninhibited control sample gave an indication of the relative activity for each cobalt-cont~inin~ sample. Controls were repeated periodically as a check for protein degradation.
.

o~ o~a8 FRI lO:Sl F-~ 212 8l8a~7~ CA 02240l83 l998-06-lO PGUlJS 9h 1 1 ~ 9 P~ ~P~ 2 JAN ~99 First the inhibition of I~LuLlLbm by 1~mn~ fir~i [Co~(acacen)(NH3)2~Cl was inv~tig~ A solution c~ 0.1 ,uM thrombin and 2 5 mM
[Com(acacen)(N~3)2lCI was ;n- 1 ~ ' at room ~ e for 24 hours. As a control, ~rombin from the same source was ;..~ t~ d without inhibitor for the 5 sarne length of time. A po tion of each ;. .~ solution was assayed at 25~C
with a~.~ e~cess of a GULLLL~;a1~ s1Tatp~ Spectrozyme IX, whose proteolysis releases a C~lLUlllU~flLUL~ p-rlitrn~ inr The pseudo-first order protuction of p-nitros-nil jnr wss lLLonLh,L~ ~l e~,1r~ t'.... h ;- _lly ant the rate constanteAL~t~;l for esch run. Thc activity of the control sarnple was normal, but the 10 cobalt~ .t,.;n;l~ sample was c , ' ~t ~y inactive (Fi~ure ~). The cobalt-free ligand had no effect.

The sa~ne ~ in~ was set up unter an inert ~ 1. .., wing a solution of wster soluble CotlI) LyLu~ Lup~' - , at si nilar ~ Iinn~ as before.
After inr~1h~ti~m, bo~h sarnple a~ad control werc exposed to air to oxidize thc 1~ cobalt. Again 109s of sctivity was fount for thc cobalt~ ;..;n~ sample.
S~rprisingly, the Co(III) ~ ....l.u1n.~ was a more effective inhibitor (0~/O
activity) ~an the Co(II) Cc)Lll~ùullt (42~/~ activity relative to its control).
Perhaps the overall positive charge on the Co(III) CCILLL~OU~d assists in a~t tirlg the inhibitor to the active site as is seen in a different system (Bagg~rl J.
20 Inorg. Biochem. 52:165 (1993)).

The [Co(IlI)acacen(NH,)21Cl ~ ~ l ~ t; ~ ' If ' werc rcpeated for a range of cob~ 1t CUILC._~3tLtLliULlS from 2.4 mM to 2.4 nM after at l:;~ hour ;-- ,~ .n at 25C
(~igure 5). Cobalt .,u. .- - .1 ~1 inn~ as low as 24 IlM were fount to inhibit the protein. ~nc11b~tinn at a lower ~ , SC, slows the onset of thrombin 25 inhibition, just as it slows the ligand ~. h~ of Co(lII). Partial inhihiti~n was observed over a range of cobalt c~n~ o..c after 3 hours at 5C.
Activity was t1PtP~in~d at 25C as before and was found to decrease with W O 97/2t431 PCT/US96/19900 higher cobalt concentrations and longer incubation times. The activity of the control sample lacking the cobalt compound was stable over time.

A crude attempt to determine the number of cobalt complexes bound to the enzyme was made by first passing the inhibited protein down a size exclusion column to remove most of the unbound cobalt complex and then determining the absorbances at two wavelengths for which the extinction coefficients of both pure thrombin and cobalt compound were known; this information gives the concentrations of eac~ based on Beer's law. The initial estimate for two separate samples is 5 - 8 cobalts per enzymc. There are only five histidines in thrombin, but binding to other residues and electrostatic binding cannot be discounted.

Inhibition of Thrombin by Peptides Containing dPheProArg The targeting approach requires attaching a recognition element to the cobalt complex for binding specifically to the active site of thrombin. The tripeptide sequence dPhe-Pro-Arg is a known inhibitor of thrombin, the arginine binding tightly to the P 1 aspartate. The peptides GGdFPR, GGGdFPR, GGFPR and GGGFPR were obtained as amides and assayed against thrombin. As expected from the crystal structures and inhibition data for similar peptides, (Bajusz etal., Int. J. Peptide Protein Res. 12:217 (1978)), the dPhe-containing peptides were found to be better inhibitors since they can access a hydrophobic binding pocket more efficiently than the natural isomers. The Kj for (Gly)3dPheProArg is about 209 ,uM.

Both purified Co(III)-peptide compounds (the two peaks from cation exchange resin) were assayed against thrombin without additional purification, the concentrations of each determined by assuming that the extinction coefficients were similar to that of [Coll~(acacen)(NH3)2]Cl (7,700 M~'cm~'). Preliminary results indicate that both inhibit at concentrations less than 1 ,~M, an order of magnitude lower than either ~Co'l'(acacen)(NH3)21CI or the peptide alone.

Example 5 Inhibition of a Zinc Finger Transcript;on Factor S In order to demostrate that cobalt compounds can disrupt the binding of a zinc finger to its consensus sequence, two model systems were used. E~uman Spl transcription factor, which contains three CCHH zinc fingers and a synthetic peptide representing the first zinc finger region of retroviral nucleocapsid protein.

Inhibition of Spl Binding Samples cont~ining 20 ul of binding buffer (25 mM Tris p~ 8.0, 100 mM KCI, 2 mM DTT, 100 uM ZnCl2, 10% glycerol) and 25 ng of Sp I (Promega) were incubated with 40 fmol of 32P labeled oligonucleotide in the presence or absence of cobalt chelate complex (Co(III)acacen with either NH3 or imidazole as the axial ligands) at various concentrations ~.001 to 0.05 mM) and evaluated by gel shift and filter binding assays.

Gel shift: Samples were run on a 4% polyacrylamide gel (80:1) at room temperature, lOOV, in 0.5 X TBE. Gels had been prerun for 30 minutes prior to loading. These experiments demonstrated that the presence of cobalt ~D complex inhibited the binding of Spl to consensus oligonucleotide.

Filter Binding Assay: The above samples were applied to nitrocellulose (0.45 um filtcrs, Schleicher and Schuell) and washed twice with washing buffer (100 mM HEPES, pH 7.5, I mM EDTA). Membranes were incubated for 15 minutes at room temperature with Filtron-X (National Diagnostics) and bound 02, 0~i~8 FRI 10:51 F.~ 212 ~18~7~ CA 0224ol83 l998-06- l~ PG~ i 013 Y lP~ 2 ~J N 1 counts were detected by liqnid ~intill~t~nn (F~er~no n I~ ~c~ ). Incre. sing a~nounts of cobalt cu~ ulJl,d resulted in decreased counts bound to ~e filter, in~ a loss of bindirlg between Spl o-nt oli~Y~ r-Jt;~

u~tion of the Structu~e of a Syn~etic Retrovi~al zinc Fin,~
5 An 18 o-n ino acid peptide ..~ ~- . c~)J~-~ to the f~rst zinc finger region of the A ~ I protein was gynth~i7~d TLt ~ .~ 7 in a sencs of s~uctural studies.

ConfirlT-otinn of Zinc Finger S1ructure: O. l mg.rnl solutions of peptide in 25 rnM
i ' ' r ' ' buffcr pH 7.0 were e ~ d by circul. r dichroism :~wt~uscupy. In the 10 abgeIAce of zinc, peptide~ displayed spectra ~h"~ 1;. of random coil structure.
In the presence of zinc, thc spcc~a changed d ~r~ll~r to one indicative of type II turn content and a:inc finger structure.

Dis.u~ti-,~. of Zinc Finger Structule: 4 mg peptide in 350 ul D20 was subjected to p~otein NM~ ~ s~o~, . Spect~a in the absence of z~nc displayed multiple 15 peaks in the aromatic region, i ~ ~ pealcs ~ g protons ~om metal-free hi~ir~in~ In the presence of zinc those peaks " . ~ i and were replacod by peaks ~ g protons rom met~l-bount hict ~in~- The prese~ce of cobalt ~ resulted in a ~3~ y altered spectra ;".1;" P~ of the structure of the peptide.

~, r~

Claims (15)

CLAIMs

1. A compound having the formula comprising:

wherein Co is either Co(II) or Co(III);
R1 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R2 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R3 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R4 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R5 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R6 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R7 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety; and R8 is hydrogen, alkyl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
wherein when Co is Co(III), at least one of R1, R2, R3, R4, R5, R6, R7 and R8 is a targeting moiety;

wherein when Co is Co(II), at least one of R1, R2, R3, R4, R5, R6, R7 and R8 is hydrophilic such that the compound is soluble in aqueous solution.

2. A compound according to claim 1 wherein Co is Co(II).

3. A compound according to claim 2 wherein at least one of R1, R2, R3, R4, R5, R6, R7 and R8 is a targeting moiety.

4. A compound according to claim 3 wherein at least one of R1, R2, R3, R4, R5, R6, R7 and R8 is polypeptide or a nucleic acid.

5. A compound according to claim 1 wherein Co is Co(III).

6. A compound according to claim 1 wherein R1, R2, R3 and R4 are each hydrogen, alkyl or aryl.

7. A compound according to claim 1 further comprises a first axial ligand.

8. A protein-cobalt compound complex comprising a protein and a compound attached thereto wherein said cobalt compound has the structure shown in claim 1.

9. A complex according to claim 8 wherein said protein is an enzyme.

10. A method of inhibiting a selected protein comprising contacting said selected protein with the compound of claim 1.

11. A method according to claim 10 wherein said protein is an enzyme.

12. A method of inhibiting a zinc finger protein comprising contacting a zinc finger protein with a compound having the structure comprising:

wherein Co is either Co(II) or Co(III);
R, is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R2 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R3 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R4 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R5 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R6 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety;
R7 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety; and R8 is hydrogen, alkyl, aryl, hydrophobic organic acid, alkyl amine, amine, alkyl alcohol, alcohol, or targeting moiety.

13. A composition according to claim 1 wherein R1 is N-hydroxypropyl, R2 is hydrogen, R3 is methyl, R4 is methyl, R5 is hydrogen, R6 is methyl, and R10 and R11 are hydrogen.

14. A composition comprising a water soluble tetradentate Schiff's base complex of Co+2.

15. A composition according to Claim 1 wherein said compound further comprises a first axial ligand.