Nuclease activity of transition metal complexes - a review

The review deals with basic elements of DNA structure, DNA interactions, importance of non-specific DNA cleavage, tran sition metal complexes that cleave DNA through redox chemistry and observations of the author and his associates on gel electrophoresis experiments using plasmid DNA with a series of 2-substilutcd heteroaromatic semicarbazones and thiosemicarbazones and their copper(II) complexes.

Nucleic acids provide exciting and difficult challenges for chemists and biochemists. They are inherently important, yet highly complex to study. Many aspects of nucleic acid biochemistry are widely appreciated throughout the scientific community.
The recognition that DNA serves as a target for natural and artificial molecules in the inhibition of cellular disorders and in therapy of certain diseases is of paramount importance in inorganic biochemistry. There has been an explosion in the research effort directed towards the isolation and evaluation of naturally occurring DNA cleaving agents and towards the design and synthesis of model compounds that can (specifically) cleave DNA.
The utility of these compounds (nucleases) is enormous and ranges from the creation of synthetic restriction enzymes for use by molecular biologists to the development of chemotherapeutic agents that may be effective against a variety of neoplastic disease. Nucleases have become the molecular scalpels of the biochemists and the indispensable tools for analyzing DNAstructure, sequencing DNA molecules and isolating and cloning genes.
Basic elements of DNA structure DNA may be viewed 1 as a double-helical assembly of the polynucleotides held together by hydrogen bonding and hydrophobic forces. Each polynucleotide includes a backbone of 2'-deoxyribose and phosphate moieties condensed together to form an alternating polymer (Fig. 1 ). The phosphodiester links involve the 5' and 3' oxygens of sugar and as a result, the polymer normally has terminal 5'-and 3'-phosphate ends. In addition, each sugar residue in the chain has a purine or a pyrimidine base covalently linked to its 1'-carbon. Under normal circumstances in aqueous solution the bases tend to stack one upon the other along the chain. In standard B-form DNA two complementary chains come together to form a duplex. The inside of the duplex contains a column of complementary base pairs. In the extended chain, the phosphate-phosphate distances is about 6 A, but the thickness of a typical base is about 3.3 A. Thus there has to be horizontal displacement of the phosphate groups in order for stacking to occur 2 . As a consequence, each sugar-phosphate backbone winds around the column of base pairs in a helical fashion. The twist angle between the long axes of adjacent base pairs is around 32°, and a stack of about II base pairs makes a complete turn of double helix. In a crude fashion, one can view B-form DNA as a twisted ladder with base pairs as the rungs and sugar-phosphate chains as the sides . The C(I')-N bonds that link the bases to the deoxyribose units occur at an angle with respect to the long axis of the base pair. Hence the C(l')-C(l')-vector that connects the two sugars lies off the axis towards one of the edges of the base pair. Consequently, the spacing between the sugar-phosphate chains is not uniform up and down the helix, and there are two different types of grooves that wind along the surface of the DNA duplex. The minor groove has a depth of about 7.5 A and a width of 5.7 A, while the major groove has a nominal depth of 8.5 A and width of 11.7 A. The deoxyribose groups form the walls of the grooves and the opposite edges of the base pairs form the floors. Although the phosphate groups attract an atmosphere of cations· that practically neutralize the charge, there is a residual electric field that is relatively strong within the minor groove.
There are two other important but less common forms of double-helical DNA. The A-form of DNA entails a different sugar conformation and represents a coiled form of a double helix. In this structure the major groove is narrower and deeper, and the bases are no longer perpendicular to the helix axis. Double helical RNA usu- ally adopts this conformation. The third structure, Z-form DNA, is left-handed. At high ionic strength sequenc~s that involve alternating runs of cytosine and guanine bases tend to adopt this structure.
Nucleases have been defined as molecule which cleave DNA either at specific sites along the strand or in an indiscriminant non-specific fashion. Therefore, the molecules, which cleave DNA, are called nucleases.

DNA interactions
The reactions performed by both site-specific and nonspecific molecules are quite remarkable and intriguingly complex. These reactions can be broken down in terms of two distinct chemical events : (i) the binding step, involving recognition of particular aspect of DNA base sequence or structure, and (ii) the cleaving step, involving a series of transformations which lead either directly or indirectly to hydrolysis of the phosphodiester linkage and scission of the DNA backbone.
A variety of nucleolytic reactions have been reported using both true enzymes and synthetic molecules, and all of which require metal ion for activity. Thus nucleases may be broadly classified into two categories : natural metallo-nucleases and artificial metallonucleases.
An enzyme which degrades nucleic acids and in doing so, has an absolute requirement for metal ion, is called 70 natural metallonuclease. A synthetic molecule also degrades nucleic acids and depends on metal ion for activity, but does not react catalytically in the true sense. For natural nucleases, numerous examples exist in which metal ion have been shown to be essential to the reactivity of the enzyme. Only in a few cases, however, the role of metal ion has been elucidatec1. For synthetic nucleases, the primary role of the metal ion has been generally to serve as a center for the generation of metal-mediated redox reactions in close proximity to the DNA helix leading to scission of the DNA strand. The reactions mediated by these redox-based nucleases, therefore, are clearly quite different from the types of reactions performed by the DNA-cleaving enzymes.

Scope of the review
This review deals with nuclease activity of transition metal complexes of synthetic molecules that cleave DNA through redox chemistry.
Importance of non-specific DNA cleavage Transition metal complexes of synthetic ligands usually accomplish non-specific DNA cleavage. The molecules which cleave DNA in a random non-specific fashion are often quite useful as tools to examine other molecules specifically bound to DNA, and therefore the lack of inherent specificity in these reagents themselves is essential.
To cleave DNA in a nondiscriminant manner requires both a binding interaction, which involves no site on sequence selectivity, and a cleaving reaction which shows no dependence on base composition or conformation. Nonspecific nucleases appear to recognize changes in the local conformation of DNA, such as groove widths and orientations about the phosphate backbone; therefore, some sequence of structural selectivity is encountered. Metal ions facilitate the non-specific cleavage of DNA. With respect to sequence neutrality in the cleavage reaction, the ideal reaction might be that involving hydrolysis of the phosphodiester. One obvious role for metal ions in hydrolysis of DNA is their function as electrophilic catalysts, i.e. in polarizing bonds and reducing the negative change on the phosphate backbone rendering the phosphodiester linkage more vulnerable to nucleophilic attack. However, model system have elucidated the importance of metal ions a source of metal-bound hydroxide ions, a potent nucleophile at neutral pH. The highest degree of non-specific cleavage of DNA with synthetic molecules is found through an innocent delivery of hydroxyl radicals to the DNA helix. Site-specific DNA cleavage has been reviewed by Basile and Barton 3 .

Transition metal complexes as chemical nucleases
As in the case of the naturally occurring enzymes 3 --6, for the synthetic complexes, the primary role of the metal rests in strand cleavage reactivity. For the enzymes, however, the metal appears to aid in hydrolysis of the phosphate ester, either by providing Lewis acidity or through the delivery of a coordinated nucleophile. In the case of synthetic molecules, the cleavage is not primarily hydrolytic. Instead, primary oxidative damage to the sugar leads eventually to scission of the sugar-phosphate backbone. In these complexes, the metal provides neither acidity nor coordinated nucleophilicity but the redox source to effect the oxidative damage, either indirectly through the delivery of reactive hydroxyl radicals or singlet oxygen, or directly through hydrogen abstraction reactions or direct binding and activation of metal oxo species. Syntheic metallonucleases are discussed below.

Methidiumpropyl-EDTA-Fe 11 [MPE-Fe 11 j :
It may be considered to be the first example of syntnetic metallonuclease to be widely used as a probe for protein-nucleic interactions as well as for the interacti n with DNA of smaller ligands, such as antiviral, antibiotic and antitumour drugs 7 -10 . These interactions can be revealed through the use of foot-printing techniques. Previously the nuclease, DNase I, had been successfully used as an enzymatic probe of the sequence preferences of DNA-binding proteins in foot-printing experiments 11 . In contrast to DNase I, however, MPE-Fe 11 offers the distinct advantage of yielding a higher resolution footprint of the interaction of small molecules, as well as proteins with DNA. This increase in resolution is due to the smaller size of the synthetic foot-printing tool compared with the enzymatic probe 12 and also because MPE-Fe 11 cleaves DNA with a lower sequence selectivity than does DNase I.
1 he synthetic nuclease MPE-Fe 11 was elegantly designed by Hertzberg and Dervan 7 . In the presence of ferrous ion and oxygen, this synthetic nuclease causes single-strand breaks in DNA in a reaction that is made more efficient upon the addition of reducing agents, such as dithiothreitol 7 . The reactive species responsible for the cleavage of DNA in this metal-mediated reaction is presumably a diffusible hydroxyl radical emanating from the ion center . For this non-metallic metallonuclease, binding is coupled to reactivity owing to the presence of the metal ion. The metal ion is essential to the activity of MPE, as it is for metalloenzyme system. The metal chelator, EDTA, found to be arranged in a slightly distorted octahedral geometry (Fig. 2) about the central ferrous ion is necessary to the reaction. The presence of an open coordination site or one that is occupied by a readily dissociable ligand, such as H 2 0 has been suggested as important to the reaction chemistry 13 • The reaction mechanisms of MPE-Fe 11 with DNA gave several key mechanistic features 14 . From product analysis, it was suggested that strand scission proceeds via oxidative degradation ofthe deoxyribose ring. The MPE cleavage reaction depends on ferrous ion and oxygen and is enhanced in the presence of reducing agents, allowing the system to become catalytic by regeneration of the reduced form of the metal. This metal-mediated reaction is also inhibited by superoxide dismutase, which converts superoxide to hydrogen peroxide and oxygen.
Both superoxide and H 2 0 2 are thought to be necessary for the generation of hydroxyl radicals in a metal catalyzed Haber-Weiss reaction 15 , Therefore, the results are consistent with the iron mediated reduction of oxygen to hydroxyl radical.

Bis(J, 1 0-phenanthroline)cuprous ion complex, [Cu 1 (phen) 2 } :
It acts as an efficient oxidative nuclease. The reaction requires binding of the complex to its nucleic acid substrate and H 2 0 2 as a coreactant 16 -21 . In this reaction, Cu(phen)~+ generated in situ is reduced to the cuprous form Cu(phen)2 in the presence ofthiol. As with MPE-Fe 11 , strand scission is seen to result from oxidative destruction of the deoxyribose moiety of DNA. The reaction involves generation of metal ion-associated radical species through a one-electron oxidation of the bound cuprous complex by H 2 0 2 21 . From analysis of the products ofthe scission reaction, it was concluded that the initial site of attack by this hydroxyl radical SB 1 like species is the C-1' hydrogen of the deoxyribose ring, with minor site of attack at the C-4' hydrogen 22 .
Two possible binding modes have been proposed for the interaction of Cu(phen)2 with DNA : ( i) interaction of one of planar phenanthroline ligands between adjacent base pairs and/or (ii) insertion of the coordination complex into a groove of duplex DNA 23,2 4 .

Metalloporphyrins :
Metal complexes of the water-soluble tetracationic porphyrin meso-tetrakis(N-methy 1-4-pyridinium )porphin (H 2 TMPyP) 25 -30 are most widely studied DNA-binding metalloporphyrins, and many of these complexes have been shown to act as metallonucleases either through chemi-cal29·30 or photochemical activation. These metallonucleases appear to cleave preferentially at AT sequences contained within a restriction fragment of pBR 322 DNA. These metalloporphyrins are, however, capable of cleaving at all four nucleotide positions suggesting that it is deoxyribose moiety which is the primary site of attack in the strand scission process. The preferential association in the minor groove with AT-rich sequences may be an indication that it is the groove-bound rather than intercalative interaction of porphyrins with DNA that leads to the cleavage chemis- try. A series of iron porphyrins tethered to an acidine ligand, in contrast, have been shown to cleave DNA in a sequence neutral manner 31 , thus it is probably the binding interaction rather than the cleavage chemistry.
Porphyrin derivatives have been used in the diagnosis and treatment of malignant diseases. A hemin derivative such as HPD (hematoporphyrin derivative) tends to accumulate specifically in neoplastic tissues and produces irreversible damages via singlet oxygen when the dye is photoactivated by visible light 32 -34 . Efforts have been made to prepare new porphyrin derivative in cancer

[Fell-EDTA} 2 -:
It is the simplest inorganic coordination complex capable of exhibiting nuclease activity. This complex provides a very simple yet powerful tool to probe the structure of DNA free in solution or bound to other molecules 52 -54 . Its mode of action as well as the function served by the metal are similar to that described for MPE-Fe 11 , since in both cases, it is the [Fe 11 -EDTA] 2 -moiety that is responsible for the DNA strand scission reaction. In binding studies of Ru(phen)J+ to a right-handed 8-DNA helix, the L1 enantiomer is favored for interaction, whereas for surface binding, it is the A enantiomer that is found to be preferred.
Nickel ~ompounds have two characteristics in common with leading antitumour drugs : (i) direct metal binding to N7 of guanine is possible and (ii) nickel complexes are able to catalyze oxidative damage to nucleic acids.
Nickel is a remarkably versatile metal in biological chemistry 74 . It is a necessary component of certain metallo-proteins and is at the same time an environmental carcinogen causing DNA damage and protein-DNA crosslinks. For the molecular biologist, nickel offers a tool for the study of nucleic acid structure in the form of tetraaza-macrocyclic complexes.

Desferal complexes :
Desferal, a well known siderophore 75 -78 and a highly effective drug in chelation therapy of iron overload diseases, forw~ a ;table octahedral Fe 111 complex. In contrast of r ~ ~::JTA, it cannot undergo a redox cycling, thus preventing iron catalyzed hydroxyl radical formation 79 , which is a useful property for its clinical applications. Indeed desferal is employed to arrest hydroxyl radical production in DNA scission reactions caused by Fell complexes 80 • 81 , while the Fe 111 complex of desferal is passive in DNA scission, the corresponding Cu 11 , Com and Ni 11 complexes of desferal actively cleave DNA 82 , 74 similar to other metallonucleases. The cleaving reaction with Cull complex requires a reducing agent such as 2mercaptoethanol, dithiothreitol or ascorbate and the reaction is made more efficient by addition of H 2 0 2 . The cleavage reactions with Colli and Ni 11 proceed even in the absence of a reducing agent. The cleavage efficiency is dependent on metal ion concen-tration, and optimal efficiency for 100% cleavage being 235, 42.5 and 10 J1M for Cu 11 , Co"', and Nil! complexes, respectively. Among the three, the Ni 11 complex showed maximum efficiency at low concentrations and the excess complex did not further degrade the DNA to the linear form.

Metal-oximates :
Copper(ll), nickei(II) and iron(II) complexes of 2acetylpyridine oxime (APO) were synthesized and characterized and their nuclease activity was investigated 83 . A gel electrophoresis experiment using pBR 322 (a circular plasmid DNA) was performed with these oximates in the absence and in the presence of H 2 0 2 as an oxidant. Only Fe 11 complex, [Fe(AP0h]S0 4 affords a discernible DNA cleavage at 30 min incubation period, as evidenced by the disappearance of form I (supercoiled) of plasmid. This is due to the stepwise conversion of form I (supercoiled) plasmid to form III (linear) through the transient formation of form II (open circular). This is consistent with the increased production of radicals mediated by ferrous ions by the well known Fenton reaction 84 . Further incubation (>40 min) results in complete degradation of DNA. This is probably due to the single cut in circular DNA giving rise to a linear form having open ends and eventual fragmentation of plasmid DNA.

Copper-hydroxamates :
The DNA nicking ability of copper(ll) complexes of the three hydroxamic acids was investigated by plasmid cleavage assay 85 . All the three metal complexes were able to convert supercoiled DNA (form I) into open circular (form II) DNA. Neither the free ligand nor the H 2 0 2 was found to have any DNA cleavage activity under identical reaction conditions. Among the three copper(II) complexes, that of p-TBHA showed maximum efficiency. Free hydroxamic acid in the presence of either an oxidizing agent or a reducing agent or both did not exhibit any DNA cleavage. The presence of hydroxyl radical scavengers such as mannitol and glycerol significantly inhibited the cleavage. All the three copper(II) complexes were able to cleave DNA only in the presence of l-1 2 0 2 indicating that the Cu 11 -DNA complex may be directly reduced by the peroxide to produce the corresponding DNA-Cuhydroperoxo species 86 • 87 thereby leading to DNA damage.

Copper(II) complexes of carbamic acids and thiocarbamic acids :
Heteroaromatic thiosemicarba.wnes and semicarbazones have been rather neglected in the main stream of chelation chemistry 88 ofbiological relevance. Perrin and Stunzi 89 have reviewed the applications ofthiosemicarbazones as antiviral agents. The report that the 2-substituted heteroaromatic thiosemicarbazones show appreciable antiviral activity 90 • 91 has evoked much interest in coordination chemistry. There is also much interest in the development of artificial nucleases. The continued interest in the copper complexes as chemical nucleases prompted us to synthesize the copper complexes 92 (Fig. 5) 92 . In overall view, all the complexes showed increased activity in the presence of oxidant may be due to the fonnation of hydroxy free radical 93 which involves oxidation of the deoxyribose moiety followed by breakage of the sugar-phosphate backbone 94 . The nuclease activity of the complexes in the absence of oxidant may be due to the binding of metal ions to DNA and these metal ions can be reduced and then oxidized by dioxygen, leading to the hydroxyl radical production close to the metal binding site which can damage DNA in site-specific reactions 23 · 95 or due to the reactive oxygen species.

Mixed ligand copper(!!) complexes :
Bis-pyridine adducts of the above mentioned parent copper(!!) complexes have been synthesized and characterized. The nucleolytic cleavage activity of the adducts 96 was carried out on a double stranded pBR 322 circular plasmid DNA by using the gel electrophoresis experiments in the presence and absence of oxidant (H 2 0 2 ). In the presence of oxidant all the adducts showed increased nuclease activity. The more pronounced nuclease activity of these adducts compared to parent complexes 92 may be due to increased production of hydroxyl radicals. In the absence of oxidant, the adducts exhibit low DNA cleavage when compared to parent complexes 92 . The low activity of these his-pyridine adducts may be due to six-coordination of copper{!!), which is consequently unable to bind with DNA due to the absence of an available site on Cu 11 • In the parent complexes, the availability of coordinated sites at the metal centre may facilitate the binding of DNA to allow nuc\eolytic cleavage.
Bis-picoline adducts 97 of the above mentioned parent copper complexes have been prepared and characterized by analytical, physicochemical and spectral techniques. All adducts showed increased nuclease activity in the presence of oxidant. The nuclease activity of picoline adducts are compared with parent copper(II) complexes.
Nucleolytic activity of natural molecules such as bleomycin and their metal complexes 98 -102 are not mentioned in this review because of intrinsic limitations of the subject. Metallo-intercalators 103 and cis-platin-DNA adducts 104 are not dealt in this review as these have been exhaustively reviewed.

Conclusions
Transition metal complexes of synthetic ligands are of paramount importance for the designing of chemotherapeutic drugs. DNA helics are chiral. They would thus be expected to interact with chiral metal complexes in an enantioselective manner. We can expect further progress in enantioselective probes.
Copper complexes of synthetic ligands appear to be promising chemical/artificial nucleases. Studies on metal complexes of nitrogen-containing ligands, especially heteroaromatic nitrogen bases are of much interest.
The present review exemplifies that simple metal complexes with suitable organic ligands and appropriate overall geometries can be tailored to produce synthetic chemical nucleases having specific nucleolytic activity. However, more work is needed to quantifY the available methods and to draw structure-function relationship and this continues to be a promising area of research in bioinorganic chemistry.