close

Вход

Забыли?

вход по аккаунту

?

CarbonЧHydrogen Bonds of DNA Sugar Units as Targets for Chemical Nucleases and Drugs.

код для вставкиСкачать
REVIEWS
Carbon-Hydrogen Bonds of DNA Sugar Units as Targets for Chemical
Nudeases and Drugs* *
Genevicve Pratviel, Jean Bernadou,* and Bernard Meunier"
This review article focuses on the molecular aspects of D N A cleavage by synthetic chemical nucleases (transition
metal complexes endowed with redox
properties and D N A affinity) and natural drugs (cytotoxic agents such as
bleomycins or enediynes) . Unlike deoxyribonucleases, which catalyze the
nucleophilic attack of water on the
phosphorus atom of a particular phosphodiester entity, these nonhydrolytic
DNA-cleavers are able to oxidize the
sugar units, generally by hydrogen atom
abstraction. Examples of oxidative attack on each of the five different C-H
bonds of deoxyribose are known, depending on the nature, structure, type of
activation, or mode of D N A interaction
of the DNA-cleaver. Further evolution
at the site of the initial lesion leads to the
release of bases, oxidized deoxyribose
units, or oxidized sugar fragments appended to the base or the terminal phosphate. In most cases the loss of a part (at
least) of a nucleoside, with the concomitant loss of one base information, primarily induces the cleavage of the D N A
strand. For both types of D N A cleavage
reagents studied within the two last
decades, the modes of activation and
1. Introduction
A large body of evidence has accumulated over the past three
decades on the role of oxidative damage to biological macromolecules in ageing phenomena, chronic inflammation, ischemia, cancer, autoimmune diseases, transcription activation
and, recently, in the induction of HIV expression."
Defenses
against oxidative stress are organized in living organisms with
nonenzymatic systems like vitamines [a-tocopherol (vitamin E),
ascorbic acid (vitamin C), or [karotene (vitamin A)], or with
enzymes like superoxide dismutases, glutathione, peroxidase,
and catalase.[zl.
Within the two last decades, much data has been published on
chemical and biochemical aspects of the structural damage generated by oxidative stress on cellular constituents (DNA,
proteins, lipids, and carbohydrates). Aerobic life can be considered as a well-organized system against injury from oxidation by
elimination of oxidants and reparation of damage to macromolecules. Repair of D N A is performed by D N A repair enzymes by several routes: nucleotide or base-excision repair. mis[*I
[**I
Dr. B. Meunier, Dr. G. Pratviel. Prof. Dr. J. Bernadou
Laboratoire de Chimie de Coordination du CNRS
205 route de Narbonne, F-31077 Toulouse cedex (France)
Telel'dx: Int code +61 55 3003
Abbreviations are listed in Section 5 at the end of the review.
D N A binding are presented, as well as
the details on the mechanism of deoxyribose oxidative degration. Because of the
need for highly efficient and highly
specific reagents, the development of
new artificial and selective D N A
cleavers, supported by an improved
knowledge of these different mechanisms of D N A cleavage, is to-day a challenging area in the rational design of
antitumoral or antiviral agents. as well
as in the field of molecular biology.
Keywords: bleomycin . D N A cleavage .
enediynes . transition metal complexes
match repair, and error-prone repair.[3] Oxidative damage to
D N A can be initiated by ionizing r a d i a t i ~ n , ' ~photooxidation
]
(UV or visible light with or without photo~entitizers),~~]
hydroperoxides activated by traces of adventitious transition
metals salts,"] hydroxyl radicals,['] or various other oxidizing
agents (for a recent review article on the oxidative degradation
of DNA, see ref. [S]). In this latter category, we find drugs like
bleomycin, cytotoxic agents with an enediyne motif, and chemical nucleases mainly based on redox-active transition metal
complexes (for recent review articles, see ref. [9]). D N A damage
can be produced by oxidation of nucleobases or sugar units.
More than 20 modified purines and pyrimidines resulting from
oxidative damage have been identified.["' This base damage
threatens genomic integrity and is the origin of lethal effects or
mutations in DNA. On the other hand, damage to deoxyribose
leads to the loss of one base information and/or a break on a
D N A strand, which might correspond to a lethal lesion, especially when an oxidation process produces a double-strand
break (DSB), that is, two close single-strand breaks (SSBs) on
opposite strands.
We will see in the present review that many highly cytotoxic
drugs or chemical nucleases are able to generate double-strand
breaks by oxidation of sugars on both strands of B-DNA (for a
textbook on D N A structures, see ref. [ I l l ) . More precisely, all
these sugar degradation processes are initiated by H abstrac-
Nonenzymatic D N A Cleavage
REVIEWS
tions, since the aliphatic nature of deoxyribose does not
favor oxidation by electron abstraction. The C-H bonds of
deoxyriboses have a well-defined arrangement, and their orientations with respect to drugs or chemical nuclease are also
controlled by D N A itself by DNA-drug interactions. In addition, the homolysis of a C-H bond depends also on the substituents on the carbon: less energy is required to remove
a H atom from a tertiary carbon than from a secondary carbon
atom. But more important is the relative orientation of the
drug to the sugar C-H bonds. For drugs interacting within
the minor groove of B-DNA, two tertiary C-H bonds are
easily accessible: C4'-H and C1'-H. The third tertiary C-H
bond ( C 3 - H ) is only accessible from the major groove. C2'
and C5' are secondary carbon atoms. In both cases, one
C-H bond is accessible from the minor groove the (pro-R) H
atom at C2' and the (pro-S) H atom at C5'; for assignments
of prochiral hydrogen atoms at C5' and C2' see ref. [12]) and
the other from the major groove. In Scheme 1 the different
arrangement of deoxyribose C-H bonds with respect to the
minor groove is depicted. It is essential for the rational
drug design of antitumoral, antiviral agents, or other pharmaceuticals based on D N A cleavers to understand the way of these
DNA sugar carbon- hydrogen bonds are activated at the
molecular level.
Before the molecular aspects of deoxyribose C-H bond activation are analyzed (Section 3), the different modes of activation and DNA binding of chemical nucleases and drugs involved in this deoxyribose oxidative damage will be described
(Section 2).
GeneiGve Pratviel II'US horn in Toulouse (Frurzce) in
1959. Shc tctt~~nded
the University Paul Subatier and
received her degree in Biochemistrj~in 1983 and her
PhD in Pharmacology in 1986 under the supervision of'
Bcrnurd Meunier. After spending post-doctorate ,fellowship in the group of Prof: A. Sartorelli at Yale Uniilersitj. shc come back to the group o f Bernard Meunier
CIS (1 C N R S researcher in 1988 at the Laboratoire de
Chimie de Coordination du C N R S . Her research interests ure the nrechcinism of D N A cleavage applied to the
dewlopmerit of biologically active compounds and genetic tool,\.
-
major groove
minor groove
Scheme 1. Among seven C - H bonds of deoxyribose that <ire posaible targets for
oxidative activation, four point into the minor groove (indicated by wedge-shaped
bonds), and three into the major groove (H, and H, refer t o the absolute configuration ( p r o - R or pro-S) of hydrogen atoms at C2' a n d C5')
2. Chemical Nucleases and Drugs Able To Cleave
DNA by Oxidation of a Sugar Unit
This section will be devoted to a survey of the nonhydrolytic
chemical nucleases and drugs able to cleave the DNA phosphodiester backbone by sugar oxidation. These D N A cleavers differ
completely from deoxyribonucleases that accomplish DNA scission by catalyzing the nucleophilic attack of a water molecule on
the phosphorus atom of a particular phosphodiester bond.[13]
In the past 15 years, several redox-active coordination complexes have been synthesized, and naturally occurring nucleolytic
agents have been discovered that are able to cleave the phosphodiester backbone of nucleic acids under physiological conditions
G. Pratviel
J. Bernadou
B. Meunier
Jean Bernudou II'US horn in Villejranche-de-Rouergue (France) in 1947, studied Chemistry, Phurmacy (doctorate in Phurmuceutical Sc,iences in 1975) and Medicine (doctorate in 1985) at the Universitji Paul Sabatier o f Toulouse, where he M U S qpointed
Profkssor ($Medicinal Chemistry in 1987. His collaboration with the group of Bernard Meunier started in 1980 with 1ii.c.research
interests in the mechanisms of action ofantitumor drugs. He is currentljj working on biological applications of tnetullo~~or~~hyrins,
in pcirticulrir on the topic discussed in the present review article.
Bernurd Meunier as born in Poitiers (France) in 1947 and obtained a first doctorate degree with Robert Corriu ui the
Universitv of' Montpellier in 1972. He ,joined Hugh Felkin at the Institut de Chimie des Substances Nuturelles rlu C N R S of
Natural Chemistry at Giflsur- Yvette and obtained a second doctorate in 1977. After one year as post-doc at 0.uford Universitj,
with K . Prout, he moved to the Laboratoire de Chimie de Coordination du C N R S , where he is presently research direc,tor. His
current rcwarch interests ,focus on metalloporphyrin-catul?.zed 0-uidations, ligninase models, oxidutions of pollutants, D N A
cleuvage hj. cationic mangunese porphyrins, and anti-HIV molecules hmsed on metalloporphj,rin clerivutives.
AnXeii. C/ic~111lnt
G I . Engl. 1995, 34. 746- 169
147
G. Pratviel et al.
REVIEWS
by oxidative attack on deoxyribose moieties. These synthetic
nucleolytic agents are complexes of transition metals, Cu(oP), ,
Fe-EDTA,[*] or metalloporphyrins that perform D N A strand
scission after activation 1 ) by reducing agents in the presence of
molecular oxygen or hydrogen peroxide, or 2) by oxygen atom
donors such as iodosylbenzene, magnesium monoperphthalate,
or potassium monopersulfate. In some cases these D N A breaks
are mediated by photoactivatable complexes (bipyridyl or
phenanthroline complexes of rhodium, ruthenium, or cobalt).
The most sophisticated nucleolytic agents are natural products
isolated from strains of Streplomyces (bleomycin, neocarzinostatin), Micrornonospora echinospora (calicheamicin), or Actinomaduva vervucospora (esperamicin) . All these chemical nucleases and drugs can be distinguished from other DNA-modifying
compounds used in nucleic acid
by their preferential deoxyribose-directed reactivity. We have also included recent information on irradiation of 5-halouracil-containing
DNAs and on photocleavage of D N A by rhodium complexes in
order to illustrate the possibility of oxidative attacks on positions 2' and 3' of deoxyribose units. We propose an almost complete description of the different ways of activating five of the
seven C - H bonds of deoxyribose: Cl'-H, one (pro-R) hydrogen atom at C2', C3'-H, C4'-H, and one (pro-S) hydrogen
atom at C5' (Scheme 1). Schemes14 to 38 each summarize one
of these five possibilities, and detail the subsequent potential
reaction pathways. These schemes also display reaction products that have been isolated or chemically characterized and that
can be used as reporters to identify at molecular level the attack
sites on deoxyriboses by D N A cleavers. (For an excellent early
review of the different routes for D N A cleavage, see ref. [9b] .)
2.1. Bis(1,lO-phenanthroline)copper, Cu(oP),
Cu(oP), (I, Scheme 2) was the first chemical nuclease demonstrated to have an efficient nucleolytic activity on doublestranded DNA.[''. 16] There are two essential coreactants, the
Cu' complex 1 and hydrogen peroxide." '1 Despite the
frequent use of 1 as a cleaving agent,
the exact nature of the reactive species
involved in this D N A cleavage is still
unknown. A popular view is that the
hydrophobic cationic complex 1 binds
to D N A and forms an intermediate
1
that reacts with H,O, to generate a
Scheme*, Structure of
bis(l.10-phenanthroline)
hydroxyl radical (HO'). However, accopper 1.
cording to reports on the detailed
chemistry of the cleavage reaction,
freely diffusible hydroxyl radicals are not responsible for the
strand scission.[17* Williams et aLii9] concluded that the
D N A damaging species can diffuse over a limited range from
the binding site of 1, which suggests that it is also unlikely to be
an 0x0-copper complex. Whether this reactive species is a hydroxyl radical coordinated to a cupric ion (called cryptoOH[lsl) is still an open question. The damage to D N A bases by
activated I is typical of processes induced by hydroxyl radi'3
@hU@
IUPAC nomenclature has not been used in this review.
748
CU"(OP),
+ 0;-
------f
+ 0,
Cul(oP),
(2)
1
2. I . I . Mechanism of Potentiation
[*I
cals,["] which suggests that at least a minor reaction pathway
involves free hydroxyl radicals.
The one-electron reduction of the starting complex Cu"(oP),
can be accomplished by adding exogeneous reducing agents
(ascorbate,["] thiols like mercaptopropionic acid.['"
231 2mercaptoethanol.[2'1 NADH[241)or by using superoxide generators (xanthine/xanthine oxidase," 71 cobalt-60 y-radiolysis).
One-electron reductants such as the superoxide anion and thiol
accomplish directly the reduction of the CU" complex, but twoelectron reductants such as N A D H are involved in a multistep
reaction in which superoxide plays a central role.[241H,O, can
be either added exogeneously, or generated in situ by spontaneous dismutation of the superoxide anion produced during
oxidation of the Cu' complex by molecular oxygen, by cobalt-60
y-radiolysis of water, o r from xanthine/xanthine oxidase.["] In
the presence of hydrogen peroxide, the D N A cleavage proceeds
as rapidly under anaerobic conditions as under aerobic conditions." 'I In the absence of added hydrogen peroxide. oxygen is
required, but its exclusive role is as a precursor for hydrogen
peroxide in the oxidation of thiol catalyzed by Cu(oP), .[171
As demonstrated by kinetic studies, the nuclease activity of
Cu(oP), proceeds by an ordered mechanism: the freely diffusing Cu"(oP), is first reduced to the Cu' complex [Eq. (I) and
( 2 ) ] , which binds reversibly to D N A [Eq. (3)]. This noncovalently bound Cu' complex is then oxidized by hydrogen peroxide to
generate the reactive species responsible for D N A strand scis[Eq. (4)]. The Cu" complex also binds to D N A but the
direct reduction of Cu''(oP),-DNA by 0;- is very slow and can
be neglected.[261
+
CU'(OP), DNA
Cu'(oP),-DNA
4
+ H,O,
CU'(OP),- DNA
[(oP),Cu"'-OH]DNA
-+
(3)
+ OH-
(4)
2.1.2. Interaction with Nucleic Acids
The pattern of D N A digestion in the presence of Cu'(oP), (1)
is consistent with its binding in the minor groove of a righthanded double helix. The reactivity of 1 in the minor groove of
DNAf2'' is confirmed by "footprinting" studies on netropsinand EcoRI-dodecamer complexes that are oxidatively degraded
by Cu'(oP), (netropsin binds within the minor groove, and Eco
RI within the major g r ~ o v e [ ' ~ ~The
' ~ reaction
~ ~ ) . mechanism
involving the H atoms at C1' and C4' of the deoxyriboses (which
are located within the minor groove) as the main targets of the
oxidative attack are also in accord with this position for the
cleaver.['5b] Initial observation of the inhibition of cleavage by
intercalators implied that the cleavage reaction proceeded
through an intercalative complex.[' 'I The nucleolytic activity
Cu'(oP), is sequence-dependent but not nucleotide
specific."".
The unsubstituted aromatic phenanthroline ligands of Cu(oP), preclude hydrogen bonding interactions with
bases as crucial factors for specificity.['5c1 The only observed
specificity is related to a predominant preference for binding at
5'-TAT triplets in the minor groove, which were cleaved most
Angrw. Chem. Int. Ed. EngI. 1995. 34, 746-769
Nonenzymatic DNA Cleavage
strongly at the deoxyadenosine sugar ring. The related sequences TGT, TAAT, TAGpyr and CAGT were moderately preferred, while CAT and TAC triplets, and polypurine and
polypyrimidine sequences were cleaved less often.[30.3 ' 1 The
preference of Cu(oP), for TAT may be related, in part, to
sequence-dependent variations of the B conformation of
Stacking interactions are favored with a propellertwisted base-pair. Changes in roll and twist angles open the
minor groove of DNA at the TA step and facilitate penetration
of the helix by an oP ring in either a partial intercalation or
groove binding mode. Cu(oP), is precluded by its tetrahedral
geometry from full intercalation, but a model in which one
phenanthroline ligand partially intercalates at a TA step is plausible.r321Since Cu(oP), attacks D N A from a binding site within
the minor groove, the geometry of this structural domain must
govern the cleavage specificity by the complex. Mutational
changes can generate additional sites detectable by the complex,
whereas DNase I, for example, did not detect substantial
changes accompanying these base substitutions.['sh. 331 Singlestranded DNA is a poor substrate.[17.2 2 . 3 3 a 1 Double-strand
scission of DNA by Cu(oP), is accomplished by two successive
single-strand nicks. DSBs have been observed with a semisynthetic nuclease prepared by chemically linking Cu(oP), to E. coli
trp
RNA is also a substrate for the nuclease activity of 1.
Thc copper complex shows a preference for single-stranded
loops relative to double-stranded A structures present
in RNA. but the detailed mechanism of cleavage is still
The C U ( ~ P ) ~ / H ,system
O,
cleaves A, B, and Z forms of D N A
at different rates: the B structure is the most easily cleaved; the
A structure. formed by RNA-DNA hybrids, is cleaved on both
strands at roughly one third the rate for B-DNA under comparable conditions; in contrast, the left-handed Z structure, with its
deep narrow minor groove, is completely resistant to Cu(oP),
degradation.[''. 2 2 1
The nuclcase activity of 1 can be aimed at specific D N A
sequences by attachment of one oP ligand to the 5' end of complementary oligodeoxynucleotides (ODNs) . Although the OPODN does not permit attack at a unique site on the DNA like
restriction ewymes. targeting of the nuclease activity by hybridization provides a method for analyzing single-stranded
DNA['.<, 3 6 . 371 and RNA.[381It is also possible to design an
oP O D N that can recognize and cleave specific sequences selectively on double-stranded D N A by the formation of a triplehelix.[", 3 9 1 Sequence-targeted cleavage of nucleic acids by
conjugates of phenanthroline complexes and oligo-sc-deoxynucleotides. which are resistant to nucleases. could be used to
regulate gene expression in v i v 0 . [ ~ ~ 1
The nucleolytic activity of I has also been directed at specific
sites of attack with DNA-binding proteins as carriers. in order
to study the interactions of DNA-binding proteins with D N A
restriction fragments or to isolate long D N A fragments for sequencing. cloning, or chromosomal mapping.[34,411 This strategy has been applied to the design of tat-based chemical nucleases
useful for site-specific cleavage of HIV TAR RNA.[42JSequence-specific scission of D N A by RNA-bound 1 has recently
been described.[431 Conjugates prepared by tethering 1,I 0phenanthroline to a dye binding in the minor groove like
REVIEWS
Hoechst 33258 have been used to increase the reactivity or to
modify the selectivity.[441
Extensive cleavage of D N A has been achieved with a chiral
system related to 1 in which copper(i1) is bound to one
chelation site of a multifunctional ligand (4,7'-phenanthrolino[5',6': 5,6]pyrazine) while the second site is occupied by
Ru"(bpy), . Cleavage is almost exclusively limited to the A isomer of this chiral bimetallic complex.[4s]
Mechanisms of DNA cleavage by 1 will be discussed in detail
in Sections 3.1.1 and 3.4.1.
2.2. Enediyne Anticancer Antibiotics
2.2.1. Mechanism of Potentiation
Naturally occurring enediyne anticancer antibiotics are prodrugs metabolized into active biradicals that can abstract hydrogen atoms from the sugar phosphate backbone of DNA in
the presence of thiols.['gJ Among enediyne antibiotics, neocarzinostatin (NCS) has been extensively studied by Goldberg et al..
and its mechanism of D N A cleavage is the best documented.['d1
Neocarzinostatin cleaves D N A through a nonprotein chromophore, NCS-chrom (2, Scheme 3). The nucleophilic addition
Scheme 3. Formation o f a biradical 4 from 2 in the presenct of a thiol, and abstraction of two hydrogen atoms from deoxyrrboses of DNA.
of a thiol (or hydride) on the C12 position of 2 (+ 3) gives rise
to the rearrangement (cycloaromatization) to an indacene biradical (4) with C2 and C6 as radial centers. This biradical is
able to abstract one or two hydrogen atoms from the D N A
backbone, leading to SSBs or DSBs, respectively. The thiol necessary for the activation of 2 must have strong reducing properties, a good affinity for DNA, or a high nucleophilicity to ensure
optimal a c t i ~ a t i o n . ' ~ ~ 1
The other natural products presented in Schemes4 and 5.
calicheamicin y: (5, CAL; a total synthesis was recently described[471),esperamicins 7-10 (ESPA, C, D, E), and dynemicin 12 (DYN), possess an enediyne unit as part of an overall
structure that also contains a DNA-binding moiety (a glycosylated aminosugar or an intercalating structure) and a triggering
749
C . Pratviel et al.
REVIEWS
gar
“‘S
6 (‘At- hiradical
L
S
0-sugar
OH
-Me
5 R = H CAI,
7 R = fucosylmthranilate ESI’A
8, 9, 10 R = O H ESPC.l>, E
0
DYNA
OH
O H i)H
OH
OH
14
13
12
1 1 ESP hiradical
Scheme4. Structures of CAL, ESPA. C. D, E. and their biradical species (for
numbering rules see refs. [Sf. 571.
N ti
./
device (a methyltrisulfide group for CAL and ESP or a quinone
for DYN). Reduction of these natural products with thiols initiates aromatization (Bergman type c y c l o a r ~ m a t i z a t i o n ,see
[~~~
Schemes 3 -5) of the enediyne moiety to a biradical species that
is responsible for the D N A damaging properties[9g*491 and the
cytotoxicity of these drugs. In the case of activated CAL or ESP,
the reducing thiol does not form a covalent adduct with the dye
and gives rise to the 1,Cphenylene biradicals 6 and 11.
D N A breaks mediated by enediyne compounds require oxygen, but neither reduced forms of dioxygen (such as the superoxide anion, hydrogen peroxide, or the free hydroxyl radical) nor
transition metals are involved in the D N A damage (2,[s01515’]).
In the presence of DNA, sulfhydryls, and O,, the enediyne
compounds undergo a single oxidation- reduction cycle, which
results in D N A cleavage and leaves an inactive form of the drug,
devoid of further redox properties. The first step of the reaction,
the generation of the biradical that abstracts H’ from DNA,
does not require 0,. In the next step with 2, D N A cleavage is
associated with the uptake of one equivalent of 0, per mole of
drug (DNA radical reacting with 0,) and is followed by the
uptake of at least one equivalent of reducing agent. Under
anaerobic conditions, misonidazole can replace dioxygen in the
cleavage step (the D N A radical is quenched by misonidazole)
and one equivalent of thiol is consumed per mole of 2.[521F o r
other enediyne molecules the stoichiometry of 0, and reductants has not been studied.
For 2 the C6 radical of the activated form 4 of the dye
(Scheme 3) performs the predominant H abstraction (C5’ chemistry) from DNA.[53] The remaining C2 radical of the drug
either reacts across the minor groove with the opposite strand of
D N A in the case of suitable special D N A sequences (Cl’, C 4 ,
or C5’ chemistry) or degenerates and leaves the D N A at the level
of a single-strand break. Selective quenching of the radical at c 2
results in the conversion of 4 to a monofunctional reacting spec i e ~ [and
~ ~ is’ due to H’ donation mainly from solvent or from
the bound thiol. This could account for the predominance of
SSBs (80%).
In the case of CAL (5, see Scheme 4), DDBs form 95% of
D N A damage.[551Both hydrogen atoms incorporated a t the C 3
and C6 positions of CAL biradical 6 originate solely from
DNA.[12”.561 For the esperamicin family 7-10 (see Scheme 4).
the two radical centers on the reactive species 11 are C7 and
c10.[9g3571
Dynemicin A (12. Scheme 5 ) can, for example, be activated
through the reduction of the quinone unit to the compound 13;
opening of the adjacent epoxide provides 14 and attack by a
nucleophile gives 15. Then, in the absence of geometrical con750
OH
OH
15
16
Scheme 5 . One proposition for the formation of the biradical 16 from DYN A (12)
by bioreduction (Nu = nucleophile).
straints imposed by the epoxide ring, this activated intermediate
undergoes a facile cycloaromatization to .he biradical 16, which
causes the classical D N A H abstraction of this series of comp o u n d ~ . [ ~So
~ .far,
~ * no
~ detailed study of H abstraction from
D N A by 12 has been reported.
and 12[601),radiolytical
Photochemical (for 2, ESP 7(2),I6l1or thermal activation (ESP[621)
are also possible and lead
to D N A cleavage.
2.2.2. Interaction with Nucleic Acids
2.2.2.1. NCS-Chrom
Strong interaction ofNCS-chrom (2) with D N A (affinity constant approximately 5 x lo6 M - ’), and proper positioning of the
activated form of the drug provide clues to its D N A cleaving
efficiency. NCS-chrom binds to D N A through intercalation of
its naphthoate moiety and through external interaction of its
protonated aminosugar residue with phosphate groups. The biradical part of the activated form is positioned in the minor
groove of DNA,[631and the intercalated part of the drug increases the distance between two base pairs within the interaction site. This detail is important when the site of the DSBs are
compared with those for other enediyne drugs. Double-strand
breaks of 2 are two base pairs (bp’s) apart in the minor groove,
compared to a 3 bp shift for other enediyne drugs. SSBs occur
mostly at T residues (T >> A >> C > G ) and show base specificity
rather than sequence specificity.[9d1Double-strand breaks are
sequence specific: 5’-AGC or 5’-ACT are the main DSB sites.
They also involve different mechanisms on each strand: H5’
abstraction by the C6 radical of 2 occurs on one strand, and for
some particular sequences, the C2 radical is properly located for
the second H abstraction on the opposite strand, where it mediates mainly either C1’ or C 4 chemistry. All SSBs are due to C5’
chemistry, and for the DSBs studied so far, at least one strand
is cleaved after H5’ abstraction.
Depending on the thiol used in D N A cleavage experiments
with 2, the amount and the nature of cleavage products are
modified.’46*641 The thiol is involved in the nucleophilic activation of the chromophore through a covalent bond, so it may
Arrjiuu. Chem. I n ( . Ed. Engl. 1995. 34. 746-769
Nonenzymatic DNA Cleavage
change the interaction of the activated drug with the minor
groove of D N A and modify the C4’ to C5’ chemistry ratio, as
well as the total amount of C4‘ chemistry observed at some
residues.[“‘] The nature of the thiol is also important for the
DSBiSSB ratio.[4h1One striking example is the generation of
direct double-strand breaks by glutathione (GSH). GSH was
shown to force the drug deeper into the minor groove of D N A
at 5’-AGC/3’-TCG sequences. which increased drug activity at
the C1’ sugar ~ a r b o n . [ ~ ~ ~ ” ]
The positioning of 2 in the minor groove of D N A is also
dependant on the local microstructure at the interacting site.
The introduclion of mismatches in double-strand cleavage sites,
or the replacement of guanine by inosine (without an NH,
group pointing into the minor groove), changes the access of the
biradical form of the drug to the CI’, C4‘, or C5’ sugar carbons
of D N A . ‘ ~ ~ ~
Mechanisms of cleavage for 2 will be discussed in Section 3.1 2.. 3.4.2.1 and 3.5.1.1 (Cl’, C4’, and C5’ chemistries,
respectively).
2.2.2.I?. 0 (her Enedijwcr
The other enediyne molecules mediate single- and a substantial number of double-strand breaks, because of the unique fit of
the drugs in DNA. The generation of a phenylene biradical
results in staggered double-strand cleavage sites through synchronous homolytic H abstractions from deoxyribose sites on
opposing strands. This series of compounds also attacks D N A
in the minor groove.
Culiclrecrmicin 1;: : CAL 5 causes double-strand D N A lesions
(more than 95%) in a plasmid or in a model of the cleavage
site.[5’] Each strand is cleaved by different mechanism. The
preferred attack sites are the 5’ penultimate pyrimidine C of a
5’-TCCT sequence and the nucleotide N, three nucleotides toward the 3’ side of the paired G in the complementary sequence
3’-NNAGGA.LS11.
In 1992 other sites of cleavage were reported: 5’-CTCT, 5’ACCT, 5’-T7TT.[’’”] CAL requires double-stranded helical
DNA at the 5‘ site of a 5’-TCCT sequence in order to anchor
efficiently. This drug binds to the minor groove of B-DNA by
orienting its oligosaccharide side chain toward the 3’ end of a
5’-TCCT sequence. For a review of CAL-DNA interactions, see
ref. [9f]. The interaction of the drug with its recognition sequence is very sensitive to changes, and truncated or modified
drugs show altered cleaving and recognition properties.[671Variations of D N A sequence within the preferential binding site also
affect the cleavage.[67b’There seems to be a real synergism of the
two parts ofthe intact drug (carbohydrate domain and aglycone
core) for the sequence-selective, double-strand cleaving efficiency of calicheamicin. CAL mediates C5’ chemistry on the TCCT
strand (Section 3.5.1.2) and C 4 chemistry on the complementary strand (Section 3.4.2.2). Some C1‘ chemistry may be involved in double-strand
but in the absence of any
precise data will not be discussed further.
E.~p~~runzicin.r:
Esperamicins 7- 10 (Scheme 4) exhibit base selectivity (T > C > A > G), but less sequence selectivity than
calicheamicin. in spite of their structural similarities. ESP A (7)
is the only member of the ESP family to mediate SSBs. ESP C,
D, and E (8-10) lack the fucosyl anthranilate moiety, and af-
ford a substantial number of DSBS.[“~]
So, the presence of two
side chains in ESPA might explain the production of SSBs,
whereas a suitable single arm might confer the ability to generate DSBs (ESP C. D, and E). In the ESP family, the trisaccharide tail seems to be very important for the cleaving efficiency.
since ESP D (9) or E (lo), with one or two sugars missing) were
considerably less efficient D N A cleavers while showing the same
sequence
h91
The binding of ESP to D N A is not as strict as for CAL ( 5 ) or
for NCS-chrom (2). ESP may bind in the minor groove in two
opposite orientations.[’71 A 3’-staggered asymmetric cleavage
pattern on the same sequences was reported for the doublestrand cleaver ESP C 8 as well as for the single-strand cleaver
ESP A 7.lS7. ESP A only mediates SSBs by targeting C5’
carbons on one or the other strand of the target DNA sequence,
while ESP C. D, o r E mediate DSBs with different mechanisms
on each strand (two possible orientations of the drug in its target
site). The authors propose that each radical position (C7 and
CIO) of the activated drug 11 has its own mechanism of deoxyribose-H abstraction: the radical at C7 would abstract a H 4 atom
and that at C10 a H5’ atom. Because the activated entity 11 can
bind in two orientations to DNA, a radical center can attack
either the one or the other strand and so promote the corresponding chemistry on either strand. SSBs generated by ESP A
can be explained by the loss of reactivity of the C7 radical center,
probably due to congestion from the nearby fucosyl anthranilate group, and to the ability of the remaining C10 radical for
H5’ abstraction. The mechanism of D N A cleavage for ESP involves both C4’ and C S chemistry (see Sections 3.4.2.3 and
3.5.1.3, respectively).
Dynemicins: Intercalation and minor groove binding have
been proposed for the interaction of this type of drug with
DNA.[58h1DYN A (12) causes single- and double-strand
breaks, but is not base-specific. Dynemicins preferentially attack the base adjacent to the 3’ side of purine bases (5’-GC.
5’-GT, 5‘-AT. and 5’-AG). They mediate strand breaks, three
bp’s apart (in the 3’ direction) in the two DNA strands.[701
General remarks on enedivne drugs: The bifunctional nature
of the enediyne biradicals make them potential double-strand
cleavers. The degeneration of the biradical species to an entity
with one active radical center may be due to altered positioning
of the activated drug, o r to steric hindrance o r quenching of one
of the radical centers by the fucosyl anthranilate moiety (for the
ESP family) or by the thiol unit (for 2). The optimal efficiency
of enediyne drugs may thus not always be achieved. because of
these surrounding tails.[s5,”1 The radical center that is not influenced by a surrounding side chain will ordinarily mediate
C5’ chemistry; the other one will react with the C4‘ position of
d e o x y r i b ~ s e . s’.s71
[ ~ ~ ~ and in some cases with Cl’[6h.’ I 1 or
C5‘.164l
2.3. Bleomycins
The bleomycins 17 (BLM, Scheme 6) are a family of glycopeptide-derived antibiotics, discovered by Umezawa and coworkers in 1966,[721that have been used clinically against nonHodgkin lymphomas, squamous cell carcinomas, and testicular
Their cytotoxic activity against mammalian cells is
REVIEWS
G. Pratviel et al.
0
+
?BLM-F~'"-OH
\ e-
--L
HO'
Fell-BLM =BLM-Fe"'-OH
+
HO'
H2O2
dismutation
t
OH
Fe-17
?
CONH,
Scheme 6. Structure of bleomycins A, and B, and proposed structure for their
chelation of an iron atom.
Scheme 7. Potential pathways for activation of Fe-BLM complexes i n the presence
of 0,. H,O,, or oxygen atom donors (PhIO. KHSO,).
troscopic characteristics of intermediates, and from observed
chemical reactivity.
Activated bleomycin can be generated by three different oxygen sources : molecular oxygen and electrons, hydrogen peroxide or alkyl hydroperoxides, and single oxygen atom donors.
The usual methods for Fe-BLM activation (either Fe"'-BLM/
0,/2e- or Fe"'-BLM/H,O,) produce active species that behave similarly toward DNA.["] Neither route distinguishes between the Fenton or the P-450 mechanism, since both are
compatible with pathways @ and @ or 8 and @ of Scheme 7.
The formation of 0x0-iron species has been well established for
cytochrome P-450 itself with single oxygen atom donors like
NaIO,, H,O,, ROOH, or iodosylbenzene,["] and for synthetic metalloporphyrins with peracids, iodosylbenzene, or
NaOC1.[yO1The association of iodosylbenzene with Fe"'-BLM
strongly supports the formation of BLM-FeV=O[911(route 0,
Scheme 7); however, in this case possible Lewis acid catalysis
which does not necessarily imply a FeV=O intermediate has
been discussed.[921Moreover, iodosylbenzene is not a benign
oxygen surrogate in reactions with BLM, especially for the long
incubation times generally used, and results should be viewed
with caution.[931With potassium monopersulfate (KHSO,), a
water-soluble oxygen atom donor at physiological pH
(Scheme 7, route @), activated Fe-BLM can give single- and
double-strand breaks on DNA[941 or transfer oxygen to
2.3.I . Mechanism of Potentiation
ole fin^.['^] When Fell'- BLM is activated by particular alkylhydroperoxides that were shown to undergo homoiytic scission of
BLM-induced D N A breaks are mediated by several redox-acthe 0-0 bond route @ in Scheme 7,'y6a1the activated species
tive metals (iron, cobalt, copper) which are strongly chelated by
is not capable of DNA cleavage.[g6b1These data support the
nitrogen atoms of the peptide moiety of BLM and which react
assignment of "activated BLM" to a high-valent bleomycin
with molecular oxygen.['" b, 8 3 1 Oxygen activation by Fell0x0-iron complex (the formation of such 0x0-iron species
BLM requires an additional electron. This can be provided by
from an alkylhydroperoxide requires a heterolytic splitting of
disproportionation of two Fe"-BLM molecules, by reductants
the peroxidic 0-0 bond). The inefficiency of radical scavsuch as ascorbic acid1841or alkyl t h i ~ l s , [ ~84c1
~ " or
, by electroe n g e r ~ [ ~in
' ] these reactions suggests that the active cleavage
chemical r e d ~ c t i o nA. ~key
~ ~step
~ in D N A lesions is the homolpathway goes predominantly through an 0x0-iron route.
ysis of the C 4 - H bond of deoxyribose rings[861by the "activatConsideration of the transformations of small molecules with
ed" b l e ~ m y c i n . [ ~871
~ "The
. exact nature of this activated BLM is
activated Fe-BLML9". b1 provides additional support for the exstill a matter of debate. All the hypotheses concerning the H 4 istence of a BLM-FeV=O species. Also, the absence of similarabstraction step can be classified in two categories (Scheme 7):
a Fenton route (HO' as reactive species, routes @ and @lM4"])ities between the DNA cleavage pattern produced by Fe-MPE,
or a P-450-like route (with BLM-FeV=O as hydroxylating
a well-known producer of hydroxyl radicals and that produced
0,
0,
first suggested by Burger et al.[87b1). by Fe-BLM (see Sections 3.4.2 and 3.4.4) indicates that Feagent, routes 8,
Evidence to support a particular pathway can be obtained from
MPE and Fe-BLM do not share a major common oxidative
the different ways of generating "activated BLM", from specpathway for D N A cleavage, and that the free hydroxylgenerally assumed to correlate with their ability to bind to and
cleave
More recently R N A has been reported as a
possible target for BLM-mediated cleavage (see Section 3.4.3
for further details). Three domains can be distinguished in the
structure of BLMs:['"-~] the bithiazole and the positively
charged chain contribute to specific binding to D N A ; the
pyrimidine, b-aminoalanine, and imidazole moieties are involved in dioxygen activation by chelating iron; finally, the
gulose and carbamoylated mannose residues are perhaps responsible for the selective accumulation of BLM in some cancer
cells, and might also participate in iron coordination and oxygen
A large variety of analogues of BLM were
snythesized to provide new insights into chemical or biochernical aspects of BLM activity. It has been possible, for example,
to establish the coordination structures of metallobleomycins
with more certainty,[761to improve the dioxygen activating capability["I or the D N A cleaving activity,[781to carry out asymmetric e p ~ x i d a t i o n , [ to
' ~ ~modify the sequence specificity,[s01to
better define the structural elements primarily responsible for
controlling sequence-specific D N A cleavage,l8 and to develop
chiral bleomycin models that discriminate according to chirality
on cleaving D N A . [ ~ ~ ]
~
752
Aigw
Chrm Inr Ed En& 1995,34,146 769
REVIEWS
Nonenzymatic D N A Cleavage
radical can be excluded as a major reactant for Fe-BLMinduced
The nondiffusible activated BLM species, two redox equivalents above the nonactivated Fe"'-BLM, is formulated as
BLM-Fe"=O, and is capable of H abstraction from deoxyribose to give a carbon-centered radical. The resulting BLM complex, BLM- FelV-OH, is still a strong oxidant (one redox state
above that of the initial Fell'-BLM) and may react with the
generated carbon-centered radical to produce a hydroxylated
derivative by a typical P-450 model route (oxygen-rebound
mechanismi"""') (Scheme 8, route a). But. as was recently
a ) oxygen transfer
H' abstraction
b) e- abstraction
I
11
Scheme 8. Thrcc possible oxidative pathways for BLM or Porph complexes
(M = Fc or M n ) . a ) P-450 route; b) diverted P-450 route: c) peroxidase route. I IS
equivalent to the so-called "compound I"; 11 is equivalent to "compound 11".
demonstrated on DNA,'"' the incorporated oxygen atom
comes from the solvent. Thus this result points to an alternative
possibility: activated BLM in the second step of the reaction can
behave like a peroxidase (see classical mechanism in Scheme 8,
route c ) ,mediating an electron abstraction rather than transfer
of an OH entity (diverted P-450 route, see Scheme 8, route b).
Then, a water molecule can react with the generated carbocation
and give the hydroxylated derivative.
Finally, little damage is detected on bases when D N A is incubated with Fe-BLM under conditions where D N A breaks are
detected.[t00]In 1993 it was reported that D N A can be cleaved
by Ni"'- BLM if potassium monopersulfate is the oxidant.""'
2.3.2. Interaction with Nucleic Acids
The association constant of Fe-BLM with duplex D N A i s on
the order of lo5
BLM-mediated D N A cleavage is
characterized by its sequence selectivity: most strand breaks
occur at 5'-G-pyr sequences. In BLM, a small difference in the
terminal amine chain and the absence of the gulose-mannose
sugar moiety have a negligible effect on the sequence specificity
of D N A cutting. Binding is mediated in part through electrostatic interactions, and probably also by intercalation or partial
intercalation of the bithiazole moiety. Metallo-BLM probably
binds in the minor groove of the B-DNA helix, where it covers
2 o r 3 bp's on DNA. The 2-amino group of the guanine base
adjacent to the 5' side of the cleaved pyrimidine base is one key
element of specific 5'-GC o r G T recognition by the metallobleomycins; the bithiazole group probably plays an important
role as an anchor on B-DNA.['031 Structurally altered D N A
such as methylated DNA['041o r platinated DNA oligomers['051
may be relevant to changes in specificity or reactivity.
BLM produces double-strand breaks in DNA with a frequency far in excess of that expected from coincidence of singlestrand breaks. The cleavage site on one strand is a pyrimidine in
a 5'-G-pyr sequence, but in the complementary strand, the
cleavage occurs at a site where single-strand cleavage is not
common.11061 Both breaks result from specific attack on
c4,['0
71
Attempts to modify the specificity of BLM have been undertaken by coupling BLM to oligonucleotides. The conjugates
efficiently cleave the D N A at the GT sequence near the oligonucleotide binding site."
Some studies have concluded that the cleavage of RNA is
unlikely to be important for therapeutic action of BLMFe,["'l but other findings suggested the opposite.[",
In
spite of recent reports that R N A was not cleaved by BLM,"
the three major classes of RNA have been shown to be degraded
by BLM-Fe.['". "*I R N A cleavage appears to be much more
selective than D N A cleavage. While 5'-G-pyr-3' sequences are
also represented to a large extent among observed R N A cleavage sites, the cleavage seems to occur predominantly in singlestranded regions and at the junctions between single- and
double-stranded regions. This suggests recognition of R N A tertiary structures, rather than recognition of specific sequences.
BLM has so Far only been involved in C 4 deoxyribose oxidation (see Section 3.4.3 for details).
2.4. Fe-EDTA and Fe-MPE
Methidiumpropyl-EDTA (18, MPE, Scheme 9) is an early example of DNA-binding molecules able to target the redox activity of a metal ion like iron and to
generate D N A breaks. It contains
H,N
the D N A intercalator methidium,
covalently bound by a short hydrocarbon tether to the metal
chelator EDTA.r"3i
2.4.1. Mechanism of Potentiation
Both MPE and EDTA in the
0
presence of iron(I1) and molecular
Scheme9. Structure of the
Oxygen are able to break the
ironfii) complex of methidiumbackbone of DNA, probably
propyl EDTA (18).
by H abstraction from deoxyribose.[' 1 3 - 'I5, 1161 Fe-MPE cleaves plasmid D N A at a concentration two orders ofmagnitude lower
M) than Fe-EDTA
M ) . Addition of reducing agents such as dithiothreitol,
sodium ascorbate, or NADH, which regenerate Fe" from Fe"',
enables D N A cleavage at concentrations comparable to those
used with bleomycin (lo-' M ) . ~ " ~ ]
Presumably in this mechanism, Fell reacts with 0, to give Fe"'
and 0;- [Eq. ( 5 ) ] .The superoxide anion in protic media rapidly
disproportionates to produce H,O, [Eq. (6)]. Fe" and H 2 0 , can
then produce hydroxyl radicals by the known Fenton-type
Fell-MPE
20;-
+ 0,
+ 2Ht
--
d
Fe"'-MPE
H,O,
+ 0,
+ 0;-
(5)
(6)
753
REVIEWS
G. Pratviel et al.
chemistry [Eq. (7)].['L31When sodium ascorbate is used as reducing agent in the presence of dioxygen, a catalytic cycle is
created, requiring lower concentrations of iron for DNA cleavage [Eqs. (5)-(8)].["s,."71
-
Fe"-MPE
+ H,O,
Fe"'-MPE
+ reductant
Fell'-MPE
--*
+ HO' + O H -
(7)
Fe"-MPE
(8)
Superoxide dismutase, which depletes the system of free superoxide, and catalase, which disproportionates hydrogen peroxide into water and oxygen, are both able to inhibit the DNA
cleavage mediated by Fe -MPE. This ability indicates the importance of free hydrogen peroxide as key intermediate in
strand scission. Furthermore, in the absence of dioxygen no
strand scission can be observed.[98.' I 3 , ' "1 Because the concentration of the hydroxyl radical species decreases with distance
from its point of formation, the decrease in the level of cleavage
with distance from the position of the EDTA group is also
consistent with a mechanism involving HO'.["81
Fe"'-MPE was also examined for DNA-cleaving ability. Its
direct oxidation with iodosylbenzene enhances cleavage of plasmid DNA. Whether this involves a perferryl species MPEFeV=O or a coordinated hydroxyl species is still a matter of
discussion,[9b,98. 1131
sequences and binding-site sizes.[98, 301 Accordingly, hybrid
molecules have been synthesized by associating EDTA with distamycin,["* 98, ' 3 1 1 a chirdl derivative of n e t r o p ~ i n , [ ' I3'l
~~.
oligopeptides,[' 19, 1331 or oligonucleotides, in order to cleave
single-stranded DNA (or RNA) under mild conditions["8. L 3 4 1
or double-stranded DNA (by triple helix formation) . [ 1 3 s 1
The chemistry of Fe-EDTA complexes with DNA will be
detailed in Section 3.4.4.
2.5. Cationic Metalloporphyrins
2.5.1. Mechanism of Potentiation
Synthetic metalloporphyrins are able to mimic oxygenation
and oxidation reactions mediated by heme
Beside
this mode of reactivity, iron or manganese porphyrins with
peripheral positive charges, for example, Fe- or Mn-mesotetra(4-N-methylpyridiniumyl)porphyrins 19 (Fe-TMPyP,
Mn-TMPyP, Scheme 10; for crystallographic data on MnTMPyP see ref. [136]), exhibit strong interaction with DNA
-
1
HC
2.4.2. Znteraction with Nucleie Acids
Fe-EDTA is a negatively charged complex which does not
associate electrostatically with DNA, so hydroxyl radicals
cleave DNA with almost no sequence dependence.[98.
lZo1
This system is a useful reagent for footprints of proteins[l17. l Z 0 , 1 2 1 ] or of small molecules (distamycin and actinomycin['221), and can also provide information about DNA
structures (unusual conformations adopted by the adenine
tracts in DNA," z31 four-arm Holliday structure^,^'^^^ and bent
DNA" l61).
Fe-MPE cleaves DNA in a relatively poor sequence-specific
manner, which is not unexpected since the methidium unit is an
intercalator of low overall base specificity. The only specificity
known for this system is a slight preference for GC rich regions.
Sequence specificity is significantly lower than that of DNase I,
and the wide use of this reagent has proved to be very useful for
many applications : for instance, the determination of the location, size, and relative importance of the binding sites of small
molecules on native D N A [ I ~ O . ''1 or proteins on restriction
fragments,['21'
and study of RNA['271and chromatin['z81
structures. For a discussion on the relative advantages of FeEDTA and Fe-MPE in footprinting experiments, see ref. [I 291.
The DNA-binding affinities of Ni" and Mg" complexes of MPE
are similar to that of ethidium bromide (about 10' M - ' ) . The
affinities of Fe"-MPE or Fe"'-MPE cannot be determined,
due to their efficiency as DNA cleavers. The binding site is
probably a 2-bp site.[981
The attachment of Fe-EDTA to sequence-specific DNA
binding molecules such as antibiotics, polypeptides, oligonucleotides, or proteins provide new classes of "DNA affinity
cleaving molecules" and is the basis for the design and construction of artificial restriction endonucleases with defined target
19 5 AcOScheme 10. Structure of M-TMPyP pentaacetate ( M
=
Fe"' or Mn"') [I361
'"9
754
that brings a powerful oxidizing species into the vicinity of the
target. The active oxidative species in the case of DNA cleavage
is probably the same as that involved in catalytic oxygenation
and oxidation reactions described for metalloporphyrins in general, namely, an oxoporphyrinato complex with the metal in a
high oxidation state and able to hydroxylate a C-H bond or to
epoxidize an ~ l e f i n . [ ~This
~ " ] active species is generated in the
presence of oxygen atom donor compounds like iodosylbenzene," 3 7 1 hydrogen peroxide,[' 38a1 potassium monopersulfate,[138a-b,
1 3 8 d l or magnesium monoperphthalate.[' 38c1 Another way to generate 0x0-metalloporphyrins is to use molecular oxygen associated with an electron source, as does
cytochrome P-450.[89"31 3 7 , 1391 The most efficient method of oxidative cleavage of DNA is the oxidation of cationic Mn-porphyrins by KHSO, . [ 1 3 8 , 1 3 0 1 Data on DNA cleavage by Mn"'TMPyP/KHSO, suggest that diffusible radical species are not
involved in the reaction, because the breaks are well-defined and
H 2 0 2is at least three orders of magnitude less efficient in generating the active species leading to cleavage.[' 38a1 The high-valent
TMPyP-MnV=O is a monoreacting DNA cleaver (unlike bireactive biradicals of enediyne drugs) .I138aJ Catalytic activity
can be observed only if it is protected from self-oxidation (leading to the chromophore bleaching) by strong interaction with its
target. For example when the DNA/Mn-porphyrin ratio was
high (75 p~ bpi5 nM Mn-porphyrin), up to five SSBs per MnTMPyP molecule were observed.[' 38a1 A proposed activation/
reaction cycle is presented in Scheme 8. The high-valent
Aiigew. C/win. In{. Ed. Engl. 1995, 34, 746- 769
Nonenzymatic D N A Cleavage
TMPyP-Mn"=O species that cleaves D N A is too reactive to be
characterized. As discusscd in Section 2.3.1 for high-valent
BLM-Fev=O, neither P-450 (route a in Scheme 8) nor "diverted P-450" (route b) reactivity has been established for this compound. Whether the activated intermediate TMPyP-M'" -OH
oxygenates the D N A substrate directly by the rebound of the
"HO equivalent" (route a) or by an electron transfer (route b) is
still unknown. Studies of the source of 0 atom incorporated in
DNA sugars would help answer this question (see Section 2.3.1).[991
Electrochemically activated Mn"'-TMPyP
and Fe"'TMPyP have been used for D N A cleavage in the presence of
0, .[l4'I The mechanism of activation and cleavage in this case
is potcntially relevant to reports on chemical activation. Photochemically activated porphyrins are also able to cleave
DNA,['421probably by singlet oxygen production.
n
,'..
5'
'.
'..
+
3"\\,!i
,'I
d(5'dd-G C
G ) ----> d(pC G )
+ I - z a +FUR
~
2.5.2. Interaction with Nucbic Acids
Mn-TMPyP belongs to a class of cationic metalloporphyrins
that do not intercalate between DNA base-pairs but
instead bind in the minor groove of AT-rich regions of
DNA.['4", 14'. 1431 The Mn-porphyrin framework is devoid of
any H-bonding donor/acceptor capacity and intercalation is
precluded by the presence of an axial ligand on manganese.[136]
Thus the ability of Mn-TMPyP to select AT-rich regions of
D N A is apparently electrostatic or steric in origin. It has been
proposed that this cationic metalloporphyrin is attracted by the
high negative potential at the surface of the minor groove['441of
A . T-rich sequences. A close contact with the minor groove is
necessary to improve the binding interaction and the cleaving
efficiency of the metalloporphyrin, as deduced from studies on
variations in total charge and in charge distribution at the
macrocycle periphery.['4""-', 1451 At high ionic strength, the
reagent is unable to bind or cleave DNA.[138",1461
Considering the size of Mn-TMPyP, it could span over five
to six base-pairs in the minor groove of B-DNA, but the preferred cleaving site consists of three consecutive AT base-pairs
creating a suitable "box" for highly selective D N A cleavage.'L37.1 3 8 d , e , 1 4 3 a * d 1 At this site Mn-TMPyP strictly mediates
C5' oxidation on nucleosides on both 3' sides of the AT box:['471
one single-strand break on each strand leads to double-strand
cleavage with a four base-pair shift to the 3' end of the opposite
D N A strand (Scheme 11). So far, there is no definite proof that
such a cleavage pattern results from two SSBs by the same
metalloporphyrin activated twice in the minor groove, or that it
is effected by two different activated metalloporphyrins. On
large D N A substrates, SSBs prevail[' 3xa1 and on oligonucleotides substrates (with one or two "AT boxes") DSBs are easily
obtained.['471 Besides cleavage at ( A . T), cleaving sites, some
secondary reaction sequences are also noted, in which any one
of the A . T pairs is replaced by a G . C b a ~ e - p a i r . [ ' ~This
~"]
secondary reactivity is particularly noticeable under drastic
cleaving conditions[1471and is one order of magnitude less than
that for (A . T)3 sites.
Mn-TMPyP is also able to degrade G . C-rich sequences
with a reactivity two orders of magnitude weaker than the
degradation of (A . T), sites. The mechanism of D N A lesions in
these regions of D N A is not fully understood and may involve
I,...........>'
Schcme 11. On using the Mn-TMPyP:KHSO, system the scission on the tn'o
strands of duplex V;VI shows a 4 bp shift (dashed doublc-headed arrow, bottom
left) in direction of the 3'-eiids. This is shown by analysis [I471 of the cleavage
products (bottom right).
sugar oxidation (C5' or CI') and/or base oxidations. Affinity
constants of up to I O 7 ~ - ' have been published for Mn
TMPyP at the (A . T), binding site.['38e1Other values (lo5 M - '
for A . T sequences and lo3 M - ' for G . C sequences) have also
been reported.['4x1
Hybrid molecules based on a metallotris(methy1pyridinioy1)porphyrin moiety attached to oligonucleotides have been prepared, and were shown to selectively cleave a single-stranded
35-mer D N A corresponding to the initiation codon region of
the tat gene of HIV-l.[1491
The mechanisms of D N A cleavage by Mn-TMPyP through
C1' and/or C5' oxidation are described in Sections 3.1.3 and
3.5.2, respectively.
2.6. Photoactivated DNA Cleaving Agents: The Case of
Phenanthrenequinonediiminerhodium(n1) Complexes
A number of studies have focused on the use of metal complexes or drugs to design spectroscopic D N A probes or photoactivated D N A cleaving agent^.['^^"-“^ Photoactivation promotes strand cleavage only for some of these
As
well as metalloporphyrins or enediynes (Sections 2.5.1 and 2.2),
a range of simple inorganic octahedral molecules are also capable of inducing single-strand scissions by photoexcitation of
different absorption bands. This cleavage is induced by metal
complexes known to bind to D N A principally in two noncovalent modes: 1 ) a groove-binding mode stabilized by hydrophobic interactions of the ligands with the major or the minor
groove of DNA[150e~1s11
and 2) inter~alation.['~']For some of
these photoactivatable complexes, the exact mode of D N A
binding is still a matter of debate.['531Rhodium,""1 rutheniand cobalt['56]complexes have been used as cleaving
agents for probing nucleic acid conforination and for the recognition of specific sites on a D N A strand.
Recently, a precise mechanism involving oxidative attack on
the C3'-H bond of deoxyribose has been deduced for phenan155
G. Pratviel et al.
REVIEWS
threnequinonediiminerhodium(Il1) complexes and will be discussed in detail in this review.
2.6.1. Mechanism of Potentiation
The production of sharp and well-defined gel bands after the
cleavage of restriction fragments or oligonucleotides suggest
that DNA strand scission by photoactivated rhodium complexes is not mediated by a diffusible oxidant. The cleavage appears
to involve a species that is generated precisely at the binding site
and induces strand scission of DNA at that site. Irradiation of
the rhodium complex at the wavelength of the ligand-to-metal
charge transfer band (LMCT) leads to an excited state with a
lifetime in the ps range. The active radical species from 20 and
21 (Scheme 12) that cleave DNA is likely to be the radical cation
of the phi ligand that is intimately bound to DNA and is formed
by the LMCT band irradiation. The position of the phi radical
cation in the site enables a fast H abstraction from the deoxyribose ring, initiating the cleaving reaction." 'I
3. Hydrogen Abstraction from Different Positions
of Deoxyribose, Leading to DNA
Each deoxyribose unit of DNA contains seven C-H bonds as
possible targets for oxidative activation; four of them are located in the minor groove and three in the major groove (see
Scheme 1). For all the chemical nucleases and drugs presented
(in Section 2) a common initial step consists of one H abstraction, leaving the relevant carbon on the sugar as a radical species. We present here the manner in which this initial lesion leads
to DNA cleavage under the different experimental conditions
used in accord with the experimental evidence for the proposed
mechanisms.
3.1. H1' Abstraction: Cu(oP), , NCS-Chrom, Mn-TMPyP,
and Halouracil-Containing Nucleotides
H1' abstraction has been described for oxidative attack on
DNA by Cu(oP), (I), NCS-chrom (2), and Mn-TMPyP (19,
M = Mn), producing a C1'-centered radical 28 (see Scheme 14).
Photoreaction of 5-halouracil-containing oligonucleotides gives
rise to C1'- or CZ-centered radicals (23 or 24 in Scheme 13,
0
'
23 Ade
24
Rh"'(phet~)~phi2 0
Rh'1'(phi)2bpy 2 1
Scheme 12. Structures of the (9,10-phenanthrenequinonediimine)rhodium(111)
complexes 20 and 21.
2.6.2. Interaction with Nucleic Acids
Upon photoactivation, the diphenanthroline complex 20
cleaves double-stranded DNA at 5'-pyrimidine-purine steps
and in particular at the consensus sequence S-pyrimidinepurine, while 21, despite containing ancillary imines for potential hydrogen bonding, cleaves DNA efficiently but in a generally sequence-neutral fashion. The asymmetry in the cleavage pattern to the 5' side on each strand suggests that the rigid complex
attacks B-DNA from the major groove of the helix. Given the
propensity of the phi ligand for intercalation, 20 should contain
the phenanthroline ligands in ancillary, nonintercalative positions along the groove, while for 21 one phi and one bpy should
occupy the ancillary sites." 54a1 NMR studies have recently provided direct structural evidence for specific intercalation of the
octahedral chiral complex 4-20, and indicate a preferential partial intercalation of the phi ligand whereas the phenanthroline
ligands bind in ancillary positions.[' 581 Intercalation through
the major groove is consistent with results of mechanistic studies
o n DNA cleavage by rhodium complexes, which demonstrate
that H3' abstraction is the initial step (see Scheme 16 and Settion 3.3). Photocleavage experiments indicated enantioselective
targeting of the 5'-CG step by 4-20,
756
Scheme 13. Proposed reaction mechanisms for oxidative attack at C I ' - H and C2'H bonds in the case of halouracil-containing oligonucleotides [159].
respectively) via 22 or, alternatively, via 25 and 26 to the 1'
cation 27 (Scheme 13; see also 29 in Scheme 14). Usually, the
next key steps of the DNA cleavage reaction (Scheme 14) are 1 )
the release of base associated with the formation of the oxidized
abasic derivative 31, 2) the first p-elimination inducing the
~--%
35
Scheme 14. Proposed IedCtlOn mechanisms when the initial site of attack by the
oxidizing agent IS at C1' of the deoxyribose [159].
1,
2,
19,
vicinai 'U. @ : vicinal "'U)
(0:
0:0: 0:
.4ngew. Chem. h t . Ed. Engl. 1995, 34* 746-769
Nonenzymatic D N A Cleavage
strand break that leaves a 5’-phosphate end and a metastable
a,P-unsaturated lactone (32) at the 3’ end, 3) the second and fast
p-elimination allowing the recovery of the 3’-phosphorylatedending DNA fragments and 5-methylene furanone (33) as the
main sugar residue. Depending on the nature of the D N A cleaving agent, different aspects of this general cleavage pathway are
more pronounced than others. Cl’ chemistry for Fe”-EDTA,
Fe” MPE, and related compounds is considered to account for
some experimental results; this point is briefly presented in Section 3.4.4.
3. I . I . Cu(oP)
Analysis of the products of D N A cleavage by Cu(oP), (1) and
hydrogen peroxide showed that phosphomonoesters are present
at both the 5’ and 3’ termini (3’-phosphomonoester termini are
the potent inhibitors of E. coli D N A polymerase activity, which
provided the first evidence for the D N A scission reaction“ 5 a . 1 5 4) and that bases are also released; the cleavage was
proposed to occur by initial oxidation of the deoxyribose to a
resonance-stabilized furan derivative.[29”]Later, under conditions where no reductant was used as coreactant, 5-methylene-2furanone (33) was identified.” 60] A possible reaction pathway
that could accommodate the formation of these products from
C1’ radical 28 is summarized in Scheme 14. The putative intermediate 32 at the 3’ terminus was observed by polyacrylamide
gel electrophoresis (PAGE) during the cleavage reaction of a
synthetic dodecamer and a restriction fragment containing the
lactose control region.[281No other information exists about
intermediate steps except that the cleaving activity does not
require the presence of molecular oxygen.[I7]
The C1’ oxidation pathway accounts for 80-90 % of the scission events induced by 1 (1 0-20 YOare related to cleavage chemistry at C4’, see Section 3.4.1).[’5bJThe relative importance of
the two pathways, suggestive of a different rate of attack at two
locations, depends on the D N A sequence,[’*] as demonstrated
by determination of the ratio of 3’-phosphoglycolate (attack at
C 4 ) to 3’-phosphate (attack at Cl’).
grated slightly faster than the full-length D N A and were transformed to the corresponding phosphate-ending fragments upon
alkali treatment. D N A fragments carrying the %$-unsaturated
lactone 32 at their 3’ end (one p-elimination from 31) were also
seen on polyacrylamide gels of 5’-[32P]-labeled oligonucleotides;[66a, they migrated ahead of the corresponding 3’-phosphate-ending fragments. After enzymatic digestion of intermediate 31, 2-deoxyribonolactone generated by the release of cyBut 5-MF (33), the
tosine was isolated and
expected end-product of this reaction chain, was never observed
because of the presence of reducing cofactors. When NaBH,
was used as an activator for 2, alkali-dependent breakage at the
C residue was not prevented, which indicated that NaBH, is not
able to reduce the oxidized intermediate 31.[l6I1In addition, by
careful examination of 5’-[32P]-labeled D N A fragments through
PAGE analyses, new bands could be observed with smaller mobility than that of the corresponding phosphate bands. These
bands, which are converted into the 3’-phosphate-ending bands
by piperidine treatment, were attributed to the adduct (Michael
addition) between thiol and the 3’-terminal, rJ3-unsaturated
lactone 32.[66bs
1 6 3 1 The mobility of this thiol adduct varies with
the structure of the thio1.166b1
Such bands were also observed in
the case of C 4 chemistry of 2 (see Section 3.4.2).
Under anaerobic conditions misonidazole can replace
dioxygen for the production of the alkali-labile lesion at
the C residue.[’611 The mechanism is not fully understood: the question whether an oxy radical (35) or a peroxy
radical (34) is involved as intermediate is still open (see
Section 3.4.2).
Direct evidence for HI‘ abstraction was also provided by
experiments where HI’ was replaced by deuterium. Values of
k,/k, zz 4 for the deuterium isotope effect were determined
from studies of alkali-induced cleavage of oxidized abasic
derivatives (31).[’ll Deuterium abstraction studies of the reacted chromophore (monitored by ’H NMR spectroscopy) also
indicate that the C2 of 2 is involved in the attack at the C residue
of a 5‘-AGC sequence.[54a1
3.1.3. Mn- TMPyP
3.1.2. NCS-Chrom
C1’ chemistry occurs particularly at the C residue of doublestrand lesion sites such as 5’-AGC/3‘-TCG sequences,Igdr’12 1611
where it represents the major pathway of attack (72%), although some minor chemistry at C4’ or C5’ was also noted.(6ha.711
It is always accompanied by a strand break on the
complementary 3’-TCG strand, where C5‘ of the T residue is
attacked. So, double-strand cleavage consists of a two-nucleotide 3’-stagger of the cleaved residues that represents the classic
3’-shift mediated by the NCS-chrom activated biradical 4
derived from 2 (see also Section 3.4.2). One important feature of
this cleavage is the strong preference for glutathione (over neutral thiols, M E or DTT) in the C1’ chemistry induced by 2 (see
Section 2.2.2.1).
Since the nucleolytic activity of 2 is dependent on the presence
of molecular oxygen. a peroxy radical (34, Scheme 14) is probably involved initially with subsequent conversion into the oxidized intermediate 31 with concomitant release of base. Such
mioligonucleotides (31), observed on polyacrylamide
Angru.. C’heni. In/. G I . Lngl. 1995. 34. 746-769
The following stable reaction products were identified from
several D N A and oligonucleotides cleavage reactions with MnTMPyP/KHSO, : 5’- and 3’-phosphorylated termini, free bases,
The mechanism proposed to
and 5-methylene furanone 33.[1641
explain the formation of the product 5-methylene furanone is
hydroxylation of C1’ as indicated in Scheme 14. The cleavage
does not require the presence of molecular oxygen (M. PitiC
et al., unpublished results). Two hypotheses (Schemes 8 and 14)
may explain the evolution after the initial lesion: 1) the radical
centered at C1’ (28) might be reoxidized by the Mn’”(0H)porphyrin by one-electron abstraction, giving the carbocation
29 that reacts with water and produces the 1’-hydroxylated
derivative 30 (“diverted P-450 route” in Scheme 8, route b) or 2)
the radical is directly hydroxylated by the Mn’”(0H)porphyrin
(cytochrome P-450 chemistry, see route a in Scheme 8 and Section 2.5.1). So far, neither the oxidized oligonucleotides 31 containing the abasic site nor oligonucleotides 32 bearing the cc$unsaturated lactone have been observed except indirectly
through their ultimate reaction product 33.
757
REVIEWS
G. Pratviel et al.
With Mn-TMPyP/KHSO, as cleaving reagent, C1' is the
main oxidation target for single-stranded or GC-rich doublestranded DNA, but only a minor target for AT-rich sequences
containing DNA (C5' becomes then the main target, see Section 3.5.2).[164b3
3.1.4. ~ a l o u ~ a c i l - c o n t a ~ nOligonucleotides
~ng
In an attractive hypothesis, formation of the 1' cation 27
could occur in the UV-photoreaction of small oligonucleotides
containing the 5'-ABrUsequence (see Scheme 13). After an intramolecular electron transfer from adenine at the 5' side to the
adjacent BrU,the resulting BrUanion radical 25 would release a
Br- anion to produce a uracilyl radical 26, which could then
abstract the adjacent Cl' hydrogen atom of the deoxyadenosine
radical cation to give rise to the 1' cation 27. Addition of another
water molecule, loss of a proton, and release of the base
(Scheme 14) would give the deoxyribonolactone derivative 31
(oxidized abasic site) which has been detected by HPLC and
transformed to the corresponding phosphate-ending fragments
upon treatment with hot NaOH. Without treatment with alkali
and after enzymatic digestion of the oligonucleotide 31, 2deoxyribonolactone (with concomitant release of Ade) was isolated and characterized by spectroscopic and chemical metho d ~ . [ ' The
~ ~ ]expected product 5-MF (33) coming from reaction
of the oxidized abasic site of 31 was not observed, probably
because of its instability under the experimental conditions (hot
alkali treatment).
Photoreaction of 5-iodouracil-containing oligonucleotides" 661 proceeds by an entirely different mechanism, which
postulates competition between C1' and C2' oxidation products
of the sugar residue at the 5' side adjacent to the iodouracil
residue. The sequence of events is indicated in Schemes 13 and
15. Photoreaction of 'U proceeds by homolytic cleavage of the
C5-I bond of uracil; the resulting uracilyl-5-yl radical (22, see
Scheme 13) competitively abstracts the C2' (+ 23) or C1' (--t 24)
hydrogen atoms of the deoxyribose at the 5' side of 'U. The C1'
and C2' oxidation products 31 and 42 were observed
(Schemes 14 and 15, respectively).
3.2. H2' Abstraction: Iodouracil-Containing Nucleotides
While the chemistry of DNA damage resulting from H abstraction at CI', C3', C4', and C5' of the DNA deoxyribose is
well documented, the DNA lesion resulting from H abstraction
at C2' has been scarcely studied. Characterization of erythrosecontaining sites arising from H' abstraction at C2' was only
reported as one of the major types of damage induced by y-radiolysis of
Very recently another example of deoxyribose C2' oxidation in DNA was described during photoreaction of various 5-iodouracil containing oligonucleotides,['661
C2' oxidation products were isolated directly for the first time.
The proposed mechanism (Scheme 15) involves the peroxy radical 36, which could be converted by homolytic cleavage of the
0-0 bond (via 37, pathway a) into the oxy radical species 38.
p-Scission leads to 39. The formation of the oxy radical from
peroxy radical in double-stranded lesions of DNA by fragmentation of an 0, bridge across the minor groove has been invoked
758
OH
I
rCHOl
Scheme 15. Proposed reaction mechanisms when initial site of attack by the oxidizing agent is a t C2' of the deoxyribose [1S9].
for enediyne compounds (Sections 3.4.2, 3.5.1). But in the
present case, combination of two C2' peroxy radicals of A
residues of 5'-A1U/3'-A'U does not seem sterically possible. Further reaction of 39 with molecular oxygen produces 42, accompanied by the release of adenine. Alternatively, heterolytic
cleavage of the 0-0 bond, perhaps favored by the formation of
an ROO1 (pathway b), would also produce 42 via intermediates
40 and 41. Under alkaline conditions the unstable compound 42
is cleaved and gives rise to fragments containing phosphoglycolaldehyde termini (43 and 44) by a retroaldol reaction. Compounds 42, 43, and 44 have been characterized, but the exact
route of their formation remains to be elucidated.
3.3. H3' Abstraction: Rhodium Complexes 20 and 21
Extensive studies have focused on DNA photocleavage by
several photoactivated DNA cleaving agents. Among them, the
phenanthrene-quinonediimine complexes of rhodium(1Ii) (20
and 21) have received particular attention in terms of molecular
mechanisms of DNA cleavage (Scheme 16).[154".]'51 While the
complexes have similar photochemical properties, overall binding modes, and affinities, their cleavage patterns indicate different recognition characteristics (see Section 2.6.1.2), which may
be understood based on the different molecular shapes of these
two complexes. The shape of these compounds also governs the
chemistry of strand scission. Because both complexes yield
oligomers containing 3'- and 5'-phosphate termini and nucleic
acid bases as primary products the sugar units seem to be reaction sites.[' 54a1 For 21 these oligomeric products account for
approximatively 70 YO of the reaction, but, in addition, base-
R03PO-pC
4s
49
so
51
52
Scheme 16. Proposed reaction mechanisms when initial site of attack by the oxidiz21,
20) is a t C3' of the deoxyribose [1S9].
ing agents
(0:0:
Angrw. Chern. Int. Ed. Engl. 1995, 31,146-169
REVIEWS
Nonenzymatic D N A Cleavage
substituted propenoic acids (54, so-called base propenoic acids)
and oligomers 53 with a 3’-phosphoglycolaldehyde terminus are
obtained (the remaining 30% of reaction).[’”] The products
obtained are consistent with photoreaction of rhodium complexes intercalated in the major groove of DNA, where the
bound complex has a direct access to H3’. So the first step
proposed is H3’ abstraction. This fact has been confirmed by
using specifically tritiated DNA.“ 5 7 1 In an oxygen-independent
pathway, C3’ radical 45 is converted into the 3’-hydroxylated
compound 46, probably by one electron abstraction by the
rhodium complex (the excited state of this complex might be
sufficiently oxidizing) and subsequent attack of a water molecule on the 3’ carbocation. Elimination of the 5’-phosphorylated
end provides 47 (with ketone end group), whieh gives rise, by
two successive /I-eliminations, to the 3’-phosphorylated end,
free base, and 2-methylene-3-furanone (48). In an oxygen-dependent pathway, the 3’ radical 45 gives, successively, peroxy
radical (49) and a hydroperoxide derivative (50), which provides
in several steps 51 and 52 (Criegee type rearrangement), the
compound with 5’-phosphorylated terminus, a phosphoglycolaldehyde (53) and a base propenoic acid (54). So far, the search
for any metastable intermediates, as have been seen for other
carbon targets of deoxyribose, was unsuccessful.
C3’ chemistry might be invoked in DNA cleavage by Fe”EDTA and some related compounds to explain some experimental results; this point will be briefly presented in Section 3.4.4.
ing between these two main pathways depends on the 0, concentration and the nature of the cleaving D N A molecules and
their activating cofactors.
3.4.1. Cu(oP)
Electrophoretic analyses of the scission of 5‘- and 3’-[”P]labeled oligomers and large DNA fragments give important
insights into the chemistry of the D N A strand scission by
Cu(oP), (1). The 3’- and 5’-phosphomonoesters are the main
termini generated by this complex and are generated by C1’
oxidation chemistry (see Section 3.1 .I and Scheme 14). In addition, the 3’-phosphoglycolate prodncts 65 are indicative of C 4
oxidative chemistry (Scheme 17) 5 h ,
Whereas “Co irradiation (another way for H1’ or H4’ abstraction) generates equivalent amounts of 3‘-phosphomonoesters and 3‘-phosphoglycolates,[’*. 16’] the amount of 3’-phosphoglycolate produced in
cleavage mediated by 1 is substantially less than that of the
3’-phosphomonoester; C4’ chemistry is thus only a minor oxidation pathway. So far, no base propeiial 64 or other fragments
from sugars with three or five carbon atoms could be isolated,
so the exact pathway from the C4’ radical 55 to the phosphoglycolate 65 is still unknown, and the mechanism remains to be
proven (see also Sections 3.4.2 and 3.4.4).
3.4.2. Enediynes
3.4.2.1. NCS-CIirom
H4’ abstraction from deoxyriboses was mainly detected at T
residues of 5’-GTsteps on synthetic ~ligonucleotides.[”~, I I‘’
Further studies revealed that this minor mechanism of D N A
oxidation predominantly occurs in staggered double-strand lesions at the Tresidue of 5’-AGT/3’-TCA s e q u e n c e ~ . [I~7 l~1 ”But
~
H4’ abstraction has been proposed as the common first step
NCS-chroni (2) may shuttle between H 4 and H5’ abstraction at
of the oxidative attack at C 4 by NCS-chrom (2), CAL (5),
the same T residue, depending on the t h i ~ l , ’”‘I~ ~ ,deuteriaBLM (17). Cu(oP), (I), and Fe-MPE (Fe-MPE) (Scheme 17).
t i ~ n , [I 7~O 1 ~ or
, the sequence of DNA.[66b1On the opposite
The resulting radical intermediate 55 can either be hydroxylated
at C4’ (57) to yield, after cleavage, a 5’-phosphorylated terminus
strand at 5’-AGT/3‘-TCA sites, the cleavage of T i s due mainly
to HS’ abstraction, but a shift to 4 chemistry[66b1cannot be
and a 3’-terminus bearing the sugar residue (59) or add an oxygen molecule to form a peroxy radical (60), which by different
excluded.
The products from C4‘ chemistry of 2 (Scheme 17) are the
pathways can give rise to base propenal 64, 3’-phosphoglycolate
4’-hydroxylated abasic derivative 58 (alkali-labile site) and the
65. and compounds with 5’-phosphorylated termini. Partition3’-phosphoglycolate 65 (direct D N A break); both are
derived from the common 4’-carbon-centered radical
55. Unlike BLM complexes (see Section 3.4.3)[’“] the
base propenals 64[9d.66h1 or malonaldehyde-like compounds containing C1‘ and C2’ of deoxyribose[”21
have been mentioned but have not yet been clearly
described. Moreover, the fate of radical 55 does not
depend on 0, concentration (0, is involved in the
production of both types of lesions) but is dependent
on the nature of the 2-activating thiol and on the type
of lesion (SSB versus DSB). More phosphoglycolates
occur at double-strand lesions sites.[4h.6”b1 Acidic or
negatively charged thiols (that have low affinity for
DNA, such as GSH and MPA) afford the greatest
68
6 9 YO,R’
extent of phosphoglycolate production, but thiols
Scheme 17. Proposed maction mechanisms when initial hitc of attack by the o x i d i h g agent
with high reducing efficiency or ability to donate H
(0
: 1. 0 ,
enediynes, @ : M-17, @ Fe-18) is a t C4’ of thedeoxyribose. Final product 64 has
(which depends on the stability of the generated sulnever been obscrved for cleavnges by 1 and Fe-18 and could not be characterized for cleavages
nith mediynec [I59]. The C4-hydroxylated 58 is represented i n its open-chain form
fdnyl(thiy1) radical) did not induce 3’-phosphoglyco-
3.4. H4‘ Abstraction: Cu(oP), , Enediynes, BLM,
and Fe - EDTA/MPE
’
A f z , q v + .Chrm. Iiri. Ed. Eiig/. 1995, 34, 146 I69
759
G. Pratviel et al.
REVIEWS
late production. So, the partitioning between the two routes
involves a reduction-sensitive D N A intermediate. When this
intermediate is reduced, the 4’-hydroxylated abasic 58 is
formed; otherwise phosphoglycolates 65 are produced. The intermediate could be the peroxy radical 60 or the oxy radical 68.
If the reduction-sensitive intermediate is the C4-peroxy radical 60, an easy reduction to the hydroperoxide 61 (by the reductant itself or by intramolecular H’ donation[641)would give the
phosphoglycolate precursor (Criegee-type rearrangement) usually proposed.[5’, 1731 But, in the presence of a strongly reducing
thiol, this peroxy radical intermediate may be preferentially converted into the 4-hydroxylated precursor 57 of the 4’-hydroxylated abasic derivative 58.
A second hypothesis, favored by Goldberg and co-workers, is
that the reduction-sensitive intermediate is the C4‘-oxy radical
68. If this radical is reduced to the C4-hydroxylated chmponnd
57, it gives rise to the 4’-hydroxylated abasic 58. If not, it
should rearrange to the p h o s p h ~ g l y c o l a t e [by
~ ~the
~ ~classical
B-scission of the oxygen radical mechanism (via 69).[1741
The
same mechanism is also invoked in Sections 3.2 and 3.5.1.1 to
explain experimental findings. Three features make the oxy
radical the better candidate for phosphoglycolate production:
1) the predominance of phosphoglycolates at DSB sites could
be explained by the formation and subsequent Russel fragmentation of a putative 0, bridge across the minor groove
[Eq. (9)];[46.64, 175,17s1 2) in the case of cleavage by 2 of DNA,
2R00’
, ROOOOR
+
2RO’
+ 0,
(9)
a Criegee-type rearrangement of the peroxide does not seem
feasible (absence of metal, nonacidic conditions),[9d,
3)
when misonidazole replaces dioxygen to trap the carbon-centered radical 55, the oxy radical 68 is formed directly, and the
percentage of phosphoglycolate 65 increases.[46b’64, 661
Experimental evidence for C4’ oxidation on D N A cleavage by
2 is provided by the following experimental evidence: 1) HPLC
monitoring of the oligonucleotide 58, that has the full length but
the 4’-hydroxylated abasic
The structure of this alkalilabile lesion was established by characterization of the pyridazine derivative that is formed by hydrazine treatment, and by
comparison of the HPLC retention time of such oligonucleotides after their reduction to diols (two diastereoisomers) with
that of a synthetic (R)isomer standard.[’6912) PAGE analysis of
5’-[32P]-labeled oligonucleotides ending with 3’-phosphate, 3’phosphoglycolates (65), and intermediate 59.[1701
This latter intermediate is not usually observed directly on polyacrylamide
gels because of its sensitivity to electrophoresis conditions.[1701
If the loading buffer pH is above 8 a second \&elimination can
occur (especially during the heat denaturation step before
sample loading), transforming the band of 59 into the corresponding 3‘-phosphate band. A nonambiguous quantification
of 4-alkali-labile oligonucleotides 58 can thus be done only after
adding hydrazine to convert the 58 into a stable 3’-(3-pyridazinylmethy1)phosphate that will migrate slower than the corresponding Maxam-Gilbert sequencing marker[46h3
641 (see
ref. [177] for details on this method). On the other hand, 3’[j2P]-labeled substrates typically afford fragments ending with
5’-phosphate groups only after treatment with alkali.[46b13)
Deuterium labeling at the 4 position of d e o x y r i b o s e ~ . [I 7~O 1~ >
760
The ratio of the kinetic isotope effect k,/k, was 4 for the glycolate formation and 3.7 for pyridazine formation (pathway via
the 4-hydroxylated abasic site) .[h41 When the kinetic isotope
effect was analyzed at T sites of various GT steps, k , / k , ranged
from 2.5 to 5.5.[17014) Gel chromatography. A minor band,
slowly moving on polyacrylamide gels, was attributed to an
adduct between the thiol used for the activation of 2 and the
3’-terminal unsaturated sugar 59 (Michael addition) .[h6h,
3.4.2.2. C‘AL
H4’ abstraction from D N A sugars by CAL (5) was observed
as part of a double-strand lesion at the frequent symmetric
cleavage site 5‘-GATCCT/3‘-CTAGGA.The oxidized bonds are
the C5’-H bo’nd of the C residue on the TCCT strand and the
s 5 * 6 7 h 3 Th e
C 4 - H bond of the Cresidue on the other
D N A cleavage products of the 5’-[”PI-labeled AGGA strand
consist of 3’-phosphoglycolates 65 (Scheme 17) and 4-hydroxylated abasic derivatives 58, which as pyridazine derivatives were
resolved by PAGE analysis. N o evidence for the production of
base propenal has been reported. Interestingly, the formation of
phosphoglycolate on one strand is always associated with C5’
chemistry on the complementary strand, supporting the involment of an 0, bridge for glycolate production (Russel fragmentation, see Section 3.4.2.1).[46h,”I Deuterium experiments followed by ‘H N M R spectroscopy showed
the deuterium
atom on C 4 of the C residue in 3’-CTAGGA strand was transfered to the C3 position of the reduced CAL biradical 6 (see
Scheme 4).
3.4.2.3. ESP
H 4 abstraction only occurs in double-strand lesions mediated
by cleavers of the ESP family 8-10, Phosphoglycolate 65, the
4-hydroxylated abasic derivative 59, and the 3‘-phosphate end
have been identified on sequencing gels from 5‘-[32P]-labeled
fragments of DNA.[s71In view of the structural similarity between ESP C (8) and CAL (S), the C7-centered radical of the
activated drug 11 is assumed to abstract the H4‘ atom (see
Scheme 4). As was mentioned for C5’ chemistry (Section 3.5.1.3). no isotope effect was observed in this C 4 chemistry.[”]
3.4.3. Blcomycin
Two sets of products of D N A strand breaks by metallobleomycins (for example, 17) have been identified, both of
which involve oxidative destruction of the deoxyribose moiety.
These products are now known in great detail: one set of products includes base propenals (64),[*6, 1 7 * ] 3’-phosphoglycolates
(65),11791
and 5’-phosphate termini; the second set consists of
free bases[1801and the evolution product from an alkali-labile
lesion (4’-hydroxylated abasic derivative 58) .I1 77ii1 The proposed mechanistic pathway is outlined as @ in Scheme 17.
As demonstrated by tritium and deuterium isotopic effects on
H 4 ab~traction,[’~.
18’] the formation of both bases and base
propenals is probably the consequence of a rate-determining
C 4 - H bond cleavage as the first step, followed by an oxygendependent partitioning of the intermediate 55. The release of
base requires no additional 0, and is accompanied by the proAngew. Chem. Int. Ed. Engngl. 1995, 34, 146-169
REVIEWS
Nonenzymatic DNA Cleavage
duction of an oxidatively damaged sugar 58, which results in
strand scission only in the presence of alkali (pH 12). The formation of base propenals 64 requires additional 0, to that required to form "activated BLM", and is accompanied by spontaneous DNA strand scission and 3'-phosphoglycolate termini
production. The magnitude of the 4 isotope effect is not sensitive to changes in the partition ratio of the two proposed D N A
cleavage pathways, but is sensitive (k,/k, = 2-7) to structural
and electronic changes in the metal-BLM complex, underscoring the importance of favorable parameters in the orientation of
the transition state.['821
Results of cleavage of D N A deuteriated at other positions of
the deoxyribose ring afford no isotope effect, suggesting high
regiospecificity for the BLM-mediated C-H bond cleavage.['821
Proposed mechanism ,for production of base propenals 64 and
3'-rcrnzind pho.spho~lycolute65: Under aerobic conditions the
4-radical55 rapidly combines with 0, to form a 4-peroxy radical 60 at what is probably a diffusion-controlled rate." 83a1 The
fate of peroxy radical has been extensively studied.r9b.1831It can
undergo a wide variety of reactions including formation of 0,containing compounds [Eq. (9)], H abstraction from reductants
such as a thiol [Eq. (lo)], and chain termination due to reduction by to low-valent transition metal ions [Eq. (1 I)].
ROO'
+ RSH
4
ROOH
+ RS'
(10)
Thus the peroxy radical gives rise principally to the 4-hydroperoxide 61, and in some cases to the 4-oxy radical 68.The
further conversion of 4-hydroperoxide 61 has been proposed to
occur by acid-catalyzed heterolytic cleavage of the peroxide 0 0 bond in a Criegee-type rearrangement. Instead of protoncatalysis, the Fe"'-BLM intermediate could act as a Lewis acid
catalyst. A molecule of water or the putative iron-bound hydroxide could trap the carbocation intermediate 62 to give 66
and ultimately the dicarbonyl derivative 67 as a precursor to 64
and 65.
Because of the stereospecific removal of the 2-(pro-R)proton
of deoxyribose and the fact that its release is kinetically more
favorable than the release of base propenal from
an
alternative pathway has been proposed: removal of the 2'-(proR) proton from the carbocation 62 and concomitant anti-elimination at 04 would generate intermediate 63. Further reaction
yields propenal 64, phosphoglycolate derivative 65, and 5'-phosphate termini. This mechanism was confirmed by "0 incorporation into the carbonyl group of 65 (in experiments carried out
with "0,) and into the aldehyde group of 64 (in experiments
carried out with H , ' s 0 ) . f ' 7 3 1
Proposed rnectrunisn?,for release of nucleohases: The C 4 radical 55 may undergo a radical rebound mechanism analogous to
that proposed by Groves and c o - w o r k e r ~for
[ ~ P-450
~ ~ ~ systems
(see also Scheme 8, pathway a) to produce the C4'-OH intermediate 57. Alternatively, the radical 55 could be oxidized by
BLM-Fe'"-OH (electron transfer, see also Scheme 8 pathway b, and Section 2.3.1) to produce a C 4 carbenium ion 56,
which could then add water to give the C 4 - O H intermediate
57. This intermediate can collapse to produce free base and
C4'-hydroxylated abasic 58, represented here in its open form.
In the presence of alkali, 58 could undergo a strand scission to
afford 5'-phosphate termini and 3' ends bearing an unsaturated
oxidized sugar residue (59). A few years ago, oligonucleotides
carrying this residue were observed in PAGE analyses of products from the BLM-mediated cleavage of a 5'-["P] DNA fragment."
Intermediate 59 was identified after 1) stabilization by
NaBH, reduction. purification, and hydrolyses with P, nuclease
and alkaline phosphatase to produce a mixture of erytliro- and
thrr0-2-deoxypentitoIs,[~~- 2) reaction with NaOH to give a
transient 2,4-dihydro~ycyclopentenone,~'
7 7 a 7 '*'I
and 3) treatment with aqueous alkylamines (for example, piperidine) to release 3'-phosphate termini.['77a1Reaction with hydrazine also
allows quantitative detection of C4-hydroxylated abasic 58 by
transformation to 3'-(3-pyridazinylmethyl) phosphate termini.[' 771
Attempts to define the source of oxygen in the oxygen carbohydrate moiety 58 (either molecular oxygen or water in the
rebound or the electron transfer mechanisms) have been undertaken with " 0 labeled dioxygen o r water.['h. '] The oxygen
atom introduced in C 4 was derived almost exclusively from
water, thus supporting the C4-carbenium route.
The cleavage patterns observed during D N A strand scission
mediated by photoactivated Co"'-BLM are dependent on the
irradiation frequency but independent of 0,. and are very
similar to those observed for Fe-BLM. They suggest that 3'phosphoglycolate-modified ends," 8h1 free bases, and alkalilabile sites resulting in 3'-phosphate-modified ends [ ' *'] are produced. All the data are consistent with C 4 oxidation of
deoxyribose, but the mechanism leading to the observed products has yet to be established unequivocally.
For the chemistry of RNA cleavage,["0-"21 it has been suggested that Fe-BLM-mediated RNA degradation follows an
oxidative pathway. Experimental data indicate release of bases
concomitantly with RNA cleavage. In PAGE analysis, end-labeled RNAs give cleavage bands believed to have 3'-phosphoglycolate and 5'-phosphate termini. Due to the sensitivity of
R N A to alkaline conditions (alkali-labile lesions cannot be detected easily) and to the difference in the sugar moiety of RNA
and D N A (the postulated 2-hydroxylated base propenals are
probably unstable if formed, and would hydrolyse fast to liberate bases), a molecular mechanism for RNA cleavage of the
same type as for D N A cleavage has yet to be conclusively
proven.
The question arises whether additional "minor" chemistry
can occur in the degradation of D N A by BLM. H I ' abstraction
analogous to that of H4' has been postulated for DNA[''*l and
also for the D N A strand of an RNA-DNA h e t e r o d ~ p l e x . ~ ' ~ ~ ~
While reasonable (in such a heteroduplex the opening of the
minor groove leads to an enhanced accessibility of activated
BLM to H l ' i n addition to H 4 ) , no evidence has been obtained
for D N A degradation by a C1' pathway.[1901Recently. modified
self-complementary octadeoxyribonucleotides containing one
ribo- or one arabino nucleoside at a position susceptible to
oxidative degradation by Fe"-BLM were found to afford products resulting from C1' chemistry.1'911
It has also been demonstrated that activated BLM can induce
damage to bases in a side reaction. The modified bases (such as
8-hydroxyguanine which was present in amounts about onehundredth that of base p r ~ p e n a l [ * ~are
] ) identical with those
761
REVIEWS
generated when hydroxyl radicals react with DNA. This finding
suggests involvement of HO' in this reaction, probably through
a partial decomposition of the 0x0 iron complex responsible for
the main route of oxidative damage to C 4 of DNA.["']
3.4.4. Fe-EDTA, Fe-MPE, and Related Compounds
The molecular mechanism of D N A cleavage by Fe-EDTA
and derivatives in the presence of molecular oxygen or hydrogen
peroxide has only been studied in detail for Fe-MPE. Little
information on the mechanism of cleavage by Fe-EDTA itself
is available;[981generally it is assumed to be similar to that of
Fe MPE. Some particular aspects of the cleavage reaction have
been described for hybrid molecules where Fe-EDTA is attached to sequence-specific DNA-binding molecules.
Although the mechanism of the D N A strand scission by
Fe-MPE is still not understood in detail, some features have
emerged.[98"]The products released from calf thymus D N A are
the four nucleotide bases in a Ade + Thy/Gua + Cyt ratio in
agreement with known D N A composition, which is consistent
with the low specificity of Fe-MPE cleavage. Base release and
strand scission are stoichiometric. The D N A termini at the
cleavage sites are 5'-phosphates and roughly equal proportions
of 3'-phosphates and 3'-phosphoglycolates. Few base propenals
were detected: since [&mercaptoethanol (and probably also
DTT) reacts with base pro penal^,[^^"^ these nucleoside residues
would probably not be observed under these conditions even if
they were formed.
All results are consistent with an oxidative degradation of
deoxyribose rings of the D N A backbone, most likely by the
hydroxyl radical which can abstract a hydrogen atom to initiate
the D N A cleavage. C 4 chemistry as predominant route, C1'
chemistry in the minor groove, and C3' chemistry in the major
groove could account for the product distribution.
The main position of attack could be the 4-hydrogen atom of
the deoxyribose in the case of Fe-MPE, but this has not been
clearly demonstrated for Fe-EDTA. H 4 abstraction as first
step of the cleavage reaction may explain the release of nucleobases and the formation of 5'-phosphate and 3'-phosphoglycolate termini (65) by an 0,-dependent pathway (see
Scheme 17). The absence of base propenals 64 (or unsuccessful
efforts to characterize them) is reminiscent of observed results
with Cu(oP), (1) and enediynes, where, in the minor C 4 pathway of oxidative cleavage, the nature of monomeric species
accompanying C3' - C 4 cleavage remains to be established (see
Sections 3.4.1 and 3.4.2).
Chemistry at C1' must be considered as a possible explanation
for the presence of significant amounts of 3'-phosphate termini
(see Scheme 14). However, due to the presence of reductants,
5-methylene-2-furanone (33),has not been characterized, so this
proposed mechanism remains unconfirmed.[9b.98a1 Both the reactions outlined above should occur in the minor groove.
Chemistry occurring in the major groove has been also postulated.[9b1H3' abstraction (see Scheme 1) produces the radical 45
(Scheme 16). Further reaction of this species in the absence of
oxygen should give rise to free bases, 2-methylene-3-furanone
(48), and 3'-phosphate and 5'-phosphate termini. In the presence of oxygen the production of 5'-phosphate termini, 3'-phosphoglycolaldehydes 53 (which would be oxidized to glycolic
-
762
G. Pratviel et al.
acid), and base propenoic acids 54 (which would not be expected to react with thiobarbituric acid, the usual reagent for base
propenal characterization) may account for some results reported by Hertzberg and Der~an.~""]This is not, however, likely to
occur in the case of derivatives in which Fe-EDTA is attached
to dispamycin,[9i. 113. 1 3 1 a l oligopeptides,[119. 1 3 1 a . 1 3 3 a l or protein~,['~']where, on the basis of PAGE analyses, the oxidative
attack clearly occurs within the minor groove. On the other
hand, when Fe-EDTA is attached to oligonucleotides[135]or
some proteins,[1931the cleavage patterns support an oxidative
attack from within the major groove, and so, in these cases,
a mechanism involving H3' abstraction has to be taken into
consideration. In a similar manner the oxidative HS'(pro-R)
abstraction (H5' points towards the entrance of the major
groove) has to be considered as another possible target (see
Scheme 18).
An alternative mode to the direct H0'-mediated hydrogen
abstraction from deoxyribose may involve an initial attack on
the base moiety. This either gives a radical base species able to
abstract a hydrogen atom from deoxyribose or creates an abasic
site leading to strand break. This alternative pathway is based
on the idea that base destruction accompanies the strand scission
3.5. H5' Abstraction: Enediynes and
Cationic Manganese Porphyrins
Until now, oxidative chemistry at C5' of deoxyribose has been
demonstrated only for enediynes and Mn-TMPyP. These compounds initiate the oxidation of the deoxyribose by H5' abstraction, giving a C5'-centered radical 70 (Scheme 18).
furfural
Scheme 18. Proposed reaction mechanisms when initial site of attack by the oxidizenediynes 2. 5 , 7-10;
19) is at CS' of the deoxyribose [159].
ing agent
(0:
0:
3.5.1. Enediynes
3.5.1 .I. NCS-Chrom
C5' chemistry represents the major mode of D N A cleavage by
NCS-chrom (2) : it causes all single-strand breaks and the cleavage of at least one strand for double-strand lesions.
Eighty percent of SSB lesions consist of breaks in which a
5'-aldehyde (73, Scheme 18) is facing a 3'-phosphate terminus,
whereas in 20% one deoxyribose moiety is lost, and release of
base accompanies 3'-phosphate and 5'-phosphate termini formation.
Angrn'. Chrin. Int. Ed. Engl. 1995. 34, 746-769
Nonenzymatic DNA Cleavage
The C5‘ radical 70 created by activated 2 combines with 0,
(as demonstrated by ‘*O, incorporation)[’94]to give the 5’-peroxy radical 76. The reduction of this radical by a excess of thiol
(or other strong reductant) could result in the recovery of the
5’-hydroxy species 72 as proposed in the literature; alternatively
72 may come from 82. Spontaneous cleavage of 72 affords fragment 73 with a 5’-aldehyde end. Nucleosides bearing the 5’-aldehyde group have been observed after enzymatic hydrolysis by
HPLC.[’951When 3’-[32P]-labeled fragments of D N A were analyzed on polyacrylamide gels, abnormal mobility of 5’-aldehydes 73 was also noted.[”‘] Aldehydes move about as fast as
the markers produced by chemical treatment in the MaxamGilbert procedure, which have a 5’-phosphate end and two
more nucleotides. Because of the presence of reducing agents.
neither intermediate 74 nor end product 75 (Scheme 18) was
observed. Some bands on polyacrylamide gels corresponded to
fragments with a 5’-COOH end; their production was probably
”‘l
attributable to the presence of traces of
Some C5’ chemistry (10-20%) gives rise to gaps in a D N A
strand with concomitant loss of a base. The fragments end with
5’-phosphate and 3 ’ - p h o ~ p h a t e . [ ’To
~ ~ explain
]
these lesions the
following two mechanisms have been proposed: 1) The peroxy
radical 76 is reduced to the hydroperoxide 77. A Criegee-type
rearrangement giving successively 78, 79, and 80, could then
account for the observed products 80 and 81. This mechanism
does not seem likely, because of the absence of any acid or metal
catalysis.‘yd, 731 2) The peroxy radical 76 is transformed to
the oxy radical 82, followed by a ,&fragmentation with cleavage
of the C4’--C5’ bond to generate an unstable 3’-formylphosphate-ending fragment 80, which hydrolyzes into formic acid
and a compound with 3’-phosphate terminus. In the case of
double strand lesions the 0, bridge may account for the formation of the oxy radical 82 [Eq. (9)].[““] When misonidazole replaces 02.these gaps constitute almost the only form of D N A
lesions.[”, ‘971 In an “0-transfer experiment the carbonyl oxygen of the activated formyl moiety (trapped as forniyl-Tris
adduct) was derived from the nitro group oxygen of misonidazole.[’y81The spontaneous transfer of formyl residues from 80
to nucleophiles (Tris or h y d r o ~ y l a m i n e ) [ ’19*]
~ ~ >and a derivative of 81 stabilized by reduction with NaBH4[1991
were evident.
Because C5’ is located at the periphery of the D N A double
strand, reduction of the sensitive intermediates 76 o r 82 to the
C5’-hydroxylated compound 72 may be readily achieved, which
explains the predominance of 5’-aldehyde 73 and the lack of
dependence on the reducing power of the thiol.
Small isotopes effects were observed in abstraction of deuterium from the 5’ position of T residues from a duplex oligonucleotide containing a TGTTTGA sequence ( k , / k , = 1.6),[641a 5’AGC/3‘-TCA sequence (1.29, or a restriction fragment of
pBR322 plasmid ( I .O -1.6 at T of 5’-GT steps; other T residues
showed a variable sensitivity to 5’ deuteriation).[’701H5’ of deoxyribose is incorporated at the C6 radical of the activated
NCS-chrom biradical 4 (see Scheme 3).1531
’
3.5.1.2. CAL
CAL 5 mediates DSBs at preferred interaction sites: the
HS’(pr0-S) abstraction that occurs on the second nucleotide
residue of the four-base-pair sequences 5’-TCCT or 5’A n g r n . Chrin. I n l . E d Et7gl. 1995, 34. 146 -769
REVIEWS
T n T [ ” , 6 7 a ] involves
.
the C6 radical of the activated CAL (6,
see Scheme 4).[””3 ’‘I In PAGE analysis of cleavage products of
the TCCT strand, the 5’-[32P]-labeled fragments exhibited mobilities identical to the corresponding Maxam and Gilbert sequencing markers, thus suggesting the presence of 3’-phosphate
ends.‘”. 67b1 Two major 3’4abeled fragments were resolved, one
ending in a 5’-aldehyde (73, Scheme 18), and one ending in a
5’-carboxylic acid (dependent on metal traces, as described for
NCS-chrom) .lsl]
3.5.1.3. ESP
Only C5’ chemistry is involved in SSBs mediated by ESPA
(7), and this chemistry also occurs in DSBs for ESP C (8) .[’ In
the latter case, the activated drug mediates H5’ and H 4 abstraction on both strands of DNA. At some residues only 5‘ chemistry is observed, at other ones both types of chemistry can take
place. PAGE analysis of the cleavage of D N A substrates labeled
with 32P at the 5’ or 3’ ends showed 3’-phosphate termini and
5’-aldehyde-ended fragments 73. The 5’-aldehyde terminus was
characterized by its usual reduction to 5‘-alcohol by NBH,
treatment or by its conversion into a 5’-phosphate end upon
piperidine treatment (this catalyzes two successive [I-eliminations on 73 with the loss of deoxyribose moiety).
Sequence-specific isotope effect ratios for the cleavage of
D N A were found to be very low (not significantly different from
1). In addition, contrary to what was observed for NCS-chrom,
deuteration at either C5’ or C 4 had no effect on the relative
partitioning of oxidative attack.r571
H5’ abstraction is due to the
C10 radical of the activated biradical drug I 1 (see Scheme 4).[”]
3.5.2. Cationic Manganese P orphjwins
Besides oxidative attack at Cl’-H, (see Section 3.1.3), oxidative activation of C5’-H[1471 constitutes the principal mechanism of activated Mn-TMPyP for cleavage of A . 7-rich sequences of double-stranded DNA[137,138d1
(0in Scheme 18).
In HPLC analysis of the products released in solution after a
heating step, free bases, oxidized nucleosides (74)
and fur~]
fragfural (75) were observed and ~ h a r a c t e r i z e d . ~ “Cleavage
ments from small oligonucleotides were analyzed by HPLC separation techniques or PAGE analyses,[’ 3 7 , 1 3 8 e .
’O’l and
their transformation by chemical and/or enzymatic treatment
allowed their identification: 3’-phosphate-ending fragments
were subjected to alkaline phosphatase to obtain the corresponding oligonucleotides with 3’-OH endings; 5’-CHO-ending
fragments were either heated (inducing two successive [I-eliminations, see below), oxidized by I, in an alkaline medium
(5’-COOH end), or reduced by NaBH, (5’-CH,OH end). Depending on the extent of cleavage and on the sequence a t the 3‘
side of (A T), site, 5’-COOH-ending fragments could be
observed immediately after cleavage of 3’-[32P]-labeled oligonucleo tides.IZ0 2021
The proposed mechanism is indicated in Scheme 18. The activated cationic metalloporphyrin initiates the oxidation of
the deoxyribose by H5’ abstraction to give a C5’ radical
(70). Even in the absence of oxygen (the same cleaving
reactivity was observed in absence or in presence of 0,: M. Pitie
et al., unpublished data), 70 can be hydroxylated directly by
163
G. Pratviel et al.
REVIEWS
TMPyP-Mn’”-OH, or in two steps by one-electron transfer
followed by reaction of water with the carbocation 71, to give
the 5’-OH derivative 72. Spontaneous cleavage follows, with
formation of a 3’-phosphate end and 5’-aldehyde 73 (direct
break of the D N A backbone). A first /Mimination then produces a second break on the D N A backbone with release of a
5‘-phosphate end and the @$-unsaturated aldehyde 74. Finally
a second p-elimination gives rise to free base and furfural (75)
as sugar degradation product. The 5’-COOH ending fragments sometimes observed[201.2021are probably due to a
second oxidation by the recycled activated Mn-TMPyP at the
same site.
The intermediate 74 was completely characterized by reduction with NaBH, (selective reduction of the aldehyde at C5’ to
the alcohol without affecting the double bond) followed by hydrogenation with palladium on charcoal, to give 2’,3’-dideoxyribonucleosides having the natural p or the nonnatural x configuration at C4.r2001The use of 3 1 P N M R spectroscopy has
confirmed the formation of two 3’- and 5’-phosphate monoester
ends in roughly equal proportions, which supports the mechanism described above.[’64d1
It should be noted that selective C5’ chemistry can take place
on both 3’ sides of the (AT), double-strand cleavage site of the
Mn-TMPyP/KHSO, system. This leads to direct strand breaks
with a four base-pair, 3’-stagger of the cleaved residues,[’471
which confirms that oxidative attack of the activated metalloporphyrin occurs from the minor groove of D N A (see
Scheme 1 1) .
3.6. Concluding Remarks on Section 3
The oxidative attack on D N A by chemical nucleases and
drugs described in the present section starts with the formation
of a carbon-centered radical on positions l’, 2’, 3’, 4’, or 5’ of
deoxyribose (Scheme 19). This initial step corresponds formally
to a one-electron oxidation (H’ = H + + e-). However, some
D N A cleavers are capable of two-electron oxidations, and only
in this case is direct hydroxylation of a carbon radical at the
sugar possible, either by oxygen or electron transfer (Scheme 8,
routes a and b). Further reaction of the unstable hydroxylated
derivative could, without cleavage of any C-C bond, give rise
to a sugar residue with five C atoms characteristic of each initial
oxidative lesion (top of Scheme 19) : 5-methylene-2-furanone
(33, attack at Cl’), 3-methylene-2-furanone (48, C3’), the C 4 hydroxylated abasic derivative (58, C4‘), furfural (75, C S ).
> c+water;
abstraction
I
transfer
\ C - OH
4 or/md p
elimnation(s)
’ 4 4
reductant
-
o*
fiscissionand
radical chemistry
-
If the rate of oxygen or electron transfer is too slow (as in the
case of BLM), or if the D N A cleaver is only capable of one-electron oxidation (as in the case of enediynes and photoactivated
halouracil oligonucleotides), the carbon radical reacts with
molecular oxygen (for recent papers on the kinetics of oxygen
rebound in hydroxylation reactions, see ref. [203]). The resulting unstable peroxy radical may be reduced to a hydroperoxide
(R = H ) by reductants [see Eqs. (20) and (II)] or by intramolecular H donation.[641A heterolytic cleavage of the peroxide
0-0 bond may induce a rearrangement and the subsequent
heterolytic cleavage of one adjacent C - C bond of the sugar (a
Criegee-type rearrangement, for example). Two sugar fragments are usually (or might be) observed. Depending on the
initially attacked carbon site and on which adjacent C-C bond
is cleaved, two sets of two sugar fragments can be generated:
one-carbon plus four-carbon fragments (C2’ chemistry: formic
acid + 42; C5’ chemistry: 80 + 81) o r two-carbon plus threecarbon fragments (C3’ chemistry: 53 + 54; C4’ chemistry:
65 + 64). Alternatively, for example when two peroxy radicals
combine (0, route), a homolytic cleavage of the peroxide 0-0
bond giving an oxy radical may lead to a homolytic cleavage of
the same adjacent C - C bond. The p-scission of the oxy radical
would afford two kinds of fragments: one is the same as produced in the preceeding pathway: 42 (C2’). 54 (C3’), 65 (C4‘).
and 80 (C5’);the other one is a radical species whose fate is
unknown and whose products have not yet been described.
In addition, the oxy and peroxy radicals can be considered
reduction-sensitive intermediates, and their possible reduction
can afford a hydroxylated sugar by competitive pathway.
All these reaction pathways are well illustrated and detailed in
Schemes 17 and 18 for C 4 and C5’ chemistry. In Scheme 15, C2’
hydroxylation (that would lead to the ribo or arabino derivative) is not shown, since neither two-electron oxidation by
halouracil derivatives nor reduction of a reduction-sensitive intermediate can be performed under the experimental conditions
used. In addition no chemical agents able to hydroxylate C2’
directly have so far been reported. In Scheme 16, in the absence
of 0, the hydroxylated product 46 must be formed after a oneelectron oxidation of the intermediate 45 (probably by the
excited state of the rhodium complex); however, the exact
nature of the species able to abstract H3‘ is still unknown.
Hydroxylation of C1’ can easily account for the observed products in Scheme 14. N o other products of D N A degradation
except oxidized abasic sites have been described, probably because the glycosidic bond is cleaved in preference to a sugar
C - C bond.
sugar residue
with 5 carbon atoms
1 sugar Fragment
1 radical species
Scheme 19. Two-electron (A) or one-electron (B) oxidation of deoxyribose mediated by drugs
can lead to three main pathways for sugar evolution (R = H, O O R ) .
764
4. Concluding Remarks
Apart from nuclease enzymes, no naturally occurring, low molecular weight compounds have been discovered that can efficiently mediate hydrolysis of the
phosphodiester bonds of the D N A backbone. The
DNA-cleavage route selected by nature for cytotoxic
drugs (bleomycin, enediyne) is a sugar oxidation and
not a hydrolysis of phosphodiesters. So it is not surprising that synthetic transition metal complexes exhibiting redox properties and D N A affinity have been
Angew. Chem. Inl.
Ed. EngI. 1995. 34, 146-769
Nonenzymatic D N A Cleavage
REVIEWS
developed a s “chemical nucleases” (a misnomer-what about
“chemical D N A sugar oxidases”?). Deoxyribose is a key target
for DNA cleavers, and current knowledge of the different
modes ofoxidation of deoxyribose units leading to D N A breaks
has been reviewed i n the present article. Development of these
artificial and selective D N A cleavers is a challenging area in the
rational design of antitumoral and antiviral agents as well as in
the field of molecular biology. The need for highly efficient
(quantitative cleavage of the target must be reached) and highly
specific (sequence specificity on large D N A fragments) cleaving
reagents is obvious, and considerable progress has already been
made. lmprovernents in several directions are still required:
1 ) Cleaving efficiency: Discovery of truly double-strand and
catalytic DNA cleavers (and also highly efficient RNA cleavers)
is required in order to create new cytotoxic agents and tools for
gene engineering. In the field of cancer chemotherapy, enediynes
represent an example where an understanding of the chemistry
and modes of action can help the design of highly efficient in
vitro drugs, such as the synthetic dynemicins analogues described by Nicolaou and
But the question whether in
vitro efficiencies transferable to in vivo conditions still remains.
2) Specificity: A specific molecular site of attack without dispersion of lesions is also highly challenging, especially if genomic double-stranded DNA is the ultimate target. This property is
not necessary in some cases: absence of selectivity allows use of
some chemical nucleases as footprinting reagents for D N A or
RNA. and they have some advantages over the DNase method.
Sequence selectivity has been drastically increased by the covalent binding of DNA-cleaving moieties to oligonucleotides or to
DNA-binding small molecules, peptides, or proteins. In particular. the attachment of an efficient D N A cleaver to a deoxyoligonucleotide should lead to a family of synthetic restriction
enzymes with tailored specificity that might be useful in the
analysis of complex genomes or could enhance the ability of an
antisense deoxyoligonucleotides to block the expression of
targeted gene. As an example of these last potential medical
applications, we recently reported that the attachment of
the cleaver motif jtris(methy1pyridiniumyl)porphyrinato)manganese(rr1). which is endowed with cytotoxic properties,12041
to a selected 19-mer vector allows selective recognition and
cleavage of a 35-mer single-stranded D N A containing the initiation codon of the fcii gene of HIV-1 (Scheme 20).”491
DNA cleaver =
targeted tnetalloporphyrin
I
’
5
Target =
I
HO~TCTCGTTCTTT
ACCTCGG”
GAGAGCAAGAAATGGAGCC
pan of ttrr gene of HIV- 1
3’
illllllllllllllll
cleavage sites
Scheme 20 Selective cicavnge by B tailored manganese porphyrin of a 35-mer single-stranded DNA containing the initiation codon of the fui gene of HIV-1 [149].
3) Selection of oxidizing agents able to leave D N A fragments
ending with suitable termini for further enzymatic workup:
Even though the cleavage of any selected D N A sequence is now
possible, no efficient chemical models of a restriction enzyme
that can hydrolyze a selected phosphodiester bond are yet available. However, in an alternative pathway, hydroxylation of C3’
or CS of deoxyribose immediately fol\owedby reduction of the
resulting 3’-ketone or 5’-aldehyde may enable this hydrolysis
step to be mimicked (oxidation step reduction step =
“pseudo-hydrolysis”) . This was recently demonstrated by using
the Mn-TMPyP/KHSO, system and its specific hydroxylating
activity at the C5’ position of deoxyribose [Eq. (12)]. The fragments may be coupled again by chemical methods, so this system might provide an alternative to the restriction endonuclease/Iigase systems.12021
+
ROPO2-0,
ROPO;
+ OHC
e
hydroxylation
e-
ROPO,
+ HOH,C
(12)
reduction
Finally, even after a decade of fruitful research on chemical
nucleases, the field is still open for discoveries of new efficient
cleavers of R N A and double-stranded D N A in order to
produce new selective cytotoxic agents and/or tools for genetic
engineering.
5. Addendum
After submission of the original manuscript several articles
have appeared in the literature. These papers are briefly mentioned below together with the appropriate sections where they
should be inserted.
Section 2.1 : Cleavage of megabase D N A substrates was performed with a DNA-binding protein modified by Cu(oP) .[’O5I
The structure of the Eschevichia coli Fis-DNA complex was
probed by a protein modified with a copper-phenanthroline
Section 2.2: The role of biradical species in D N A cleavage by
neocarzinostatin chromophore (2) has been discussed recentIY.~’~’] Moreover, there is a critical role of the iodine substituent
of the calicheamicin oligosaccharide and the guanine C2-NH,
group within the minor groove of the target DNA.[’os1
Section 2.3: The stability of the oxygenated form of FellBLM has been investigated.[20g1A recent study by electrospray
mass spectrometry of activated bleomycin suggested that the
iron peroxide BLM-Fe“’-OOH is the active species.12’01The
same iron peroxide was also proposed for bleomycin models.[’’ ’] A hybrid molecule containing bleomycin and the DNAbinding domain of Hin recombinase was used as DNA
cleaver.[’12] Both strands of a DNA-RNA heteroduplex were
cleaved by Fe-BLM.12131
Section 2.4: Fe-MPE and Fe-EDTA were used for mapping
RNA regions in eukaryotic ribosomes. RNA fragments were
analyzed by primer extension with reverse transcriptase using
oligodeoxynucleotide primers.[2141The tertiary structure of
tRNAPhe has been investigated by RNA-EDTA-Fe“ autocleavage.[’ ”1
Section 2.5: The transfer in water of an oxygen atom from the
high-valent TMPyP-MnV=O species to a water-soluble olefin
was investigated. A “redox tautomerism” involving an axial
water molecule was proposed to explain the incorporation of the
oxygen atoms from water and from the peroxide in a 50:50
The interaction of cationic porphyrin derivatives with
calf thymus DNA was studied by linear dichroism and fluorescence energy transfer.[” ’1
Section 2.6: Binding of A - and A-[Ru(phen),]’’ with a selfcomplementary decamer was investigated by N M R spec765
G. Pratviel et al.
REVIEWS
troscopy.[*’‘1 Enantioselective recognition of DNA by
Rh”’(phen),(phi) (20) was found to be controlled by the opening
in the major groove that results from changes caused by propellerlike twisting of a base pair.[2191
(ARC, Villejuif) , the Agence Nationale de Recherches sur le Sida
(ANRS. Paris) , the RPgion Midi-PyrPnPes and Genset (Paris)
for financinl support. We thank Lisa Rosenberg (University o j
British Columbia) jor her assistance in preparing the manuscript.
Received: September 3, 1993
Revised version: February 12, 1994
Addendum: October 31. 1994 [A 21 IE]
German version: A n ~ e w Chem.
.
1995, 107. 819
Appendix: Important Abbreviations Used in the Text
A:
Ade :
bPY:
BLM :
bp :
C:
CAL:
Cyt:
DSB:
DTT :
DYN:
EDTA :
ESP :
G:
Gua:
GSH :
ME :
5-MF:
MPA:
MPE:
N:
NADH:
NCS :
NCS-chrom:
ODN :
OP:
PAGE :
phi :
phen :
Porph :
pu :
Py :
PYr:
TMPyP:
SSB :
T:
TAR :
tat:
Thy:
Tris :
Ura:
2’-deoxyadenosine
adenine
bipyridine
bleomycin
base-pair
2‘-deoxycytidine
calicheamicin
cytosine
double-strand break
dithiothreitol
dynemicin
ethylenediaminetetraacetate
esperamicin
2’-deoxyguanosine
guanine
ghtathione
mercaptoethanol
5-methylene-2-furanone
3-mercaptoproprionic acid
methidiumpropyl EDTA
nucleo tide
nicotinamide adenine dinucleotide, reduced
form
neocarzinostatin
neocarzinostatin chromophore
oligodeoxynucleotide (The deoxyribo prefix
was omitted in this review, since it focuses on
DNA.)
ortho-phenanthroline = 1,lo-phenanthroline
(The classical abbreviation used is phen (see J.
Reedijk, Coniprehensive Coordination Chemisfry, Vol. 2 (Ed.: G. Wilkinson, Pergamon,
1987, pp. 73-98); however we decided to keep
the formulation Cu(oP), for the bis(1,lOphenanthro1ine)copper complex popularized
by D. Sigman in oxidative DNA cleavage research.)
polyacrylamide gel electrophoresis
phenanthrenequinonediimine
1,IO-phenanthroline
a porphyrinato ligand
purine
pyridine
pyrimidine
meso-tetrakis(N-methylpyridinio4-y1)porphyrinato
single-strand break
thymidine
trans-activation responsive region
trans-activator of transcription
thymine
tris(hydroxymethy1)aminomethane
uracil
The aulhors are deeply indebted to Ihe many co-workers and
collaborators whose names are listed in the references. We are also
grateful to the Centre National de la Recherche Scientifique
(CNRS) , the Association pour la Recherche contre le Cancer
766
[ I ] B. N. Ames, Science 1983,221. 1256-1264; W. Troll. R. Wiesner, Annu. Rev.
Phurmacol. Toxfiol.1985,25,509-528: P. A. Cerutti, Science 1985,227,37538 1 ; M. K. Shigenaga. C. J. Gimeno. B. N. Ames, Proc. Nut/. Arud. Sci. U S A
1989,86,9697-9701; Methods Enzymol. 1990, 186 (A and B) (both volumes
are dedicated to oxygen radicals in biological systems); T. Kalebic, A. Kinter.
G . Poli. M. E. Anderson. A. Melster. A. Fauci. Proc. Natl. Acud. Sci. U S A
1991,88,986-990: J. R. Wagner, C. C. Hu. B. N. Ames, ihid. 1992,89,33803384; and references therein; M. Becana. R. Kiucas, !bid. 1992, 89, 89588962; A. W. Khan, Y. H. Mei. T. Wilson. ihid. 1992, 89, 11426-11427; S.
Xanthondakis, G . Miao, F. Wang. Y. C. E. Pan, T. Curran. EMBO J. 1992-11,
3323-3335; D. I. Feig, L. A. Loeb. Biochemistry 1993, 32, 4466-4473: S.
Legrand-Poels, M. Hoeheke. D. Vaira, B. Rentier. J. Piette, J. Photochem.
Photohid. B 1993, 17, 229-237.
H. Sies. Angeu,. Chem. 1986, 98, 1061 -1075; Angeu. Chem. In/. Ed. Engl.
1986,25,1058-1071; 1. Fridovich, Annu. Rev. Phurmacol. Toxicol. 1983.23,
239- 257.
E. C. Friedberg, DNA Repair. Freeman. New , irk. 1985; M . M. Cox, I . R.
Lehman. Annu. R F ~Biochem.
.
1987. 56, 229-262; A. Sancar, G. B. Sancar.
ihid. 1988,57.29-67; R. M . Story, I. T. Weber, T. A. Steitz. Nature (London)
1992, 355, 318-325; S. Boiteux, E. Gajewski. J. Laval, M. Dizdaroglu, Biochemistry 1992, 31, 106-110.
J. Cadet, M. Berger, C. Decarroz. J. R. Wagner, J. E. van Lier, Y. M. Gmot,
P. Vigny, Bfochimie 1986, 68. 813-834: C . von Sonntag. The Chemical Busi.s
qf’Radiation Bk~logy,Taylor and Francis, Philadelphia, 1987; C. von Sonntag,
H.-P. Schuchmann, Methods Enzymol. 1994. 233. 3-19.
H. Urata. K. Yamamoto, M . Akagi, H . Hiroaki, S. Uesugi, Biochemistry
1989.28.9566- 9569; J. Piette. J. Photorhrm. Phoiohiol. B 1991, 11, 241 -260;
H. Kasai, 2. Yamaizumi, M. Berger. J. Cadet, J. Am. Chem. Soc. 1992, 114.
9692-9694; L. P. Candeias, S. Steenken, ihid. 1993,115,2437-2440; N. PailIous, P. Vicendo, J. Photochem. Photobiol. B 1993, 20, 203-209.
B. Demple, J. Halbrook, Nuturr (London) 1983. 304, 466-468; J. F. Mouret,
F. Odin, M. Polverelli, J. Cadet. Chem. Rrs. Toxicol. 1990, 3. 102-110.
S. Steenken. Chem. Rev. 1989.89, 503-520; A. J. S. C. Vieira, S. Steenken, J.
Am. Chem. Soc. 1990, 112,6986-6994.
D. G . Knorre, 0. S. Federova. E. I. Frolova, Russ. Chem. Rev. (Engl. Trans/.)
1993.62, 65-86.
a) S. M. Hecht, Acc. Chem. Re.>.1986,19. 383-391; b) J. Stubbe, J. W Kozarich, Chem. Rev. 1987,87,1107-1136;~) D. H. Petering, R. W. Byrnes, W. E.
Antholine, Chern. B i d . Interact. 1990, 73, 133-182; d) 1. H . Goldberg, Ace.
Chem. Rrs. 1991. 24, 191 198; e) Free Radical B i d . Med. 1987, 3,41-54; f )
M. D. Lee. G . A. Ellestad. D . B. Borders. Acc. Chem. Res. 1991,24,235-243:
g) K. C. Nicolaou, W. M. Dai, Angew. Chem. 1991, 103, 1453-1481; Angew.
C h r m . I n ~E. d Engl. 1991,30. 1387-1416; K. C.Nico1aou.A. L.Smith, E. W.
Yue, Proc. Null. Acad. Sci. U S A 1993. 90, 5881-5888; h) K. C. Nicolaou,
A. L. Smith, Arc. Chem. Res. 1992,25,497-503; i) J. S. Taylor, P. G . Schultz,
P. B. Dervdn. 7i.trahedron 1984,40,457-465; j ) I. Lee. J. K . Barton, Biochemi.sIry 1993, 32. 6121 -6127. and references therein; k) D. S. Sigman, T. W.
Bruice. A. Mazumder, C. L. Sutton, Arc. Chem. Res. 1993.26, 98-104; I) T.
Tanaka. M. Hirama, K.-I. Fujita, M. Ishiguro. J. Chem. Soc. Chem. Cummun.
-
1993. 1205-1207.
A. P. Grollman. M. Moriya. Trends Gene{. 1993, 9. 246-249.
W. Saeuger. Principles of Nucleic Acid Structure. Springer, New York. 1984.
a) A. Mazumder, J. A. Gerlt. M. J. Absdlon, J. Stubbe, R. P. Cummingham,
J. Withka, P. H. Bolton, Biochemistry 1991. 30, 1119-1226; b) J. J. Hangeland. J. J. De Voss, J. A. Heath, C. A. Townsend. W. D. Ding, J. S. Ashcroft.
G. A. Ellestad, J. Am. Chrm. Soc. 1992. 114. 9200-9202.
D. Suck, C. Oefner, Nature (London) 1986, 321, 620-625; J. A. McClarin,
C . A. Frederick, B. C. Wang. P. Greene. H. W. Boyer, J. Grable. J. M. Rosenberg. Science 1986. 234. 1526-1541.
[141 A. M . Maxam. W. Gilbert, Method.\ Enz.ymo1. 1980, 65, 499-560.
[151 a ) D. S. Sigman. D . R. Graham, V. DAurora, A. M. Stern, J. B i d . Chem.
1979, 254,12269-12272; h) D. S. Sigman, Acc. Chem. Res. 1986. 19, 180186: c) C. Yoon, M . D. Kuwabara, R. Law, R. Wall, D. S . Sigman, J. B i d .
Chem. 1988,263,8458-8463; d) D. S. Sigman, Eiochemislry 1990,29,90979105; e ) J. M. Veal, R. L. Hill, &id. 1991,30, 1132-1140; f) A. Spassky, ibid.
1992.31. 10502-10509.
[I61 v. M. Downey. B. G. Que. A. G . So, Biochem. Biophys. Res. Commun. 1980,
93, 264.
71 L. E. Marshall. D. R. Graham, K. A. Reich, D. S. Sigman, Biochemistry
1981. 20, 244-250.
Angew C h m i In! Ed Eunl 1995, 34, 746-769
NonenLymatic D N A Cleavage
iir Johnson. N. B. Nazhat. J. An7. C k n i . Soc. 1987, 109. 199019Y4.
[lY] L. I). Willi;iiiis. J. Thivierge. I. H . Goldberg, Nucleic Acids Re.% 1988. 16,
11 607-1 1615.
[20] M . Dizd;iro:Iu. 0. 1. Aruoma. B. Halliwell, Biuch(vni,rrry 1990. 2Y. 8447843 1.
1211 J. M. Veal. ti. Merchant, R. L. Rill, Nuckic Acids Rrs. 1991. 1Y. 3383 3388.
[22] L. E. Popc. D. S . Sigman. Proc. Nurl. Acud. Sci. U S A 1984. 81, 3-7
[23] J.-c'. Franqois. T. Saison-Behmoaras. M. Chassignol. N. T. Thuong. J.-S. Sun.
c'. HL:Ene. Uio~~lieini.~ti:i.
1988, 27. 2272-2276.
(241 ti. A . Reich. L. E. Marshall. D. R. Graham. D. S. Sigman. J. A m . Cheni. Soc.
1981. 10.1. 3582Z3584.
(251 T. B. Thcdcrxhn, M. D. Kuwabars. T. A. Larsen. D. S. Sigman. J. Am. Chmn.
S o ( . 1989. 1i1. 4Y41-4946.
[26] S . (ioldstein. G . Czapski. J. A m . Chem. Sue. 1986, 108. 224442250,
1 9 ~ 6 . 2 ~ 7 4 0740s.
1
[29] a ) L. Marshal Pope, K. A . Reich. D. R. Graham, D. S. Sigman, J. Bid. Chem.
1982. 3,?.
12 121 12128: b) D. S. Sigman. C. B. Chen, Metul-DNA Chemi.s/rj (Ed.: T. D. Tullius) ( A C S Symp. Ser. 1989, 402. pp. 24-~47).
[30] J. M . Veal. R. L. Rill. Biochcmi.s/ry 1988. 27, 1822-1827.
[31] C. Yoon. M . D. tiuwabara. A. Spassky. D. S. Sigman. Biuchrmi.vtr.L;1990, 29.
2116 2111.
[37] J. M. Vwl. R. L. Rill. Bio(./irnii,s/r~
1991, 30. 1132-1140.
[33] a ) D. S. Sigman. A. Spassky, S. Rimsky. H. Buc. Biopu/i~rner.s1985, 24. 183~
N u ~ . l ( , i, ~A . d s RLJS.1989. 17. 5361 . 5375.
1361 C. B. Chcn. D. S Sigmun. Proc. Not/. Acud Sci. USA 1986. 83. 7147-7151.
: . FraiiFois. T. Saison-Behmoaras. M. Chassignol, N . T. Thuong. C. Hele1371 JC
ne. .I B i d . Chon. 1989. 264. 5891 - 5898.
[38] C. B. Chen. D. S. Sigman. J. Aiii. C'heni. Soc. 1988, 110. 6570-6572.
139) J.-C'. Fr;inqois. T. Saison-Behmoaras. C. Barbier. M. Chassignol, N. T.
Thuong, C. Hel&nc. Proc. Nut/. Acad. Sci. U S A 1989, 86, 9702-9706.
[40] J.-S. Sun. J:C. Fraqois. R. Lavery, T. Saison-Behmoaras, T, Montenay-Ga-
REVIEWS
[62] Y. Uesawa, Y. Sugiura. Bm~hemistry1991. 30. 9242-Y746.
[63] D. Dasgupta. 1. H. Goldberg, Biuchemisrrj~1985, 24. 6913 -6920.
[64] L. S. Kappen. 1. H. Goldberg. B. L. Frank. L. Worth. Jr.. D. F. Christner,
J. W. Kozarich, J. Stubbe, Biochmnis/r:l,1991. 30. 2034 -2042.
[65] A. Galat. 1. H. Goldberg, Nudeic Acids Res. 1990. 18. 2093-209').
[66] a) L. S. Kappen. I. H. Goldberg, Proc. Nurl. Acud. S i . U S A 1992.89, 6706-
167) a ) S. Walker, R. Landovitz. W. D. Ding. G . A. Ellestad. 0 . Kahne. Proc. Null.
Acud. Sci. U S A 1992.89,4608-4612: b) N . Zein. M. Poncin, R. Nilakantan,
G. A. Ellestad. Science 1989.244, 697-699: c) J. Drak. N . Iwasawa. S. Danishefsky. D. M. Crothers, Proc. Nu//. Acud. Sci. USA 1991. 88. 7464-7468; d)
1. Aiyar. S. J. Danishefsky, D. M. Crothers, J. Am. ( % ~ r n . Soc. 1992. 114.
7552-7554; e ) K. C . Nicolaou. S.-C. Tsay. T. Suzuki. G . F. Joyce. ihid. 1992,
114, 7555-7557.
1681 a) Y. Sugiura, Y. Uesawa, Y Takahashi. J Kuwaharii. .!. Golik. T. W. Doyle.
Proc. Narl. Acad. Sci. GSA 1989,86. 7672-7676: h) 8. H. Long. J. Golik. S.
Forenza. B. Ward. R. Rehfuss, J. C. Dabrowiak, J. J. Catino, S. T. Musial.
K. W. Brookshire. T. W. Doyle. ihid. 1989. 86. 2-6; c ) H. Kishikawa. Y-P.
Jiang. J. Goodisman. J. C. Dabrowiak. J. An]. C/irni. Soc. 1991. 113. 54345440.
1691 M. Uesugi. Y Sugiura, Biocheriiistr~1993. 32, 4622 4627.
[70] Y. Sugiura. T. Shiraki, M. Konishi. T. Oki. Proc. Nuti Acad. Sci. U S A 1990.
87, 3831 -3835.
[71] L. S. Kappen. I. H. Goldberg, S. H. Wu. J. Stubbe. L. Worth, Jr.. J. W. Kozarich, J. Am. C%Pnm.Sue. 1990. 112. 2797-2798.
I721 H. Umezawa. K. Maeda. T. Takeuchi. Y.0 k i m i . J Anrihior. Sr,i-. A . 1966. 19.
200-209.
[73] H. Umezawa in Bleunmwin: Currmr Srutus und N w Dci&pn'nlmts (Eds. :
S . K. Carter. S. T. Crooke. H. Umerawa). Academic Press. New York, 1978,
pp. 15- 19.
[74] H. Suruki, K. Nagai, H. Yakami, N. Tanaka. H. Umezawa, J. Antrhior.
1969, 22. 446-448: Y. Fujiwara. T. Kondo. Bioe/irm Pliurmacul. 1973, 22.
323-333.
[42] S. D. J a y s e n a , B. H. Johnston. Proc. Null. Acud. Sci. U S A 1992. 89. 3526-
3530.
. M. B. Gorin. D. S. Sigman, Proc. Nut/. Acad. Sci. U S A 1993, 90.
1441 C. B. ('hen. A. Maromder, J:F.
Constant. D. S. Sigman. B i u c u n j u ~ u Chmm.
t~~
1993. 4, 69 77.
[45] A. D. Baker. R. J. Morgan, T. C . Strekas, J. Clzem. Soc. Chem. Cunimun. 1992,
1099 1100.
1461 i i ) P.C. Dedon. 1. H. Goldberg, Biochen?is/rr 1992. 31. 1909-1917; b) P. C.
Dedon. Z.-W. Jiang. I. H. Goldberg, ihid. 1992, 31. 1917-1927.
1471 R. D . Groncberg. T. Miyazaki. N. A. Stylianides, T. J. Schulze. W. Stah]. E. P.
Schreiner. T. Suzuki. Y Iwabuchi, A. L. Smith. K. C . Nicolaou, J. Ain. Chen.
. S U C 1993. f f S , 7593 -7611: A. L. Smith. E. N. Pitsinos, C.-K. Hwang. Y.
Mizuno. H. Saimoto. G. R. Scarlato. T. Suzuki. K . C. Nicolaou. ibid. 1993.
115.7612 7624; K . C. Nicolaou. C. W. Hummel, M. Nakada. K. Shibayama.
E. N. Pitsinos. H. Saimoto. Y. Mizuno, K.-U. Baldenius. A. L. Smith, ihid.
1993. 11.5, 7625-~7635:see also K. C. Nicolaou. Angew. Chem. 1993. 105.
1462-1471: A n g e w Chenf. I r i r . Ed. Engl. 1993, 32, 1377-1385.
[48] R. G. Bcrgman, Ace. Chrin. Rex. 1973. 6. 25-31.
(4Y] N. Zein. W. J. McGahren. G. 0. Morton, J. Ashcroft. G . A. Ellestad. J. A m
C / i m . .S'n(.. 1989. 11 1. 6888--6890.
[SO] D.-H. C'hin. I. H . Goldberg, Biochenii.~rry1986. 25. 1009-1015.
[ S l ] N Zein. A M . Sinha. W. J. McGahren. G. A . Ellestad. Sciencr 1988. 240.
1iY8 1201.
1521 I<.
S. Kappen, 1. H . Goldberg, Pruc. Nut/. Acud. Sci. U S A 1984. 81. 3312~3316.
[53] S. M. Meschwitz, 1. H. Goldberg. Proc. Null. Acud. Sci. USA 1991.85, 30473051
[79] Y. Kaku. M. Otsuka, M. Ohno, Chem. Lerr. 1989. 611 614; A. Suga. T.
Sugiyama, M. Otsuka. M. Ohno, Y. Sugiura. K. Maeda. Telruhedron 1991,47.
1191-1204.
[SO] M. Otsuka. T. Masuda, A. Haupt, M. Ohno, T. Shiraki, Y Sugiura. K. Maeda, J. Am. Chein. Soc. 1990, 112, 838-845.
1811 B. J. Carter. K. S. Reddy. S. M. Hecht, Tetrdwrlroii 1991. 47, 24632474.
[82] K . Shinozuka. H. Morishita, T. Yamazaki. Y. Sugiura. H. Sawai. 7t1rulredron
LEI!.1991, 32. 6869-6872.
(Eds.: S . Neidle.
[83] L. F. Povirk in Muleculur Apects qf Anri-Cuncer Drug A~.rion
M. J. Waring). Verlag Chemie. Weinheim. 1983. pp. 157-181
[84] a) E. A. Sausville, 3. Peisach, S. B. Horwitz. Biuchc,niistry 1978, 17, 27402746; b) H. Sugiyama. R. E. Kilkuskie. L.-H. Chang. L.-T. Ma, S. M. Hecht,
G. A. van der Marel, J. H. van Boom, J. Am. Chew. Soc. 1986, 108. 3852; c)
G. R. Buettner, P. L. Moseley. Biuchrinisiry 1992, 31. Y784-9788.
[85] a) R. B. Van Atta. E. C. Long, S. M . Hecht. G. A. van der Marel, J. H.
van Boom, J. Am. Chrm. Sue. 1989,111.2722-2724: h) K. Kohda. H. Kasai,
T. Ogawa. T. Suruki. Y Kawazoe. Chrnm. Pharm. Bull. 1989, 37. 10281030.
[86] L. Giloni, M. Takeshita. F. Johnson. C. Iden, A. P. Grollman, J. Bid. Chew.
1981. 256. 8608 -8615.
1871 a) Y Sugiura. J. Am. Cheni. Suc. 1980. 102, 5208-5215; b) R. M. Burger, J.
Peisach. S. B. Horwitz, J. B i d . Chrm. 1981. 256, 1 1636-11 644: c) R. M.
Burger, J. S. Blanchard. S. B. Horwitz. J. Peisach. ihid 1985, 260. 15406
15409.
[XX] A. Natrajan. S. M. Hecht. G. A. van der Marel. J. H. v a n Boom. J. Am. Clienm.
Soc. 1990. 112. 3997-4002.
1891 a) T. J. McMurry, J. T. Groves in [89d]. pp. 1-28: b ) V. Ullrich. Tup. Curr.
Chrm. 1979. 83. 67-104: c) R. E. White. M. J. Coon. Anmr. Rev. Biuc/irm.
1980. 49. 315- 356; d ) C:rtochromc P-450:Srructirrf,. M~~chunism
ond Binchenii.stry (Ed.: P. Ortiz de Montellano). Plenum. New York. 1985;e) A. Sorokin,
A. Robert, B. Meunier, J. Am. Chem. Soc. 1993, 11.5. 7293-7299,
[YO] a) B. Meunier, Clirm. Ree. 1992. 92.1413 -1456; b) J. T. Groves. R. C. Haushalter, M. Nakamura. T. E. Nemo. B. J. Evans, J. Am Chcn7. Srrc. 1981, 103,
2884-2886: c) D. Mandon, R. Weiss. K . Jayaraj. A. Gold, J. Terner, E. Bill,
~
[56] .I.J. De Vosa. C. A . Townsend. W:D. Ding, G. 0. Morton. G. A. Ellestad, N .
Zein. A . B. Tabor. S. L. Schreiber, ./. Am. Chein. Sue. 1990, 112, 9669-9670.
[57] D. F C'hristner, B. L. Frank, J. W. Kozarich. J. Stubbe, J. Golik, T. W. Doyle.
I . E. Rosenberg, B. Krishnan, J. Am. Chem. Sue. 1992, 114. 8763 -8767.
[58] a) M. K Semmelhack, 1. Gallagher. D. Cohen. Tetruhedrun LPII.1990, 31.
1521 - 1522: b) P. A. Wender. R. C. Kelly, S. Beckam. B. L. Miller. Proc. Nut/.
4 ~ i i d5 i . C'SA 1991. 88. 8835-8839; c) Y. Sugiura. T. Arakawa. M. Uesugi.
T. Shiraki. H. Ohkuma. M. Konishi, Biochemistry 1991, 30, 2989-2992.
Angew. Chrm Jnt. Ed. Engl. 1995, 34. 746-769
761
G. Pratviel et al.
REVIEWS
A. X. Trautwein, Inorg. Chen. 1992.31.444-4409: d ) 0. Bortolini. B. Meunier. J. Chwii. Sol. Chem. Commun. 1983, 1364-1365: e) 0 . Bortolini. M.
Ricci. B. Meunier. P. Friant, I. Ascone. J. Goulon, Nous. J. Chim. 1986, 10.
39-49.
[9i] a) N. Murugesan, G. M. Ehrenfeld. S. M. Hecht. J. Biol. Chcm. 1982, 257,
8600-8603; h) N. MUrUgesan, S. M. Hecht, J. Ani. Chem. Soc. 1985. 107.
493-500; c) Inorg. Chem. 1987. 26. 3835-3836.
[92] R. M. Moriarty, R. Penmasta, I. Prakash. Tetruhedron Lett. 1985. 26.46994702; E. C. Long, S. A. Hecht. ihid. 1988, 29. 6413-6416.
[93] R. S. Magliozzo. J. Peisach. Inorg. Chcm. 1989. 28, 608-611.
[94] G. Pratviel. J. Bernadou. B. Meunier, Biochem. Biophys. Res. Commun. 1986,
136, 1013- 1020: Biochrm. Phurmucol. 1989, 38. 133 -140: B. Meunier, N P I I .
J. Chem. 1992, 16. 203 - 21 I .
[95] M. Girardet, B. Meunier. Tetruhedron Lelr. 1987. 28. 2955-2958.
[96] a) G. Padbury, S. G. Sligar, R. Labeque. L. J. Marnett, Biochemistry 1988,27,
7846-7852; h) A. Natrajan, S. M. Hecht. G. A. van der Marel. J. H. van
Boom. J. Am. Chem. Soc. 1990. 112,4532-4538.
[97] R. E. Shepherd, T. J. Lomis, R. R. Koepsel. J. Chiwi. Soc. Chon. Commun.
1992, 222-224.
[98] R. P. Hertzberg, P. B. Dervan, B~ochernisrry 1984. 23, 3934-3945: D. W.
Celander, T. R. Cech. ibid. 1990. 29. 1355-1361: R. V. Prigodich. C. T. Martin. rhid. 1990. 29. 8017-8019.
[Y9] L. E. Rabow, G. H. McGall, J. Stuhbe, J. W. Korarich, J. Am. Chem. Sot.
1990. 112, 3203-3208.
[loo] E. Gajewski. 0. I . Aruoma. M. Dizdaroglu, B. Halliwell, Biochrmi.sfry 1991.
30, 2444-2448.
[loll L. L. Guan. J. Kuwahara, Y Sugiura, Bioc/iemisrr,l:1993, 32, 6141-6145.
[lo21 J. C. Dabrowiak in Ad~'unce.sin Inorgunic Biochiwi.rtrv (Eds.: G. L. Eichhorn.
L. G. Marzilli). Elsevier, New York, 1982, pp. 69-1 13.
[lo31 J. Kuwahara, Y. Sugiura, Proc. Null. Acud. Scr. U S A 1988, R5. 24592463.
[lo41 R. P. Hertzherg, M. J. Caranfa, S. M. Hecht, Biochenlistri~1988. 27, 31643174: E. C. Long. S. M. Hecht. G. A. van der Marel. J. H. van Boom, J. Am.
Cheni. Soc. 1990, 112? 5272- 5276.
[I051 B. Gold. V. Dange, M. A. Moore. A. Esastman. G. A. van der Marel. J. H.
van Boom, S. M. Hecht. J. Am. Chern. Soc. 1988, 110. 2347-2349.
11061 L. F. Povirk, Y.-H. Han, R. J. Steighner. Biochemisrr,- 1989, 28, 5808- 5814.
[I071 R. J. Steighner. L. F. Povirk. Proc. Null. Acud. Sci. USA 1990. 87,8350-8354.
[lox] D. S. Sergeyev. V. F. Zarytova, S. V. Mamaev. T. S. Godovikova. V. V. Vlassov, Auti.\cwsr Res. Dev. 1992. 2. 235-241: V. F. Zarytova, D. S. Sergeyev,
T. S. Godovikova. BlUciJnjlL&?~J/P
Chenl. 1993. 4. 189.- 193.
(1091 A. Hiittenhofer. S. Hudson. H. F. Noller. P. K. Mascharak. J. B ~ o l .Chem.
1992, 267, 24471 -24475.
[110] B. J. Carter, E. de Vroom. E. C. Long, G. A. van der Marel, J. H. van Boom,
S. M. Hecht, Proc. Nurl. Acud. Sci. USA 1990. 87, 9373--9377.
[ I l l ] C. E. Holmes, B. J. Carter. S. M. Hecht, Biochrmistry 1993. 32. 4293-4307,
and references therein.
[112] R. S. Magliozzo. J. Peisach. M. R. Ciriolo, Mol. Phurmucol. 1989, 35. 428432.
[113] R. P. Hertzberg, P. B. Dervan. J. Am. Chem. Soc. 1982. 104, 313-315.
[I141 C. H. L. Kennard, Inorx. Chin?.Actu 1967. 1. 347-354.
[I151 T. D. Tullius, B. A. Dombroski. Science 1985. 230. 679 681.
(1161 T. D. Tullius, Trmd, Biochem Sci. 1987. 297-300.
[117] T. D. Tullius. Nuture (Lotidon) 1988. 332. 663-664.
[118] S:B. Lin, K. R. Blake. P. S. M~ller.P. 0. P. Ts'o, Biochemistrj. 1989. 28,
1054-1061.
[I 191 R. S. Youngquist, P. B. Dervan. PUK. Null. Acod. S(,i. USA 1985. 82. 25652569.
[120] R. P. Hertzherg, S. M. Hecht. V. L. Reynolds. I. J. Molineux, L. H. Hurley.
Biochemistry 1986. 25, 1249 - 1258.
[121] S. I. Gunderson, K. A. Chapman. R. R. Burgess. Biochiwisrrv 1987, 26.
1539-1546.
[122] M. E. A. Churchill. J. J. Hayes. T. D. Tullius, Biochemistry 1990. 29, 60436050.
11231 M. A. Price. T. D. Tullius. Biuchemi.\rr~~1993. 32. 127-136; A. M. Burkhoff.
T. D. Tullius, Cell 1987. 48, 935-943.
[I241 M. E. A. Churchill, T. D. Tullius. N. R. Kallenhach. N. C. Seeman, Proc
Nuit. Acud. Sci. U S A 1988,85. 4653-4656.
I1251 M. W. Van Dyke. R. P. Hertzherg, P. B. Derwan, Proc. Nutl. A u d . Sci. USA
1982,79,5470-5474: L. H. Hurley. C . 3 . Lee. J. P. McGovren. M. A. Warpehoski. M. A. Mitchell. R. C. Kelly. P. A. Aristoff. Biochemi.str)~1988. 27.
388663892,
11261 M. Sawadogo. R . G. Roeder, Cdl 1985. 43, 165-175: H.-M. Jantren, U.
Strihle, B. Gloss, F. Stewart. W. Schmid, M. Boshart, R. Miksicek, G. Schiitz.
ihid. 1987. 49, 29-38; N. F. Landolfi. X.-M. Yin. J. D. Capra. P. W. Tucker.
Biotecliniques 1989. 7. 500- 504.
[127] C. P. H . Vary. J. N . Vournakis. Proc. Null. Acurl. SC.I.USA 1984. 81, 697X6982.
[128] 1. L. Cartwright. R. P. Hertzherg, P. B. Dervan. S. C. Elgin. Proc. Nuil. Acud.
Sci. U S A 1983.80.3213-3217:H. Richard-F0y.G. L. Hager, EMBOJ. 1987.
6. 2321 -2328.
768
[I291 T. D. Tullius, B. A. Domhroski, M. E. A. Churchill. L. Kam. Methods E n q mol. 1987. 155, 537-558.
[I301 P. B. Dervan. Science 1986. 232, 464-471.
[I311 a) R. S. Youngquist. P. B. Dervan. J. Ani. Chem. Soc. 1985, 107. 5528-5529:
h) W. S. Wade. P. B. Dervan. ihid. 1987, 109.1574- 1575: c) P. G. Schultz, J. S.
Taylor. P. B. Dervan, ihid 1982, 104, 6861 -6863.
[132] J. H. Griffin. P. B. Dervan, J. Am. Chein. Sor. 1986. 108. 5008-5009.
[133] a) J. P. Sluka, S. J. Horvath, M. F. Bruist, M. I. Simon, P. B. Dervan, Science
1987,238. 11 29 - 1132: b) R. S. Youngquist, P. B. Dervan, J. An?. Chem. Soc.
1987, 109, 756447566,
[I341 a ) A. S. Boutorin. V. V. Vlassov, S. A. Kazakov, I. V. Kutiavin, M. A. Podyminogin, FEBS Lett. 1984, 172. 43-46; h) B. C. F. Chu, L. E. Orgel. Proc.
Null. Arud. Sci. U S A 1985.82.963-967: c) G. B. Dreyer, P. B. Dervan, ihid.
1985. 82, 968-972: d) M. Boidot-Forget, N. T. Thuong. M. Chassignol, C.
Helene, C. R. Acud. Sci. Ser. 2 1986. 302, 75-80.
[I351 H. E. Moser. P. B. Dervan. Scicwce 1987.238. 645-650: S . A. Strohel, H. E.
Moser, P. B. Dervan. J. A m . Chem. Sue. 1988. 110. 7927-7929: T. J. Povsic,
P. B. Dervan. ihid. 1989. I l l , 3059-3061; S. A. Strohel, P. B. Dervan, ibid.
1989, 111, 728667287; D. A. Horne. P. B. Dervan. ibid. 1990, 112, 24352437: S. A. Strohel, P. B. Dervan, Science 1990, 249. 73-75: L. C. Griffin,
L. L. Kiessling, P. A. Beal, P. Gillespie. P. B. Dervan, J. Am. Chem. Soc. 1992,
114. 7976- 7982; H. Han. P. B. Dervan, Proc. Nut/. Arud. Sci.U S A 1993. 90,
3806-3810.
[136] S. Prince. F. Korher. P. R. Cooke, J. R. Lindsay Smith, M. A. Mazid, Actu
Crystultugr. Seri. C 1993, 49, 1158 - 1160.
[I371 B. Ward. A. Skorobogaty, J. C. Dabrowiak. Biochmiisrry 1986, 25, 68756883.
[138] a) J. Bernadou, G . Pratviel, F. Bennis, M. Girardet, B. Meunier. Biocheniirlry
1989.28. 7268-7275: h) E. Fouquet. G. Pratviel, J. Bernadou, B. Meunier, J.
Chrfii.Sue. Chrm. Commun. 1987. 1169-1171;c)G. Pratviel, J. Bernadou. M.
Ricci. B. Meunier. Biochem. Bioph.vs. Res. Commun. 1989,160, 1212-1218; d)
B. Ward. R. Rehfuss, J. C. Dabrowiak. J. Biomo/. Slruci. Dyn. 1987, 4, 685695: e) J. C. Dabrowiak. B. Ward, J. Goodisman, Biochemistry 1989, 28.
3314-3322.
[139] R. J. Fiel, T. A. Beerman, E. H. Mark. N. Datta-Gupta, Biochein. Biuphys.
Rij.s. Cuminun. 1982. 107. 1067 -1074.
[I401 R. W. Byrnes, R. J. Fiel, N. Datta-Gupta, ('hem. Biol. Interucr. 1988, 67.
225- 241
[I411 M. Rodriguez. T. Kodadek. M. Torres, A. J. Bard, Biuconjugute Chem. 1990.
I . 123 -131.
11421 J. M. Kelly. M. J. Murphy. D. J. McConnell. C. OhUigin, Nucteic Acid$ Res.
1985, 13, 167-184; D. Praseuth. A. Gaudemer, J. B. Verlhac, 1. Kraijic, I.
Sissoeff, E. GuillC., Phorocheni. Photohiol. 1986. 44,717-724: C. Sentagne, B.
Meunier. N . Paillous. Photochem. Phorohiol. 1992, 16.47-59: D. T. Croke.
L. Perrouault, M. A. Sari, J. P. Battioni. D . Mansuy, C. Helene. T. Le Doan,
ihid. 1993, 18. 41 -50; T. Le Doan. L. Perrouault, M. Rougee, R. Bensasson,
C. Heline in Pliolodq'namic Thrrupy of ~ ~ m o u rund
, s Orher Diseuses (Eds: G.
Jori, C. Perria). Libreria Progetto, Padova. Italy. 1985, pp. 56-58.
11431 a) S D. Bromley, B. Ward. J. C. Dabrowiak, Nuclei Acids Re.,. 1986, 14,
9133-9148; h) B. Ward. A. Skorohogaty, J. C. Dabrowiak, Biochemistry
1986.25.7827-7833; c j G . Raner. B. Ward, J. C. Dabrowiak, J. Coord. Chem.
1988, 19, 17-23; d ) Porphyrins (11. Probm of DNA Structure und Drug-DNA
Inreroc /ions G. Raner, J. Goodisman. J. C. Dabrowiak, ACS Symp. Ser.
1989. 402, 74-89: e) R. F. Pasternack, E. J. Gihbs. ibid. 1989, 402, 59-73; f)
R. F. Pasternack. E. J. Gihhs, J. J. Villafranca, Biochemistry 1983, 22, 5409 5417: g) J. A. Strickland, L. G. Marzilli, K. M. Gay, W. D. Wilson, h d . 1988.
27. 8870 8878; h) R. .I.Fiel, J. Biomol. Slruct. Dyn.1989, 6. 1259-1273: i)
L. G. Marzilli, New J. Chem. 1990, 14, 409-420.
[144] X. Hui. N. Gresh, 8. Pullman, Nucleic Acids Re.s. 1990, 18. 1109-1114: R.
Lavery, B. Pullman,J Biomol. Slrnci. Dyn. 1985.2.1021 -1032: P. K. Weiner.
R. Langeridge, J. M. Blaney. R. Schaefer, P. A. Kollman, Pruc. Nurl. Acud.
Sci. USA 1982, 79, 3754-3758.
[I451 L. G. Marzilli, G. Petho, M. Lin. M. S. Kim. D. W. Dixon, J. Am. Chem. Soc.
1992. 114. 7575- 7577; M. A. Sari. J. P. Battioni. D. Duprk, D. Mansuy. J. B.
Le Pecq. Biochenii.stry 1990. 39, 4205-4215.
[146] R. B. Van Atta, J. Bernadou. B. Meunier. S. M. Hecht. Biochemistry 1990,29.
4783 -4789.
[147] M. Pitie. G. Pratviel. I. Bernadou. B. Meunier. Proc. Null. Acud. Sci. USA
1992. 89. 3967-3971
[I481 L. Ding, J. Bernadou. B. Meunier. Bioconjitgurr Chem. 1991, 2, 201-206.
[149] C . Casas, C. J. Lacey. B. Meunier. Bioconjugute Chem. 1993, 4. 366--371;
M. Pitie, C. Casas. C. J. Lacey, G. Pratviel, J. Bernadou. B. Meunier,
Anpew. Chrm. 1993. 105, 607-609: Angczn,. Chem. Inl. Ed. Engl. 1993, 32,
557-559.
[ISO] a) A. M. Pyle, J. K. Barton, Prog. Inorg. Cliern. 1990, 38, 413-475; hj N. J.
Turro, J. K. Barton, D. A. Tomalia. Arc. Chem. Res. 1991, 24, 332-340:
c) J. K . Barton. Si,kncr 1986. 233. 727-734: d ) C. V. Kumar, J. K. Barton,
N. J. Turro, J. Am. C h ~ m Soc.
.
1985, 107, 5518-5523: e) M. B. Fleisher,
K. C. Waterman. N. J. Turro, J. K . Barton. Inorg. Cliein. 1986. 25, 35513552.
I1511 E. C. Long. J. K. Barton, Arc. Chtvn. Res. 1990, 23, 273 279.
'
A n g w . Chem. Inl. Ed. Engl. 1995, 34, 746-769
Nonenzymatic
DNA Cleavage
[152] J. P. Rehmann, J. K. Barton. Biochemistry 1990, 29. 1701-1709, 1710-1717.
(1531 S. Satyanarayana. J. C. Dabrowiak. J. B. Chaires, Biochemistry 1993, 32,
2573-2584; S. Satyanarayana, J. C. Dabrowiak, J. B. Chaires, ibid. 1992.31,
9319-9324: C . Hiort. B. Norden, A. Rodger, J. Am. Chem. Soc. 1990, 112.
1971- 1982: C. Hiort. P. Lincoln, B. Norden, [bid. 1993, 115, 3448-3454;
A. E.Friedman, C. V. Khumar. N. J. Turro, J. K. Barton, Nucleic Acids Res.
1991. IY. 2595-2602
[154] a) A M. Pyle. E. C. Long. J. K. Barton, J. Am. Chem. Soc. 1989. 111. 45204522; b) M. R. Kirshenbaum. R. Tribolet, J. K. Barton, Nucleic Acids Res.
1988, 16. 7943- 7960.
[155] N . Grover, H. H . Thorp, J. Am. Chem. Soc. 1991, 113, 7030-7031; N. Grover. N. Gupta. P. Singh, H. H. Thorp, Inorg. Chem. 1992, 31, 2014-2020;
H . Y Mei. J. K. Barton, Proc. Natl. Acad. Scr. U S A 1988, 85, 1339-1343;
J. M. Kelly, D. J. McConnell, C. OhUigin, A. B. Tossi. A. Kirsch-De Mesmaeker. A . Masschelein, J. Nasielski. J. Chem. SOC.Chem. Commun. 1987,
1821 -1823; H. Y. Me1.J. K . Bart0n.J. Am. Chem. SOC.1986, 108.7414-7416.
[156] B. C. Miiller. A L. Raphael. J. K . Barton, Proc. Narl. Acad. Sci. U S A 1987,
84.1764--3768; R. J. Baseman. D . W. Pratt. M. Ch0w.P. D0wd.J. A m . Chem.
Soc. 1976. 98% 5726.
[157] A. Sitlani, E C Long. A. M. Pyle, J. K . Barton, .IAm. Chem. Soc. 1992, 114,
2303 -2312.
[158] S. S. David. J. K. Barton. J. Am. Chem. Soc. 1993, 115, 2984-2985.
[159] In Schemes 13 to 18 all products in frames have been detected and fully (or at
least partially) characterized. A negative charge has been omitted for clarity
on phosphodiesters and terminal phosphates.
[160] T. E. Goyne. D . S. Sigman, J. Am. Chem. Soc. 1987, 109, 2846-2848.
[I611 L. S. Kappen. C . Chen. I. H. Goldberg, Biochemisiry 1988, 27,4331-4340.
11621 L. S. Kappen. I H . Goldberg, Biochemistry 1989.28, 1027-1032.
[163] M Manoharan. A. Mazumder, S. C. Ransom, J. A. Gerlt. J. Am. Chem. Soe.
1988, 110. 2690-2691; V. Bailly. W. G. Verly. Nucleic Acids Res. 1988. 16,
9489- Y496.
[164] a) J. Bernadou. B. Lauretta. G. Pratviel. B. Meunier. C.R. Acad. Sci. Ser. 3
1989.309, 409-414: b) G. Pratviel, M. Pitie, J. Bernadou, B. Meunier, Angew.
Chrm. 1991, 103. 718-720; Angew. Chem. Int. Ed. Engl. 1991. 30, 702-704;
c) G. Pratviel. M. Pitie, J. Bernadou, B. Meunier, Nucleic Acids Res. 1991, 19,
6283-6288. d ) G. Gasmi, M. Pasdeloup. G. Pratviel, M. Pitie, J. Bernadou,
B. Meunier. ibid. 1991, 19, 2835-2839.
11651 H. Sugiyama. Y. Tsutsumi, I. Saito. J. A m . Chem. Soe. 1990,112,6720-6721.
[166] H. Sugiyama. Y. Tsutsumi. K. Fujimoto. I. Saito.2 Am. Chem. Soc. 1993,115.
4443-4448.
[167] M. Dizdaroglu. D. Schulte-Frohlinde. C. von Sonntag, Z. Naturfbrsch. C
1977, 32. 1021 1022.
[l68] W. D Hemmer. S. M. Grunberg. W. A. Haseltine. J. Biol. Chem. 1982, 257,
1 1 750- 11 754: W. D. Henner, L. 0 .Rodriguez. S. M. Hecht, W. A. Haseltine,
ihid 1983. 258, 711-713.
[169] I Saito. H. Kawabata. T. Fujiwara. H. Sugiyama, T. Matsuura, J. Am. Chem.
S o c 1989, 111, 8302-8303.
[170] B. L Frank. L. Worth, Jr.. D. F. Christner. J. W. Kozarich, J. Stubbe, L. S .
Kappen. I. H . Goldberg. J. Am. Chem. Soc. 1991. 113. 2271-2275.
[I711 P. C. Dedon. I . H. Goldberg, J. Biol. Chem. 1990. 265. 14713-14716.
[172] T. Hatdyama. 1. H. Goldberg, Biochemistry 1980, 19. 5890-5898.
[173] G. H. McGall. L. E. Rabow, G. W. Ashley, S. H . Wu. J. W. Kozarich, J. Stubbe, J Am. Chiwr. Soc. 1992, 114. 4958-4967.
[174] D. Schulte-Frohlinde. C. von Sonntag in Oxidative Srress (Ed.: H. Sies),
Academic Press. London, 1985, pp. 11 -40.
[175] C . von Sonntag. Adr. Curbohydr. Chem. Biochem. 1980, 37. 7-77.
[176] G. H. McGall, J. Stubbe, J. W. Kozarich. J. Org. Chrm. 1991, 56, 48-55.
11771 a ) H. Sugiyama. C. Xu, N. Murugesan. S. M . Hecht, G. A. van der Marel.
J. H. van Boom. Bio&misiry 1988, 27. 58-67; b) H. Sugiyama, H.
Kawabata. T. Fujiwara, Y. Dannoue, I. Saito, .I Am. Chem. Sac. 1990, 112.
5 2 5 2 ~5257.
[178] R. M. Burger. S . J. Prolan. S. B. Horwitz. J. Peisach, J. Biol. Chem. 1986,261,
15955- 35959.
[I791 N.Murugesan. C. Xu. G. M. Ehrenfeld. H . Sugiyama, R. E. Kilkuskie, L. 0.
Rodriguez. L. H. Chang, S. M. Hecht. Biorhemistr?; 1985, 24. 5735-5744.
[I801 J. C . Wu. J. W. Kozarich, J. Stubbe, J. Biol. Chem. 1983. 258.4694-4697.
[181] J. W. Kozarich. L. Worth. Jr.. B. L. Frank, D . F. Christner, D . E. Vanderwall.
J. Stubbe. S c i r w e 1989, 245. 1396-1398.
I1821 L. Worth. Jr.. 9. L. Frank. D. F. Christner, M. J. Absdlon, J. Stubbe. J. W.
Kozarich. Biochemistrr 1993, 32, 2601 -2609.
~
An.qcw. Chmi. Int. Ed. Engl. 1995, 34, 746-769
REVIEWS
[183] a) K. U. Ingold, Arc. Chem. Res. 1969, 2, 1-9; b) C. von Sonntag, Inr. J.
Radial. Relat. Stud. Phys. Chem. Med. 1984, 46, 507-518; c) E. Bothe, G.
Behrens, E. Bohm, B. Sethuram. D. Schulte-Frohlinde,ibid. 1986,49. 57-66.
I
Am. Chem. So(. 1990, 112, 3169[184] L. E. Rabow, J. Stubbe, J. W. Kozarich, .
3203.
[185] H. Sugiyama, C. Xu, N. Murugesan, S. Hecht, J. Am. Chrm. Soc. 1985, 107,
4104-4105.
[I 861 T. G. Wenzel, C. H. Chang, C. F. Meares, Biochemrsrr.v 1985, 24, 3060-3069.
[1871 I. Saito, T. Morii, H. Sugiyama, T. Matsuura. C. F. Meares, S. M. Hecht, J.
Am. Chem. Soe. 1989, 1 1 I , 2307-2308.
[ 1881 S. M. Hecht in Bleomyan ;Chemical, Biochemical and Biologzcal Aspects (Ed. :
S . M. Hecht), Springer, New York, 1979, pp. 1-23.
[I891 C. R. Krishnamoorthy, D . E. Vanderwall, J. W. Kozarich. J. Stubbe, J. Am.
Chem. SOC.1988, 110, 2008-2009.
[190] M. J. Absalon, C. R. Krishnamoorthy, G. McGall. J. W. Kozarich, J. Stubbe,
Nucleic Acids Res. 1992. 20, 4179-4185.
[191] R. J. Duff, E. deVroom, A. Geluk, S. M. Hecht.J Am. Chem. Soc. 1993. 115.
3350-3351.
[192] J. P. Slukd, S. J. Horvath, A. C. Glasgow, M. I. Simon, P. B. Dervan. Biochemisfr.? 1990, 29,6551 -6561 ; D . P. Mack, J. P. Sluka, J. A. Shin, J. H . Griffin,
M. I. Simon. P. B. Dervan. ihid. 1990, 29. 6561-6567.
[193] M . G. Oakley, P. B. Dervan, Scienre 1990,248, 847-850; M. G. Oakley, M.
Mrksich, P. B. Dervan, Biochemistry 1992, 31, 10969-10975.
[194] D.-H. Chin, S. A. Carr, I. H. Goldberg, 1 Biol. Chem. 1984,259. 9975-9978.
[195] L. S. Kappen. I. H. Goldberg, J. M. Liesch, Proc. Natl. Acad. Sci. U S A 1982,
79, 744-748.
[196] L. S. Kappen. I. H . Goldberg, Biochemistry 1983, 22. 4872-4878.
[197] D.-H. Chin, L. S. Kappen, I. H . Goldberg, Proc. Narl. Acad. Sei. U S A 1987.
84,7070- 7074.
[198] L. S. Kappen, T. R. Lee, C. Yang, I. H. Goldberg, Biochemistry 1989, 28,
4540-4542.
[199] H. Kawahata, H. Takeshita, T. Fujiwara. H. Sugiyama, T. Matsuura, I. Saito,
Tetrahedron Lett. 1989, 30, 4263-4266.
[200] G. Pratviel, M. Pitie, C. Perigaud, G. Gosselin, J. Bernadou, B. Meunier, J.
Chem. Sor. Chem. Commun. 1993, 149-151.
[201] M. Pitie, G. Pratviel, J. Bernadou, B. Meunier in The Activation ofDioxygen
andHomogeneous CaralyficOxidation (Eds.: D . H . R. Barton, A. E. Martell,
D . T. Sawyer), Plenum, New York, 1993, pp. 333-346.
f2021 G. Pratviel, V. Duarte. J. Bernadou, B. Meunier, J. Am. Chem. Soc. 1993, 115,
7939 - 7943.
[203] V. W Bowry, K . U. Ingold. .
IAm. Chem. Soc. 1991, 113. 5699-5707; J. K.
Atkinson. K. U. Ingold, Biochemistry 1993, 32, 9209-9214.
[204] L. Ding, G. Etemad-Moghadam, B. Meunier, Biochemistry 1990,29, 78687875: L. Ding, G. Etemad-Moghadam, S. Cros, C. Auclair, B. Meunier, J.
Med. Chem. 1991,34,900-906.
12051 P. S. Pendergrast, Y. W. Ebright, R. H. Ebright, Science 1994,265,959-962.
[206] C. Q. Pan, J. A. Feng, S. E. Finkel, R. Landgraf. D. Sigmdn, Proc. Natl. Acad.
Sci. U S A 1994, 91, 1721-1725.
[207] 0. D. Hensens, D.-H. Chin. A. Stassinopoulos, D. L. Zink, L. S. Kappen.
1. H . Goldberg, Proc. Nafl. Acad. Sci. USA 1994, 91. 4534-4538: A. G.
Myers, S . B. Cohen. B. M. Kwon, J. Am. Chem. Soc. 1994, 116,1670-1682.
[208] T. Li. Z. Zeng, V. A. Estevez, K . U. Baldenius. K. C. Nicolaou, G. F. Joyce,
J. Am. Chem. Sor. 1994, 116, 3709-3715.
[209] P. Fulmer, D. H . Petering, Biochemistry 1994, 33, 5319-5327.
[210] J. W. Sam. X. J. Tang, J. Peisach, J. Am. Chem. SOC.1994, 116, 5250-5256.
[2111 R. J. Guajardo, S. E. Hudson, S. J. Brown, P. K . Mascharak, .I Am. Chem.
Soc. 1993, 115, 7971 -7977.
[212] M. G. Oakley, K . D. Turnbull, P. B. Dervan, Bioconpgate Chem. 1994. 5.
242-241.
[213] M. A. Morgan, S. M. Hecht, Biochemistry 1994, 33, 10286-10293.
[214] H. Han, A. Schepartz, M. Pellegrini, P. B. Dervan, Biochemistry 1994, 33.
9831 -9844.
[215] H. Han. P. B. Dervan, Proc. Natl. Acad. Sci. USA 1994. 91,4955-4959.
[216] J. Bernadou, A. S. Fahiano, A. Robert, B. Meunier. J. Am. Chem. Soc. 1994,
116,9375-9376.
[217] U. Sehlstedt, S. K. Kim, P. Carter, J. Goodisman, J. F. Vallano. B. Norden,
J. C. Dabrowiak, Blochemistry 1994,33,417-426.
I2181 M. Eriksson, M. Leijon, C. Hiort, B. Norden, A. GrCslund, Biochemistry
1994, 33, 5031 -5040.
[219] D. Campisi, T. Morii. J. K. Barton, Biochemistry 1994, 33, 4130-3139.
769
Документ
Категория
Без категории
Просмотров
4
Размер файла
2 945 Кб
Теги
unit, bond, drug, chemical, target, dna, sugar, carbonчhydrogen, nuclease
1/--страниц
Пожаловаться на содержимое документа