Steric and Electronic Effects, Enantiospecificity, and Reactive Orientation in DNA Binding/ Cleaving by Substituted Derivatives of [SalenMnIII]+

Article abstract of DOI:10.1021/ic960196x

Twenty-six derivatives of [SalenMn^III]⁺ (1) bearing halogen, nitro, amino, ether, alkyl, or aryl substituents on the aromatic rings and/or at the imine positions or containing 1,3-propylene-, 1,2-phenylene-, 1,2-cyclohexane-, or 1,2-diphenylethylenediamine in place of ethylenediamine as the bridging moiety have been synthesized. The DNA binding/cleaving properties of these complexes in the presence of terminal oxidants have been examined using DNA affinity cleaving techniques. Active derivatives produced DNA cleavage from the minor groove at sites containing multiple contiguous A:T base pairs. For aryl-substituted derivatives, DNA cleavage efficiency was found to vary with both the identity and position of attachment of substituents. The precise patterns of cleavage at A:T target sites varied with the position of attachment of substituents, but not with the identity of the substituents. The results suggest that substituents alter specificity through both steric and electronic effects. The 3,3′-difluoro and -dichloro derivatives produced cleavage patterns that match those of the parent complex, suggesting that the activated form of 1 produces cleavage from an orientation in which the concave edge of the complex faces away from the floor of the DNA minor groove. Bridge modifications yield complexes with reduced DNA cleaving activity relative to 1. DNA cleaving efficiency was found to vary with both the structure and stereochemistry of the bridge. Cleavage efficiency for the complex derived from (R,R)-cyclohexanediamine was 5 times greater than that for the (S,S) enantiomer. Cleavage patterns produced by the enantiomeric complexes at A:T rich target sites were different, demonstrating enantiospecific recognition and cleavage of right-handed double-helical DNA.

Full text of DOI:10.1021/ic960196x

Inorg. Chem. 1996, 35, 4837-4847
4837
Steric and Electronic Effects, Enantiospecificity, and Reactive Orientation in DNA Binding/
Cleaving by Substituted Derivatives of [SalenMnIII]+
Dennis J. Gravert and John H. Griffin*
Department of Chemistry, Stanford University, Stanford, California 94305-5080
ReceiVed February 22, 1996X
Twenty-six derivatives of [SalenMnIII]+ (1) bearing halogen, nitro, amino, ether, alkyl, or aryl substituents on the
aromatic rings and/or at the imine positions or containing 1,3-propylene-, 1,2-phenylene-, 1,2-cyclohexane-, or
1,2-diphenylethylenediamine in place of ethylenediamine as the bridging moiety have been synthesized. The
DNA binding/cleaving properties of these complexes in the presence of terminal oxidants have been examined
using DNA affinity cleaving techniques. Active derivatives produced DNA cleavage from the minor groove at
sites containing multiple contiguous A:T base pairs. For aryl-substituted derivatives, DNA cleavage efficiency
was found to vary with both the identity and position of attachment of substituents. The precise patterns of
cleavage at A:T target sites varied with the position of attachment of substituents, but not with the identity of the
substituents. The results suggest that substituents alter specificity through both steric and electronic effects. The
3,3′-difluoro and -dichloro derivatives produced cleavage patterns that match those of the parent complex, suggesting
that the activated form of 1 produces cleavage from an orientation in which the concave edge of the complex
faces away from the floor of the DNA minor groove. Bridge modifications yield complexes with reduced DNA
cleaving activity relative to 1. DNA cleaving efficiency was found to vary with both the structure and
stereochemistry of the bridge. Cleavage efficiency for the complex derived from (R,R)-cyclohexanediamine was
5 times greater than that for the (S,S) enantiomer. Cleavage patterns produced by the enantiomeric complexes at
A:T rich target sites were different, demonstrating enantiospecific recognition and cleavage of right-handed double-
helical DNA.
Introduction
cleaving properties of metalloporphyrins,6 octahedral cobalt and
rhodium complexes,7 bis(1,10-phenanthroline):copper com-
plexes,8 and macrocycle:nickel complexes,9 have prompted us
to consider and probe the effects that structural modifications
have on the DNA binding/cleaving properties of 1.
We recently reported that [SalenMnIII]+ (1, Salen ) N,N ′-
ethylenebis(salicylideneaminato)) mediates the cleavage of right-
handed double-helical DNA in the presence of terminal oxi-
dants.1 The combination of 1 and oxidant produces single-strand
cleavage as well as double-strand cleavage which derives from
independent, proximal nicks on opposite DNA strands. Cleav-
age is effected selectively from the DNA minor groove in
segments rich in A:T base pairs and is consistent with a
mechanism involving deoxyribose C-H activation by an
oxidatively activated intermediate such as [SalenMnVO]+.2,3
Complex 1 produces cleavage in regions of DNA that are
targeted by netropsin-like affinity cleaving agents and neocarzi-
nostatin but only weakly cleaved by bleomycin:Fe and methidi-
umpropyl-EDTA:Fe.4 These findings, when coupled with the
knowledge that substituents modulate the properties of 1 as a
catalyst for hydrocarbon oxidation2,5 and that changes in ligand
structure and stereochemistry can alter the DNA binding/
In this paper, we report the synthesis of 26 derivatives of the
crescent-shaped, cationic parent complex bearing symmetrical
substitutions on the aryl rings, at the imine positions, or on the
bridge and the analysis of their DNA binding/cleaving proper-
ties. We find that the addition of substituents to 1 has large
effects on DNA binding/cleaving efficiency and induces changes
in cleavage patterns within A:T rich target sites. The results
have a number of implications for the role(s) of substituent
effects in DNA recognition and cleavage by [SalenMnIII]+
(6) (a) Bromley, S. D.; Ward, B. W.; Dabrowiak, J. C. Nucleic Acids
Res. 1986, 14, 9133-9148. (b) Raner, G.; Ward, B; Dabrowiak, J. C.
J. Coord. Chem. 1988, 19, 17. (c) Sehlstedt, U.; Kim, S. K.; Carter,
P.; Goodisman, J.; Vollano, J. F.; Norde´n, B.; Dabrowiak, J. C.
Biochemistry 1994, 33, 417-426.
(7) (a) Raphael, A. L.; Barton, J. K. J. Am. Chem. Soc. 1984, 106, 2466-
2468. (b) Raphael, A. L.; Barton, J. K. Proc. Natl. Acad. Sci. U.S.A.
1985, 82, 6460-6464. (c) Pyle, A. M.; Long, E. C.; Barton, J. K. J.
Am. Chem. Soc. 1989, 111, 4520-4522. (d) Pyle, A. M.; Morii, T.;
Barton, J. K. J. Am. Chem. Soc. 1990, 112, 9432-9434. (e) Sitlani,
A.; Long, E. C.; Pyle, A. M.; Barton, J. K. J. Am. Chem. Soc. 1992,
114, 2303-2312. (f) Krotz, A. H.; Kuo, L. Y.; Shields, T. P.; Barton,
J. K. J. Am. Chem. Soc. 1993, 115, 3877-3882. (g) Krotz, A. H.;
Hudson, B. P.; Barton, J. K. J. Am. Chem. Soc. 1993, 115, 12577-
12578. (h) Sitlani, A.; Dupureur, C. M.; Barton, J. K. J. Am. Chem.
Soc. 1993, 115, 12589-12590. (i) Campisi, D.; Morii, T.; Barton, J.
K. Biochemistry 1994, 33, 4130-4139. (j) Sitlani, A.; Barton, J. K.
Biochemistry 1994, 33, 12100-12108.
* To whom correspondence should be addressed: phone (415) 723-3625;
fax (415) 725-0259; email jgriffin@leland.stanford.edu.
X Abstract published in AdVance ACS Abstracts, July 15, 1996.
(1) Gravert, D. J.; Griffin, J. H. J. Org. Chem. 1993, 58, 820-822.
(2) Srinivasan, K.; Michaud, P.; Kochi, J. K. J. Am. Chem. Soc. 1986,
108, 2309-2320.
(3) In preliminary studies of the cleavage mechanism, we have determined
that the combination of [SalenMnIII]+ and oxidant produces direct DNA
strand breaks and base-labile lesions through processes that do not
involve dioxygen. DNA cleavage releases free, unmodified nucleotide
bases. The DNA termini produced consist of 3′ phosphate groups as
well as 5′ phosphate and 5′ nonphosphate, nonhydroxyl groups that
are converted to phosphates by base treatment. Addition of substituents
to [SalenMnIII]+ does not substantially alter the observed reaction
pathways.
(8) (a) Sigman, D. S. Acc. Chem. Res. 1986, 19, 180-186. (b) Thederahn,
T. B.; Kuwabara, M. D.; Larsen, T. A.; Sigman, D. S. J. Am. Chem.
Soc. 1989, 111, 4941-4946.
(4) Griffin, J. H. Bioorg. Med. Chem. Lett. 1995, 5, 73-76.
(5) Jacobsen, E. N.; Zhang, W.; Gu¨ler, M. L. J. Am. Chem. Soc. 1991,
113, 6703-6704.
(9) (a) Muller, J. G.; Chen, X.; Dadiz, A. C.; Rokita, S. E.; Burrows, C.
J. J. Am. Chem. Soc. 1992, 114, 6407-6411. (b) Burrows, C. J.;
Rokita, S. E. Acc. Chem. Res. 1994, 27, 295-301.
S0020-1669(96)00196-6 CCC: $12.00 © 1996 American Chemical Society

4838 Inorganic Chemistry, Vol. 35, No. 17, 1996
Gravert and Griffin
Table 2. (Chiral) Bridge-Substituted [SalenMnIII]+ Derivatives
Table 1. Ring- and Imine-Substituted [SalenMnIII]+ Derivatives
Scheme 1
derivatives. Most significantly, our data (1) provide evidence
that both steric and electronic effects alter the DNA binding/
cleaving specificity exhibited by derivatives of 1, (2) demon-
strate that chiral derivatives of 1 exhibit enantiospecific
recognition of DNA, and (3) indicate orientations relative to
DNA from which 1 and its derivatives produce DNA cleavage.
Results
Synthesis and Characterization. Substituted derivatives of
the parent ligand SalenH2 were prepared by condensation of
salicylaldehyde derivatives with ethylenediamine. Where not
commercially available, substituted salicylaldehydes were pre-
pared via the Reimer-Tiemann reaction10 or through lithium
aluminum hydride reduction followed by manganese dioxide
oxidation of a substituted salicylic acid. Stirring the Schiff base
ligands with Mn(OAc)2 under aerobic conditions afforded the
requisite Mn(III) complexes.11 One set of derivatives studied
were those bearing symmetrical substitutions of fluoro, chloro,
bromo, nitro, methyl, tert-butyl, methoxy, or diethylamino
groups at the 3,3′-, 4,4′-, 5,5′-, or 6,6′-positions (2-14, 19, and
20; Table 1), phenyl or methyl substituents at the 7,7′-imine
positions (15-17), and the naphthalene analog 18. In another
set of derivatives, the 1,2-ethane bridge was replaced with 1,3-
propane (21), 1,2-phenyl (22), (R,R)-, (S,S)-, or meso-1,2-
cyclohexane (23), or (R,R)- or (S,S)-1,2-diphenylethane (24)
linkages (Table 2). The complexes were characterized by
spectroscopic methods (FT-IR, UV-vis, FAB-MS), melting
(decomposition) point measurements, magnetic susceptibility
measurements, elemental analysis, and optical rotation where
appropriate.
DNA Affinity Cleaving.12 Scheme 1 outlines our general
approach to characterizing the DNA binding/cleaving properties
of [SalenMnIII]+ derivatives. DNA double-strand affinity cleav-
ing was used to screen for DNA binding/cleaving activity, to
determine relative cleavage efficiencies, and to define cleavage
(10) Postmus, C., Jr.; Kaye, I. A.; Craig, C. A.; Matthews, R. S. J. Org.
Chem. 1964, 29, 2693-2698.
(11) Zhang, W.; Jacobsen, E. N. J. Org. Chem. 1991, 56, 2296-2298.
(12) (a) Schultz, P. G.; Taylor, J. S.; Dervan, P. B. J. Am. Chem. Soc.
1982, 104, 6861-6863. (b) Taylor, J. S.; Schultz, P. G.; Dervan, P.
B. Tetrahedron 1984, 40, 457-465. (c) Dervan, P. B. Science 1986,
232, 464-470.
specificity at a resolution of 25-50 base pairs. This was
followed by high-resolution analysis of cleavage patterns

[SalenMnIII]+ Derivatives in DNA Binding/Cleaving
Inorganic Chemistry, Vol. 35, No. 17, 1996 4839
produced on restriction fragments containing loci of intense
double-strand cleavage.
Substrates for the affinity cleaving studies were derived from
plasmid pBR322 (4363 base pairs, bp). For the double-strand
cleavage assays, pBR322 was linearized with StyI, which cleaves
this plasmid at position 1369 within the sequence 5′-C/CTTGG-
3′. Given the nonpalindromic nature of this cleavage site, it
was possible to independently label the linearized plasmid at
each of the resulting unique recessed 3′ ends by filling them in
with either R-32P dATP (bottom strand) or R-32P TTP (top
strand) using the Klenow fragment of DNA polymerase.
Following incubation of these substrates with 1 and its deriva-
tives in the presence and absence of terminal oxidants, products
of DNA double-strand cleavage were resolved by electrophoresis
on 1% agarose gels and visualized by autoradiography. The
substrates for the high-resolution cleavage assays were 3′- and
5′-32P end labeled, 517 bp EcoRI/RsaI restriction fragments from
pBR322 (spanning positions 3849-4361 on pBR322). The
products of cleavage of these substrates were separated to
nucleotide resolution by electrophoresis on denaturing poly-
acrylamide sequencing gels and visualized by autoradiography.
Double-Strand Cleavage Competency and Efficiency of
Ring- and Imine-Substituted Derivatives. It was found that
complexes 11 and 13-18, which bear ether, amine, or aryl ring
substituents and/or substituents at the imine positions, did not
produce observable double-strand DNA cleavage by themselves
or in the presence of millimolar concentrations of magnesium
monoperoxyphthalate or potassium peroxymonosulfate. (These
terminal oxidants do not produce significant DNA cleavage in
the absence of metal complexes.1) Double-strand cleavage was
effected by combinations of terminal oxidant with complexes
1-10, which bear net electron-withdrawing halogen or nitro
groups, and by 12, which bears methoxy groups at the 4,4′-
positions. The efficiency with which complexes 1-10 and 12
produced double-strand DNA cleavage varied significantly, as
illustrated by the concentrations required to achieve comparable
levels of cleavage in the presence of 1 mM magnesium
monoperoxyphthalate after a reaction time of 1 h at 37 °C
(Figure 1). Complexes 9 (5,5′-Br2), 6 (5,5′-Cl2), and 5 (4,4′-
Cl2) cleave with greater efficiency than does the parent complex
1, complex 10 (5,5′-(NO2)2) cleaves with efficiency similar to
that of 1, and complexes 3 (5,5′-F2), 2 (3,3′-F2), 7 (6,6′-Cl2), 4
(3,3′-Cl2), 8 (3,3′,5,5′-Cl4), and 12 (4,4′-(OCH3)2) cleave with
diminishing efficiency. For the series of dichloro derivatives
4-8, substituents at the 5,5′- or 4,4′-positions afford enhanced
DNA binding/cleaving efficiencies relative to that of 1, while
substitution at the 6,6′-, 3,3′-, or (3,3′ + 5,5′)-positions leads
to cleaving efficiencies that are lower than that of 1 and decrease
in the order listed. For derivatives 3, 6, 9, 10, and 13, which
bear different substituents at the 5,5′-positions, DNA binding/
cleaving efficiency is greatest for the dibromo and dichloro
derivatives, similar to 1 for the dinitro derivative, somewhat
reduced relative to 1 for the difluoro derivative, and greatly
reduced (undetectable) for the dimethoxy derivative. The
relationship between the identity of the substituent at a given
position and the ordering of DNA binding/cleaving efficiencies
is not general. The 3,3′-difluoro derivative 2 cleaves DNA more
efficiently than does the 3,3′-dichloro derivative 4. Complexes
19 and 20 were not tested for DNA double-strand cleavage
activity, but they did produce efficient single-strand cleavage
(see below).
Figure 1. Autoradiographs of DNA double-strand cleavage patterns
produced by ring- and imine-substituted [SalenMnIII]+ complexes on
StyI-linearized, 32P end labeled pBR322 plasmid DNA in the presence
of 1.0 mM magnesium monoperoxyphthlate and resolved by electro-
phoresis on a 1% agarose gel. Top: 3′-32P dATP end labeled DNA.
Bottom: 3′-32P TTP end labeled DNA. Lane 1, DNA incubated in the
absence of metal complex or oxidant; lane 2, pBR322 molecular weight
markers from pBR322 4365, 3371, 2994, 2368, 1998, 1768, 1372, 995,
and 666 bp in length; lane 3, 5,5′-Br2 (9, 12.5 µM); lane 4, 5,5′-Cl2 (6,
12.5 µM); lane 5, 4,4′-Cl2 (5, 15 µM); lane 6, [SalenMnIII]+ (1, 20
µM); lane 7, 5,5′-(NO2)2 (10, 30 µM); lane 8, 5,5′-F2 (3, 50 µM); lane
9, 3,3′-F2 (2, 50 µM); lane 10, 6,6′-Cl2 (7, 50 µM); lane 11, 3,3′-Cl2
(4, 100 µM); lane 12, 3,3′,5,5′-Cl4 (8, 100 µM); lane 13, 4,4′-(OCH3)2
(12, 100 µM).
from that of 1. The observed cleavage loci correspond to
regions of pBR322 containing sites of multiple, contiguous A:T
base pairs, and the loci of greatest cleavage intensity correspond
to the most A:T rich regions of the plasmid (Table 3).
The specificity of DNA cleavage by active derivatives of 1
was then probed at nucleotide resolution using EcoRI/RsaI
restriction fragments from pBR322 which contained the series
of intense cleavage loci spanning positions 4164-4331 on
pBR322. In the autoradiograph shown in Figure 2, complexes
1-10 and 12 have been grouped in order to emphasize the
phenomenon that substituents alter the precise patterns of
cleaVage produced by [SalenMnIII]+ deriVatiVes in a position
specific but substituent independent fashion. Complexes 2 and
Cleavage Specificity of Ring- and Imine-Substituted
Derivatives. The identity and position of attachment of ring
and imine substituents had no observable effect on DNA double-
strand cleavage specificityspatterns of cleavage produced by
2-10 and 12 were, at this level of resolution, indistinguishable

4840 Inorganic Chemistry, Vol. 35, No. 17, 1996
Gravert and Griffin
Figure 2. Autoradiograph of DNA single-strand cleavage patterns produced by ring- and imine-substituted [SalenMnIII]+ complexes on 3′- and
5′-32P end labeled 517 bp restriction fragments (EcoRI/RsaI) from pBR322 plasmid DNA in the presence of 1.0 mM magnesium monoperoxyphthlate.
Cleavage patterns were resolved on a 1:20 cross-linked 8% polyacrylamide/45% urea denaturing gel. Left set of lanes: 3′-32P end labeled DNA.
Right set of lanes: 5′-32P end labeled DNA. For each set of lanes: lane 1, DNA incubated in the absence of metal complex or oxidant; lanes 2,
Maxam-Gilbert chemical sequencing G reaction; lane 3, [SalenMnIII]+ (1, 20 µM); lane 4, 3,3′-Cl2 (4, 100 µM); lane 5, 3,3′-F2 (2, 50 µM); lane
6, 3,3′,5,5′-Cl4 (8, 100 µM); lane 7, 5,5′-Br2 (9, 12.5 µM); lane 8, 5,5′-Cl2 (6, 12.5 µM); lane 9, 5,5′-F2 (3, 50 µM); lane 10, 5,5′-(NO2)2 (10, 30
µM); lane 11, 4,4′-Cl2 (5, 15 µM); lane 12, 4,4′-(OCH3)2 (12, 100 µM); lane 13, 6,6′-Cl2 (7, 50 µM).
4, with fluoro or chloro substituents in the 3,3′-positions,
produce a cleavage pattern that matches that of 1. Complexes
3, 6, 9, and 10, which bear halogen or nitro substituents in the
5,5′-positions, produce a common cleavage pattern which differs
from that produced by 1, 2, or 4. Tetrachloro derivative 8, with
substituents in the 3,3′- and 5,5′-positions, produces a pattern
of cleavage which matches that of 3, 6, 9, and 10. Complexes
5 and 7, with chloro substituents in the 4,4′- and 6,6′-positions,
respectively, generate unique cleavage patterns. Some subtle
differences are noted in the cleavage patterns produced by 5
and the 4,4′-dimethoxy analog 12, but these patterns are more
similar to one another than they are to that of any other
derivative. Relative DNA binding/cleaving efficiencies ob-
served in these experiments, which report single-strand cleavage,
matched those determined by DNA double-strand affinity
cleaving.
The efficiency and specificity of DNA cleavage by the 5,5′-
dimethyl derivative 19 and the 5,5′-dibromo-3,3′-di-tert-butyl
derivative 20 were also determined by high-resolution affinity
cleaving (Figure 3). Complex 19 cleaves DNA with an
efficiency similar to that of complexes 4 and 8 (i.e., with
significantly less efficiency than 1). Complex 19 bears electron-
donating methyl substituents and produces a pattern of cleavage
which is similar but not identical to the patterns of cleavage

[SalenMnIII]+ Derivatives in DNA Binding/Cleaving
Inorganic Chemistry, Vol. 35, No. 17, 1996 4841
Figure 4. Representative histograms of single-strand DNA cleavage
produced by ring- and imine-substituted [SalenMnIII]+ derivatives in
the presence of 1.0 mM magnesium monoperoxyphthlate on the 517
bp restriction fragments (EcoRI/RsaI) from pBR322 plasmid DNA.
Histograms were derived from densitometric analysis of the lower
regions of the autoradiographs shown in Figure 2. Lengths of arrows
correspond to the relative amounts of cleavage as determined by optical
densitometry. Sequence positions of cleavage were determined by
comparing the electrophoretic mobilities of DNA cleavage fragments
to bands in Maxam-Gilbert chemical sequencing lanes.
Table 3. Loci of Greatest Double-Strand Cleavage Intensity
Produced on StyI-Linearized Plasmid pBR322 by [SalenMnIII]+ and
Derivatives
positiona,b
sequence
3231
TTTTAAATTAAAAATGAAGTTTTAAAT-
CAATCTAAAGTATATAT
TTTTTCAATATTATTGAA
TATTTAGAAAAATAAACAAATA
ATTAACCTATAAAAATA
Figure 3. Autoradiograph of DNA single-strand cleavage patterns
produced by ring-substituted [SalenMnIII]+ complexes on 3′- and 5′-
32P end labeled 517 bp restriction fragments (EcoRI/RsaI) from pBR322
plasmid DNA in the presence of 1.0 mM magnesium monoperoxyph-
thalate. Cleavage patterns were resolved on a 1:20 cross-linked 8%
polyacrylamide/45% urea denaturing gel. Left set of lanes: 3′-32P end
labeled DNA. Right set of lanes: 5′-32P end labeled DNA. For each
set of lanes: lane 1, DNA incubated in the absence of metal complex
or oxidant; lane 2, Maxam-Gilbert chemical sequencing G reaction;
lane 3, [SalenMnIII]+ (1, 20 µM); lane 4, 5,5′-CH3 (19, 100 µM); lane
5, 5,5′-Cl2 (6, 12.5 µM); lane 6, 5,5′-Br2-3,3′-tert-Bu2 (20, 20 µM).
4164
4228
4315
a Assignment made by interpolation from the sizes and migration
distances of cleavage products relative to those of molecular weight
standards. b Lowest numbered base pair of the underlined sequences,
pBR322 numbering.
of the autoradiographs (Figures 4 and 5). The histograms show
that complexes 2-10, 12, 19, and 20, similar to the unsubstituted
parent complex, produce cleavage within and adjacent to sites
of multiple contiguous A:T base pairs. However, intensities
of cleavage at each nucleotide within the A:T target sequences
produced by derivatives substituted in different ring positions
differ by up to a factor of 4. The histograms also show that
the cleavage patterns are shifted to the 3′ side on one DNA
strand relative to the other and that single sites of cleavage are
observed at some locations while at other sites cleavage extends
symmetrically or unsymmetrically over many nucleotides. The
minimal sequence requirement for specific cleavage by 1 and
its derivatives appears to be two contiguous A:T base pairs.
produced by 6 and other complexes substituted with electron-
withdrawing substituents in the 5,5′-positions. Complex 20
cleaves DNA more efficiently than 1 but less efficiently than
the 5,5′-Br2 derivative 9. As was observed with the 3,3′,5,5′-
tetrachloro derivative 8, complex 20 exhibits DNA cleavage
specificity which differs from that of 1 and the 3,3′-disubstituted
derivatives 2 and 4, yet matches that of complexes 3, 6, 9, and
10.
The autoradiographs shown in Figures 2 and 3 were quantified
by densitometry, and these data were used to construct
histograms of the DNA cleavage observed in the lower regions

4842 Inorganic Chemistry, Vol. 35, No. 17, 1996
Gravert and Griffin
Figure 5. Histograms of single-strand DNA cleavage produced by
ring-substituted [SalenMnIII]+ derivatives in the presence of 1.0 mM
magnesium monoperoxyphthlate on the 517 bp restriction fragments
(EcoRI/RsaI) from pBR322 plasmid DNA. Histograms were derived
from densitometric analysis of the lower regions of the autoradiographs
shown in Figure 3. Lengths of arrows correspond to the relative amounts
of cleavage as determined by optical densitometry. Sequence positions
of cleavage were determined by comparing the electrophoretic mobilities
of DNA cleavage fragments to bands in Maxam-Gilbert chemical
sequencing lanes.
Figure 6. Autoradiographs of DNA double-strand cleavage patterns
produced by bridge-substituted [SalenMnIII]+ complexes on StyI-
linearized, 32P end labeled pBR322 plasmid DNA in the presence of
1.0 mM magnesium monoperoxyphthlate and resolved by electrophore-
sis on a 1% agarose gel. Top: 3′-32P dATP end labeled DNA. Bottom:
3′-32P TTP end labeled DNA. Lane 1, DNA incubated in the absence
of metal complex or oxidant; lane 2, pBR322 molecular weight markers
from pBR322 4365, 3371, 2994, 2368, 1998, 1768, 1372, 995, and
666 bp in length; lane 3, [SalenMnIII]+ (1, 20 µM); lane 4, (R,R)-23(20
µM); lane 5, (S,S)-23 (100 µM); lane 6, meso-23 (50 µM); lane 7, (R,R)-
23 (50 µM); lane 8, (S,S)-23 (50 µM); lane 9, meso-23 (50 µM); lane
10, [SalophenMnIII]+ (22, 50 µM); lane 11, [SalpnMnIII]+ (21, 100 µM).
Cleavage Competency and Efficiency of Bridge-Substi-
tuted Derivatives. Diphenyl-substituted derivatives (R,R)-24
and (S,S)-24 did not generate observable DNA cleavage. The
bridge-substituted derivatives 21-23 were found to produce
DNA cleavage in the presence of terminal oxidants, albeit with
reduced efficiency relative to the parent complex 1. This can
be illustrated by the concentrations required to achieve similar
levels of DNA double-strand cleavage in the presence of 1 mM
magnesium monoperoxyphthalate after a reaction time of 1 h
at 37 °C or by the differing extents of cleavage observed using
equal concentrations of the agents (Figure 6). The ability to
cleaVe DNA was found to Vary with the configuration of the
complex, as cleaVage efficiency decreased in the order (R,R)-
23 > meso-23 > (S,S)-23. In addition, it was found that 21
([SalpnMnIII]+) exhibited greatly reduced DNA binding/cleaving
efficiency relative to 1, while 22 ([SalophenMnIII]+) produced
DNA cleavage at a level intermediate between that of 1 and
21. Considering all of the bridge-substituted derivatives,
cleavage efficiency decreases in the order 1 > (R,R)-23 > 22
> meso-23 > (S,S)-23 > 21 . (R,R)-24, (S,S)-24.
A:T rich regions of pBR322 (Table 3). The 3′- and 5′-32P end
labeled, 517 bp EcoRI/RsaI restriction fragments from pBR322
and DNA sequencing gels were then used to probe the
specificity of single-strand DNA cleavage by 21-23 at nucle-
otide resolution. The DNA affinity cleaving data shown in
Figure 7 demonstrate that the patterns of cleaVage produced
by [SalenMnIII]+ deriVatiVes depend on both the structure and
stereochemistry of the bridge. Complexes 21-23 produce
single-strand DNA cleavage patterns that do not match one
another, those produced by the parent complex 1, or those of
the other complexes examined.
Histograms of DNA cleavage patterns produced by 21-23
(Figure 8) reveal that complexes 21-23, like 1, produce
cleavage within and adjacent to sites of multiple contiguous
A:T base pairs. Also, cleavage patterns produced by 21-23
are shifted to the 3′ side on one DNA strand relative to the
other and consist of single sites of cleavage at some positions
Cleavage Specificity of Bridge-Substituted Derivatives. As
with the ring-substituted derivatives, 21-23 produced patterns
of DNA double-strand cleavage that were indistinguishable from
that of 1, with cleavage loci of greatest intensity mapping to

[SalenMnIII]+ Derivatives in DNA Binding/Cleaving
Inorganic Chemistry, Vol. 35, No. 17, 1996 4843
Figure 7. Autoradiograph of DNA single-strand cleavage patterns produced by bridge-substituted [SalenMnIII]+ complexes on 3′- and 5′-32P end
labeled 517 bp restriction fragments (EcoRI/RsaI) from pBR322 plasmid DNA in the presence of 1.0 mM magnesium monoperoxyphthlate. Cleavage
patterns were resolved on a 1:20 cross-linked 8% polyacrylamide/45% urea denaturing gel. Left set of lanes: 3′-32P end labeled DNA. Right set of
lanes: 5′-32P end labeled DNA. For each set of lanes: lane 1, DNA incubated in the absence of metal complex or oxidant; lanes 2, Maxam-Gilbert
chemical sequencing G reaction; lane 3, [SalenMnIII]+ (1, 20 µM); lane 4, (R,R)-23 (20 µM); lane 5, (S,S)-23 (100 µM); lane 6, meso-23 (50 µM);
lane 7, (R,R)-23 (50 µM); lane 8, (S,S)-23 (50 µM); lane 9, meso-23 (50 µM); lane 10, [SalophenMnIII]+ (22, 50 µM); lane 11, [SalpnMnIII]+ (21,
100 µM).
while at other sites cleavage extends symmetrically or unsym-
metrically over many nucleotides. Unlike ring-substituted
[SalenMnIII]+ derivatives wherein complexes containing dif-
ferent substituents at the same position display nearly identical
cleavage patterns, the cleavage patterns of complexes 21-23
are each unique. Within a given A:T target sequence, cleavage
intensities at individual nucleotides differ by up to a factor of
10 for 1 and 21-23.
All of the [SalenMnIII]+ derivatives exhibited UV-vis spectra
typical of this family of complexes.2 The spectra of some
derivatives showed little change after incubation for 1 week in
the aqueous buffer used for DNA cleavage reactions, while
others (complexes 2, 4, 10, 12, 13, 17, 18, 21, and 22) showed
more pronounced changes. For example, after 1 week the
phenyl-bridged complex 22 displayed greater than 30% loss of
absorbance throughout most of the UV-vis region with the
emergence of a new absorbance maximum at 258 nm. The 3,3′-
UV-Vis Analysis of Complex Stability and Activation.

4844 Inorganic Chemistry, Vol. 35, No. 17, 1996
Gravert and Griffin
activated species mediate DNA cleavage. These changes could
in turn be manifested as changes in the properties that we have
addressed in the present study: DNA cleavage efficiency and
specificity. The contributions of individual effects to the
observed behavior, and the steric and electronic factors from
which they are ultimately derived, may be difficult to distinguish
since they may act in concert or in opposition through multiple
mechanisms. However, consideration of results obtained with
a series of related derivatives, including stereoisomers, can
provide insight into the determinants involved.6-9 This ap-
proach has been facilitated in the present study by the ease of
synthesis of [SalenMnIII]+ derivatives.
Binding Location and Cleavage Mode. When examined
at high resolution, the cleavage patterns produced by all active
derivatives were found to be shifted to the 3′ end on one DNA
strand relative to the other. These results indicate that cleavage
occurs from the minor groove of right-handed double-helical
DNA.12 The observed nonuniform patterns of cleavage and
structure dependent differences in cleavage patterns are con-
sistent with a process involving a specifically bound rather than
a freely diffusible reactive species such as hydroxyl radical.12,13
This latter point is compellingly demonstrated by the finding
that the enantiomeric complexes (R,R)-23 and (S,S)-23 cleave
DNA with different efficiencies and generate different cleavage
patterns.
Cleavage Efficiency: Ring- and Imine-Substituted De-
rivatives. Seventeen of the 26 derivatives of complex 1 studied
produced specific cleavage of DNA in the presence of terminal
oxidants. No clear correlation was found between the UV-
vis absorption behavior in the absence and the presence of
oxidant and the DNA binding/cleaving properties of these
complexes. Complexes 11 and 13 do not produce DNA
cleavage, but UV-vis studies indicate that they are stable in
reaction buffer, undergo activation in the presence of oxidant,
and are not especially prone to decomposition upon activation.
Thus, it appears that the inability of these complexes to effect
DNA cleavage is due to the inability of their oxidatively
activated forms to interact and/or react with DNA.
The results clearly indicate that the position of attachment
and identity of substituents affect the ability of [SalenMnIII]+
derivatives to effect DNA cleavage. The position of attachment
is obviously important, since the regioisomeric dichloro deriva-
tives exhibit both enhanced (5 and 6) and diminished (4 and 7)
DNA cleavage efficiencies relative to that of 1. With regard
to substituent identity, the finding that 9 (5,5′-Br2), 6 (5,5′-Cl2),
and 10 (5,5′-(NO2)2) cleave DNA with efficiencies equal to or
greater than that of the parent complex 1 while 13 (5,5′-(OCH3)2)
is inactive is in general accord with the results of Kochi et al.,
who found that 6 and 10 exhibited increased efficiency as
catalysts for olefin epoxidation and carbon-hydrogen activation
relative to the efficiencies of 1 and 13.2 The data for this series
of derivatives cannot be explained by simple models in which
either steric repulsion or favorable interactions between 5,5′-
substituents and DNA dictate cleavage efficiency. Substituents
at the 5,5′-positions strongly affect the redox properties of
[SalenMnIII]+ derivatives,2,5 and it is likely that the great
differences in DNA binding/cleaving efficiencies between
complexes 9, 6, and 10 on the one hand and complexes 13 and
19 (5,5′-(CH3)2) on the other derive primarily from electronic
effects associated with the presence of net electron-withdrawing
bromo, chloro, or nitro groups versus net electron-donating
methoxy or methyl groups at these positions. However, among
Figure 8. Histograms of single-strand DNA cleavage produced by 1
and chiral bridge-substituted [SalenMnIII]+ complexes in the presence
of 1.0 mM magnesium monoperoxyphthlate on the 517 bp restriction
fragments (EcoRI/RsaI) from pBR322 plasmid DNA. Histograms were
derived from densitometric analysis of the lower regions of the
autoradiographs shown in Figure 7, lanes 3-5. Lengths of arrows
correspond to the relative amounts of cleavage as determined by optical
densitometry. Sequence positions of cleavage were determined by
comparing the electrophoretic mobilities of DNA cleavage fragments
to bands in Maxam-Gilbert chemical sequencing lanes.
dichloro derivative 4 displayed reduced absorbance below 300
nm and increased absorbance at 400 nm with the loss of an
absorbance maximum. When treated with magnesium mono-
peroxyphthalate, each of the complexes exhibiting DNA cleav-
age activity (except 21) showed the increased absorption in the
visible region which characterizes oxidatively activated 1.2
Among the derivatives which do not exhibit DNA binding/
cleaving activity, complexes 13, 15, and 24 showed spectral
changes consistent with oxidative activation in the presence of
magnesium monoperoxyphthalate, while the spectra of com-
plexes 16-18 and 22 showed little change upon addition of
the oxidant. The spectral changes observed upon addition of
the oxidant decayed over a period of minutes to afford spectra
similar to those of the unactivated complexes, though of lesser
intensity. In the case of complex 21, this reduction in absorption
intensity was observed immediately after addition of oxidant.
Discussion
Substituents alter both the steric and electronic properties of
metal complexes, and in the case of DNA binding/cleaving metal
complexes such as 1, these could lead to changes in (1) the
stability in aqueous buffered solution of the complexes or
oxidatively activated intermediates derived from them, (2) the
ability of activated intermediates to bind to target sites for
cleavage in a productive fashion, and/or (3) the rates at which
the complexes are oxidatively activated or at which oxidatively
(13) (a) Hertzberg, R. P.; Dervan, P. B. J. Am. Chem. Soc. 1982, 104, 313-
315. (b) Hertzberg, R. P.; Dervan, P. B. Biochemistry 1984, 23, 3934-
3945. (c) Tullius, T. D.; Dombroski, B. A.; Churchill, M. E. A.; Kam,
L. Methods Enzymol. 1987, 155, 537-558.

[SalenMnIII]+ Derivatives in DNA Binding/Cleaving
Inorganic Chemistry, Vol. 35, No. 17, 1996 4845
complexes bearing electron-withdrawing substituents in the 5,5′-
positions, DNA binding/cleaving efficiency is not an unvarying
function of the increasing electron-withdrawing character of the
substituent. pKa values and oxidation potentials of para-
substituted methoxy-, methyl-, fluoro-, bromo-, chloro-, and
nitrophenols14 suggest that oxidation potentials for the Salen
complexes will increase in the order 13 < 19 < 3 < 1 < 9 <
6 < 10, yet DNA cleavage efficiency is observed to increase
in the order 13 < 19 < 3 < 1 ≈ 10 < 9 ≈ 6. A similar
phenomenon was observed by Muller et al., who found that
DNA oxidation by nickel complexes of tetraazamacrocyclic
ligands increases with redox potential only up to a certain point.9
For the series of 3,3′-substituted derivatives, DNA cleavage
efficiency decreases with increasing size of hydrogen, fluoro,
chloro, and methoxy substituents: 1 > 2 > 4 . 11. However,
the fact that complex 20, with bulky tert-butyl groups in the
3,3′-positions, exhibits DNA cleavage efficiency greater than
that of 1 suggests that the DNA cleavage properties of these
derivatives are not dictated by the steric nature of 3,3′-
substituents. This notion is further supported by consideration
of the cleavage specificity exhibited by these derivatives (see
below).
Cleavage Efficiency: Bridge-Substituted Derivatives.
Modifications on the bridge of [SalenMnIII]+ also significantly
affect the DNA cleavage efficiency. Differences between
complexes 1 and 21 ([SalpnMnIII]+) may derive from the
additional flexibility imparted by the extra methylene unit. In
general, Salen complexes form trans isomers with the phenolate
oxygens and the imine nitrogens forming a plane; however, 21
can form trans or cis â conformers.15 The cis â conformation
of 21 readily forms the stable dioxo dimer [SalpnMnIVO]2 in
the presence of base and oxidant. Formation of this or related
species may account for the decreased cleavage efficiency of
21 relative to that of 1.16
Due to the phenyl linkage, complex 22 should exhibit
increased rigidity relative to that of 1. Complex 22 also has
increased aromatic surface area which may participate in
interactions with DNA by groove binding or perhaps even
intercalation. ESR spectra of substituted Salen copper com-
plexes bound to DNA fibers are consistent with a change in
DNA binding mode from groove binding to intercalation upon
changing the bridging ethylene group to phenylene or naph-
thalene groups.17 However, manganese porphyrin complexes
require axial ligands which sterically restrict the metallopor-
phyrin to DNA groove binding modes.18 This combined with
the finding that 22 displays A:T selective cleavage similar to
that of 1 and the other derivatives leads us to conclude that
minor groove binding rather than intercalation is the principal
mode through which 22 produces DNA cleavage. The reduced
cleaving efficiency of 22 versus 1 may arise from differential
electronic factors associated with conjugating versus noncon-
jugating linkers and/or from steric interactions associated with
the phenyl bridge which hinder approach of 22 to the DNA
double helix in a cleavage competent orientation. The cleavage
specificity results suggest that at least some derivatives of 1
produce DNA cleavage from an orientation in which the bridge
of the complex is oriented toward or along the minor groove
(see below). It is therefore possible that repulsive interactions
between the phenyl bridge of activated 22 and the DNA target
could result in diminution of cleavage efficiency for 22 relative
to that for 1.
(R,R)-, (S,S)-, and meso-23 exhibit different DNA cleavage
efficiencies, indicating that DNA recognition by activated forms
of these complexes is both stereo- and enantioselective. Because
the electronic properties of unbound (R,R)-23 and (S,S)-23 (and
activated intermediates derived from them) are identical, the
observed differences in cleavage efficiencies (5-fold for these
enantiomeric complexes) must have their basis in diastereomeric
interactions with DNA. If activated intermediates derived from
23 interact with DNA in a bridge first or side first fashion, then
steric interactions between their cyclohexano bridges and the
DNA minor groove which hinder the initiation of strand scission
could account for the reduced cleavage efficiency for 23 relative
to that for 1 as well as the observed stereoselectivity.
The diphenyl derivatives (R,R)-24 and (S,S)-24 do not produce
observable DNA cleavage under the reaction conditions em-
ployed. However, since UV-vis analyses demonstrate that
these complexes are stable in reaction buffer and yield spectral
changes consistent with oxidative activation without consider-
able decomposition, it appears that DNA cleavage is inhibited
by the inability of oxidatively activated forms of 24 to interact
and/or react with DNA.
Cleavage Specificity: General. The substituents present
in the cleavage competent complexes do not alter the proclivity
of 1 to produce cleavage from the minor groove of right-handed
double-helical DNA at sites containing multiple, contiguous A:T
base pairs. This suggests a general shape selection mecha-
nism7,19 for DNA recognition by [SalenMnIII]+ derivatives in
which these planar cationic agents can be sandwiched within
the electron rich,20 relatively narrow minor groove found in A:T
rich regions. Here they form stabilizing van der Waals
interactions with deoxyribose residues of the DNA backbone
but not highly specific contacts such as hydrogen bonds with
the base pairs themselves. The protruding N2 amino group of
guanine will preclude deep penetration into the minor groove
by 1 and its derivatives at sequences containing G:C base pairs.
While substituents do not alter the general A:T specificity of
1, substituents do change the patterns of cleavage produced
within A:T rich target sites. These differences reflect substitu-
ent-induced changes in the relative rates of DNA cleavage at
each nucleotide within A:T rich target sequences. The data do
not allow us to distinguish whether these differences derive from
changes in the relative affinities of activated intermediates for
different binding/cleaving sites and/or from changes in rate
constants for cleavage at the various sites.21
Cleavage Specificity: Chiral Bridge-Substituted Deriva-
tives. Chiral derivatives of [SalenMnIII]+ and structurally related
species have been shown to be highly efficient catalysts for
asymmetric epoxidation,5,11,22 aziridination,23 aldol,24 sulfide
oxidation,25 and cyclopropanation reactions.26 The observation
that the enantiomeric complexes (R,R)-23 and (S,S)-23 produce
different patterns of cleavage demonstrates enantiospecificity
in DNA binding/cleaving. We believe that both the observed
differences in binding/cleaving efficiencies and specificities arise
from differences in nonbonded interactions between the dia-
stereomeric complexes of activated (R,R)-23:DNA and activated
(S,S)-23:DNA at individual sites of cleavage. While models
indicate that (R,R)-23 is twisted in a sense that is more
(14) Arnett, E. M.; Amarnath, K.; Harvey, N. G.; Venimadhavan, S. J.
Am. Chem. Soc. 1990, 112, 7346-7353.
(15) Larson, E. J.; Pecoraro, V. L. J. Am. Chem. Soc. 1991, 113, 3810-
3818.
(16) We have found that preformed [SalpnMnIVO]2 does not produce
nicking of supercoiled plasmid DNA at 100 µM concentration in either
the absence or presence of additional oxidant.
(17) Sato, K.; Chikira, M.; Fujii, Y.; Komatsu, A. J. Chem. Soc., Chem.
Commun. 1994, 625-626.
(18) (a) Pasternack, R. F.; Gibbs, E. S.; Villafranca, J. J. Biochemistry 1983,
22, 2406-2414. (b) Ward, B.; Skorobogaty, A.; Dabrowiak, J. C.
Biochemistry 1986, 25, 7827-7833.
(19) (a) Lee, M. D.; Ellestad, G. A.; Borders, D. B. Acc. Chem. Res. 1991,
24, 235-243. (b) Uesugi, M.; Sugiura, Y. Biochemistry 1993, 32,
4622-4627. (c) Kahne, D. Chem. Biol. 1995, 2, 7-12.
(20) Pullman, B.; Lavery, R.; Pullman, A. Eur. J. Biochem. 1982, 124,
229-238.

4846 Inorganic Chemistry, Vol. 35, No. 17, 1996
Gravert and Griffin
complementary to the minor groove of B form DNA, further
analysis and modeling studies will be needed to understand the
detailed basis for these effects.
There is considerable precedent for differential DNA cleavage
properties based on diastereomeric interactions between chiral
small molecules and DNA. For example, enantiomers of
intercalating tris(chelate) cobalt and rhodium complexes exhibit
enantiospecific photocleavage of double-helical DNA.7 In these
systems, enantiospecificity is based on steric interactions
between the DNA receptor and nonintercalated ligands on the
metal complexes. The intrinsic site selectivity of these com-
plexes may be significantly enhanced by addition of substituents
to the nonintercalated ligands which participate in specific
hydrogen bonding and van der Waals interactions with groups
present in the major groove of B DNA.7d-f These enhancements
in specificity were not observed in the present study, indicating
that the specificity of the [SalenMnIII]+ derivatives examined
derives from general shape selective recognition alone.
Cleavage Specificity: Ring- and Imine-Substituted De-
rivatives. The patterns of cleavage produced by derivatives of
1 are dependent upon the position of substituents, but nearly
independent of substituent identity. Complexes substituted in
the 5,5′-positions with electron-withdrawing groups (3, 6, 9,
and 10) produce patterns of cleavage that are indistinguishable
from one another, subtly different from the pattern of cleavage
produced by the complex (19) which bears electron-donating
methyl substituents in the 5,5′-positions, and more substantially
different from the cleavage pattern produced by 1, which has
hydrogen substituents in the 5,5′-positions. Since 3, 6, 9, and
10 bear substituents that are smaller than (fluoro), similar to
(chloro), and larger than (bromo, nitro) a methyl group in size,
we are led to conclude that the slight differences in specificity
observed with 3, 6, 9, and 10 versus 19 are due to electronic
factors, but that the more substantial differences between 3, 6,
9, 10, and 19 versus 1 arise from steric factors. It is reasonable
to expect that steric interactions between substituents and the
DNA minor groove will perturb the rates of hydrogen abstraction
at and/or thermodynamic partitioning among DNA binding/
cleaving sites. It is somewhat puzzling that the smallest change
in size from hydrogen (van der Waals radius ) 1.20 Å) to
fluorine (van der Waals radius ) 1.47 Å) at the 5,5′-positions
alters cleavage specificity, but that substitution of much larger
Figure 9. Models for possible reactive orientations for DNA cleavage
by [SalenMnVO]+. Circles with two dots represent lone pairs of
electrons on adenine N3 or thymine O2 atoms at the edges of the base
pairs on the floor of the minor groove of B form DNA. Top: Bridge
first approach of the complex to DNA. Middle: Side first approach of
the complex to DNA. Bottom: Concave edge first approach of the
complex to DNA. The data with derivatives substituted in the 3,3′-
positions is inconsistent with the latter model.
chloro, bromo, or nitro groups does not lead to further changes
in cleavage specificity. However, we note that a series of
manganese and cobalt porphyrins exhibited no change in
cleavage specificity when the steric bulk of substituents lying
in the porphyrin plane was increased.6b,c
The regioisomeric dichloro derivatives 4-7 generate different
patterns of cleavage. We consider it likely that steric effects
are also important in determining these differences, but more
data will be needed to demonstrate this with certainty.
Reactive Orientation. The results obtained with derivatives
bearing substituents in the 3,3′-positions (2, 4, 8, and 20) provide
insight into the orientation of cleavage competent intermediates
derived from these complexes and 1 relative to DNA. The
patterns of cleavage produced by 2 and 4 match those produced
by 1, suggesting that activated intermediates derived from these
three complexes adopt a common reactive orientation when
bound to DNA in which their concave edges, from which
protrude the 3,3′-substituents, are not oriented toward the double
helix. Sigman and co-workers observed no change in cleavage
patterns produced by bis(1,10-phenanthroline):copper complexes
substituted in the 5- or 6-positions and used these results to
support a model for DNA complexation in which substituents
at these positions extend away from the DNA helix.8
Three possible reactive orientations for the putative interme-
diate [SalenMnVO]+ are depicted in Figure 9: convex edge
(bridge) first, side first, and concave edge first. The data for 1,
2, 4, 8, and 20 are consistent with either of the first two models
but not with the third. That tetrasubstituted complexes 8 and
20 produce cleavage patterns similar to those produced by the
5,5′-disubstituted derivatives 3, 6, 9, and 10 suggests that
activated intermediates derived from the latter complexes also
do not bind with their concave edges oriented toward the double
helix. Bridge first or side first approaches are also consistent
(21) (a) Baker, B. F.; Dervan, P. B. J. Am. Chem. Soc. 1985, 107, 8266-
8268. (b) Baker, B. F.; Dervan, P. B. J. Am. Chem. Soc. 1989, 111,
2700-2712. (c) De Voss, J. J.; Hangeland, J. J.; Townsend, C. A. J.
Am. Chem. Soc. 1990, 112, 4554-4556. (d) Hurley, L. H.; Warpehoski,
M. A.; Lee, C.-S.; McGovren, J. P.; Scahill, T. A.; Kelly, R. C.;
Mitchell, M. A.; Wicnienski, N. A.; Gebhard, I.; Johnson, P. D.;
Bradford, V. S. J. Am. Chem. Soc. 1990, 112, 4633-4649. (e) Wender,
P. A.; Kelly, R. C.; Beckham, S.; Miller, B. L. Proc. Natl. Acad. Sci.
U.S.A. 1991, 88, 8835-8839. (f) Myers, A. G.; Harrington, P. M.;
Kwon, B.-M. J. Am. Chem. Soc. 1992, 114, 1086-1087. (g) Walker,
S.; Landovitz, R.; Ding, W. D.; Ellestad, G. A.; Kahne, D. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 4608-4612. (h) Mah, S. C.; Townsend,
C. A.; Tullius, T. D. Biochemistry 1994, 33, 614-621. (i) Myers, A.
G.; Cohen, S. B.; Kwon, B.-M. J. Am. Chem. Soc. 1994, 116, 1255-
1271. (j) Myers, A. G.; Cohen, S. B.; Kwon, B.-M. J. Am. Chem.
Soc. 1994, 116, 1670-1682. (k) Chatterjee, M.; Cramer, K. D.;
Townsend, C. A. J. Am. Chem. Soc. 1994, 116, 8819-8820.
(22) (a) Jacobsen, E. N.; Deng, L.; Furukawa, Y.; Martinez, L. E.
Tetrahedron 1994, 50, 4323-4334. (b) Deng, L.; Jacobsen, E. N. J.
Org. Chem. 1992, 57, 4320-4323.
(23) Li, Z.; Conser, K. R.; Jacobsen, E. N. J. Am. Chem. Soc. 1993, 115,
5326-5327.
(24) Carreira, E. M.; Singer, R. A.; Lee, W. J. Am. Chem. Soc. 1994, 116,
8837-8838.
(25) Sasaki, C.; Nakajima, K.; Kojima, M.; Fujita, J. Bull. Chem. Soc. Jpn.
1991, 64, 1318-1324.
(26) (a) Nozaki, H.; Takaya, H.; Moriuti, S.; Noyori, R. Tetrahedron 1968,
24, 3655-3669. (b) Aratani, T.; Yoneyoshi, Y.; Nagase, T. Tetrahe-
dron Lett. 1977, 2599-2602. (c) Aratani, T.; Yoneyoshi, Y.; Nagase,
T. Tetrahedron Lett. 1982, 685-688.

[SalenMnIII]+ Derivatives in DNA Binding/Cleaving
Inorganic Chemistry, Vol. 35, No. 17, 1996 4847
UV-Vis Analysis. UV-vis spectra of complexes 1-24 were
measured at a concentration of 50 µM in 40 mM Tris acetate pH 7.0
immediately after fresh preparation and after incubation at room
temperature in the dark for 1 week. For the analysis of complex
activation, spectra were taken before and at several time points after
addition of 10 equiv of magnesium monoperoxyphthalate.
with the observation that bridge structure and stereochemistry
affect both DNA binding/cleaving efficiency and specificity by
complexes 21-23 and provide a rationale for the lack of
cleavage by diphenyl derivatives (R,R)-24 and (S,S)-24. Finally,
when considered in light of concave edge away models, it is
apparent that the diminished DNA binding/cleaving efficiency
observed for complexes 2 and 4 relative to that of 1 derives
from electronic rather than steric factors.
General Procedure for DNA Cleavage Reactions with Substituted
]
[SalenMnIII + Complexes. Three microliters of a substituted [Salen-
MnIII]+ complex solution was added to 9 µL of a solution containing
32P end labeled DNA restriction fragment29 (10000-30000 cpm) and
calf thymus DNA in Tris acetate buffer pH 7.0. DNA cleavage was
initiated by the addition of 3 µL magnesium monoperoxyphthlate
(MgMPP) solution. Final concentrations in 15 µL total volume were
the following: substituted [SalenMnIII]+ complex, as given in the figure
legends; calf thymus DNA, 100 µM bp; Tris, 40 mM; MgMPP, 1.0
mM. DNA cleavage was allowed to proceed 1 h at 37 °C before DNA
cleavage products were separated by gel electrophoresis. For studies
of DNA double-strand cleavage, cleavage reactions were treated with
3 µL of 25% Ficoll loading buffer, loaded onto a 25 cm long by 4 mm
thick 1% agarose gel, and electrophoresed at 120-160 V until the BPB
(bromophenol blue) tracking dye reached the bottom of the gel. The
gel was then dried and autoradiographed. For high-resolution DNA
cleavage studies, the cleavage reactions were terminated by precipitation
with NaOAc/EtOH, and the precipitate was washed with 70% EtOH
and dried. The resulting pellets were resuspended in 5 µL of 80%
formamide loading buffer, heat denatured 2 min at 95 °C, loaded onto
a 40 cm long by 0.25-0.75 mm thick 1:20 cross-linked 8% polyacry-
lamide/45% urea gel, and electrophoresed at 1200-2000 V until the
BPB tracking dye reached the bottom of the gel. The gel was then
dried and autoradiographed.
Bridge first and side first orientations bear some analogy to
the geometries proposed for minor groove binding by enanti-
omers of tris(1,10-phenanthroline):ruthenium(II).27 In contrast,
the apparent preference for bridge first or side first orientations
by activated intermediates derived from 1 and at least some of
its derivatives differs from that observed with the crescent-
shaped minor groove binding agents netropsin and distamycin.
However, in the case of those bis- and tris(N-methylpyrrole-
carboxamides), orientation of their concave edges toward the
minor groove allows for formation of a series of hydrogen bonds
between amide NHs and lone pair electrons from O2 atoms of
thymine and N3 atoms of adenine.12,28 No hydrogen bond donor
groups are found on the concave edge of 1; rather, this edge of
the molecule displays hydrogen bond acceptor groups (i.e., lone
pair electrons on the phenolate oxygen atoms that are ligated
to the manganese ion).
Experimental Section
Synthesis of Substituted Salicylaldehydes. 5-Fluorosalicylaldehyde
and 3-, 4-, and 6-chlorosalicylaldehyde were made by the Reimer-
Tiemann method from the corresponding substituted phenols.10 5-Me-
thylsalicylaldehyde was obtained by LiAlH4 reduction of 5-methylsal-
icylic acid and subsequent MnO2 oxidation.
Densitometry. Optical densitometry was performed using an LKB
Bromma Ultroscan XL laser densitometer operating at 633 nm. Sites
of significant double-strand cleavage reported in Table 3 were
interpolated from the sizes and migration distances of cleavage products
relative to those of molecular weight standards. Positions of single-
strand cleavage shown in the histograms were obtained by reference
to the Maxam-Gilbert G lanes.30 Relative peak areas for each cleavage
band were equated to the relative cleavage efficiencies at those sites
and are reflected in the lengths of arrows.
General Method for Synthesis of Substituted [SalenMnIII]OAc
Salts. The general method for synthesis of the Mn(III) complexes is
typified by the procedure for the 3,3′-(OCH3)2 derivative 11. To a
solution of o-vanillin (326 mg, 2.14 mmol) in ethanol (2 mL) was added
ethylenediamine (63.5 mg, 1.06 mmol) in ethanol (1 mL). The solution
turned bright yellow, and a yellow precipitate was formed. The slurry
was stirred and heated until homogeneous, whereupon a solution of
manganese(II) acetate (224 mg, 0.91 mmol) in warm ethanol (2 mL)
was added. The reaction mixture immediately became dark. After
refluxing in the presence of air for 1 h, the mixture was cooled to room
temperature and allowed to crystallize overnight (some complexes
required the addition of ether to induce crystallization or precipitation).
The product was filtered, washed with cold ethanol and diethyl ether,
and dried in vacuo to afford 319 mg (0.85 mmol, 80%) of [N,N ′-
ethylenebis(3-methoxysalicylideneaminato)]manganese(III) acetate (11)
as brown microcrystals, mp 198-200 °C (dec). FT-IR (cm-1, film):
1629 (s), 1591 (s), 1566 (s), 1554 (s), 1536 (s), 1429 (s), 1395 (s),
1335 (m), 1296 (s), 1226 (m), 1190 (m), 1144 (s), 1087 (m), 1044
(m), 986 (w), 938 (m), 918 (w), 879 (w), 857 (m), 846 (m), 780 (m),
742 (m), 689 (m), 676 (m). UV-vis (H2O) λmax (ꢀ, M-1 cm-1): 402
(3700), 323 (9300), 301 (9400), 262 (13800), 225 (34200). UV-vis
(buffer, fresh) λmax (ꢀ, M-1 cm-1): 405 (4800), 323 (11100), 295
(13500), 262 (14600), 228 (37000). UV-vis (buffer, >1 week) λmax
(ꢀ, M-1 cm-1): 405 (4900), 323 (8400), 295 (11000), 262 (9300),
228 (29400). Anal. Calcd for C21H26N2O7.5Mn: C, 52.40; H, 5.44;
N, 5.82. Found: C, 52.25; H, 5.47; N, 5.83. FAB-MS: calcd for
C18H18N2O4Mn ([3,3′-(OCH3)2SalenMnIII]+) 381.1; found 381.1. µeff
) 4.81.
Acknowledgment. This work was supported by Stanford
University, the Arnold and Mabel Beckman Foundation, the
Camille and Henry Dreyfus Foundation, the Shell Foundation,
and an American Cancer Society Institutional Research Grant.
D.J.G. was supported by a NSF Predoctoral Fellowship and a
Franklin Veatch Fellowship. The UCSF Mass Spectrometry
Facility (A. L. Burlingame, Director) is supported by the
Biomedical Research Technology Program of the National
Center for Research Resources, NIH NCRR BRTP RR04112
and RR01614. We are grateful to Professor Eric Jacobsen
(Harvard) for kindly providing a sample of the aldehyde needed
to prepare the 5,5-dibromo-3,3′-di-tert-butyl derivative 20.
Supporting Information Available: Spectroscopic and analytical
data for metal complexes and procedures for preparation of DNA
substrates (16 pages). Ordering information is given on any current
masthead page.
IC960196X
(27) (a) Rehmann, J. P.; Barton, J. K. Biochemistry 1990, 29, 1701-1709.
(b) Rehmann, J. P.; Barton, J. K. Biochemistry 1990, 29, 1710-1717.
(c) Eriksson, M.; Leijon, M.; Hiort, C.; Norde´n, B.; Gra¨slund, A. J.
Am. Chem. Soc. 1992, 114, 4933-4934.
(28) (a) Kopka, M. L.; Yoon, C.; Goodsell, D.; Pjura, P.; Dickerson, R. E.
Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 1376. (b) Kopka, M. L.; Yoon,
C.; Goodsell, D.; Pjura, P.; Dickerson, R. E. J. Mol. Biol. 1985, 183,
553-563.
(29) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular Cloning, A
Laboratory Manual, 2nd ed.; Cold Spring Harbor Laboratory: New
York, 1989.
(30) Maxam, A. M.; Gilbert, W. S. Methods Enzymol. 1980, 65, 499-
560.