Multinuclear non-heme iron complexes for double-strand DNA cleavage

Article abstract of DOI:10.1002/chem.200801409

The cytotoxicity of the antitumor drug BLM is believed to be related to the ability of the corresponding iron complex (Fe-BLM) to engage in oxidative double-strand DNA cleavage. The iron complex of the ligand N4Py (Fe-N4Py; N4Py = N,N-bis(2-pyridyl)-N-bis

Full text of DOI:10.1002/chem.200801409

FULL PAPER
DOI: 10.1002/chem.200801409
Multinuclear Non-Heme Iron Complexes for Double-Strand DNA Cleavage
Rik P. Megens, Tieme A. van den Berg, A. Dowine de Bruijn, Ben L. Feringa,* and
Gerard Roelfes*[a]
Abstract: The cytotoxicity of the anti-
tumor drug BLM is believed to be re-
lated to the ability of the correspond-
ing iron complex (Fe-BLM) to engage
in oxidative double-strand DNA cleav-
age. The iron complex of the ligand
N4Py (Fe-N4Py; N4Py=N,N-bis(2-pyr-
idyl)-N-bis(2-pyridyl)methylamine) has
proven to be a particularly valuable
spectroscopic and functional model for
Fe-BLM. It is also a very active oxida-
tive DNA-cleaving agent. However,
like all other synthetic Fe-BLM
mimics, it gives only single-strand
DNA cleavage. Since double-strand
DNA cleavage requires the delivery of
two oxidizing equivalents to the DNA,
it was envisaged that multinuclear iron
complexes might be capable of effect-
ing double-strand cleavage. For this
purpose, a series of ditopic and tritopic
N4Py-derived ligands has been synthe-
sized and the corresponding iron com-
plexes have been evaluated for their ef-
ficacy in the oxidative cleavage of su-
percoiled pUC18 plasmid DNA. The
dinuclear iron complexes showed sig-
nificantly enhanced double-strand
cleavage activity compared to mononu-
clear Fe-N4Py, which was relatively in-
dependent of the structure of the link-
ing moiety. Covalent attachment of a
9-aminoacridine intercalator to a dinu-
clear complex did not give rise to im-
proved double-strand DNA cleavage.
The most efficient oxidative double-
strand cleavage agents proved to be
the trinuclear iron complexes. This is
presumably the result of increased
probability of the simultaneous deliv-
ery of two oxidizing equivalents to the
DNA.
Keywords: artificial nucleases
DNA cleavage double-strand
multitopic
·
·
cleavage
·
iron
·
complexes
Introduction
It is important to recognize that single-strand DNA
(ssDNA) cleavage and double-strand DNA (dsDNA) cleav-
age represent different pathways. Single-strand DNA cleav-
age is obtained when one oxidation equivalent is delivered
to DNA and, through a cascade of reactions, causes a single
strand of DNA to break: a nick. When the DNA-cleaving
agent does not exhibit sequence selectivity, that is, it cleaves
the DNA at random positions, it will continue to nick the
DNA until a new nick is sufficiently close to an existing nick
in the opposite strand to cause the DNA to separate. Since
this is a statistical process, a DNA ds break will only occur
after extensive ss cleavage. In vivo, a single-strand DNA
break is rapidly repaired by the cellular repair mecha-
nisms.[7]
Bleomycins (BLMs) are a class of glycopeptide antibiotics
that are isolated from Streptomyces Verticillus.[1] This class
consists of 200 structurally related compounds, of which
bleomycins A2 and B2, under the name Blenoxane, are used
in the clinical treatment of head, neck, and testicular can-
cers.[2–5] Both in vivo and in vitro, the iron complex of bleo-
mycin (Fe-BLM) is capable of inducing both single-strand
and double-strand DNA cleavage.[1,6] The cytotoxicity of
BLM is believed to be mainly the result of the double-
strand DNA cleavage activity and only to a lesser extent
caused by single-strand DNA cleavage.[7]
Double-strand DNA cleavage is the result of two succes-
sive strand scissions in opposite strands in close proximity to
each other, resulting in a double-strand break.[8] This path-
way requires the delivery of two oxidizing equivalents to the
DNA helix. BLM is capable of doing this, although the pre-
cise nature of the oxidizing species is still under debate.[9]
Although repair of double-strand DNA cleavage is possi-
ble,[10] repair of single-strand DNA cleavage is much more
likely to occur.[7]
[a] R. P. Megens, Dr. T. A. van den Berg, A. D. de Bruijn,
Prof. Dr. B. L. Feringa, Dr. G. Roelfes
Stratingh Institute for Chemistry
University of Groningen
Nijenborgh 4, 9747 AG Groningen (The Netherlands)
Fax : (+31)50-363-4296
E-mail: b.l.feringa@rug.nl
j.g.roelfes@rug.nl
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801409.
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1723

Much effort has been devoted to mimicking the chemistry
of Fe-BLM with synthetic model complexes.[11–22] Many of
these complexes have proven to be capable of cleaving
DNA in the presence of a metal source and O2.[15,16] Howev-
er, all of these synthetic mimics have proven to be capable
only of effecting single-strand DNA cleavage.[23]
In our group, the pentadentate ligand N,N-bis(2-pyridyl-
methyl)-N-bis(2-pyridyl)methylamine (N4Py, 1; Figure 1)
has been designed as a mimic of the metal-binding domain
of BLM.[24–26] Its acridine-tethered derivative has proven to
be an efficient DNA-cleaving agent when using O2 as termi-
nal oxidant, without requiring addition of a sacrificial reduc-
tant.[24] However, as with other synthetic BLM mimics, only
single-strand DNA cleavage was achieved.
In a related study, a relationship was observed between
DNA-cleavage activity and the relative orientation of the
copper complexes. Covalent attachment of a third copper
complex, that is, formation of a trinuclear copper complex,
led to reduced activity.[37]
Here, we present a study on the design of di- and tritopic
ligands and their effect on iron-catalyzed oxidative DNA
cleavage. Furthermore, efforts have been made to further in-
crease the double-strand cleavage activity by covalent at-
tachment of a 9-aminoacridine DNA binding moiety to a di-
nuclear complex.
Results and Discussion
Since double-strand cleavage requires the delivery of two
oxidizing equivalents to the DNA, we hypothesized that co-
valently linking two single-strand cleaving agents, for exam-
ple, Fe-N4Py, might effectively give rise to double-strand
cleaving agents.[27] Recently, we reported on a binuclear iron
complex, based on N4Py, that showed significantly enhanced
double-strand cleavage activity.[28]
Multinuclear copper complexes have also been investigat-
ed in relation to DNA cleavage.[29–37] It has been reported
that dinuclear copper complexes show increased activity
compared to their mononuclear counterparts. Moreover,
with these dinuclear complexes, selective cleavage at junc-
tions between single- and double-strand DNA was observed.
Synthesis of the ligands: The ligands employed in this study
are shown in Figure 1. In all cases, the ligands were pre-
pared starting from a common precursor, that is, N4Py
propyl amine 9.[24] The ditopic ligands 2–4 and the tritopic
ligand 7 were prepared in moderate to good yields by reac-
tion of 9 with the appropriate N-hydroxysuccinimide-activat-
ed esters 10–13, following a procedure reported previously
(Scheme 1).[28]
For the synthesis of the tritopic ligand 8 with a flexible
core, a slightly modified approach was used (Scheme 2). The
core moiety was prepared starting from 4-ketopimelic acid.
Benzyl-protection of the carboxylic acid groups was fol-
Figure 1. Ligands used in this study.
1724
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1723 – 1733

Double-Strand DNA Cleavage with Iron Complexes
FULL PAPER
Scheme 1. Synthesis of ditopic N4Py ligands 2–4, 7. Reagents and conditions: a) CH2Cl2, overnight.
lowed by
a
Horner–Wads-
worth–Emmons reaction be-
tween 14 and triethyl phospho-
noacetate to give compound 15.
Hydrogenation of 15 resulted in
reduction of the double bond
with concomitant removal of
the two benzyl protecting
groups. Finally, hydrolysis of 16
yielded the triacid 17. Coupling
of triacid 17 with 9 in the pres-
ence of EDC, HOBt, and N,N-
diisopropylethylamine (DIEA)
in
dichloromethane[38]
was
found to be most efficient; the
target tritopic ligand 8 was ob-
tained in an excellent yield of
90%.
The first step towards the ac-
ridine-containing target ligands
5 and 6 was the synthesis of the
corresponding alkylated iso-
phthalic acid derivatives 18 and
19. The synthesis of 18 required
the use of dimethyl 5-hydroxy-
isophthalate as starting materi-
al, since the free diacid (5-hy-
droxyisophthalate) was found
to be unreactive with regard to
alkylation (Scheme 3). The re-
Scheme 2. Synthesis of tritopic ligand 8. a) Cs2CO3, benzyl bromide, DMF, 24 h; b) triethylphosphonoacetate,
NaH, THF, 08C, then 14, RT, 1 h, reflux overnight; c) Pd/C, H2, MeOH, overnight; d) LiOH, MeOH/H2O,
overnight; e) EDC, HOBt, DIEA, CH2Cl2, 1 h at 08C, then RT overnight.
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1725

G. Roelfes, B. L. Feringa et al.
Complexation: The correspond-
ing iron(II) complexes 29–36 of
the ligands 1–8 were generated
in situ by complexation with
(NH4)2FeII
ACHTNUTGRNEUNG(SO4)2·6H2O imme-
diately prior to use (Scheme 6).
The uptake of one iron(II)
ion per N4Py moiety was con-
Scheme 3. Synthesis of alkylated isophthalic acid 18. a) K2CO3, acetone, reflux overnight; b) LiOH, THF/H2O
1:1, 48 h.
firmed
by
titration
of
(NH4)2FeII
AHCTUNGTERN(GNUN SO4)2 against a se-
lection of ligands, with monitor-
1
action between dimethyl 5-hydroxyisophthalate and Boc-
protected aminopropyl bromide 20 in acetone in the pres-
ence of K2CO3 furnished 21, hydrolysis of which yielded the
diacid 18.
The synthesis of isomer 19 proved to be somewhat more
laborious (Scheme 4). First, 2-methoxyisophthalic acid was
converted into the corresponding dimethyl ester 22. Remov-
al of the methyl of the methoxy group by reaction with BBr3
in dry CH2Cl2 at ꢀ808C yielded dimethyl 2-hydroxyisoph-
thalate (50% over two steps; 23). Finally, alkylation of the
alcohol with Boc-protected bromopropylamine 20 in DMF
in the presence of K2CO3 yielded 24, which was hydrolyzed
to the desired diacid 19.
ing by H NMR, comparing the relative area of the absorp-
tion at d=25.2 ppm with that of acetone, as well as by UV/
Vis spectrophotometry. A typical example is shown in Fig-
ures 2 and 3, indicating binding of three FeII ions to the
ligand 7.[39]
DNA cleavage with the dinuclear complexes: The DNA
cleavage activities of the FeII complexes 29–36 were studied
using supercoiled pUC18 plasmid at 378C. Although DNA
cleavage activity was also observed in the absence of reduc-
ing agents, the experiments were carried out in the presence
of 1 mm dithiothreitol (DTT) as a reductant to increase the
rate of DNA cleavage. The final concentration of the cleav-
ing agent was 1 mm based on FeII, with a stoichiometry of
1:150 with respect to DNA base pairs. This corresponds to a
catalyst loading of 0.6 mol%.
Figure 4 shows the time dependences of the DNA cleav-
age activities for the iron complexes 29 (mononuclear) and
31 (dinuclear). Extensive DNA cleavage was observed with
all of the complexes. The main difference was in the amount
of linear DNA that was formed during the reaction. With
the dinuclear complexes 30–32, a sudden increase in the
amount of linear DNA was observed after about 10 min,
which was not seen with FeII-N4Py (29). After 60 min, the
fractions of linear DNA were 5% and 30% for 29 and 31,
respectively. Beyond this time, the complexes remained
active and continued to cleave DNA, but the DNA cleavage
activity could no longer be quantified.[6,40] The same reactivi-
ty pattern has been observed previously with other N4Py-
based dinuclear complexes having different linking moiet-
ies.[28] In the case of complexes 30–32, approximately equal
amounts of linear DNA were produced,[39] indicating that
these complexes show comparable activities towards double-
strand DNA cleavage.
The numbers of single-strand and double-strand cuts were
determined for each sample by statistical analysis using the
Poisson distribution.[41] Figure 5 shows a plot of the number
of double-strand breaks (m) versus the number of single-
strand breaks (n) for the complexes 29–32. From this plot, it
is quite clear that complex 29 is a typical single-strand cleav-
ing agent since it follows the Freifelder–Trumbo relation-
ship, which describes a purely single-strand cleaving path-
way.[8] In other words, double-strand cleavage occurs only
after extensive nicking of supercoiled DNA has taken place.
The plots for complexes 30–32 initially follow the Freifeld-
er–Trumbo relationship, but then suddenly show a strong in-
Scheme 4. Synthesis of alkylated isophthalic acid 19. a) conc. H2SO4,
MeOH, overnight; b) BBr3, CH2Cl2, ꢀ788C, 45 min; c) 20, K2CO3, DMF,
reflux overnight; d) LiOH, THF/H2O 1:1, 24 h.
Peptide coupling conditions (EDC/HOBt/N,N-diisopropy-
lethylamine) for the reactions between the dialkylated iso-
phthalic acids 18 and 19 and N4Py propylamine 9 furnished
25 and 26 in quantitative and 70% yields, respectively
(Scheme 5). The Boc protecting group was removed by
overnight treatment with a 10% solution of TFA in CH2Cl2.
The isolation of 27 and 28 was complicated by the high solu-
bility of these compounds in water. The final step involved
coupling of the free amines with 9-chloroacridine.[24] The re-
sulting ditopic ligands 5 and 6 were isolated as their HCl
salts in quantitative yields. After treatment with aqueous
NaOH the free base was obtained, which was used for char-
acterization. The HCl salts of 5 and 6 showed good stability;
no signs of degradation were observed after prolonged stor-
age. In contrast, the free base showed signs of degradation
within a few days.
1726
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1723 – 1733

Double-Strand DNA Cleavage with Iron Complexes
FULL PAPER
tween these complexes were ra-
tionalized by assuming a coop-
erative interaction between the
two copper centers to activate
dioxygen (or hydrogen perox-
ide) in the case of the 1,3-sub-
stitution pattern. The fact that
the DNA-cleavage results with
the present dinuclear iron com-
plexes are not dependent on
ligand topology suggests that
the two iron centers activate di-
oxygen independently.
In an attempt to further in-
crease the activity and double-
strand cleavage selectivity, di-
topic Fe-N4Py complexes 33
and 34 containing a DNA-inter-
calating 9-aminoacridine moiety
were prepared and tested. In
the case of mononuclear iron
complexes, this gave rise to sig-
Scheme 5. Synthesis of ligands 5 and 6. a) 9, HOBt, DIEA, EDC, CH2Cl2, overnight; b) 10% TFA in CH2Cl2,
overnight; c) 9-chloroacridine, phenol, 808C, 3 h.
nificantly
enhanced
DNA-
cleavage activity.[24]
Scheme 6. Complexation with FeII salts.
crease in the number of double-strand cuts (m) with respect
to the number of single-strand cuts (n). This is a clear indi-
cation that more linear DNA is formed via a direct double-
Figure 2. 1H NMR titration of 7 in D2O with (NH4)2FeII
ACTHNUGTRNEGNU(SO4)2 (c)
strand DNA cleavage pathway. Such a sudden increase in ds
cleavage activity has been observed previously, whereupon
it was demonstrated that the dinuclear iron complexes were
more efficient in bringing about double-strand cleavage of
nicked DNA than that of supercoiled DNA.[28] In view of
their structural similarity to the present complexes, this is
most likely the case here as well.
0 equiv FeII, (c) 1 equiv FeII, (c) 2 equiv FeII, (c) 3 equiv FeII,
and (c) 4 equiv FeII. Inset: Plot of the relative area of the absorption
at d=25.2 ppm compared to the acetone peak against the number of
equivalents of (NH4)2FeACTHNUTGRNEUNG(SO4)2.
Nearly identical m/n ratios were calculated for 30–32,
within the uncertainty limits of the data. Apparently, the
topology of the dinuclear complexes has little influence on
the DNA oxidation activity or on the type of cleavage path-
way. In this respect, the dinuclear iron complexes behave
differently to the previously studied dinuclear copper com-
plexes. Guo and co-workers found clear differences in the
DNA oxidation activities between 1,3- and 1,4-disubstituted
phenyl-based dicopper complexes.[37] The differences be-
Depending on the substitution pattern of the central
phenyl core, significantly different results were obtained in
the DNA-cleavage experiments. A much slower consump-
tion of supercoiled DNA was observed for 33 (t1/2 ꢁ25 min)
compared to 34 (t1/2 ꢁ10 min). The rate of consumption of
supercoiled DNA by the latter complex was comparable to
that by dinuclear complexes not bearing the acridine moiety
(see above). Furthermore, hardly any linear DNA was pro-
duced within 1 h when 33 was employed as catalyst (Fig-
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1727

G. Roelfes, B. L. Feringa et al.
Figure 3. UV/Vis spectra of ligand 7 (0.06mm) after addition of (d)
1 equiv, (g) 2 equiv, (a) 3 equiv, and (c) 4 equiv of (NH4)2FeII-
Figure 5. Number of double-strand cuts per DNA molecule (m) as a func-
tion of the number of single-strand cuts per DNA molecule (n) for 29
ACHTUNGTRENNUNG
!
~
&
*
(
), 30 ( ), 31 ( ), and 32 ( ). The dotted line is the Freifelder–Trumbo
AHCTUNGTRENNUNG
relationship.[8] Error bars represent the uncertainty of the data, based on
a Monte Carlo simulation, taking into account a standard deviation s of
0.03.[39]
Figure 4. Temporal evolution of the aerobic oxidation of supercoiled plas-
~
&
*
mid DNA ( ) to nicked DNA ( ) and linear DNA ( ) under catalysis by
a) [(1)FeII]2+ (29) and b) [(3)FeII
]
(31). Error bars represent the root-
4+
2
mean-square (rms) error based on three runs. A correction factor of 1.31
was used to compensate for the reduced ethidium bromide uptake ca-
pacity of supercoiled DNA.[28]
Figure 6. Temporal evolution of the aerobic oxidation of supercoiled plas-
~
&
*
mid DNA ( ) to nicked DNA ( ) and linear DNA ( ) under catalysis by
4+
4+
a) [(5)FeII
]
(33) and b) [(6)FeII
]
2
(34). Error bars represent the root-
2
mean-square (rms) error based on three runs. A correction factor of 1.31
was used to compensate for the reduced ethidium bromide uptake ca-
pacity of supercoiled DNA.
ure 6a), whereas more than 30% of linear DNA was formed
within the same period with 34 (Figure 6b).
A plot of the average number of double-strand versus
single-strand DNA cuts for 33 and 34 (Figure 7) shows a
similar pattern for both complexes; initially, the Freifelder–
Trumbo relationship (dashed line) is followed, implying a
single-strand cleavage pathway. In the later stages of the re-
action, double-strand cleavage activity is observed, similar
to what was seen with 30–32, albeit with less efficiency.
Interestingly, complex 34 displayed double-strand DNA
cleavage even in the early phase of the reaction (n>0.5),
whereas complex 33 required a higher number of single-
strand cuts (nꢁ2) before double-strand DNA cleavage com-
menced.
The observed difference between the acridine complexes
is most likely related to the relative positioning of the acri-
dine moiety with respect to the two Fe
ACHTUNGTREN(NGNU N4Py) moieties.
When the acridine of complex 33 (with a 1,3,5-substitution
1728
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1723 – 1733

Double-Strand DNA Cleavage with Iron Complexes
FULL PAPER
Figure 7. Number of double-strand cuts per DNA molecule (m) as a func-
tion of the number of single-strand cuts per DNA molecule (n) for 33
[8]
&
(&) and 34 ( ). The dotted line is the Freifelder–Trumbo relationship.
Error bars represent the uncertainty of the data, based on a Monte Carlo
simulation, taking into account a standard deviation s of 0.03.
pattern at the central phenyl moiety) intercalates with the
DNA, the two active iron moieties are pointing away from
the DNA, thus reducing the DNA-cleavage activity. In con-
trast, due to the 1,2,3-substitution pattern, the iron centers
of complex 34 are pointing towards the DNA and can
engage in its oxidation. From these results, it is clear that no
beneficial effect is obtained from attaching a DNA interca-
lator; at best the ds cleavage activity is slightly less than that
of the binuclear complexes without an acridine moiety. Pre-
sumably, the highly cationic binuclear iron complex already
has a very strong affinity for DNA and, hence, an additional
DNA-binding moiety has no added advantage. However,
more research is required to further substantiate these find-
ings.
Figure 8. Temporal evolution of the aerobic oxidation of supercoiled plas-
~
&
*
mid DNA ( ) to nicked DNA ( ) and linear DNA ( ) under catalysis by
6+
6+
a) [(7)FeII
]
(35) and b) [(8)FeII
]
3
(36). Error bars represent the root-
3
mean-square (rms) error based on three runs. A correction factor of 1.31
was used to compensate for the reduced ethidium bromide uptake ca-
pacity of supercoiled DNA.
son of the results with those obtained for a dinuclear com-
plex, namely the 1,3-substituted phenyl dinuclear complex
31, revealed a similar trend (Figure 9).
However, for the trinuclear complexes, the introduction
of double-strand breaks in the DNA was observed at an ear-
lier stage of the reaction and generally higher m/n ratios
were found. This suggests that the trinuclear complexes are
DNA cleavage with the trinuclear complexes: A similar
cleavage pattern was observed for [(7)FeII ]6+ (35) and
3
[(8)FeII ]6+ (36) (Figure 8). Initially, only nicked DNA was
3
formed, and linear DNA was only formed after a short lag
period, as was also observed with the dinuclear complexes.
It seems that 35, which has a phenyl core, is more effective
in producing linear DNA than the trimer 36 possessing a
more flexible core. In the case of 35, 30% of the DNA is in
the linear form after just 30 min. Within the same time, only
20% linear DNA is obtained with 36; a reaction time of
60 min is required for the fraction of linear DNA to reach
30% (Figure 8). Taken together, these results suggest that
both complexes display double-strand cleavage, but that the
more rigid complex 35 has higher activity.
As before, a plot of double-strand cuts versus single-
strand cuts shows that, initially, only single-strand cuts are
observed for both complexes (Figure 9). In the course of the
reaction, double-strand cuts are observed as well. Although
the trinuclear complex 35 produces double-strand cuts
somewhat more quickly than trinuclear complex 36, this is
not reflected in the m/n ratio. This points to a similar cleav-
age pathway, albeit with different reaction rates. Compari-
Figure 9. Number of double-strand cuts per DNA molecule (m) as a func-
&
tion of the number single-strand cuts per DNA molecule (n) for 35 ( )
~
and 36 (&). Data for a dinuclear complex are also included ( , 31). The
dotted line is the Freifelder–Trumbo relationship.[8] Error bars represent
the uncertainty of the data, based on a Monte Carlo simulation, taking
into account a standard deviation s of 0.03.
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1729

G. Roelfes, B. L. Feringa et al.
and immediately frozen in liquid nitrogen. The samples were analyzed by
gel electrophoresis. Results are the average of three independent runs.
Statistical analysis was employed to calculate the amounts of ssc and dsc
as described previously (see below).[28]
more efficient double-strand cleaving agents, but it should
be emphasized that the increase in double-strand cleavage
activity is not as large as was observed on going from mono-
nuclear complexes to dinuclear complexes. The higher
double-strand cleavage activity is most likely the result of
the increased probability of simultaneous delivery of multi-
ple oxidizing equivalents to both DNA strands within 16
base pairs of each other.
Calculation of the uncertainty in the values by a Monte Carlo method:
The uncertainties in the values of n and m were calculated by a Monte
Carlo simulation using the software program Mathematica ver-
sion 5.2.0.0. An algorithm was used to calculate the values of both n and
m from 5000 pseudo-random generated fraction sizes of supercoiled,
nicked, and linear DNA using the following equations:[6,8]
fm ¼ m ꢂ eꢀm
fI ¼ eꢀðmþnÞ
ð1Þ
ð2Þ
Conclusion
In the present study we have shown that oxidative double-
strand DNA cleavage activity can be achieved by using mul-
tinuclear iron complexes based on the N4Py ligand scaffold.
Although these complexes are still not as efficient as Fe-
BLM, they are nevertheless are the most efficient synthetic
iron-based double-strand DNA-cleaving agents known to
date. Compared to mononuclear iron complexes, binuclear
iron complexes show significantly enhanced double-strand
cleavage activity, which appears to be affected only to a
minor extent by the structure of the linking moiety. Cova-
lent attachment of a DNA-binding moiety did not give rise
to improved DNA-cleaving agents. However, trinuclear iron
complexes did exhibit significantly enhanced ds cleavage ac-
tivity, with complex 35, containing a rigid core, being the
most efficient. Taken together, the present results demon-
strate that by appropriate design of multinuclear non-heme
iron complexes, efficient oxidative double-strand DNA
cleavage can be achieved.
n2ð2h ¼ 1Þ
ð3Þ
m ¼
4L
Equation (3) is the Freifelder–Trumbo relationship. In these equations, fI
is the fraction of supercoiled DNA, fIII is the fraction of linear DNA, m is
the average number of double-strand cuts, n is the average number of
single-strand cuts, h is the maximum distance in base pairs between nicks
on opposite strands to generate a double-strand cut (i.e. 16), and L is the
total number of base pairs of the DNA used (2686 bp for pUC18 plasmid
DNA).
These pseudo-random generated fraction sizes are based on the experi-
mentally obtained average value (based on three independent runs) and
generated within a normal distribution based on a standard deviation of
0.03. This standard deviation was determined independently by 24 identi-
cal DNA oxidation experiments with supercoiled pUC18 plasmid DNA
and FeACTHUNRTGNEUNG(N4Py) (vide supra; reaction quenched at t=30 min) and is the
largest standard deviation in the experiment.
Synthesis of ligands 2–4 and 7 general procedure: Two equivalents of the
propylamine-substituted N4Py 9 were added to a solution (0.1 mm in
CH2Cl2) of one equivalent of the corresponding N-hydroxysuccinimide-
activated diacid 10, 11, or 12. The mixture was stirred overnight at room
temperature. It was then washed with water (3ꢁ10 mL) and dried over
Na2SO4. After filtration, the filtrate was concentrated and the product
was precipitated by adding Et2O. Further purification was achieved by
Experimental Section
size-exclusion chromatography on
a Sephadex LH20 column using
Chemicals and methods: All reactions were performed under an inert at-
mosphere (N2 or Ar). The N4Py-propylamine 9 was synthesized accord-
ing to literature procedures and all data were in agreement with those
published.[24,42] Syntheses of 10–24 can be found in the Supporting Infor-
mation.
MeOH as the mobile phase. The fractions were analyzed by HPLC using
method A (see the Supporting Information). After evaporation of the
MeOH, the oily product was taken up in the minimum volume of CH2Cl2
and the product was precipitated by adding Et2O.
N1,N2-Bis(3-{[(6-{[[di(2-pyridinyl)methyl](2-pyridinylmethyl)amino]meth-
yl}-3-pyridinyl)carbonyl]amino}propyl)phthalamide (2): Starting from 10
(39.3 mg, 0.109 mmol) and 9 (0.102 g, 0.218 mmol), 2 was obtained as a
pale light-brown solid (0.101 g, 95 mmol, 94%). 1H NMR (400 MHz,
CD3OD): d=8.78 (d, J=2.2 Hz, 2H), 8.45 (m, 4H), 8.35 (m, 2H), 8.04
(dd, J=8.1, 2.2 Hz, 2H), 7.79–7.64 (m, 14H), 7.55 (m, 2H), 7.49 (m, 2H),
7.25 (m, 4H), 7.18 (m, 2H), 5.33 (s, 2H), 3.99 (s, 4H), 3.95 (s, 4H), 3.44
(m, 8H), 1.85 ppm (m, 4H); 13C NMR (50.3 MHz, CD3OD): d=171.8,
167.7, 163.8, 160.7, 160.3, 150.0, 149.5, 148.7, 138.5, 138.4, 137.2, 136.9,
131.3, 129.9, 128.8, 125.6, 124.8, 124.1, 124.0, 123.7, 74.2, 58.7, 58.3, 38.3,
38.2, 29.8 ppm; MS (ESI+): m/z: 1065.5 [M+H]+, 895.6
pUC18 plasmid DNA, isolated from E. coli XL1 Blue, was purified using
QIAGEN maxi kits. Concentrations were determined by means of both
dilution gels and UV/Vis (A260). Restriction enzymes and restriction buf-
fers were purchased from New England Biolabs (NEB). Agarose used
for the agarose slab gels was purchased from Invitrogen. A solution of
number IV dye, consisting of 0.08% bromophenol blue and 40% sucrose
(6ꢁ conc.) was added to the samples prior to electrophoresis. All gel ex-
periments were run on 1.2% agarose slab gels for at least 90 min at 70 V.
Gels were stained in an ethidium bromide bath (1.0 mgmLꢀ1) for 45 min.
Pictures of the gel slabs were taken with a Spot Insight CCD camera and
handled with the software program Spot (version 3.4). The bands on the
film were quantified using the software program Gel-Pro Analyzer ver-
sion 4.0.00.001.
[MꢀPyCH2Py+H]+, 533.5 [M+2H]2+, 356.1 [M+3H]3+
.
N1,N3-Bis(3-{[(6-{[[di(2-pyridinyl)methyl](2-pyridinylmethyl)amino]meth-
yl}-3-pyridinyl)carbonyl]amino}propyl)isophthalamide (3): Starting from
11 (38 mg, 0.11 mmol) and 9 (0.10 g, 0.21 mmol), 3 was obtained as a pale
light-brown oil (65 mg, 0.06 mmol, 56%). 1H NMR (400 MHz, CD3OD):
d=8.83 (d, J=1.5 Hz, 2H), 8.47 (d, J=4.7 Hz, 4H), 8.37 (d, J=4.8 Hz,
2H), 8.33 (m, 1H), 8.10 (dd, J=8.1, 2.2 Hz, 2H), 7.98 (dd, J=7.7, 1.8 Hz,
2H), 7.81–7.66 (m, 15H), 7.54 (m, 2H), 7.27 (m, 4H), 7.21 (t, J=5.9 Hz,
2H), 5.36 (s, 2H), 4.01 (s, 4H), 3.98 (s, 4H), 3.49 (m, 8H), 1.93 ppm (m,
4H); 13C NMR (50.3 MHz, CD3OD): d=169.42, 167.89, 163.86, 160.78,
149.95, 149.50, 148.58, 138.42, 138.35, 136.87, 136.14, 131.16, 129.93,
127.30, 125.62, 124.81, 124.16, 124.00, 123.64, 74.40, 58.77, 58.37, 54.80,
DNA oxidation reactions: (NH4)2FeII
ACTHNUGTRENNUG(SO4)2 (1 equiv per N4Py moiety)
was added to solutions of ligands 1–8 in H2O. The respective mixtures
and a solution of dithiothreitol (DTT) were simultaneously added to a
buffered solution (Tris-HCl, 10 mm, pH 8.0) of supercoiled pUC18 plas-
mid DNA (0.1 mgmLꢀ1; 150 mm in base pairs) to give a final concentration
of 1.0 mm iron complex, based on iron, and 1.0 mm DTT. The mixture,
with a final volume of 50 mL, was incubated at 378C. At the times indi-
cated, a sample (2 mL) was taken from the reaction mixture, diluted in
water (15 mL; containing 1000 equiv of NaCN) and loading buffer (3 mL)
1730
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1723 – 1733

Double-Strand DNA Cleavage with Iron Complexes
FULL PAPER
38.43, 30.22 ppm; MS (ESI+): m/z: 1065.5 [M+H]+, 895.6
Et2O was decanted off and fresh Et2O (10 mL) was added. The yellow
suspension was stirred overnight. After decanting off the Et2O once
more, further fresh Et2O (10 mL) was added and the suspension was
placed in an ultrasonic bath for 10 min. After stirring for an additional
1 h, the Et2O was removed and the yellow solid was isolated and dried.
This HCl salt of the product was recovered in quantitative yield. A por-
tion of this yellow solid (100 mg) was taken up in 2m aqueous NaOH
(10 mL) and the solution was extracted with CH2Cl2 (3ꢁ10 mL). The
combined organic phases were washed with 2m NaOH (3ꢁ10 mL) and
dried over Na2SO4. Filtration and evaporation of the solvent yielded a
yellow oil (95 mg, quantitative). An optional purification step by prepara-
tive RP-HPLC yielded pure material. 1H NMR (400 MHz, CD3OD): d=
8.80 (dd, J=2.3 Hz, 0.5 Hz, 2H), 8.44 (m, 4H), 8.33 (ddd, J=5.0, 1.4,
0.9 Hz, 2H), 8.29 (d, J=8.7 Hz, 2H), 8.07 (dd, J=8.2, 2.3 Hz, 2H), 7.83
(t, J=1.4 Hz, 1H), 7.81–7.55 (m, 20H), 7.35 (d, J=1.4 Hz, 2H), 7.30
(ddd, J=8.7, 6.9, 1.4 Hz, 2H), 7.24 (ddd, J=6.0, 4.9, 2.6 Hz, 4H), 7.16
(ddd, J=6.9, 5.0, 1.5 Hz, 2H), 5.33 (s, 2H), 4.15 (t, J=5.6 Hz, 2H), 4.11
(t, J=6.6 Hz, 2H), 3.99 (s, 4H), 3.94 (s, 4H), 3.46 (m, 8H), 2.28 (m, 2H),
1.89 ppm (m, 4H); 13C NMR (100 MHz, CD3OD): d=167.9, 166.7, 162.7,
159.6, 159.1, 159.0, 153.7, 148.8, 148.3, 147.4, 137.22, 137.16, 136.2, 135.7,
130.7, 128.7, 126.5–126.0, 124.4, 123.9, 123.6, 123.0, 122.8, 122.5, 122.4,
118.2, 116.02, 115.97, 73.2, 65.9, 57.6, 57.2, 48.7, 37.24, 37.18, 30.3,
29.0 ppm (signals missing due to peak overlap); MS (ESI+): m/z: 1337.5
[M+Na]+, 1315.6 [M+H]+, 680.6 [M+2Na]2+, 669.6 [M+Na+H]2+, 658.6
[MꢀPyCH2Py+H]+, 533.5 [M+2H]2+, 356.1 [M+3H]3+
.
N1,N4-Bis(3-{[(6-{[[di(2-pyridinyl)methyl](2-pyridinylmethyl)amino]meth-
yl}-3-pyridinyl)carbonyl]amino}propyl)terephthalamide (4): Starting from
12 (38 mg, 0.11 mmol) and 9 (0.10 g, 0.21 mmol), 4 was obtained as a pale
light-brown oil (80 mg, 75.4 mmol, 71%). 1H NMR (400 MHz, CD3OD):
d=8.83 (d, J=1.5 Hz, 2H), 8.47 (d, J=4.8 Hz, 4H), 8.36 (d, J=4.8 Hz,
2H), 8.19 (dd, J=8.1, 1.8 Hz, 2H), 7.91 (s, 4H), 7.81–7.66 (m, 14H), 7.35
(m, 4H), 7.28 (t, J=5.5 Hz, 2H), 5.36 (s, 2H), 4.03 (s, 4H), 3.98 (s, 4H),
3.48 (dd, J=6.6, 6.2 Hz, 8H), 1.92 ppm (t, J=6.6 Hz, 4H); 13C NMR
(50.3 MHz, CD3OD): d=169.3, 167.9, 163.9, 160.8, 160.3, 145.0, 149.5,
148.6, 138.4, 136.9, 129.9, 128.5, 125.6, 124.8, 124.2, 124.01, 123.7, 74.4,
58.8, 58.4, 38.4, 30.2 ppm; MS (ESI+): m/z: 1065.5 [M+H]+, 895.5
[MꢀPyCH2Py+H]+, 533.5 [M+2H]2+
.
5-(N-tert-Butyloxycarbonyl-3-aminopropoxy)-N1,N3-bis(3-{[(6-{[[di(2-pyr-
idinyl)methyl](2-pyridinylmethyl)amino]methyl}-3-pyridinyl)carbonyl]
amino}propyl)isophthalamide (25): EDC (0.187 g, 2.04 mmol) and N,N-
diisopropylethylamine (89.9 mg, 115 mL, 0.696 mmol) were added to a
mixture of propylamine-N4Py derivative 9 (0.324 g, 0.693 mmol), HOBt
(94 mg, 0.696 mmol), and diacid 18 (0.107 g, 0.315 mmol) in CH2Cl2
(40 mL). The reaction mixture was stirred overnight at room tempera-
ture. Saturated aqueous NaHCO3 solution (20 mL) was then added and
the layers were separated. The organic phase was washed with saturated
aqueous NaHCO3 solution (2ꢁ20 mL) and water (20 mL). Drying over
Na2SO4, filtration, and evaporation of the solvent yielded a yellow foam
(0.417 g, 0.337 mmol, quant.). Optional purification was achieved by size-
exclusion chromatography (Sephadex LH 20/MeOH) or preparative RP-
HPLC. 1H NMR (400 MHz, CD3OD): d=8.81 (dd, J=2.3, 0.7 Hz, 2H),
8.46 (m, 4H), 8.35 (ddd, J=4.9, 1.7, 0.9 Hz, 2H), 8.09 (dd, J=8.2, 2.3 Hz,
2H), 7.88 (m, 1H), 7.82–7.61 (m, 14H), 7.52 (d, J=1.3 Hz, 2H), 7.26
(ddd, J=6.3, 4.9, 2.4 Hz, 4H), 7.19 (ddd, J=7.2, 4.9, 1.4 Hz, 2H), 5.34 (s,
2H), 4.09 (t, J=6.0 Hz, 2H), 4.01 (s, 4H), 3.96 (s, 4H), 3.47 (m, 10H),
3.23 (t, J=6.8 Hz, 2H), 1.92 (m, 6H), 1.40 ppm (s, 9H); 13C NMR
(100 MHz, CD3OD): d=168.1, 166.7, 162.7, 159.6, 159.5, 159.1, 157.3,
148.8, 148.3, 147.4, 137.3, 137.2, 136.2, 135.7, 128.8, 124.5, 123.6, 123.0,
122.8, 122.5, 118.1, 116.2, 78.8, 73.2, 65.9, 57.6, 57.2, 37.3, 37.2, 29.4, 29.0,
27.6 ppm; MS (ESI+): m/z: 1260.5 [M+Na]+, 1238.6 [M+H]+, 1068.5
[MꢀPyCH2Py+H]+, 620.0 [M+2H]2+, 536.0 [MꢀPyCH2Py+2H]2+, 413.9
[M+3H]3+, 395.2 [MꢀBoc+2Na+H]+, 380 [MꢀBoc+3H]+.
[M+2H]2+
,
570.1 [Mꢀacridine+2H]2+
, , 446.8
454.0 [M+2Na+H]3+
[M+Na+2H]3+, 439.5 [M+3H]3+; HRMS (ESI+): calcd for C78H76N16O5
[M+2H]2+
: m/z 658.309; found: 658.308, calcd for C78H77N16O5
[M+3H]3+: m/z 439.209; found 439.210.
2-(N-tert-Butyloxycarbonyl-3-aminopropoxy)-N1,N3-bisisophthalamide (26):
EDC (0.180 g, 1.16 mmol) and N,N-diisopropylethylamine (0.14 mL,
0.825 mmol) were added to a mixture of propylamine-N4Py derivative 9
(350 mg, 0.75 mmol), HOBt (112 mg, 0.825 mmol), and diacid 19
(127 mg, 0.375 mmol) in CH2Cl2 (10 mL). The reaction mixture was
stirred overnight at room temperature. The solvent was then removed
and fresh CH2Cl2 (20 mL) was added. The solution was washed with satu-
rated aqueous NaHCO3 solution (3ꢁ20 mL) and water (3ꢁ20 mL), and
then dried over Na2SO4. Evaporation of the solvent gave 26 as a brown
solid (322 mg, 0.261 mmol, 70%). 1H NMR (400 MHz, CD3OD): d=8.81
(m, 2H), 8.47 (m, 4H), 8.35 (m, 2H), 8.10 (m, 2H), 7.68 (m, 16H), 7.27
(m, 4H), 7.21 (m, 3H), 5.35 (s, 2H), 4.01 (s, 4H), 3.96 (s, 4H), 3.49 (t, J=
6.7 Hz, 4H), 3.30 (m, 8H), 3.18 (t, J=6.9 Hz, 2H), 1.92 (m, 6H),
1.33 ppm (s, 9H); 13C NMR (50.3 MHz, CD3OD): d=167.8, 166.7, 162.7,
159.6, 159.0, 154.4, 148.8, 148.4, 147.5, 137.2, 135.8, 132.0, 130.2, 128.8,
124.5, 124.2, 123.7, 123.0, 122.9, 122.6, 78.8, 74.1, 73.2, 57.6, 57.2, 37.3,
5-(3-Aminopropoxy)-N1,N3-bis(3-{[(6-{[[di(2-pyridinyl)methyl](2-pyridi-
nylmethyl)amino]methyl}-3-pyridinyl)carbonyl]amino}propyl)isophthala-
mide (27): The Boc-protected di-N4Py ligand 25 (0.417 g, 0.337 mmol)
was taken up in CH2Cl2 (9 mL), and then TFA (1 mL) was carefully
added. After stirring the mixture overnight at room temperature, 2m
aqueous NaOH solution was added until pH>10. The resulting solution
was stirred for an additional 10 min and then extracted with CH2Cl2 (3ꢁ
20 mL). The combined organic layers were washed with water (3ꢁ
20 mL) and dried over Na2SO4. Filtration and evaporation of the solvent
yielded a yellow-brown oil, which was purified on a Sephadex LH-20
column using MeOH as the mobile phase. A light-brown solid was ob-
tained (0.166 g, 0.146 mmol, 43%). 1H NMR (400 MHz, CD3OD): d=
8.81 (d, J=2.3 Hz, 2H), 8.45 (m, 4H), 8.35 (ddd, J=4.9, 1.6, 0.89 Hz,
2H), 8.09 (dd, J=8.2, 2.3 Hz, 2H), 7.89 (t, J=1.5 Hz, 1H), 7.82–7.62 (m,
14H), 7.53 (d, J=1.5 Hz, 2H), 7.26 (ddd, J=6.3, 4.9, 2.4 Hz, 4H), 7.19
(ddd, J=7.2, 4.9, 1.4 Hz, 1H), 5.34 (s, 2H), 4.15 (t, J=6.1 Hz, 2H), 4.01
(s, 4H), 3.96 (s, 4H), 3.47 (m, 8H), 2.89 (t, J=7.0 Hz, 2H), 1.99 (m, 2H),
1.90 ppm (dd, J=6.4, 6.3 Hz, 4H); 13C NMR (75.5 MHz, CD3OD): d=
168.0, 166.7, 162.7, 159.6, 159.4, 159.1, 148.8, 148.4, 147.4, 137.24, 137.18,
136.3, 135.7, 128.8, 124.5, 123.7, 123.0, 122.8, 122.6, 122.5, 118.2, 116.2,
73.3, 66.3, 57.6, 57.2, 38.2, 37.3, 37.2, 31.1, 29.0 ppm; MS (ESI+): m/z:
1160.5 [M+Na]+, 1138.8 [M+H]+, 968.6 [MꢀPyCH2Py+H]+, 570.5
30.3, 29.12, 7.7 ppm; MS (ESI+): m/z: 621 [M+2H]2+, 415 [M+3H]3+
2-(3-Aminopropoxy)-N1,N3-bis[3-(6-{[(dipyridin-2-ylmethyl)(pyridin-2-yl-
methyl)amino]methyl}nicotinamido)propyl]isophthalamide (28): 26
.
(320 mg, 0.26 mmol) was dissolved in a mixture of CH2Cl2 (10 mL) and
TFA (1 mL). After stirring overnight at room temperature, 2m aqueous
NaOH was added until pH >9. The resulting mixture was stirred for
15 min and then extracted with CH2Cl2 (3ꢁ20 mL). The combined ex-
tracts were washed with water (3ꢁ20 mL) and dried over Na2SO4. Evap-
oration of the solvent gave 28 as a yellow solid (200 mg, 0.176 mmol,
68%). 1H NMR (400 MHz, CD3OD): d=8.81 (m, 2H), 8.47 (m, 4H),
8.35 (m, 2H), 8.10 (m, 2H), 7.68 (m, 16H), 7.27 (m, 4H), 7.21 (m, 3H),
5.35 (s, 2H), 4.05 (t, J=6.1 Hz, 2H), 4.02 (s, 4H), 3.96 (s, 4H), 3.49 (m,
8H), 2.78 (t, J=6.9 Hz, 2H), 1.92 ppm (m, 6H); 13C NMR (50.3 MHz,
CD3OD): d=167.9, 166.8, 162.7, 159.6, 159.1, 154.3, 148.8, 148.4, 147.4,
137.2, 135.7, 131.9, 130.4, 128.8, 124.5, 123.7, 123.0, 122.9, 122.5, 73.3,
57.6, 57.2, 53.6, 37.7, 29.1 ppm; MS (ESI+): m/z: 380 [M+3H]3+
.
2-[3-(9-Acridinylamino)propoxy]-N1,N3-bis(3-{[(6-{[[di(2-pyridinyl)meth-
yl](2-pyridinylmethyl)amino]methyl}-3-pyridinyl)carbonyl]amino}propyl)-
isophthalamide (6): A mixture of 28 (150 mg, 0.13 mmol), 9-chloroacri-
dine (30 mg, 0.13 mmol), and phenol (1 g, 10 mmol) was heated at 808C
for 2 h under an N2 atmosphere. After cooling to room temperature,
Et2O (5 mL) was added and the suspension was stirred for 15 min. The
Et2O was then decanted off and fresh Et2O (5 mL) was added. The sus-
pension was stirred for a further 5 min. This procedure was repeated
[M+2H]2+, 380.5 [M+3H+]3+
.
5-[3-(9-Acridinylamino)propoxy]-N1,N3-bis(3-{[(6-{[[di(2-pyridinyl)meth-
yl](2-pyridinylmethyl)amino]methyl}-3-pyridinyl)carbonyl]amino}propyl)-
isophthalamide (5): A mixture of 27 (92 mg, 81 mmol), 9-chloroacridine
(17.3 mg, 81 mmol), and phenol (1.0 g) was heated at 808C for 3 h. After
cooling to room temperature, Et2O (10 mL) was added, which resulted in
the formation of a yellow-brown precipitate. After stirring for 5 min, the
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1731

G. Roelfes, B. L. Feringa et al.
three times. The Et2O was finally removed and the yellow solid was
taken up in CH2Cl2 (5 mL). The resulting solution was washed with 2m
aqueous NaOH (3ꢁ3 mL) and dried over Na2SO4. Evaporation of the
solvent gave 6 as a fine yellow solid (170 mg, 0.13 mmol, 99%). A small
fraction was purified by preparative RP-HPLC. 1H NMR (400 MHz,
CD3OD): d=8.73 (m, 2H), 8.44 (s, 4H), 8.32 (m, 2H), 7.98 (m, 2H), 7.74
(m, 16H), 7.67 (m, 8H), 7.25 (m, 4H), 7.16 (m, 3H), 5.31 (s, 2H), 4.20 (t,
J=5.8 Hz, 2H), 4.08 (t, J=6.5 Hz, 2H), 3.95 (s, 2H), 3.91 (s, 2H), 3.31
(m, 8H), 2.25 (m, 2H), 1.80 ppm (m, 4H); 13C NMR (50.3 MHz,
CD3OD): d=168.0, 166.6, 162.5, 159.5, 159.0, 154.3, 148.8, 148.3, 147.3,
137.2, 135.6, 131.7, 130.6, 128.6, 124.4, 123.6, 122.9, 122.8, 122.5, 74.3,
73.2, 57.6, 57.1, 53.7, 37.1, 31.1, 29.1 ppm; MS (ESI+): m/z: 658
Acknowledgement
We thank Dr. R. M. Scheek for useful discussions, and the Netherlands
Research School on Catalysis (NRSC-Catalysis) for financial support.
[1] H. Umezawa, K. Maeda, T. Takeuchi, Y. Okami, J. Antibiot. 1966,
19, 200.
[2] R. M. Burger, Chem. Rev. 1998, 98, 1153.
[3] S. M. Hecht, Bleomycin: Chemical, Biochemical and Biological As-
pects, Springer, New York, 1979.
[4] S. M. Hecht, J. Nat. Prod. 2000, 63, 158.
[5] W. K. Pogozelski, T. D. Tullius, Chem. Rev. 1998, 98, 1089.
[6] L. F. Povirk, W. Wꢂbker, W. Kçhnlein, F. Hutchinson, Nucleic Acids
Res. 1977, 4, 3573.
[7] J. Y. Chen, J. Stubbe, Nat. Rev. Cancer 2005, 5, 102.
[8] D. Freifelder, B. Trumbo, Biopolymers 1969, 7, 681.
[9] A. Decker, M. S. Chow, J. N. Kemsley, N. Lehnert, E. I. Solomon, J.
Am. Chem. Soc. 2006, 128, 4719.
[10] L. Thompson, C. Limoli, C. Origin, Recognition, signaling and repair
of DNA double strand mammalian cells, Springer, New York, 2003.
[11] R. J. Guajardo, S. E. Hudson, S. J. Brown, P. K. Mascharak, J. Am.
Chem. Soc. 1993, 115, 7971.
[12] C. Hemmert, M. Pitiꢃ, M. Renz, H. Gornitzka, S. Soulet, B. Meuni-
er, J. Biol. Inorg. Chem. 2001, 6, 14.
[13] R. P. Hertzberg, P. B. Dervan, J. Am. Chem. Soc. 1982, 104, 313.
[14] P. Mialane, A. Nivorojkine, G. Pratviel, L. Azꢃma, M. Slany, F.
Godde, A. Simaan, F. Banse, T. Kargar-Grisel, G. Bouchoux, J. Sain-
ton, O. Horner, J. Guilhem, L. Tchertanova, B. Meunier, J. J. Girerd,
Inorg. Chem. 1999, 38, 1085.
[15] M. Pitiꢃ, C. Boldron, G. Pratviel, Adv. Inorg. Chem. 2006, 58, 77.
[16] E. L. M. Wong, G. S. Fang, C. M. Che, N. Y. Zhu, Chem. Commun.
2005, 4578.
[M+2H]2+, 439 [M+3H]3+
.
N1,N3,N5-Tris(3-{[(6-{[[di(2-pyridinyl)methyl](2-pyridinylmethyl)amino]-
methyl}-3-pyridinyl)carbonyl]amino}propyl)benzene-1,3,5-tricarboxamide
(7): N4Py-propylamine 9 (93 mg, 0.20 mmol) and 13 (30 mg, 0.067 mmol)
were dissolved in CH2Cl2 (20 mL) and the mixture was stirred overnight.
Saturated aqueous NaHCO3 solution (10 mL) was then added and the
layers were separated. The aqueous layer was extracted with CH2Cl2 (3ꢁ
25 mL) and the combined organic fractions were washed with saturated
aqueous NaHCO3 solution (2ꢁ10 mL) and dried over Na2SO4. Filtration
and evaporation of the solvent yielded a light-brown solid, which was fur-
ther purified on a size-exclusion column (Sephadex LH 20 with MeOH;
200 mL column). It was obtained as a yellow oil (77.87 mg, 0.05 mmol,
1
75%). H NMR (400 MHz, CD3OD): d=8.81 (d, J=1.8 Hz, 3H), 8.45 (d,
J=4.6 Hz, 6H), 8.35 (d, J=4.7 Hz, 3H), 8.32 (m, 3H), 8.09 (m, 3H), 7.96
(m, 3H), 7.80–7.63 (m, 15H), 7.52 (m, 3H), 7.26 (m, 6H), 7.19 (m, 3H),
5.26 (s, 3H), 4.01 (s, 6H), 3.96 (s, 6H), 3.40 (m, 12H), 1.97–1.83 ppm (m,
6H); 13C NMR (50 MHz, CD3OD): d=168.67, 167.88, 163.83, 160.77,
160.24, 149.98, 149.52, 148.61, 138.36, 136.88, 136.67, 129.93, 125.63,
124.80, 124.18, 124.02, 123.63, 74.40, 58.75, 58.36, 38.52, 38.34, 30.17 ppm;
MS (ESI+): m/z: 1558.3 [M+H]+, 780.1 [M+2H]2+, 520.5 [M+3H]3+
.
[17] J. J. Hayes, Biochemistry 1996, 35, 11931.
[18] C. Marchand, C. H. Nguyen, B. Ward, J. S. Sun, E. Bisagni, T. Gar-
estier, C. Hꢃlꢄne, Chem. Eur. J. 2000, 6, 1559.
[19] M. Roy, B. Pathak, A. K. Patra, E. D. Jemmis, M. Nethaji, A. R.
Chakravarty, Inorg. Chem. 2007, 46, 11122.
[20] G. C. Silver, W. C. Trogler, J. Am. Chem. Soc. 1995, 117, 3983.
[21] Q. Jiang, N. Xiao, P. F. Shi, Y. G. Zhu, Z. J. Guo, Coord. Chem. Rev.
2007, 251, 1951–1972.
[22] R. P. Hertzberg, P. B. Dervan, Biochemistry 1984, 23, 3934–3945.
[23] Oxidative double-strand cleavage has been reported with Cu com-
plexes, albeit with high copper/DNA base pair ratios: a) F. V. Pama-
tong, C. A. Detmer, J. R. Bocarsly, J. Am. Chem. Soc. 1996, 118,
5339; b) Y. Jin, J. A. Cowan, J. Am. Chem. Soc. 2005, 127, 8408; c) J.
He, P. Hu, Y. J. Wang, M. L. Tong, H. Z. Sun, Z. W. Mao, L. N. Ji,
Dalton Trans. 2008, 3207.
[24] G. Roelfes, M. E. Branum, L. Wang, L. Que, B. L. Feringa, J. Am.
Chem. Soc. 2000, 122, 11517.
[25] M. Lubben, A. Meetsma, E. C. Wilkinson, B. Feringa, L. Que,
Angew. Chem. 1995, 107, 1610; Angew. Chem. Int. Ed. Engl. 1995,
34, 1512.
4-{2-(3-{[(6-{[[Di(2-pyridinyl)methyl](2-pyridinylmethyl)amino]methyl}-
3-pyridinyl)carbonyl]amino}propyl)-2-oxoethyl}-N1,N7-bis(3-{[(6-{[[di(2-
pyridinyl)methyl](2-pyridinylmethyl)amino]methyl}-3-pyridinyl)carbonyl]-
amino}propyl)heptanediamide (8): A solution of propylamine-N4Py de-
rivative 9 (0.10 g, 0.214 mmol), 4-(carboxymethyl)heptanedioic acid 17
(15.08 mg, 0.069 mmol), and HOBt (28.94 mg, 0.214 mmol) in CH2Cl2
(7.5 mL) was cooled to 08C. EDC (59.84 mg, 0.312 mmol) and N,N-diiso-
propylethylamine (27.8 mg, 0.215 mmol) were then added and the reac-
tion mixture was stirred for 1 h at 08C and at room temperature over-
night. The mixture was then washed with saturated aqueous NaHCO3 so-
lution (2ꢁ10 mL) and water (10 mL). The organic phase was dried over
Na2SO4, filtered, and concentrated to afford a light-yellow solid, which
was further purified on a size-exclusion column (Sephadex LH 20 with
MeOH; 200 mL column) to yield
8 (97.1 mg, 0.062 mmol, 90%).
1H NMR (300 MHz, CD3OD): d=8.88 (dd, J=4.6, 1.5 Hz, 3H), 8.48 (d,
J=4.1 Hz, 6H), 8.38 (d, J=3.2 Hz, 3H), 8.27–8.15 (m, 6H), 8.07 (t, J=
7.7 Hz, 6H), 7.70–7.35 (m, 27H), 7.17–6.95 (m, 6H), 5.33 (s, 3H), 3.97 (s,
6H), 3.92 (s, 6H), 3.48–3.31 (m, 12H), 1.74–1.47 ppm (m, 6H); 13C NMR
(126 MHz, CD3OD): d=176.2, 175.3, 167.8, 163.8, 160.8, 160.2, 150.0,
149.5, 148.6, 138.6, 138.5, 137.0, 130.0, 128.3, 126.6, 125.7, 124.9, 124.3,
124.1, 123.8, 74.6, 58.8, 58.5, 41.6, 38.4, 37.9, 36.1, 34.3, 30.7, 30.2 ppm;
[26] G. Roelfes, V. Vrajmasu, K. Chen, R. Y. N. Ho, J. U. Rohde, C. Zon-
dervan, R. M. la Crois, E. P. Schudde, M. Lutz, A. L. Spek, R. Hage,
B. L. Feringa, E. Munck, L. Que, Inorg. Chem. 2003, 42, 2639.
[27] For hydrolytic double-strand cleavage with a dinuclear iron com-
plex, see: R. X. Q. Chen, X. J. Peng, J. Y. Wang, Y. Wang, S. Wu,
L. Z. Zhang, T. Wu, Y. K. Wu, Eur. J. Inorg. Chem. 2007, 5400.
[28] T. A. van den Berg, B. L. Feringa, G. Roelfes, Chem. Commun.
2007, 180.
MS (ESI+): m/z: 523.2 [M+3H]3+, 392.7 [M+4H]4+
Titration studies of (NH4)2FeII(SO4)2 and 3: For 1H NMR titrations, 0, 1,
2, or 3 equivalents of (NH4)2FeII
(SO4)2 were added to a solution of 3 in
.
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
D2O (with acetone as internal standard) (Figure 2a). The 1H NMR spec-
tra were referenced with respect to the acetone peak (d=2.22 ppm),
since the residual solvent peak of D2O tends to shift after the addition of
[29] K. J. Humphreys, K. D. Karlin, S. E. Rokita, J. Am. Chem. Soc.
2001, 123, 5588.
the (NH4)2FeII
ACHTUNGTRENNUNG
[30] K. J. Humphreys, A. E. Johnson, K. D. Karlin, S. E. Rokita, J. Biol.
Inorg. Chem. 2002, 7, 835.
[31] K. J. Humphreys, K. D. Karlin, S. E. Rokita, J. Am. Chem. Soc.
2002, 124, 8055.
[32] K. J. Humphreys, K. D. Karlin, S. E. Rokita, J. Am. Chem. Soc.
2002, 124, 6009.
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
added to a solution of 3 in H2O (Figure 3a. The relative extinction coeffi-
cient (e/emax) was plotted against the number of equivalents of (NH4)2FeII-
ACHTUNGTRENNUNG(SO4)2 (Figure 3b).
1732
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1723 – 1733

Double-Strand DNA Cleavage with Iron Complexes
FULL PAPER
[33] T. Ito, S. Thyagarajan, K. D. Karlin, S. E. Rokita, Chem. Commun.
2005, 4812.
[40] The first term of a Poisson distribution predicts that the amount of
linear DNA will reach a maximum at around 37%; in practice, this
means that significant amounts of a smear are produced in the gel,
which precludes quantitative analysis.
[41] See the Experimental Section: Equations (1) and (2).
[42] G. Roelfes, Ph.D. Thesis, Rijksuniversiteit Groningen, 2000.
[34] L. Li, K. D. Karlin, S. E. Rokita, J. Am. Chem. Soc. 2005, 127, 520.
[35] L. Li, N. N. Murthy, J. Telser, L. N. Zakharov, G. P. A. Yap, A. L.
Rheingold, K. D. Karlin, S. E. Rokita, Inorg. Chem. 2006, 45, 7144.
[36] S. Thyagarajan, N. N. Murthy, A. A. N. Sarjeant, K. D. Karlin, S. E.
Rokita, J. Am. Chem. Soc. 2006, 128, 7003.
[37] Y. M. Zhao, J. H. Zhu, W. J. He, Z. Yang, Y. G. Zhu, Y. Z. Li, J. F.
Zhang, Z. J. Guo, Chem. Eur. J. 2006, 12, 6621.
[38] L. C. Chan, B. G. Cox, J. Org. Chem. 2007, 72, 8863.
[39] See also the Supporting Information.
Received: July 11, 2008
Revised: September 29, 2008
Published online: January 7, 2009
Chem. Eur. J. 2009, 15, 1723 – 1733
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1733