Triplex-directed interstrand DNA cross-linking by diaziridinylquinone- oligonucleotide conjugates

Article abstract of DOI:10.1021/ja973819u

Triplex-forming oligonucleotides (TFOs) bind in the major groove to specific double-helical DNA sequences and have been shown to inhibit the function of targeted genes. Diaziridinylquinones are bifunctional alkylating agents that can form interstrand cros

Full text of DOI:10.1021/ja973819u

VOLUME 120, NUMBER 38
SEPTEMBER 30, 1998
©
Copyright 1998 by the
American Chemical Society
Triplex-Directed Interstrand DNA Cross-Linking by
Diaziridinylquinone-Oligonucleotide Conjugates
Michael W. Reed,* Ansel Wald, and Rich B. Meyer
Contribution from Epoch Pharmaceuticals Incorporated, 1725 220th Street SE, #104,
Bothell, Washington 98021
ReceiVed NoVember 6, 1997
Abstract: Triplex-forming oligonucleotides (TFOs) bind in the major groove to specific double-helical DNA
sequences and have been shown to inhibit the function of targeted genes. Diaziridinylquinones are bifunctional
alkylating agents that can form interstrand cross-links in DNA at 5′-GNC sites. To demonstrate the feasibility
of targeted interstrand cross-linking, a diaziridinylquinone-TFO conjugate was prepared from a diaziridi-
nylquinone intermediate bearing an activated ester linker. The first demonstration of triplex-directed interstrand
DNA cross-linking by a single targeted bifunctional alkylating agent is described. Up to 38% interstrand
cross-linking was observed at pH 6.2.
Introduction
specific mutations in plasmid DNA models.4 Interstrand cross-
links are believed to be especially cytotoxic, but the mutagenic
effects of these lesions are not clear. Although alkylation of
Triplex-forming oligonucleotides (TFOs) bind in the major
groove to specific double-stranded DNA (dsDNA) sequences
and have been shown to inhibit function of targeted genes. DNA
alkylating agents have been attached to TFOs to stabilize the
triple-stranded complex, and photoactivated DNA cross-linking
by psoralen-TFO conjugates has been shown to induce sequence
5
one strand of dsDNA has been demonstrated using a variety of
TFO-directed electrophilic agents, this report is the first case
of targeted DNA interstrand cross-linking by a single bifunc-
1
2
3
6
tional alkylating agent.
Diaziridinylquinones are bifunctional alkylating agents that
7
form interstrand cross-links in DNA. A variety of sym-
*
Corresponding author. Telephone: 425-485-8566. Fax: 425-486-8336.
E-mail: mreed@epochpharm.com.
1) For reviews on TFOs, see: (a) Thuong, N. T.; Helene, C. Angew.
metrically substituted diaziridinylbenzoquinone analogues have
8
(
been examined for DNA cross-linking properties. In general,
Chem., Int. Ed. Engl. 1993, 32, 666-690. (b) Mirkin, S. M.; Frank-
Kamenetskii, M. D. Annu. ReV. Biophys. Biomol. Struct. 1994, 23, 541-
(3) The term “cross-linking” as used here refers to interstrand alkylation
of the targeted dsDNA. Alkylation of a single strand in the dsDNA target
is defined as “monoalkylation”.
(4) (a) Havre, P. A.; Gunther, E. J.; Gasparro, F. P.; Glazer, P. M. Proc.
Natl. Acad. Sci. U.S.A. 1993, 90, 7879-7883. (b) Wang, G.; Glazer, P. M.
J. Biol. Chem. 1995, 270, 22595-22601. (c) Raha, M.; Wang, G.; Seidman,
M. M.; Glazer, P. M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 2941-2946.
(5) Sanderson, J. S.; Shield, A. J. Mutat. Res. 1996, 355, 41-57 and
references therein.
(6) Interstrand DNA cross-linking by a 3′,5′-bischlorambucil TFO was
described in ref 2f.
(7) (a) Szmigiero, L.; Kohn, K. W. Cancer Res. 1984, 44, 4453-4457.
(b) Szmigiero, L.; Erickson, L. C.; Ewig, R. A. G.; Kohn, K. W. Cancer
Res. 1984, 44, 4447-4452.
5
76. (c) Sun, J. S.; Garestier, T.; Helene, C. Curr. Opin. Struct. Biol. 1996,
6
, 327-333.
(2) (a) Vlassov, V. V.; Gaidamakov, S. A.; Zarytova, V. F.; Knorre, D.
G.; Levina, A. S.; Nikonova, A. A.; Podust, L. M.; Federova, O. S. Gene
1
988, 72, 313-322. (b) Povsic, T. J.; Dervan, P. B. J. Am. Chem. Soc.
1
990, 112, 9428-9430. (c) Takasugi, M.; Guendouz, A.; Chassignol, M.;
Decout, J. L.; Lhoumme, J.; Thuong, N. T.; Helene, C. Proc. Natl. Acad.
Sci. U.S.A 1991, 88, 5602-5606. (d) Shaw, J.-P.; Milligan, J. F.; Krawczyk,
S. H.; Matteucci, M. J. Am. Chem. Soc. 1991, 113, 7765-7766. (e) Gruff,
E. S.; Orgel, L. E. Nucleic Acids Res. 1991, 19, 6849-6854. (f) Kutyavin,
I. V.; Gamper, H. B.; Gall, A. A.; Meyer, R. B. Jr. J. Am. Chem. Soc.
1
993, 115, 9303-9304. (g) Colombier, C.; Lippert, B.; Leng, M. Nucleic
Acids Res. 1996, 24, 4519-4524.
S0002-7863(97)03819-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/11/1998

9730 J. Am. Chem. Soc., Vol. 120, No. 38, 1998
Reed et al.
the cross-linking efficiency of these agents is increased by
reduction of the quinone to the hydroquinone or by lowering
pH. Substituents can alter electron density in the benzoquinone
ring and reactivity of the aziridines. For example, 2,5-
diaziridinyl-1,4-benzoquinone (DZQ) can cross-link DNA at pH
Scheme 1a
7
.2, whereas the 3,6-dimethyl-substituted analogue (MeDZQ)
8b
required enzymatic reduction at the same pH.
Reductive
activation can also alter the sequence selectivity of these agents.
For example, the hydroquinone form of DZQ cross-links
preferentially at 5′-GC or 5′-GNNC sites, whereas the hydro-
quinone of MeDZQ prefers 5′-GNC sites.8 These sequence
c
8
d
preferences have been rationalized by molecular modeling.
Diaziridinylquinone cross-links DNA in the major groove
between the N-7 atoms of dG residues on opposite strands of
9
the duplex, an ideal location for targeted delivery by TFOs.
To demonstrate the feasibility of targeted interstrand cross-
linking, we prepared a diaziridinylquinone-TFO conjugate (7)
from an activated ester intermediate (6) as shown in Scheme 1.
Results and Discussion
Synthesis. We have developed methods to selectively couple
alkylating agents bearing activated ester linkers to amine-
modified oligonucleotides.1
0,11
Anhydrous conditions allow
selective acylation of the oligo without competing side reactions
of the DNA reactive agent. A suitable ester linker arm was
first introduced into the diaziridinylquinone precursor (2) by
palladium-catalyzed coupling12 of 4-iodo-2,5-dimethoxytoluene
(1) and methyl 4-pentynoate followed by hydrogenation of the
alkyne to give the aliphatic linker arm in 3. Ceric ammonium
nitrate (CAN) oxidation13 gave an intermediate quinone (4) that
was reacted with excess ethyleneimine to give the diaziridi-
nylquinone methyl ester (5). Careful saponification and treat-
14
ment with 2,3,5,6-tetrafluorophenyl trifluoroacetate (TFP-TFA)
gave the desired diaziridinylquinone TFP ester (6). As shown
in Figure 1, reaction of the triethylammonium salt of a
a
(
3 4 2 2
a) Methyl pentynoate, Pd(PPh ) , CuI, EtN(iPr ), DMF; (b) H ,
5
′-hexylamine-modified TFO with excess 6 in DMSO gave
Pd-C, triethylammonium formate (pH 6.5), THF, EtOH; (c)
NH Ce(NO , CH CN; (d) ethyleneimine, Cu(OAc) , MeOH; (e)
LiOH (aq), MeOH; (f) 2,3,5,6-tetrafluorophenyl trifluoroacetate, Et N,
CH Cl ; (g) TFO-NH , Et N, DMSO.
clean conversion to diaziridinylquinone-TFO (7). Reverse
phase HPLC purification and precipitation gave 56% recovery
of conjugate 5 (97% pure by HPLC).
(
4
)
2
)
3 6
3
2
3
2
2
2
3
Triplex Model System. For the DNA alkylation studies
described here, a 21-mer G/A motif TFO was chosen which
targets a homopurine run in an endogenous human gene as
shown in Figure 2. We used an analogous nitrogen mustard-
TFO conjugate to alkylate this specific sequence in genomic
1
5
DNA with 90% efficiency at a concentration of 0.5 µM. In
(
8) (a) Hartley, J. A.; Berardini, M.; Ponti, M.; Gibson, N. W.; Thompson,
A. S.; Thurston, D. E.; Hoey, B. M.; Butler, J. Biochemistry 1991, 30,
1719-11724. (b) Lee, C.-S.; Hartley, J. A.; Berardini, M. D.; Butler, J.,
Siegel, D.; Ross, D.; Gibson, N. W. Biochemistry 1992, 31, 3019-3025.
c) Berardini, M. D.; Souhami, R. L.; Lee, C. S.; Gibson, N. W.; Butler, J.;
1
(
Hartley, J. A. Biochemistry 1993, 32, 3306-3312. (d) Haworth, I. S.; Lee,
C.-S.; Yuki, M.; Gibson, N. W. Biochemistry 1993, 32, 12857-12863.
(
9) (a) Alley, S. C.; Brameld, K. A.; Hopkins, P. B. J. Am. Chem. Soc.
Figure 1. Reaction of 5′-hexylamine-modified TFO with TFP ester
. Progress of the reaction was monitored by C18 HPLC as described
in the Experimental Procedures. The upper chromatogram shows
starting hexylamine-modified TFO. The middle chromatogram shows
the reaction mixture after 1 h. The lower chromatogram shows HPLC-
purified diaziridinylquinone-TFO (7).
1
1
994, 116, 2734-2741. (b) Alley, S. C.; Hopkins, P. B. Chem. Res. Toxicol.
6
994, 7, 666-672.
(10) Reed, M. W.; Lukhtanov, E. A.; Gorn, V.; Kutyavin, I.; Gall, A.;
Wald, A.; Meyer, R. B., Jr. Bioconjugate Chem. 1998, 9, 64-71.
11) Lukhtanov, E. A.; Kutyavin, I. V.; Gorn, V. V.; Reed, M. W.;
Adams, A. D.; Lucas, D. D.; Meyer, R. B. J. Am. Chem. Soc. 1997, 119,
(
6
214-6225.
earlier work,16 we showed that G/A motif TFOs formed more
stable triplexes than the C/T or G/T motif for this G-rich target,
(
12) Hobbs, F. W., Jr. J. Org. Chem. 1989, 54, 3420-3422.
(13) Jacob, P., III; Callery, P. S.; Shulgin, A. T.; Castagnoli, N. J. Org.
Chem. 1976, 41, 3627-3629.
14) Gamper, H. B.; Reed, M. W.; Cox, T.; Virosco, J. S.; Adams, A.
D.; Gall, A. A.; Scholler, J. K.; Meyer, R. B., Jr. Nucleic Acids Res. 1993,
1, 145-150.
15) Belousov, E. S.; Afonina, I. A.; Podyminogin, M. A.; Gamper, H.
B.; Reed, M. W.; Meyer, R. B. Nucleic Acids Res. 1997, 25, 3440-3444.
(
(16) Gamper, H. B.; Kutyavin, I. V.; Rhinehart, R. L.; Lokhov, S. G.;
Reed, M. W.; Meyer, R. B. Biochemistry 1997, 36, 14816-14826. As
described in ref 15, an incorrect sequence recorded in GenBank was used
for these initial studies. The correct dsDNA sequence is used for the work
described here.
2
(

Triplex-Directed DNA Cross-Linking
J. Am. Chem. Soc., Vol. 120, No. 38, 1998 9731
Figure 2. Design of model triplex system for interstrand cross-linking.
The diaziridinylquinone linker structure (Q) in TFO 7 is shown in
Scheme 1. Only the hybridization region is shown for the 65-bp dsDNA
target. The alkylated dG residues in the purine-rich strand (Pu) and
pyrimidine-rich strand (Py) are indicated by arrows. Since cell culture
experiments were ultimately planned, a 3′-hexanol modification was
incorporated into TFO 7 to provide resistance to nuclease degradation.
especially in the presence of the triplex-specific intercalating
17
agent coralyne. There is a potential 5′-GAC cross-linking site
just downstream of the homopurine run in the dsDNA target
that we hoped to alkylate with the diaziridinylquinone conjugate
group. Despite the preferred18 5′-GXC mustard cross-linking
sequence just downstream of the homopurine run, only monoalky-
1
9
lation products were observed for the mustard-TFO. Under
the conditions used here (100:1 TFO:dsDNA, 10 µM coralyne),
no more than 48% monoalkylation was observed for the
mustard-TFO analogue of TFO 7.10
DNA Alkylation Assay. Sequence-specific alkylation by
TFO 7 of a 65-mer dsDNA fragment containing the target site
was examined by denaturing polyacrylamide gel electrophoresis
(
PAGE). The time course of DNA alkylation at 37 °C was
32
determined for each 5′- P-labeled target strand at pH 6.2 as
shown in Figure 3. Three bands in each of the imaged gels
were observed which correspond to an interstrand cross-linked
species, a monoalkylated species, and the unmodified labeled
single-stranded target. At longer time points, a band due to
spontaneous thermal depurination and cleavage of alkylated
products appeared.
Figure 3. Alkylation of dsDNA target by TFO 7 at pH 6.2. Preformed
duplex (20 nM labeled strand and 40 nM unlabeled strand) was
incubated for 24 h at 37 °C with TFO 7 (2 µM) in 20 mM PIPES (pH
6
2
.2), 140 mM KCl, 10 mM MgCl , 1 mM spermine, and 10 µM
coralyne chloride. Denaturing gel electrophoresis analysis of the
products from each 32P-labeled strand vs time is shown. Upper bands
are interstrand cross-linked products, intermediate bands are monoalky-
lation products, and lower bands are unmodified labeled strands. The
lowest bands (at long time points) correspond to cleavage of alkylated
products. The CL lane shows strand cleavage by heat/piperidine
treatment of the 24-h sample (63% cleavage). MeDZQ lane shows 24-h
incubation of the labeled duplex with 10 µM MeDZQ (no coralyne).
For each labeled DNA target, the slow-moving cross-linked
species increased to a maximum at 6 h, then slowly degraded.
As discussed below, the labeled Pu strand showed an intermedi-
ate monoalkylated product, whereas the labeled Py strand
showed only a trace of monoalkylation. Lane CL for the Py-
labeled target shows the short (15-mer) DNA fragments that
result from heat/piperidine strand cleavage after 24 h of
alkylation. A single major band appeared, indicative of the
specificity of alkylation. This band (63% of the labeled Py
strand) represents the sum of interstrand cross-linked products
that had formed (and/or degraded) over 24 h. A control (lane
MeDZQ) with 10 µM of Me-DZQ showed no interstrand DNA
cross-linking, thus demonstrating the increased DNA reactivity
of the diaziridinylquinone upon targeting with the TFO.
The dynamics of the interstrand cross-linking process is
shown in Figure 4. The upper panel shows how monoalkylated
product builds rapidly in the Pu strand, then slows as some of
these products are converted to cross-links. For the labeled Py
strand (lower panel), the cross-linked product forms at the same
rate as for the Pu strand, but only a trace of monoalkylated
product is observed. Apparently the N-7 dG target in the Py
strand can react only after prior alkylation of the N-7 dG site
in the Pu strand. Modeling studies show that alkylation of dG
in the Pu strand (by the less hindered aziridine next to the methyl
group) allows the remaining aziridine (next to the linker) to
stretch across the major groove and react with dG in the Py
strand. The TFO is held to the Pu strand (by reverse Hoogsteen
H bonds) which may facilitate the initial alkylation event.
Degradation of the cross-linked DNA presumably occurs by
initial depurination of the alkylated Py strand (since no
monoalkylation products appear in the labeled Py strand).
Finally, depurination of the (more stable) alkylated Pu strand
occurs to give the short cleaved fragments apparent at 24 h.
Effect of pH on Interstrand Cross-Linking Efficiency.
Since the mechanism of DNA alkylation by (nonreduced)
aziridinylquinones requires protonation of the aziridine, it was
expected that alkylation by TFO 7 would be slower at neutral
pH. As shown in Figure 5, this was the case. DNA alkylation
by TFO 7 was measured over time at pHs 6.2, 6.7, and 7.2
using carefully prepared PIPES buffers. Only the labeled Pu
strand was examined since we were also interested in monoalky-
lation rates. At pH 7.2, the diaziridinylquinone-TFO gave up
to 17% interstrand cross-linking under the assay conditions but
required extended reaction time with a plateau at 24 h.
Monoalkylation at pH 7.2 was 47% at 24 h (data not shown).
At pH 6.7, the interstrand cross-linking reached a maximum
25% at 720 min (42% monoalkylation). At pH 6.2, a plateau
of 38% interstrand cross-linking was reached at 360 min (36%
monoalkylation). It makes sense that alkylation rates are faster
at lower pH, but it is not clear why the maximum percent
(
17) Lee, J. S.; Latimer, L. J. P.; Hampel, K. J. Biochemistry 1993, 32,
591-5597. This reference describes stabilization of C/T motif triplexes
by coralyne.
5
(
18) (a) Ojwang, J. O.; Grueneberg, D. A.; Loechler, E. L. Cancer Res.
1
989, 49, 6529-6537. (b) Millard, J. T.; Raucher, S.; Hopkins, P. B. J.
Am. Chem. Soc. 1990, 112, 2459-2460.
(
19) The mustard-TFO analogue of 7 alkylated primarily the dG residue
in the purine-rich strand.

9732 J. Am. Chem. Soc., Vol. 120, No. 38, 1998
Reed et al.
Figure 6. Aqueous stability of TFO 7 at pH 6.2 (37 °C). Degradation
was monitored by C18 HPLC as described in the Experimental
Procedures. Starting 7 (top chromatogram) was analyzed at 3, 6, 12,
and 24 h (bottom chromatogram).
important factor for cross-linking efficiency. The electrophilic
aziridine rings can react with other nucleophiles or water to
give degradation products that compete with intact TFO and
prevent targeted cross-linking. The G-rich nature of TFO 7 is
also a potential liability since self-alkylation is possible. For
example, a G/A motif mustard-TFO in a related target was
found to self-associate as parallel stranded duplexes and form
2
0
alkylated products. A reverse phase HPLC assay was used
to determine the aqueous half-life of TFO 7 at pH 6.2 and 7.2
(
37 °C). As shown in Figure 6, 7 decomposes at pH 6.2 to a
2
1
complex mixture of faster eluting products. At pH 6.2, only
1
7
1% intact 7 remained after 24 h (t1/2 ) 6 h). After 24 h at pH
.2, 64% intact TFO 7 remained.
Conclusion
The interstrand DNA cross-linking observed for the 3,6-
dialkyl-substituted diaziridinylquinone when linked to the TFO
shows the advantage of sequence-specific targeting. The
analogous dimethyl-substituted compound (MeDZQ) required
Figure 4. Dynamics of DNA alkylation by TFO 7 at pH 6.2. Results
of the PAGE assay shown in Figure 3 were quantitated by phospho-
rimaging. The upper panel shows the fate of the labeled Pu strand, and
the lower panel shows the fate of the labeled Py strand.
8
c
enzymatic reduction for DNA cross-linking. This hybridiza-
tion-driven reactivity is presumably due to the higher effective
local concentration of alkylating agent at the targeted site, but
other interactions in the major groove may contribute. Molec-
ular modeling studies show that interstrand cross-linking by
diaziridinylquinones at the 5′-GAC flanking sequence can be
accomplished with little distortion of the B-DNA structure. This
is in contrast with nitrogen mustard-mediated cross-linking that
requires a severe reduction in helical twist at the preferred cross-
1
8
link site.
The linker system in TFO 7 has not been optimized, and self-
alkylation with the G/A motif is a competing side reaction. We
are currently exploring structural modifications of the diaziri-
dinylbenzoquinone cross-linkers and determining the sequence
specificity of DNA cross-link formation in other triplex systems.
An understanding of the geometric constraints for cross-link
formation should allow us to optimize the TFO linker structure
for various genetic targets. By analogy with MeDZQ, reduction
of the diaziridinylquinone to the hydroquinone should increase
DNA reactivity.
Figure 5. Effect of pH on interstrand DNA cross-linking by TFO 7.
The pH 6.2 PAGE assay described in Figure 3 was repeated at pH 6.7
and pH 7.2. The percent interstrand cross-link was quantitated for each
time point by phosphorimaging.
Diaziridinylquinone-TFOs may have applications in gene
therapy if cellular (chromatinized) dsDNA can be modified.
They may also be used as rare-cutting DNA cleavage agents.
We showed up to 63% targeted DNA cleavage by heat/
interstrand cross-linking decreases at higher pH. Competing
side reactions (such as alkylation of spermine in the triplex
buffer by the aziridine groups in 7) or less efficient triplex
formation at neutral pH may be possible explanations.
Aqueous Stability/Self-Alkylation of TFO 7. The aqueous
stability of diaziridinylquinone-TFO conjugates may be an
(
20) Lampe, J. N.; Kutyavin, I. V.; Rhinehart, R.; Reed, M. W.; Meyer,
R. B.; Gamper, H. B. Nucleic Acids Res. 1997, 25, 4123-4131.
21) Similar mixtures of degradation products were observed for
decomposition of mustard-TFOs (see ref 10).
(

Triplex-Directed DNA Cross-Linking
J. Am. Chem. Soc., Vol. 120, No. 38, 1998 9733
piperidine treatment. As shown by others,22 N-7 alkylation sites
in the DNA target give dsDNA fragments with residual
phosphates that can be used as substrates for ligation enzymes.
In this sense, these agents are the first example of synthetic
mixture was filtered through diatomaceous earth and washed with
ethanol. The filtrate was concentrated in vacuo, and the residue was
partitioned between 3% sodium bicarbonate and ethyl acetate. The
organic layer was evaporated to dryness to give 1.08 g (96%) of 3 as
1
a colorless liquid: TLC (2:1 hexanes-ethyl acetate) R
NMR (CDCl
H), 3.67 (s, 3 H), 2.60 (t, 2 H, J ) 7.7 Hz), 2.36 (t, 2 H, J ) 7.1 Hz),
.8-1.6 (m, 4 H). Anal. Calcd for C15 : C, 67.65; H, 8.33.
Found: C, 67.39; H, 8.10.
,4-Cyclohexadiene-1-pentanoic Acid, 4-Methyl-3,6-dioxo-, Meth-
f
) 0.80; H
“
restriction enzymes”. In addition, the diaziridinylquinone
3
) δ 6.67 (s, 1 H), 6.64 (s, 1 H), 3.80 (s, 3 H), 3.77 (s, 3
activated ester 6 is suitable for conjugation to a wide variety of
amine-containing targeting ligands for evaluation as possible
anticancer agents.
1
22 4
H O
1
Experimental Procedures
yl Ester (4). To a solution of 1.08 g (4.07 mmol) of 3 in 20 mL of
acetonitrile was added a solution of 4.69 g (8.56 mmol) of ceric
ammonium nitrate in 10 mL of water. The aqueous solution was added
to the stirred organic solution, dropwise, with stirring, until the transient
green color that occurs during addition is no longer visible (about 2
equiv of CAN were used). After the mixture was stirred for 45 min,
the solvents were removed and the residue was partitioned between
water and ethyl acetate. The organic layer was dried over sodium
sulfate and evaporated to dryness. The residue was purified by flash
chromatography (5 × 50 cm silica) using 4:1 hexanes-ethyl acetate.
The yellow band was collected and evaporated, and the residue was
recrystallized from aqueous methanol solid to give 0.69 g (71% yield)
1H NMR spectra were run on a Varian Gemini 300-MHz spectrom-
eter. Elemental analyses were performed by Quantitative Techonologies
Inc. (Boundbrook, NJ). Melting points were determined on a Mel-
Temp melting point apparatus in open capillary tubes and are
uncorrected. All air- and water-sensitive reactions were carried out
under a slight positive pressure of argon. Flash chromatography was
performed on 230-400 mesh silica gel. Analytical thin-layer chro-
matography was carried out on EM Science F254 aluminum-backed,
fluorescent indicator plates. Triethylammonium bicarbonate (TEAB)
(0.1 M) was prepared by sparging an aqueous mixture of the appropriate
amine with CO until the organic layer disappeared. MeDZQ was
2
of the quinone 4 as yellow crystals: mp ) 46-48 °C; TLC (4:1
prepared by treatment of 2,5-dimethyl-p-benzoquinone (Acros Organics)
1
1
hexanes-ethyl acetate) R
f
) 0.45; H NMR (CDCl
3
) δ 6.59 (s, 1 H),
with excess ethyleneimine and was characterized by H NMR.
6
.55 (s, 1 H), 3.67 (s, 3 H), 2.43 (t, 2 H, J ) 7.4 Hz), 2.35 (t, 2 H, J
4
-Iodo-2,5-dimethoxytoluene (1). To a stirred solution of 7.19 g
)
for C13
7.1 Hz), 2.04 (s, 3 H), 1.69 (m, 2 H), 1.54 (m, 2 H). Anal. Calcd
: C, 66.09; H, 6.83. Found: C, 66.12; H, 6.75.
,4-Cyclohexadiene-1-pentanoic Acid, 2,5-Bis(1-aziridinyl)-4-
(47 mmol) of 2,5 dimethoxytoluene (Aldrich) in 50 mL of ether was
H
16
O
4
added a solution of 7.66 g (47 mmol) of iodine monochloride in 20
mL of chloroform over 30 min. Stirring was continued for another 3
h. The mixture was partitioned between 250 mL of ether and 200 mL
of 10% sodium thiosulfate with 6 g of sodium bicarbonate. The
decolorized organic layer was washed again with a similar portion of
the thiosulfate solution and dried over sodium sulfate. Removal of
solvents gave a white solid residue that was recrystallized from 100
mL of methanol to yield 9.28 g (71%) of 1 as a white solid: mp 81-
1
methyl-3,6-dioxo-, Methyl Ester (3). To a stirred solution of 0.686
g (2.90 mmol) of quinone 4 in 18.8 mL of methanol was added 105
mg (0.594 mmol) of cupric acetate. Ethyleneimine23 (1.53 mL, 29.0
mmol) was added, and an air-filled balloon was placed over the neck
of the flask to provide a reservoir of oxygen while controlling
evaporation of ethyleneimine. After the mixture was rapidly stirred
overnight, TLC showed nearly complete reaction. The mixture was
evaporated, and the residue was purified by flash chromatography (5
8
2 °C; TLC (methylene chloride) R
f
) 0.87 (0.83 for starting material);
) δ 7.18 (s, 1H), 6.68 (s, 1H), 3.82 (s, 3H), 3.78 (s,
H), 2.19 (s, 3H). Anal. Calcd for C I: C, 38.87; H, 3.99; I,
5.63. Found: C, 39.20, H, 3.76, I, 45.33.
-Pentynoic Acid, 5-(2,5-Dimethoxy-4-methylphenyl), Methyl
1
H NMR (CDCl
3
4
3
×
50 cm silica) using 2:1 hexanes-ethyl acetate (2% triethylamine).
9 11 2
H O
A red band was collected and evaporated to give 827 mg (90% yield)
of 5 as orange-red crystals: mp ) 97-99 °C; TLC (2:1 hexanes-
4
1
ethyl acetate) R
f
) 0.33; H NMR (CDCl
3
) δ 3.66 (s, 3 H), 2.53 (t, 2
Ester (2). To 3.88 g (14.0 mmol) of aryl iodide 1 and 1.88 g of methyl
pentynoate (16.7 mmol) in 50 mL of dry DMF were added 0.277 g
H, J ) 7.7 Hz), 2.35 (t, 2 H, J ) 7.4 Hz), 2.29 (s, 4 H), 2.27 (s, 4 H),
.71 (m, 2 H), 1.53 m, 2 H). Anal. Calcd for C17 : C, 64.13;
H, 6.96; N, 8.80; O, 20.10. Found: C, 64.04; H, 6.85; N, 8.61.
,4-Cyclohexadiene-1-pentanoic Acid, 2,5-Bis(1-aziridinyl)-4-
1
22 2 4
H N O
(
(
1.45 mmol) of cuprous iodide and 0.80 g (0.69 mmol) of tetrakis-
(triphenylphosphine)palladium (0)) (Lancaster). The flask was flushed
1
with argon and sealed with a septum, and 3.6 mL of N,N-diisopropy-
lethylamine was added. After the mixture was stirred overnight, TLC
showed a trace of unreacted starting material. The solution was
evaporated, and the residue was purified by flash chromatography (5
methyl-3,6-dioxo-, 2,3,5,6-Tetrafluorophenyl Ester (6). To a solution
of 135 mg (0.424 mmol) of methyl ester 5 in 2 mL of methanol was
added 0.5 mL of 1 N LiOH. The mixture was stirred at 50-60 °C for
4
h, cooled to room temperature, and concentrated in vacuo. The
residue was purified by flash chromatography (2 × 25 cm silica) using
9:1 methylene chloride-methanol (2% triethylamine). The product
×
40 cm silica) using a gradient of 1:1 hexanes-methylene chloride
to 100% methylene chloride. Evaporation gave a dark brown solid.
Despite good NMR purity, a second silica gel column was run using
1
eluted with a gradient up to 10% methanol. The major red band was
collected and evaporated to give 187 mg (109% yield) of the product
4
:1 hexanes-ethyl acetate to eliminate color. Appropriate fractions
were evaporated to give 1.88 g (51% yield) of the desired product as
1
as a dark red syrup: H NMR showed residual triethylamine; TLC
an off-white solid: mp ) 66-67 °C; TLC (2:1 hexanes-ethyl acetate)-
1
1
(98:2 ethanol-triethylamine) R
f
) 0.28; H NMR (CDCl
3
) δ 2.53 (t,
R
f
) 0.63; H NMR (CDCl
3
) δ 6.77 (s, 1 H), 6.62 (s, 1 H), 3.78 (s, 3
2
H, J ) 8.0 Hz), 2.31 (t, 2 H, J ) 5 Hz), 2.27 (s, 4 H), 2.26 (s, 4 H),
.99 (s, 3 H), 1.71 (m, 2 H), 1.50 (m, 2 H).
H), 3.73 (s, 3 H), 3.67 (s, 3 H), 2.75 (t, 2 H, J ) 6.3 Hz), 2.63 (t, 2 H,
J ) 6.3 Hz), 2.16 (s, 3 H). Anal. Calcd for C15 : C, 68.69; H,
.92. Found: C, 68.59; H, 6.53.
Benzenepentanoic Acid, 2,5-Dimethoxy-4-methyl-, Methyl Ester
3). To 186 mg of 10% palladium on carbon were added 3.7 mL of
1
18 4
H O
6
The intermediate acid was dissolved in 10 mL of acetonitrile, and
.3 mL of dry triethylamine was added (1.03 mmol). The flask was
0
flushed with argon and cooled in an ice bath. A solution of 315 µL
(
1
4
(
1.67 mmol) of 2,3,5,6-tetrafluorophenyl trifluoroacetate in 10 mL
ethanol and 3 drops of formic acid. The mixture was warmed enough
to product slight hydrogen evolution. After 30 min, this mixture was
flushed with hydrogen and left under a balloon of hydrogen. In a
separate flask, 3.7 mL of 4 M aqueous triethylammonium formate (pH
of acetonitrile was added dropwise to the flask. After 30 min, the
solvents were evaporated and the product was purified by flash
chromatography (2 × 25 cm silica) using 4:1 hexanes-ethyl acetate.
The orange band was collected and evaporated to give 50 mg (26%
6.5) was mixed with an equal volume of ethanol, flushed with hydrogen,
yield, based on methyl ester 5) of the desired product 6 as a red solid:
then added to the activated catalyst. A solution of 1.12 g (4.27 mmol)
of alkyne 2 in 7.5 mL of dry THF and 3.7 mL of ethanol was introduced
into the hydrogenation flask. After the mixture was stirred at room
temperature for 19 h, TLC showed no residual starting material. The
1
mp ) 110-114 °C; TLC (2:1 hexanes-ethyl acetate) R
NMR (CDCl
f
) 0.56; H
3
) δ 7.01 (m, 1 H), 2.75 (t, 2 H, J ) 7.2 Hz), 2.61 (t, 2 H,
J ) 8.1 Hz), 2.33 (s, 4 H), 2.30 (s, 4 H), 2.03 (s, 3 H), 1.89 (m, 2 H),
(
22) Povsic, T. J.; Strobel, S. A.; Dervan, P. B. J. Am. Chem. Soc. 1992,
(23) Allen, C. F. W.; Spangler, F. W.; Webster, E. R Org. Synth., Coll.
Vol. IV 1963, 433-435.
1
14, 5934-5941 and references therein.

9734 J. Am. Chem. Soc., Vol. 120, No. 38, 1998
Reed et al.
1
6
.59 (m, 2 H). Anal. Calcd for C22
.19. Found: C, 58.47; H, 4.86; N, 5.88.
Synthesis and Purification of Oligodeoxynucleotides. Oligo-
H
20
F
4
N
2
O
4
: C, 58.41; H, 4.46; N,
Reverse Phase HPLC Analysis of TFO 7. Purity of the starting
TFO-NH
2
, extent of reaction, and purity of TFO 7 were analyzed by
C
18 HPLC (see Figure 1). Analyses were performed using a Rainin
nucleotides were prepared on an Applied Biosystems Model 394
synthesizer using the 1-µmol protocols supplied by the manufacturer.
Protected â-cyanoethyl phosphoramidites, CPG supports, deblocking
solutions, cap reagents, oxidizing solutions, and tetrazole solutions were
purchased from Glen Research (Sterling, VA). A 3′-hexanol modifica-
tion was introduced into the TFO using a hexanol-modified CPG
Gradient HPLC system (Emeryville, CA) equipped with 5-mL pump
heads and a Rainin Dynamax PDA-1 photodiode array detector. The
samples were injected on a 4.6 × 150 mm Rainin Microsorb C18 column
and eluted using a gradient of 5-65% acetonitrile in 0.1 M triethy-
lammonium acetate (pH 7.5) over 20 min (flow rate ) 1 mL/min).
Oligonucleotide products were detected by UV absorbance at 260 nm,
and data were integrated and analyzed using Rainin Dynamax software.
The UV spectrum of 7 showed a characteristic absorbance at 320 nm
due to the aziridinylquinone conjugate group.
2
4
support, and a 5′-aminohexyl linker was introduced using an N-
monomethoxytritylhexanolamine phosphoramidite (Glen Research).
Preparative HPLC purification, detritylation, and butanol precipitation
of the oligos were carried out as usual. Aliquots (0.2 mg) of the 65-
mer oligos for triplex cross-linking experiments were further purified
by preparative gel electrophoresis. The concentrations of all oligos
were determined from the UV absorbance at 260 nm in phosphate-
buffered saline (pH 7.2). An extinction coefficient for each oligo was
Electrophoresis Assay for DNA Alkylation by TFO 7. The duplex
used in these studies was 65 bp’s long and contained a 33-bp
homopurine-homopyrimidine run.26 Either the purine-rich strand (Pu)
or the pyrimidine-rich strand (Py) of the duplex was 5′-end-labeled by
32
treatment with T4 polynucleotide kinase and [γ- P]ATP under standard
conditions. Labeled DNA was purified using a Nensorb column (NEN
Research Products) and had a specific activity of ∼6000 dpm/fmol.
Duplexes were formed by annealing 20 nM of the Pu strand with 40
nM of the complementary Py strand in 20 mM of PIPES buffer (pH
25
estimated using a nearest-neighbor model. For the TFO used in these
studies, a value of 29.1 µg/mL per A260 unit was used.
Synthesis of Diaziridinylquinone-TFO (7). HPLC-purified TFO-
2
NH was detritylated and precipitated as usual. The dried pellet was
dissolved in 0.5 mL of water and injected on a 300 × 7 mm PRP-1
2
6.2) with 140 mM KCl, 10 mM MgCl , and 1 mM spermine using an
column (Hamilton, Reno, NV) that was equilibrated with 0.1 M TEAB
incubation profile of 1 min at 95 °C and 30 min at 37 °C. After
annealing, coralyne chloride (Sigma) was added to give a final
concentration of 10 µM. Triplexes were formed in capped and
siliconized polypropylene microcentrifuge tubes (1.7 mL) at 37 °C in
a final volume of 25 µL. The labeled duplex (22.5 µL) was combined
with 2.5 µL of TFO 7, and the solutions were incubated at 37 °C. The
concentration of labeled duplex was 20 nM, and the concentration of
TFO was 2 µM. Aliquots (2.5 µL) were removed at 0, 30, 60, 120,
240, 360, 720, and 1440 min and stored frozen in 4 µL of loading
buffer (80% formamide, 0.01% xylene cyanol, and bromphenol blue).
The aliquots were thawed, and cross-linked products were electro-
phoretically resolved in a denaturing 8% polyacrylamide gel. The
labeled bands were visualized by autoradiography (see Figure 3) and
quantified using a Bio-Rad GS-250 Phosphorimager. The percent of
each 32P-labeled band was plotted vs time (see Figure 4).
Effect of pH on DNA Alkylation by TFO 7. A 100 µM solution
of PIPES buffer was prepared with 700 µM KCl, 50 mM MgCl, and
5 mM spermine hydrochloride. These solutions were carefully titrated
with 3 M HCl to provide solutions of pH 6.2, 6.7, or 7.2 after 5-fold
dilutions. These solutions were used for preparation of Pu-labeled
duplexes and triplexes as described above. Kinetics of DNA alkylation
were determined by PAGE, and the results are plotted in Figure 5.
Aqueous Stability of TFO 7. A 50 µM solution of 7 (100 µL)
was prepared in 20 mM appropriate PIPES buffer (pH 6.2 or 7.2) with
(pH 7.2). The TEA salt of the ODN was eluted from the column using
a gradient of 0-60% acetonitrile/30 min. The desired peak (∼15 min)
was collected and dried in vacuo on a centrifugal evaporator. The
residue was dissolved in 0.5 mL of water, and the concentration was
determined. A 705-µg aliquot (0.1 µmol) was redried in a 1.7-mL
Eppendorf tube, and the residue was dissolved in 0.1 mL of DMSO
(
(
Aldrich sure-seal) with 10 µL of triethylamine. A solution of 1.4 mg
3.1 µmol) of TFP ester 4 in 70 µL of DMSO was prepared and added
to the ODN. After shaking for 1 h, C18 HPLC analysis showed
complete reaction. After 1.5 h, the ODN products were precipitated
by adding to 6 mL of 2% NaClO /acetone in a 14-mL polypropylene
4
tube. The mix was centrifuged at 3000 rpm for 5 min, and the pellet
was sonicated (2 min) with 1 mL of acetone and recentrifuged. The
pellet was dried in vacuo for 15 min, dissolved in 0.2 mL of water,
and stored frozen. Purification by C18 HPLC used a 4.6 × 250 mm
Rainin Dynamax C18 column (10-µm particle size, 300-Å pore size)
and eluted using a gradient of 5-65% acetonitrile in 0.1 M triethy-
lammonium acetate (pH 7.5) over 30 min (flow rate ) 1 mL/min).
The peak eluting at 13 min (collected in ∼1 mL of TEAA/acetonitrile)
was concentrated to ∼0.2 mL by adding ∼3 mL of cold butanol. The
lower aqueous layer was removed, and the purified conjugate (7) was
precipitated by adding 100 µL of 3 M sodium acetate and 6 mL of
absolute ethanol. The mixture was centrifuged at 3000 rpm for 5 min,
and the pellet was sonicated with 2 mL of ethanol and recentrifuged.
The pellet was dried in vacuo for 15 min, and the purified product was
dissolved in 0.20 mL of water. A 5-µL aliquot was removed for C18
HPLC analysis, and another 5-µL aliquot was removed for concentration
determination. The bulk solution was immediately stored at -20 °C
2
140 mM KCl and 10 mM MgCl . The 0.1 mM stock solution was
immediately aliquoted to Eppendorf tubes and heated in a 37 °C oven.
Aliquots were frozen at 3, 6, 12, and 24 h and immediately thawed
just prior to HPLC analysis as described above. Stacked chromatograms
for pH 6.2 are shown in Figure 6.
for future use. HPLC analysis showed 97% purity.
showed a concentration of 1.97 mg/mL (56% recovery based on starting
TFO-NH ).
A260 analysis
Acknowledgment. This work was supported in part by NIH
Contract N01-AR-6-2230.
2
(
24) Petrie, C. R.; Reed, M. W.; Adams, A. D.; Meyer, R. B.
Bioconjugate Chem. 1992, 3, 85-87.
25) Cantor, C. R.; Warshaw, M. M.; Shapiro, H. Biopolymers 1970, 9,
059-1077.
JA973819U
(
(26) The full sequence of the 65-mer dsDNA target can be found in ref
10.
1