Recognition and cleavage at the DNA major groove

Article abstract of DOI:10.1016/S0968-0896(01)00219-X

DNA recognition agents based on the indole-based aziridinyl eneimine and the cyclopent[b]indole methide species were designed and evaluated. The recognition process involved either selective alkylation or intercalating interactions in the major groove. DN

Full text of DOI:10.1016/S0968-0896(01)00219-X

Bioorganic & Medicinal Chemistry 9 (2001) 2445–2459
Recognition and Cleavage at the DNA Major Groove
Edward B. Skibo,* Chenggou Xing and Thomas Groy
Department of Chemistry and Biochemistry, Arizona State University, Tempe, AZ 85287-1604, USA
Dedicated to Professor Peter Dervan for his Contributions to the Field of Bioorganic Chemistry
Received 20 March 2001; accepted 5 June 2001
Abstract—DNA recognition agents based on the indole-based aziridinyl eneimine and the cyclopent[b]indole methide species were
designed and evaluated. The recognition process involved either selective alkylation or intercalating interactions in the major
groove. DNA cleavage resulted from phosphate backbone alkylation (hydrolytic cleavage) and N(7) -alkylation (piperidine clea-
vage). The formation and fate of the eneimine was studied using enriched 13C NMR spectra and X-ray crystallography. The azir-
idinyl eneimine specifically alkylates the N(7) position of DNA resulting in direction of the aziridinyl alkylating center to either the
30- or 50-phosphate of the alkylated base. The eneimine species forms dimers and trimers that appear to recognize DNA at up to
three base pairs. The cyclopent[b]indole quinone methide recognizes the 30-GT-50 sequence and alkylates the guanine N(7) and the
thymine 6-carbonyl oxygen causing the hydrolytic removal of these bases. In summary, new classes of DNA recognition agents are
described and the utility of 13C-enrichment and 13C NMR to study DNA alkylation reactions is illustrated. # 2001 Elsevier Science
Ltd. All rights reserved.
Introduction
new reactive species capable of recognizing and cleaving
at specific sites in the DNA major groove. This report
describes the chemistry and DNA alkylating/cleaving
properties of the aziridinyl eneimine and of the cyclo-
pent[b]indole methide species shown in Chart 1. Either
species has the capacity to trap two nucleophiles result-
ing in a crosslinking reaction.
The recognition of DNA sequences and subsequent
cleavage by chemical nucleases have been the subject of
intense interest. The pyrrolopeptide-based recognition
agents of the DNA minor groove were pioneered in
Professor Dervan’s laboratory.1ꢀ3 Attachment of the
.
radical cleaving moiety, MPE Fe (II), to the pyrrolo-
peptide-based agents led to the design of sequence
These reactive species were considered to be responsible
for the cytotoxic and antitumor properties of indole and
cyclopent[b]indole based aziridinyl quinones.22,23 This
report confirms the formation and build-up of these
reactive species in solution as well as the cleavage
mechanism. Novel aspects of this report include:
4ꢀ6
specific DNA cleaving agents.
Largely due to Pro-
fessor Dervan’s work, the field of DNA minor groove
recognition has grown with many investigators and with
many variations of the pyrrolopeptide structure.7ꢀ12
The DNA major groove has also been utilized for the
design of sequence specific recognition agents. Usually
DNA strands or synthetic analogues thereof recognize
the major groove by Hoogsteen base pairing to afford a
triple helix structure.13ꢀ18 Because DNA recognition at
the minor and major groove has value in cancer
chemotherapy and molecular biology, efforts in this
area will no doubt continue into the new millennium.16
. Rapid determination of hydrolytic DNA cleavage
chemistry using the 13C-labeling developed in this
laboratory.24
. Crosslinking of the G-base by N(7) and 30-phos-
phate alkylation.
. Recognition of 30-GT-50 and 30-GGA-50 with small
molecules.
This laboratory has been involved in the design of
major groove recognition agents that cleave DNA at the
phosphate backbone.19ꢀ21 Currently, we are developing
Eventually, more functionalized versions of these indole
and cyclopent[b]indole-based aziridinyl quinones will
permit the recognition and cleavage of specific sequences
at the major groove of DNA.
*Corresponding author. Tel.: +1-480-965-3581; fax: +1-480-965-
2747; e-mail: eskibo@asu.edu
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00219-X

2446
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
Experimental Section. The conversion of 6 to 7 and the
eventual preparation of 2ꢀ-13C-1a were carried out
employing procedures previously reported for the unla-
beled analogues.22 The incorporation of the 13C-label at
the 3a-position was carried out by Vilsmeier formyla-
tion of 8 with 13C-DMF to afford 9. Multistep conver-
sion of 9 to 3ꢀ-13C-1a–d was carried out employing
reported procedures (Scheme 2).22
Eneimine formation and fate
The 3a-13C labeled derivatives of 1c and 1d were used
to verify that the eneimine can build-up in aqueous
media and undergo complex reactions. Both of these
compounds do not possess the aziridinyl group, so that
only eneimine formation and fate can be studied.
Shown in Figure 1 is the 100 MHz 13C NMR spectrum
(obtained on a 400 MHz instrument) of two-electron
reduced 3ꢀ-13C-1c obtained after 4 h of acquisitions.
The presence of the methylene center of the eneimine is
apparent at 98 ppm, along with starting material at 58
ppm and a reaction product at 18 ppm. The reaction
product was isolated from a preparative reaction
described below.
Chart 1.
Results and Discussion
The alkylating systems investigated
The indole (1) and cyclopent[b]indole (2) systems shown
in Chart 2 were previously prepared and evaluated as
antitumor agents possessing DNA alkylating cap-
ability.22 The cytotoxicity of both systems prompted the
detailed study of their DNA recognition and alkylation
properties described in this report. Both systems require
two-electron reduction to the hydroquinone in order to
activate the alkylating centers. Hydroquinone forma-
tion activates the aziridinyl group as an alkylating cen-
ter
by
permitting
nitrogen
protonation
at
neutrality.20,21,25 Reduction also activates the acetate as
a leaving group resulting in eneimine or quinone
methide formation. The quinone methide species is an
alkylating agent known to trap nucloephiles on
DNA.26ꢀ32 The eneimine species is related to the mito-
sene iminium ion known to alkylate DNA nucleo-
philes.24,33,34 The metabolism of 3-methylindole in fact
involves the formation of the eneimine species (also
called the methylene imine) that traps nucleophiles.35,36
Thus 1a is potentially a triple alkylating agent capable
of forming a quinone methide, an eneimine, and a pro-
tonated aziridine while 1b and 1c each have one or two
alkylating center removed. Similarily, 2a and 2b possess
either two or one alkylating centers respectively. The
parallel study of these compounds will provide insights
into how these multiple alkylating agents interact with
DNA.
The 13C-labeled analogues shown in the inset of Chart 2
were also prepared in order to assess the hydrolytic and
DNA alkylation chemistry of 1. The placement of 13C at
reacting centers permits the rapid assessment of the fate
of highly reactive species.24 The number of enriched 13
C
resonances and their chemical shifts provide insights
into the number and type of compounds formed in
alkylation reactions. Incorporation of the 13C-label at
the 2a-position to afford 2ꢀ-13C-1a involved the proce-
dure outlined in Scheme 1. Procedures for the conver-
sion of 5-methoxyindole to 6 are provided in the
Chart 2.

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2447
Reduction of 3ꢀ-13C-1c and incubation in methanol
afforded the complex mixture of products whose struc-
tures are shown in Scheme 3. An internal redox reaction
(transfer of H2 from the hydroquinone to the eneimine)
results in formation of 10 and its transesterification
product 11. Product 10 is the same as that observed in
the NMR tube reaction described above. The eneimime
also traps methanol solvent to afford 12 as its hydro-
quinone, that then affords the corresponding quinone
upon aerobic workup. However, most of the eneimine
traps hydroquinone nucleophiles to form dimers 13–15
through bond formation at C-5 or C-7.
Reduction of 3ꢀ-13C-1d and incubation in methanol
likewise afforded the complex mixture of products
whose structures are shown in Scheme 4. The products
arising from the internal redox reaction, 16 and 17, are
minor components of the mixture. The hydroquinone is
the major nucleophile in solution resulting in the for-
mation of dimers (19–20 and 22–24) and trimer 21 in
addition to methanol trapping product 18.
The structure of the dimers and trimer 21 in Schemes 3
1
and 4 were derived from H NMR, 13C NMR, COSY,
HMQC, and HMBC. Furthermore the structure of tri-
mer 21 was confirmed by X-ray crystallography. The
incorporation of 13C into the 3a position proved valu-
able in structural determinations, as illustrated by Fig-
ures 2–4. The 13C NMR of dimers 23 and 20 shown in
Figures 2 and 3 reveal the presence of the acetoxymethyl
(57 ppm) as well as the linking methylene (38 ppm). The
trimeric structure of 21 was readily assessed from the
13C NMR shown in Figure 4 that shows two linking
methylenes as well as the terminal acetoxymethyl.
Methylene links resulting from direct carbonyl addition
possess 13C chemical shifts in the 35–38 ppm range.
Previous studies of polymeric species derived from imi-
nium ions showed that methylene links resulting from a
Michael addition, such as those in 19 and 20, possess
13C chemical shifts in the 25–30 ppm range.24
The dimer and trimer structures shown in Schemes 3
and 4 can be readily determined spectroscopically, in
spite of their complexity, as summarized:
Scheme 1.
Scheme 2.
Figure 1. The 13C NMR spectrum of eneimine formation from two-electron reduced 3ꢀ -13C-1c. The reaction contained 2 mg of starting material,
0.5 mL of DMSO-d6, and 0.5 mL pD=7.2 0.1 M phosphate D2O buffer.

2448
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
. Either the direct carbonyl or the Michael addition
The 13C carbon splits the doublets of doublets into
two sets with a J value typically of 130 Hz.
processes afford a stereocenter adjacent to the 13C-
methylene center. As a result, the H NMR spec-
trum of this methylene center shows two doublet
of doublets. The methylene protons are chemically
nonequivalent due to the adjacent stereocenter
resulting in gem splitting (doublet of doublets).
. The 13C-enriched HMQC spectrum of the dimer or
trimer product readily identifies the H NMR che-
mical shifts of the 13C-methylene center. Inspec-
tion of the COSY spectrum will show two adjacent
‘boxes’ corresponding to the doublets of doublets.
A COSY spectrum of a mitosene dimer showing
this feature was published recently.24
. The 13C- enriched HMBC spectrum shows two-
bond coupling between the 13C-methylene center
and a methylene center adjacent to a carbonyl
indicating a Michael addition product.
. The indole NH protons are readily apparent
between 9 and 10 ppm in the H NMR spectrum
indicating that alkylation of this center had not
occurred as well as the number of indole units (two
for dimer and three for trimer).
The chemistry of the eneimine and the parent indole
system can be readily explained by referring to the elec-
tron density of the indole ring based on molecular
orbital calculations,37 inset of Scheme 5. These chemical
explanations are outlined below in conjunction with
Schemes 5 and 6
. Electron density for the indole system is higher at
the 3-position than at the 2-position resulting
almost exclusive eneimine formation. The forma-
tion of the quinone methide species by acetate
elimination from the 2-position contributes only if
eneimine formation is not possible.
. Excess electron density resides on positions 5 and
7 of the indole ring resulting in electrophilic attack
by the eneimine at these positions. The mechanism
for the formation of the 5-linked dimers 13, 19, 20
is illustrated in Scheme 6. Trapping of the enei-
mine electrophile by the 3- and 8-positions was not
observed in this study. A previous study reported
Scheme 3.
Scheme 4.

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2449
trapping of the iminium ion by the 3-position fol-
lowed by loss of formaldehyde.24 Trapping by the
8-position has never been observed probably
because the product can never aromatize.
proved to be a highly selective alkylating and cleaving
agent of DNA.
Eneimine reaction with DNA
. The high specificity of the eneimine for nucleo-
philic sites on the indole system suggests that this
electrophilic species will selectively trap nucleo-
philes on DNA.
The most reactive nucleophilic center of DNA is the
guanine N(7) position,38ꢀ41 and the eneimine electro-
phile should selectively alkylate DNA at this position.
In order to determine the DNA alkylation site, hexamer
DNA with the sequence ATGCAT was treated with
two-electron reduced 3ꢀ-13C-1b. The DNA hexamer
was first purified by 20% PAGE as described in the
Experimental and then reductively alkylated. The
reduction was carried out catalytically and the DNA
was recovered from precipitation. The 13C NMR spec-
trum of this DNA was compared with that of the native
hexamer. The alkylated DNA spectrum shown in Figure 5
has a strong peak at 42 ppm indicating that the 13C-
methylene is bound to a nitrogen. This result is consistent
Cyclopent[b]indole quinone methide
The formation of this reactive species has not been
documented as well as the eneimine because 13C incor-
poration at the electrophilic center has yet to be
accomplished. Reduction of 2a in anaerobic buffer
produced the dimeric species 25a,b suggesting that the
quinone methide species can form in solution (see
mechanism in the inset of Scheme 7). Modeling studies
indeed suggested that the quinone methide would be
stable.23 The cyclopent[b]indole quinone methide
Figure 2. The 13C NMR spectrum of dimer 23 from the hydrolysis of 3ꢀ-13C-1c.
Figure 3. The 13C NMR spectrum of dimer 20 from the hydrolysis of reduced 3ꢀ-13C-1c.
Figure 4. The 13C NMR spectrum of trimer 21 from the hydrolysis of reduced 3ꢀ-13C-1c.

2450
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
with alkylation of adenine or guanine N(7) as well the 2-
amino of guanine. PAGE studies described below indi-
cated that the N(7) positions are the alkylation sites.
the 3a-position rather than from the 2a- position as
observed in the hydrolysis study (Scheme 6).
The PAGE gel shown in Figure 7 reveals that the
reductive cleavage of DNA by 1a occurred at both the
guanine (N-7) and the 30 guanosine phosphate. The
alkylation of either the adenine or guanine N(7) was
determined by piperidine treatment according to
Maxam and Gilbert.42 The piperidine treatment also
hydrolyzed the phosphotriester resulting from phos-
phate alkylation without detectable backbone cleavage.
In contrast, the treatment of the alkylated DNA with
basic loading buffer resulted in only phosphate back-
bone cleavage without detectable cleavage via N(7)
adducts.20
The hexamer was also treated with reduced 3ꢀ-13C-1a
and 2ꢀ-13C-1a and 13C NMR spectra of the alkylation
products were obtained and compared with the 13C
natural abundance spectrum of the hexamer (Fig. 6).
The series of spectra in this figure illustrate the limita-
tion of using enriched 13C NMR to study DNA alkyla-
tion reactions. The formation of multiple alkylation
products as well as low yields could dilute the incorpo-
rated 13C labels to the point that they are as intense as
the natural abundance 13C spectrum. The solution is to
compare the natural abundance and enriched spectra
and note differences. Comparison of the natural abun-
dance 13C NMR spectrum (part A) with the 13C spec-
trum of 2a-labeled alkylating agent 2ꢀ-13C-1a (part B)
reveals that the oxygen of acetate was still attached at
the 2a-13C center. In contrast, the 3a-labeled alkylating
agent 3ꢀ-13C-1a afforded a 13C spectrum showing that
nitrogen alkylation had occurred at this position. These
results are consistent with elimination of acetate from
The first two lanes of Figure 7 show the results of a
limited alkylation reaction followed by piperidine and
hydrolytic cleavage respectively. The ladder in the first
lane shows the presence of specific guanine N(7) alkyla-
tion while the ladder in the second lane shows phos-
phate alkylation in the G+1 direction (of 50-labeled
DNA). The PAGE results and the 13C NMR spectrum
Scheme 5.
Scheme 6.

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2451
shown in Figure 5 support a process where the eneimine
intermediate selectively reacts with guanine N(7) and
the aziridinyl center is then directed to the 30-guanosine
phosphate )Scheme 8).
When borohydride reduction was employed, both ade-
nine and guanine N(7) alkylation occurred (lanes 3 and
5) along with alkylation of the corresponding 30-pho-
phates (G +1 and A +1 cleavage ladders). Reductive
activation with borohydride results in the rapid forma-
tion of the eneimine alkylating agent. In contrast, the
slower catalytic reduction affords limiting amounts of
eneimine and hence limited cleavage (lanes 1 and 2).
The PAGE gel in Figure 8 compares the reductive clea-
vage by 1a with cleavage by 1b,d and an analogue
Figure 5. The 13C NMR spectrum of alkylated hexamer 50-ATGCAT-
30 obtained by treatment of the DNA with two-electron reduced
3a-13C-1c..
Scheme 7.
Figure 6. 13C NMR Spectra of native hexamer 50-ATGCAT-30 (part A) and alkylation products obtained by treatment of the hexamer with two-
electron reduced 2ꢀ-13C-1a (part B) and 3ꢀ-13C-1a (part C).

2452
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
without a 3a leaving group. Lanes 1 and 2 show piper-
idine and hydrolytic cleavage, respectively, by reduced 1a.
These lanes were obtained for comparative purposes and
possess the same cleavage patterns as those in Figure 7.
attachment of the 1b eneimine to the N(7) center per-
mits access of the aziridinyl alkylating center to the 50
phosphate or to neither phosphate. Note that DNA
cleavage by N(7) alkylation and piperidine treatment
affords the same products as 50 phosphate hydrolysis.
The change from the 2-acetoxymethyl (1a) to the 2-
ethoxycarbonyl group (1b) does not influence eneimine
trapping of the adenine and the guanine N(7) nucleo-
philes (compare lanes 1 and 3). However, the hydrolysis
ladder in lane 4 does not show the G+1 and the A+1
ladder observed for 1a (lane 2). In some sequences, the
piperidine and hydrolysis cleavage patterns of lane 4 are
identical while in other sequences there is no corre-
sponding hydrolytic cleavage. These results suggest that
the piperidine and hydrolysis cleavage ladders shown in
lanes 5 and 6, respectively, were determined for an azir-
idinyl indoloquinone without a 3a leaving group. the
weak hydrolysis ladder in lane 5 in fact indicates that
phosphate alkylation is not an important process when
the eneimine species cannot form. these results are con-
sistent with the mechanism whereby the eneimine is the
initial dna alkylating species followed by aziridinyl
alkylation of the phosphate backbone.
The cleavage ladders shown in lanes 7 and 8 show the
unusual multibase recognition capability of 1d. The
hydrolysis ladder in lane 8 shows that recognition
occurs only at 3-GGA-50 with phosphate alkylation and
cleavage occurring largely at the center G base. The
reason for multibase recognition by such a small mole-
cule may be the presence of dimers and trimers. Recall
that the hydrolysis of reduced 1d afforded a variety of
such species in good yields, 19–24 in Scheme 4. Any of
these species could form a quinone methide or eneimine
species capable of alkylating the phosphate backbone or
N(7) centers. We are currently investigating if the trimer
21 and dimers could actually recognize and cleave
DNA.
Figure 7. Autoradiogram of the 1a-mediated reductive cleavage of a
50-32P end-labeled 514 bps restriction fragment from pBR322 DNA
(Ecor I/Rsa I).
Compound
Conditions
Cleavage Pattern
G
Lane 1
1a
Catalytic reductive
activation
piperidine
cleavage
Lane 2
Lane 3
Lane 4
Lane 5
1a
1a
1a
1a
Catalytic reductive
activation, hydrolytic
cleavage
NaBH4 reductive
activation, piperidine
cleavage
NaBH4 reductive
activation, hydrolytic
cleavage
G+1 (30 direction)
G & A
G+1 (30 direction)
A+1 (30 direction)
NaBH4 reductive
activation,
piperidine
G & A
cleavage
NaBH4 reductive
activation,
hydrolytic cleavage
Piperdine cleavage
Lane 6
Lane 7
1a
G+1 (30 direction)
A+1 (30 direction)
DMS
G & A
Scheme 8.

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2453
Cyclopent[b]indole quinone methide reaction with DNA
ring, or even its removal, resulted in the same cleavage
pattern. Furthermore, the weak or nonexistent hydro-
lytic ladders in Figure 9 reveal the absence of phosphate
alkylation. The reduced cyclopent[b]indole recognizes
30-GT-50 and cuts at both the G and T bases upon
treatment with piperidine.
The DNA cleavage ladders obtained with reduced
cyclopent[b]indoles 2a,b and an isomer of 2a are shown
in Figure 9. These ladders clearly show that the azir-
idinyl group was not involved in the alkylation and
cleavage process. Thus repositioning of the aziridinyl
The mechanism of cyclopent[b]indole DNA recognition
and cleavage is as yet unclear, but insights into these
processes were arrived at from the PAGE results shown
in Figure 9 and molecular modeling. A cyclo-
pent[b]indole quinone methide would have to recognize
Figure 8. Autoradiogram of the reductive cleavage of a 50-32P end-
labeled 514 bps restriction fragment from pBR322 DNA (Ecor I/Rsa
I) by 1a,b,d and the reported 3-methyl analogue.22
Compound
Conditions
NaBH4
Cleavage Pattern
G & A
Lane 1
Lane 2
1a
Figure 9. Autoradiogram of the reductive cleavage of a 50-32P end -
labeled 514 bps restriction fragment from pBR322 DNA (Ecor I/ Rsa
I) by 2a,b and the reported 6- aziridinyl isomer of 2a.22
reductive activation,
piperidine cleavage
NaBH4 reductive
1a
G+1
activation, hydrolytic (30 direction)
cleavage
A+1
30 direction)
G & A
Compound
Conditions
Cleavage Pattern
Lane 3
Lane 4
Lane 5
1b
1b
NaBH4 reductive
activation, piperidine Preference for A
cleavage
NaBH4 reductive
activation, hydrolytic
cleavage
NaBH4 reductive
activation, piperidine
cleavage
Lane 1
Lane 2
Lane 3
Lane 4
Lane 5
2a
NaBH4 reductive
activation, piperidine
cleavage
NaBH4 reductive
activation, hydrolytic
cleavage
NaBH4 reductive
activation, piperidine
cleavage
NaBH4 reductive
activation, hydrolytic
cleavage
G & T
of 30-GT-50
Some G & A
Some G & A
2a
2b
2b
Weak Cleavage
at G
G & T of
30-GT-50
No Significant
Cleavage
Lane 6
NaBH4 reductive
activation, hydrolytic
cleavage
Weak Cleavage
G T
NaBH4 reductive
activation, piperidine
cleavage
G & T of
30-GT-50
Lane 7
Lane 8
Lane 9
1d
1d
NaBH4 reductive
activation, piperidine (50’ direction)
cleavage
NaBH4 reductive
activation, hydrolytic at the center G
cleavage
30-GGA-50
Lane 6
Lane 7
NaBH4 reductive
activation, hydrolytic
cleavage
No Significant
Cleavage
DMS
Piperdine cleavage
G & A
DMS
Piperdine cleavage
G & A

2454
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
the 30-GT-50 sequence by interacting with the G C and
AT base pairs in the major groove so as to place the
alkylating center near the guanine N(7) and the adjacent
to the oxygen near the thymine 6-carbonyl. Alkylation of
either of these centers would result in hydrolytic removal
of either base followed by piperidine-mediated strand
cleavage. When the methide center was docked near the
guanine N(7) in the 30-GT-50 sequence and the resulting
complex minimized, the methide had intercalated
between the base pairs without any distinctive hydrogen
bonding interactons. In addition, the model indicated
that placement of the methide near the guanine N(7)
will permit reaction with the thymine oxygen.
DNA nucleophiles [gunaine N(7) and the thymine
6-carbonyl oxygen]. The aziridinyl group is not
required for alkylation and cleavage of DNA;
these processes are mediated entirely by the qui-
none methide species.
The finding enumerated above are starting points for
further studies that will lead to new classes of DNA
cleaving and recognition agents.
Experimental
The 13C NMR spectra was taken on Varian Inova 400
either in Nolorac dual PFG for hexamer or Varian
indirect PFG for hydrolytic products at 100 MHz.
Conclusions
DNA recognition agents based on the aziridinyl
eneimine and the cyclopent[b]indole methide species
were designed and evaluated utilizing 13C NMR and
PAGE. The recognition process involved either selective
alkylation or intercalating interactions in the major
groove. DNA cleavage resulted from phosphate back-
bone alkylation (hydrolytic cleavage) and N(7)-alkyl-
ation (piperidine cleavage). The following conclusions
are made.
Acetylation procedure for 1 and 2
The reported compounds in Chart 2 (1a, 1b, and 2a)
were prepared as previously described.22 Compounds
1c, 1d, and 2b were prepared by acetylation of the pre-
viously reported alcohol derivatives.22
To a solution of 0.2 mmol of the alcohol in 5 mL of
CH2Cl2 and 1 mL of acetone containing 50 mg of
DMAP was added 100 l acetic anhydride. The solution
was stirred at room temperature for 5 min and directly
flash chromatographed using ethyl acetate as the eluent.
The eluted solution containing the product was washed
with NaHCO3, dried over Na2SO4, and vacuum dried.
The product was recrystallized from ethyl acetate and
hexane to afford a near quantitative yield of the acety-
lated product. Physical properties of these acetylated
products are provided below for 1c, 1d, and 2b.
. The aziridinyl eneimine, resulting from elimination
of acetate from the indole 3a position, specifically
alkylates the N(7) position of DNA resulting in
direction of the aziridinyl alkylating center to
either the 30- or 50- phosphate of the alkylated
base. One reported aziridinyl eneimine was found
to be specific for guanine N(7) and this base’s 30-
phosphate.
. The nonaziridinated indole eneimine was studied
in order to gain insights into eneimine chemistry in
the absence of other reacting centers. Spectro-
scopic evidence was obtained for the buildup of
the eneimine in solution. Both dimers and trimers
are formed by electrophilic attack of the eneimine
at either the 5- or 7- positions of the indole ring.
NMR and X-ray crystallography were employed
to characterize these novel products. The eneimine
trimers and dimers appear to be capable of multi-
base recognition of DNA.
. Generation of an indole-base quinone methide by
acetate elimination from the 2a-position did not
occur either in hydrolysis reactions or in DNA
alkylation reactions. Since the indole 3-position is
more electron rich than the 2-position, eneimine
formation rather than quinone methide formation
is the predominate reaction. This conclusion is
relevant to the mechanism of indole -based anti-
tumor agents that rely on leaving group elimina-
tion from the 2-position, such as WV-1543 and
EO9.44,45 Our results suggest that the postulated
quinone methides or extended quinone methides
resulting from leaving group elimination may not
be important intermediates.
Ethyl 3-acetoxymethyl-5-methoxyindol-4,7-dione-2-car-
boxylate (1c). Mp 194–196 ꢁC; TLC (dichloromethane/
MeOH 98:2), Rf 0.55; 1H NMR (CDCl3) d 9.86 (1H, bs,
indole proton), 5.79 (1H, s, 6-proton), 5.84 and 5.28
(2H, d, J=168.3 Hz, 3-methylene protons), 4.41 (2H, q,
J=7.2 Hz, 2-methylene of ethyl), 3.87 (3H, s, 5-meth-
oxy), 2.06 (3H, s, 3-acetoxylmethyl), 1.39 (3H, t, J=7.2
Hz, 2-methyl of ethyl); 13C NMR (CDCl3) d 55.48 ppm;
IR (KBr pellet) 3491, 3354, 3302, 3211, 3041, 2943,
2862, 1675, 1558, 1431, 1384, 1200, 1173, 1128, 919
cmꢀ1; MS (EI) 322 (M+), 307 (M+-CH3), 280, 268, 251,
234, 207. Anal. calcd: C, 56.21; H, 4.69; N, 4.34. Found:
C, 56.03; H, 4.79; N, 4.21.
2,3-Diacetoxymethyl-5-methoxyindol-4,7-dione (1d). Mp
179–181 ꢁC. TLC (dichloromethane/MeOH 98:2), Rf
1
0.50. H NMR (CDCl3) d 9.74 (1H, bs, indole proton),
5.72 (1H, s, 6-proton), 5.79 and 5.24 (2H, d, J=165.6
Hz, 3-methylene protons), 5.20 (2H, s, 2-methylene
protons), 3.84 (3H, s, 5-methoxy), 2.11 and 2.04 (6H, 2s,
2,3-acetoxymethyls). 13C NMR (CDCl3) d 56.47 ppm.
IR (KBr pellet) 3443, 3371, 3342, 3260, 2993, 2918,
2827, 1674, 1634, 1569, 1404, 1374, 1241, 1139, 919, 827
cmꢀ1. MS (EI) 322 (M+), 280 (M+ꢀacetyl), 269, 251,
231, 216, 207, 179. Anal. calcd: C, 56.21; H, 4.69; N,
4.34. Found: C, 56.18; H, 4.72; N, 4.33.
. The cyclopent[b]indole quinone methide forms in
solution and can trap itself (dimerize) and trap

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2455
3-Acetoxy-7-methoxy-1,2,3,4-tetrahydrocyclopent[b]in-
dole-5,8-dione (2b). Mp 173–175 ꢁC. TLC: (dichloro-
methane/MeOH 95:5), Rf 0.55. 1H NMR (CDCl3) d
10.03 (1H, bs, indole proton), 5.68 (1H, s, 6-protons),
5.64 (1H, m, 3-hydroxymethylene proton), 3.82 (3H, s,
7-methoxy), 2.96, 2.73 and 2.25 (4H, m, methylenes of
cyclopentyl), 2.04 (3H, s, 3-acetyl methyl). IR (KBr
pellet) 3452, 3206, 3035, 2949, 2892, 1701, 1623, 1577,
1502, 1433, 1322, 1219, 1167, 1024, 955, 834 cmꢀ1. MS
(EI) 275 (M+), 247 (M+ꢀCO), 232, 214, 202, 187, 160,
131. Anal. calcd: C, 61.08; H, 4.76; N, 5.09. Found: C,
60.94; H, 4.81; N, 5.00.
dry THF. The resulting solution was stirred for 15 min
and then the reaction allowed to warm to room tem-
perature after the addition of 0.5 g NH4Cl. After the
liquid NH3 was gone, the solution was mixed with 20
mL of methanol containing 300 mg of NaBH4. The
solution was stirred for 10 min, diluted with 30 mL of
water, and then extracted 4ꢂ with 100 mL portions of
methylene chloride. The extracts were dried over
Na2SO4 and evaporated to a residue. The residue was
dissolved in 10 mL of methylene chloride containing
100 mg of dimethylaminopyridine. To this solution was
added 300 l acetic anhydride. The solution was stirred
for 10 min and directly purified by a flash chromato-
graphy using methylene chloride as the eluent. The
product was disolved in 10 mL of methylene chloride
and this solution was added to a solution of 200 l
POCl3 and 600 l DMF in 10 mL of methylene chloride
at 0 ꢁC. The solution was stirred overnight and neu-
tralized with saturated NaHCO3 followed by extraction
6ꢂ with 50 mL portions of methylene chloride. The
extract was dried over Na2SO4 and concentrated to a
residue, which was further purifed by flash chromato-
graphy using methylene chloride as the eluent: Yield 123
mg (25%); Mp 189–192 ꢁC; TLC (dichloromethane,
methanol 98:2), Rf 0.68; 1H NMR (CDCl3) d 10.27
(1H, s, 3-formyl proton), 9.00 (1H, bs, indole pro-
ton), 7.74 (1H, d, J=2.7 Hz, 4-proton), 7.29 (1H,
d, J=8.7 Hz, 7-proton), 6.95 (1H, dd, J=2.7, 8.7
Hz, 6-proton), 5.79 and 5.29 (2H, d, J=150.9 Hz,
2-acetoxylmethyl), 3.89 (3H, s, 5-methoxy), 2.16
5-Methoxy-1-phenylsulfonylindole (3). To 2.5 g (17
mmol) of 5-methoxyindole in 13 mL of dry THF,
cooled by dry ice/acetone bath over a nitrogen atmo-
sphere, was added 12 mL of 1.6 M n-butyllithium in
hexane over 10 min. The reaction solution was stirred
for 1.2 h at 0 ꢁC and then chilled in a dry ice acetone
bath. To this chilled reaction was added 2.6 mL of
PhSO2Cl over 15 min and the resulting mixture stirred
for 12 h at room temperature. The reaction mixture was
then mixed with 50 mL of 3% NaHCO3 and stirred for
30 min. The solution was then extracted 4ꢂ with 100
mL portions of CH2Cl2. The extracts were dried over
Na2SO4 and vacuum dried to a yellow oil. The oil was
then mixed with 4 mL of THF followed by addition of
10 mL of hexane and kept at ꢀ10 ꢁC for 5h. The crys-
tallized product was then filtered and collected: 3.99g
(82%) yield; mp 54–56 ꢁC; TLC (CH2Cl2) Rf 0.75; IR
(KBr pellet) 3342, 3211, 3017, 2939, 2902, 2878, 1622,
(3H, s, 2-acetoxylmethyl). 13C NMR (CDCl3)
d
1547, 1301, 1187, 1125, 1003, 842 cmꢀ1
;
1H NMR
56.10 ppm; IR (KBr pellet) 3494, 3337, 3169, 3088,
2962, 2812, 1665, 1616, 1537, 1518, 1198, 1174, 1102,
1031 cmꢀ1; MS (EI) 248 (M+), 206 (M+ꢀCOCH2),
188, 172, 159, 145, 116. Anal. calcd (C13H13NO4): C,
63.30; H, 5.28; N, 5.64. Found: C, 62.99; H, 5.39; N,
5.52.
(CDCl3) d 7.86, 7.52, 7.45, 6.93, and 6.59 (10H, m, aro-
matic protons), 3.80 (3H, s, 5-methoxy); MS (EI mode)
m/z 287 (M+), 272 (M+ꢀCH3), 147. Anal. calcd
(C15H13NO3S) : C, 62.71; H, 4.56; N, 4.88. Found: C,
62.43; H, 4.71; N, 4.80.
2-Formyl-5-methoxy-1-phenylsulphonylindole (4). To
1.44 g of 3 in 10 mL of THF, cooled at ꢀ78 ꢁC under a
nitrogen atmosphere, was added 4 mL of 1.6 M n-
butyllithium in hexane over 10 min. The reaction solu-
tion was stirred for 25 min followed by addition of 0.7
mL 13C-DMF and stirring at 0 ꢁC for 30 min. The
completed reaction was mixed with 100 mL of saturated
NH4Cl solution and extracted 5ꢂ100 mL portion of
CH2Cl2. The extracts were dried over Na2SO4 and dried
to brown solid, which was further purifed by flash
chromatography employing dichloromethane as the
eluant. Yield: 1.21 g (76%); mp 123–124 ꢁC; TLC (di-
chloromethane), Rf 0.44; 1H NMR (CDCl3) d 10.82 and
10.19 (1H, d, J=189 Hz, 2-formyl protons), 8.13, 7.73,
7.53, 7.39, 7.13, and 6.99 (9H, m, aromatic protons),
3.82 (3H, s, 5-methoxy); 13C NMR (CDCl3) d 182.7; IR
(KBr pellet) 3354, 3326, 2917, 2889, 1679, 1532, 1514,
1347, 1236, 1218, 1036, 945, 834 cmꢀ1; MS (EI) 316
2-Acetoxymethyl-3-formyl-5-methoxy-4-nitroindole (6).
To a solution of 80 mg of 5 in 10 mL CH2Cl2 was
added 50 l of 70% HNO3. The reaction was stirred for
0.5 h and made basic with NaHCO3 saturated solution.
The basic solution was then extracted 4ꢂ with 40 mL
portions of methylene chloride. The extracts were dried
over Na2SO4 and dried, which was further purifed by
flash chromatography. Yield: quantitative. Mp: 201–
203 ꢁC. TLC (dichloromethane/MeOH 98:2), Rf 0.58.
1H NMR (CDCl3) d 10.04 (1H, s, 3-formyl proton), 9.29
(1H, bs, indole proton), 7.51 and 7.09 (2H, 2d, J=8.7
Hz, 6, 7-protons), 5.88 and 5.38 (2H, d, J=152.4 Hz, 2-
acetoxylmethyl), 3.96 (3H, s, 5-methoxy), 2.19 (3H, s, 2-
acetoxylmethyl). 13C NMR (CDCl3) d 57.53 ppm. IR
(KBr pellet) 3467, 3323, 3198, 3001, 2983, 2822, 1701,
1643, 1502, 1471, 1304, 1229, 1137, 1039, 927, 864 cmꢀ1
.
MS (EI) 293 (M+), 251 (M+ꢀCOCH2), 233, 215, 201,
185, 172, 156, 146, 116. Anal. calcd (C13H12N2O6) C,
53.58; H, 4.12; N, 9.55. Found: C, 52.99; H, 4.28; N,
9.32.
13
(M+), 301 (M+-CH3), 287 (M+ꢀ CO), 274, 257, 172,
146. Anal. calcd (C16H13N O4S): C, 61.07; H, 4.14; N,
4.43. Found: C, 60.94; H, 4.20; N, 4.39.
13C NMR chemical shifts (ꢁ) of labeled products.
2ꢀ-13C-1a, 56.28 ppm; 3ꢀ-13C-1a, 55.98 ppm; 3ꢀ-13C-
1b, 55.5 ppm; 3ꢀ-13C-1c, 55.5 ppm; 3ꢀ-13C-1d, 56.5
ppm.
2-Acetoxymethyl-3-for mLy-5-methoxyindole (5). To a
solution of 200 mg Na metal in 10 mL of NH3, cooled
at ꢀ78 ꢁC, was added 630 mg of 4 dissolved in 10 mL of

2456
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
1
13C NMR study for the eneimine formation. A solution
of 2 mg of 3ꢀ-13C-1c in 0.5 mL DMSO-d6, containing 2
mg of 5% Pd on carbon, was then mixed with 0.5 mL of
0.1M pD=7.2 phosphate D2O buffer. The solution was
degassed with argon for 5 min followed by purging with
H2 until the quinone color of the solution disappeared.
The solution was then degassed again with argon for 5
min, and the catalyst was removed by centrifuge in a
nitrogen box. The supernatant was then transferred into
a 3 mm NMR tube for the 13C NMR spectrum obtained
on a 100 MHz instrument after 4 h of scanning.
acetone 95:5), Rf 0.70; H NMR (CDCl3) d 9.99 and
9.53 (2H, 2bs, indole nitrogen protons), 5.84 (1H, s, 60-
proton), 5.53 (1H, s, 6-proton), 4.28 and 4.16 (4H, m,
methylenes of ethyls), 3.87 (6H, s, 5 and 50-methoxy
methyls), 3.74 (2H, dd, J1=12.8 Hz, J2=133.6 Hz, 30-
methylene), 1.69 (3H, d, J=124 Hz, 30-methyl), 1.33
and 1.21 (6H, m, methyls of ethyls); 13C NMR (CDCl3)
d 38.38 and 10.56; MS (EI) 528 (M+). Anal. calcd
(C26H26N2 O10) : C, 59.09; H, 5.31; N, 5.33. Found: C,
58.68; H, 5.46; N, 5.12.
Ethyl 3-(ethyl 50-methoxy-30-acetoxymethyl-indole-40-
one-70-hydroxy-70-ly) methyl-5-methoxy-indol-4,7-dione-
2-carboxylate (15). Yield 7.03 mg (10.0%); TLC
(CH2Cl2/acetone 95:5), Rf 0.68; 1H NMR (CDCl3) d
9.88 and 9.74 (2H, 2bs, indole nitrogen protons), 5.85
(1H, s, 6-proton), 5.67 (1H, s, 60-proton), 4.72 (2H, dd,
J1=6.8 Hz, J2=144.8 Hz, 3-methylene), 4.30 and 4.14
(4H, m, methylenes of ethyls), 3.87 (6H, s, 5 and 50-
methoxy methyls), 3.85 (2H, dd, J1=12.8 Hz, J2=164.8
Hz, 30-methylene), 1.92 (3H, s, 3-acetate methyl), 1.32
and 1.22 (6H, m, methyls of ethyls); 13C NMR (CDCl3)
d 56.64 and 38.53; MS (EI) 584 (M+). Anal. calcd
(C28H28N2O12): C, 57.53; H, 4.83; N, 4.79. Found: C,
57.17; H, 5.02; N, 4.68.
Hydrolysis of 3ꢀ-13C-1c. To a solution of 80 mg SM in
50 mL MeOH was added 60 mg 5% Pd on carbon. The
solution was degassed with argon for 5 min, followed by
passing H2 through for 3 min. The solution was then
degassed with argon for another 5 min and opened to
air. The catalyst was filtered off through Celite and the
filtrate was dried. The solid residue was purified by
preparative TLC plate employing different eluents. The
following products were isolated and characterized.
Ethyl 5-methoxy-3-methylindole-4,7-dione-2-carboxylate
(10). Yield 2.36 mg (3.59%); TLC (CH2Cl2), Rf 0.62;
1H NMR (CDCl3) d 9.61 (1H, bs, indole nitrogen pro-
ton), 5.79 (1H, s, 6-proton), 4.39 (2H, q, J=7.2 Hz, 2-
ethyl methylene), 3.86 (3H, s, 5-methoxy methyl), 2.63
(3H, d, J=129.6 Hz, 3-methyl), 1.40 (3H, t, J=7.2 Hz,
2-ethyl methyl); 13C NMR (CDCl3) d 10.99; MS (EI)
264 (M+), 245, 234, 217, 202.
To a solution of 80 mg SM in 50 mL MeOH was added
60 mg 5% Pd on carbon. The solution was degassed
with argon for 5 min, followed by passing H2 through
for 3 min. The solution was then degassed with argon
for another 5 min and opened to air. The catalyst was
filtered off through Celite and the filtrate was dried. The
solid residue was purified by preparative TLC plate
employing different eluents. The following products
were isolated and characterized (reaction in MeOH with
NaBH4 as the reductant affords the same hydrolysis
products).
Methyl 5-methoxy-3-methylindole-4,7-dione-2-carboxy-
late (11). Yield 0.31 mg (0.52%); TLC (CH2Cl2), Rf
1
0.61; H NMR (CDCl3) d 9.61 (1H, bs, indole nitrogen
proton), 5.79 (1H, s, 6-proton), 3.86 and 3.76 (6H, 2s, 2,5-
methyls), 2.61 (3H, d, J=129.6 Hz, 3-methyl); 13C NMR
(CDCl3) d 11.21; MS (EI) 250 (M+), 234, 217, 202.
Ethyl 5-Methoxy-3-methoxymethyl-lindole-4,7-dione-2-
carboxylate (12). Yield 1.73 mg (2.4%); TLC (CH2Cl2),
Rf 0.56; H NMR (CDCl3) d 9.61 (1H, bs, indole nitro-
gen proton), 5.78 (1H, s, 6-proton), 4.76 (2H, d,
J=132.3 Hz, 3-methylene), 4.37 (2H, q, J=7.5 Hz, 2-
ethyl methylene), 3.86 (3H, s, 5-methoxy methyl), 3.82
(3H, d, J=6.6 Hz, 3-methoxy methyl), 1.40 (3H, t,
J=7.5 Hz, 2-ethyl methyl); 13C NMR (CDCl3) d 64.73;
MS (EI) 294 (M+), 262, 250, 233, 202.
2,3-Dimethyl-5-methoxy-indol-4,7-dione (16). Yield 1.16
1
mg (2.27%); TLC (CH2Cl2/acetone 98:2), Rf 0.70; H
1
NMR (CDCl3) d 9.37 (1H, bs, indole nitrogen proton),
5.61 (1H, s, 6-proton), 3.81 (3H, s, 5-methoxy methyl), 2.24
(3H, d, J=128.1 Hz, 3-methyl), 2.23 (3H, s, 2-methyl); 13C
NMR (CDCl3) d 9.69; MS (EI) 206 (M+), 191.
2-Acetoxy-5-methoxy-3-methylindol-4,7-dione (17). Yield
1.10 mg (1.68%); TLC (CH2Cl2/acetone 98:2), Rf 0.68;
1H NMR (CDCl3) d 9.42 (1H, bs, indole nitrogen pro-
ton), 5.69 (1H, s, 6-proton), 5.05 (2H, s, 2-methyl), 3.82
(3H, s, 5-methoxy methyl), 2.35 (3H, d, J=128.7 Hz, 3-
methyl), 2.09 (3H, s, 2-acetate methyl); 13C NMR
(CDCl3) d 9.58; MS (EI) 264 (M+), 249, 209, 202, 149.
Ethyl 3-(ethyl 30-methyl-indol-40,70-dione-20-carboxylate-
50-yl)
(13). Yield 2.37 mg (4.00%); TLC (CH2Cl2/acetone
methyl-5-methoxy-indol-4,7-dione-2-carboxylate
1
98:2), Rf 0.73; H NMR (CDCl3) d 9.83 and 9.58 (2H,
2bs, indole nitrogen protons), 5.96 (1H, t, J=1.2 Hz, 60-
proton), 5.83 (1H, s, 6-proton), 4.38 (4H, m, methylenes
of ethyls), 4.35 (2H, doublet of triplet, J1=1.2 Hz,
J2=132.8 Hz, 3-methylene), 3.84 (3H, s, 5-methoxy
methyl), 2.65 (3H, d, J=129.6 Hz, 30-methyl), 1.40 and
1.31 (6H, m, methyls of ethyls); 13C NMR (CDCl3) d
24.58 and 10.90; MS (EI) 496 (M+).
2-Acetoxy-5-methoxy-3-methoxymethylindol-4,7-dione (18).
Yield 3.16 mg (4.33%); TLC (CH2Cl2/acetone 98:2), Rf
1
0.65; H NMR (CDCl3) d 9.63 (1H, bs, indole nitrogen
proton), 5.71 (1H, s, 6-proton), 5.20 (2H, s, 2-methyl-
ene), 4.71 (2H, d, J=143.7 Hz, 3-methylene), 3.83 (3H,
s, 5-methoxy methyl), 3.42 (3H, d, J=5.7 Hz, 3-meth-
oxy methyl), 2.02 (3H, s, 2-acetate methyl); 13C NMR
(CDCl3) d 64.57; MS (EI) 294 (M+), 279, 261, 250, 233,
218, 204, 190. Anal. calcd (C14H15NO6) : C, 57.13; H,
5.14; N, 4.76. Found: C, 56.97; H, 5.32; N, 4.73.
Ethyl 3-(ethyl 50-Methoxy-30-methyl-indole-40-one-70-
hydroxy-70-ly) methyl-5-methoxy-indol-4,7-dione-2-car-
boxylate (14). Yield 5.28 mg (7.73%); TLC (CH2Cl2/

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2457
2-Acetoxymethyl-3-(20-acetoxymethyl-50-methoxy-30-meth-
yl-50,60-dihydroindol-40,70-dione-50-yl) methyl-5-methoxy-
indol-4,7-dione (19). Yield 0.75 mg (1.15%); TLC
(CH2Cl2/acetone 95:5), Rf 0.56; 1H NMR (CDCl3) d
9.54, 9.19 (2H, 2bs, indole nitrogen protons), 5.76 (1H,
s, 6-proton), 4.94 (2H, m, 2-methylene), 4.52 (2H, m, 20-
methylene), 3.87 (3H, s, 5-methoxy methyl), 3.85 and
3.02 (2H, m, 3-methylene), 3.63 (3H, s, 50-methoxy
methyl), 3.62 and 2.64 (2H, m, 60-protons), 2.06 and
2.02 (6H, 2s, acetate methyls), 1.78 (3H, d, J=128.1 Hz,
30-methyl); 13C NMR (CDCl3) d 29.36, 8.03; MS (EI)
528 (M+).
5.10 (2H, m, 20-methylene), 4.66 (2H, d, J=147.0 Hz, 3-
methylene), 4.56 (2H, m, 2-methylene), 3.85 and 3.81
(6H, 2s, 5 and 50-methoxy methyls), 3.98, 3.54, 3.29 and
2.89 (2H, m, 30-methylene), 2.07, 2.04 and 1.99 (9H, 3s,
acetate methyls); 13C NMR (CDCl3) d 57.02 and 38.58;
MS (EI) 586 (M+). Anal. calcd (C28H28N2O12) : C,
57.33; H, 4.81; N, 4.80. Found: C, 57.18; H, 4.87; N,
4.69.
2-Acetoxymethyl-3-(20, 30-Diacetoxymethyl-50-methoxy-
indole-40-one-70-hydroxy-70-ly-) methyl-5-methoxy-indol-
4,7-dione (24). Yield 0.65 mg (1.00%); TLC (CH2Cl2/
1
acetone 90:10), Rf 0.51; H NMR (CDCl3) d 9.78 and
2-Acetoxymethyl-3-(20,30 -diacetoxymethyl-50 -methoxy-
50,60-dihydroindol-40,70-dione-50-yl) methyl-5-methoxy-in-
dol-4,7-dione (20). Yield 1.03 mg (1.42%); TLC
(CH2Cl2/acetone 95:5), Rf 0.54; 1H NMR (CDCl3) d
9.49, 9.31 (2H, 2bs, indole nitrogen protons), 5.72 (1H,
s, 6-proton), 5.01 (2H, m, 2-methylene), 4.70 (2H, m, 20-
methylene), 4.28 (2H, m, 30-methylene), 3.84 (3H, s, 5-
methoxy methyl), 3.85 and 3.08 (2H, m, 3-methylene),
3.63 (3H, s, 50-methoxy methyl), 3.61 and 2.72 (2H, m,
60-protons), 2.07, 2.04 and 1.97 (9H, 3s, acetate
methyls); 13C NMR (CDCl3) d , 56.94, 29.36; MS (EI)
586 (M+).
9.60 (2H, 2bs, indole nitrogen protons), 5.70 (1H, s, 60-
proton), 5.41 (1H, s, 6-proton), 5.09 (2H, m, 20-methyl-
ene), 4.58 (2H, m, 2-methylene), 4.46 (2H, d, J=92.4
Hz, 3-methylene), 3.84 and 3.79 (6H, 2s, 5 and 50-meth-
oxy methyls), 3.89, 3.54, and 3.00 (2H, m, 30-methyl-
ene), 3.23 (3H, d, J=5.7 Hz, 3-methoxy methyl), 2.07
and 2.04 (6H, 2s, acetate methyls); 13C NMR (CDCl3) d
65.05 and 38.15; MS (EI) 558 (M+).
50-Hydroxy-7,30,70-trimethoxy-1,2,3,4,20,30,40,50-octahydro-
10H-[3,50]bi[cyclopent[b] indolyl]-5,8,80-trione (25a). TLC
1
(CH2Cl2) Rf 0.55; H NMR (CDCl3) d 9.47 and 9.04
(2H, 2bs, 4 and 40 protons), 5.65 and 5.46 (2H, 2s, 6 and
60 protons), 4.77 (1H, m, 30-proton), 3.87 (1H, m, 3-
proton), 3.85, 3.48, and 3.38 (9H, 3s, 7, 30, and 70
methoxys), 2.89, 2.75, 2.47, 2.28, 2.09, 1.88, and 0.87
(8H, mm, 1, 2, 10, and 20 protons).
2-Acetoxymethyl-3-(20-acetoxymethyl-30-(200,300-diacetoxy-
methyl-500 -methoxy-indole-400 -one-700 -hydroxy-700 -ly-)
methyl-50 -methoxy-indole-40 -one-70 -hydroxy-70 -ly-)
methyl-5-methoxy-indol-4,7-dione (21). Yield 4.28 mg
(6.27%); TLC (CH2Cl2/acetone 90:10), Rf 0.32; 1H
NMR (CDCl3) d 10.37, 9.96 and 9.75 (3H, 3bs, indole
nitrogen protons), 5.77 (1H, s, 600-proton), 5.55 and 5.34
(2H, 2s, 6, 60 -protons), 5.60 and 5.40 (2H, 2bs, 7, 70 -
hydroxy protons), 5.06 (2H, m, 20-methylene), 4.54 (2H,
d, J=93.6 Hz, 3-methylene), 4.41 (2H, m, 200 -methyl-
ene), 3.89, 3.84 and 3.65 (9H, 3s, 5, 50 and 500-methoxy
methyls), 4.12, 3.92, 3.74, 3.00 and 2.59 (4H, m, 30and
300-methylenes), 2.06, 2.01, 1.97 and 1.93 (12H, 4s, ace-
tate methyls); 13C NMR (CDCl3) d 56.84, 39.52 and
37.99; MS (EI) 828 (M+). Anal. calcd (C39H41N3O17):
C, 56.52; H, 5.34; N, 5.10. Found: C, 56.31; H, 5.52; N,
5.02.
30-Acetoxy-50-hydroxy-7,70-dimethoxy-1,2,3,4,20,30,40,50-
octahydro-10H-[3,50]bi[cyclopent[b] indolyl]-5,8,80-trione
1
(25b). TLC (CH2Cl2) Rf 0.52; H NMR (CDCl3) d 9.66
and 9.40 (2H, 2bs, 4 and 40 protons), 5.61 and 5.40 (2H,
2s, 6 and 60 protons), 4.60 (1H, m, 30-proton), 3.84 (1H,
m, 3-proton), 3.85 and 3.34 (6H, 2s, 7 and 70 methoxys),
3.59 (1H, bs, 50-hydroxy proton), 2.61, 2.45, 2.01, 1.73,
1.10, and 0.87 (8H, mm, 1, 2, 10, and 20 protons).
Preparation of the 13C-labeled indoloquinone–DNA
adduct
The DNA hexamer d(ATCGAT)2 was prepared by the
phosphorimidate method and purified by 20% pre-
parative polyacrylamide gel. The DNA adducts were
prepared by mixing 8.0 mg (2.0 mmol of strand) of
d(ATCGAT)2 in 0.05 M of Tris buffer (pH 7.4) with 1.0
mg (3.0 mmol) of the quinone in 0.25 mL of DMSO. To
the mixture was added 0.2 mg of 5% Pd on carbon. The
resulting mixture was degassed under argon for 30 min
and followed by purging with H2 at 1 atm for 20 min.
The mixture was then purged with argon for 10 min and
incubated at 30 ꢁC for 2 h. The reaction was opened to
the air and the catalyst was centrifuged off. To the
supernatant was added 0.45 mL of 7.5 M ammonium
acetate, 10 mL of cold ethanol and left in a ꢀ20 ꢁC
freezer for 24 h. The mixture was centrifuged at 12,000g
for 20 min and the supernatant was removed. The DNA
pellet was redissolved in water, ethanol precipitated
again and dried. The pellet was then dissolved in D2O
and lyophilized twice before diluting in 0.7 mL 100%
D2O.
2-Acetoxymethyl-3-(20-acetoxymethyl-50-methoxy-30-meth-
yl-indole-40-one-70-hydroxy-70-ly) methyl-5-methoxy-in-
dol-4,7-dione (22). Yield 0.78 mg (1.20%); TLC
1
(CH2Cl2/acetone 90:10), Rf 0.53; H NMR (CDCl3) d
9.57, 9.22 (2H, 2bs, indole nitrogen protons), 5.74 (1H,
s, 60-proton), 5.42 (1H, s, 6-proton), 4.99 (2H, m, 20-
methylene), 4.36 (2H, 2m, 2-methylene), 3.87 and 3.83
(6H, 2s, 5, 50-methoxy methyls), 3.82, 3.62, 3.18
and 2.75 (2H, m, 30-methylene), 2.07 and 2.03 (6H,
2s, acetate methyls), 1.72 (3H, d, J=126 Hz, 3-
methy); 13C NMR (CDCl3) d 38.46, 8.87; MS (EI) 528
(M+).
2-Acetoxymethyl-3-(20-acetoxymethyl-50-methoxy-30-meth-
oxymethyl-indole-40-one-70-hydroxy-70-ly-)
methyl-5-
methoxy-indol-4,7-dione (23). Yield 8.84 mg (12.2%);
TLC (CH2Cl2/acetone 90:10), Rf 0.52; 1H NMR
(CDCl3) d 9.59 and 9.48 (2H, 2bs, indole nitrogen pro-
tons), 5.73 (1H, s, 60-proton), 5.43 (1H, s, 6-proton),

2458
E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
Materials for sequencing studies
Diego) was used for modeling studies and consistent
valence force field (CVFF) was utilized for all mini-
mization protocols.
Quinone stock solutions were made up in DMSO at 10,
1, and 0.1 mM. Electrophoresis-grade acryalmide and
bis (acrylamide) were purchased from Sigma, ultrapure
urea and agarose from GIBCO-BRL Life Technologies
(Grand Island, NY, USA), and piperidine from Sigma
(St Louis, MI, USA). The pBR322 plasmid DNA,
restriction enzymes EcoRI and RsaI, T4 polynucleotide
kinase (PNK), and bacterial alkaline phosphatase
(BAP) were purchased from Gibco BRL Life Technol-
ogies. The 50-[g-32P]ATP were purchased from New
England Nuclear Research Products (Wilmington, DE,
USA). For sequencing labeled DNA, 32P-end labeled
514 base pair restriction fragments were prepared by
first digesting supercoiled pBR322 plasmid DNA with
EcorR I restriction endonuclease. 50-end labeling was
achieved by the treatment of the EcorR I pre-cleaved
DNA with alkaline phosphatase, [g-32P]ATP and T4
polynucleotide kinase. Following the end-labeling, the
DNA was further digested with RsaI to yield the 514
base pair fragment that was purified by 6% preparative
non-denaturing gel electrophoresis with TBE buffer
(TBE electrophoresis buffer is 90mM Tris, 90mM boric
acid, and 2mM EDTA at pH 8.3).
The cyclopent[b]indole quinone methide and the B-
DNA sequence were constructed utilizing fragments
available in fragment libraries of BUILDER module
and minimized utilizing consistent valence force field
(500–1000 iterations). The quinone methide reaction
center was contrained to within a bond length of either
the guanine N(7) or thymine O(6) nucleophiles. The
best-fit structure was achieved by using the DOCKING
module of INSIGHT II employing the consistent
valence forcefield (CVFF). The orientation with the
lowest energy (van der Waals and electrostatic) was
chosen as the optimal DNA-bound structure.
Acknowledgements
The authors gratefully acknowledge support from the
National Institutes of Health, National Science Foun-
dation, and the Arizona Disease Control Research
Commission.
Cleavage of 32P-end labeled DNA. Indoloquinone-
induced cleavage reactions were carried out in two dif-
ferent reaction conditions. For catalytic reduction, the
reaction was carried out in 200 mL total volumes con-
taining calf thymus DNA (500 mM nucleotide con-
centration) 2ꢂ104 cpm 32P-end labeled restriction
fragment and 100 mM of drug in DMSO in 25 mM
phosphate buffer (pH 7.4) with 0.05 mg 5% Pd on car-
bon as catalyst. Reaction mixtures were degassed under
argon for 5 min before being initiated by the bubbling
of H2 at 1 atm for 30 s. The reaction was then degassed
under argon for 5 min and incubated at 37 ꢁC for 30
min. For sodium boromhydride reduction, the reaction
was carried out in 200 mL total volumes containing calf
thymus DNA (500 mM nucleotide concentration) 2ꢂ104
cpm 32P-end labeled restriction fragment and 100 mM of
drug in DMSO in 25 mM phosphate buffer (pH 7.4).
Reaction mixtures were degassed under argon for 5 min
before being initiated by the addition of 100 mM
sodium boromhydride in water. The reaction incubated
at 37ꢂC for 30 min.Under both reaction condition, the
reactions were quenched by ethanol precipitation in the
presence of 0.3M sodium acetate. The DNA pellet was
washed with 70% ethanol, dried, and resuspended in 3
mL of 80% formamide loading dye. Piperidine treatment
was also carried out by dissolving the DNA pellet in 20
mL of fresh 0.2 M piperidine, heated at 90 ꢁC for 10 min
and lyophilized. The samples, along with the Maxam-
Gilbert A+G sequencing reaction mixtures, were then
loaded onto 12% polyacrylamide/7.5 M urea denaturing
gels and electrophoresed at 1200 V for 5 h in the TBE
buffer described above. Autoradiography of the gel was
conducted at ꢀ70 ꢁC using Kodak X-omat film.
References and Notes
1. Wemmer, D. E.; Dervan, P. B. Curr. Opin. Struct. Biol.
1997, 7, 355.
2. White, S.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.;
Dervan, P. B. Nature 1998, 391, 468.
3. Dervan, P. B.; Burli, R. W. Curr. Opin. Chem. Biol. 1999, 3,
688.
4. Herzberg, R. P.; Dervan, P. B. J. Am. Chem. Soc. 1982,
104, 313.
5. Van Dyke, M. W.; Herzberg, R. P.; Dervan, P. B. Proc.
Natl. Acad. Sci. U.S.A. 1982, 79, 5470.
6. Taylor, J. S.; Schultz, P. G.; Dervan, P. B. Tetrahedron
1984, 40, 457.
7. Iida, H.; Jia, G. F.; Lown, J. W. Curr. Opin. Biotechnol.
1999, 10, 29.
8. Fox, K. R.; Yan, Y. F.; Gong, S. Anti-Cancer Drug Des.
1999, 14, 219.
9. Cozzi, P. Farmaco 2000, 55, 168.
10. Denny, W. A. Expert Opin. Ther. Patents 2000, 10, 459.
11. Hamy, F.; Albrecht, G.; Florsheimer, A.; Bailly, C. Bio-
chem. Biophys. Res. Commun. 2000, 270, 393.
12. Guelev, V. M.; Harting, M. T.; Lokey, R. S.; Iverson, B. L.
Chem. Biol. 2000, 7, 1.
13. Nielsen, P. E. Antisense properties of peptide nucleic acid.
In Antisense Technology, Pt A; Phillips, M. I., Ed.; Academic:
San Diego, 2000; p 156.
14. Nielsen, P. E. Annu. Rev. Biophys. Biomol. Struct. 1995,
24, 167.
15. Majumdar, A.; Khorlin, A.; Dyatkina, N.; Lin, F. L. M.;
Powell, J.; Liu, J.; Feiz, Z. Z.; Khripine, Y.; Watanabe, K. A.;
George, J.; Glazer, P. M.; Seidman, M. M. Nat. Genet. 1998,
20, 212.
16. Thurston, D. E. Brit. J. Cancer 1999, 80, 65.
17. Ren, J. S.; Chaires, J. B. Biochemistry 1999, 38, 16067.
18. Crooke, S. T. Bba Gene Struct. Express 1999, 1489, 31.
19. Huang, X.; Suleman, A.; Skibo, E. B. Bioorg. Chem.,
2000, 28, 324.
Modeling into DNA
20. Schulz, W. G.; Nieman, R. A.; Skibo, E. B. Proc. Natl.
Acad. Sci. U.S.A. 1995, 92, 11854.
INSIGHT II from Molecular Simulations, Inc. (San

E. B. Skibo et al. / Bioorg. Med. Chem. 9 (2001) 2445–2459
2459
21. Skibo, E. B.; Schulz, W. G. J. Med. Chem. 1993, 36, 3050.
22. Xing, C.; Skibo, E. B.; Dorr, R. T. J. Med. Chem., sub-
mitted for publication.
23. Xing, C. G.; Wu, P.; Skibo, E. B.; Dorr, R. T. J. Med.
Chem. 2000, 43, 457.
24. Ouyang, A.; Skibo, E. B. Biochemistry 2000, 39, 5817.
25. Gutierrez, P. L. Front Biosci. 2000, 5, D629.
26. Kang, H. M.; Rokita, S. E. Nucleic Acids Res. 1996, 24, 3896.
27. Zeng, Q. P.; Rokita, S. E. J. Org. Chem. 1996, 61, 9080.
28. Rokita, S. E.; Yang, J. H.; Pande, P.; Greenberg, W. A. J.
Org. Chem. 1997, 62, 3010.
29. Lemus, R. L.; Skibo, E. B. J. Org. Chem. 1988, 53, 6099.
30. Skibo, E. B. J. Org. Chem. 1992, 57, 5874.
31. Skibo, E. B. J. Org. Chem. 1986, 51, 522.
36. Ruangyuttikarn, W.; Skiles, G. L.; Yost, G. S. Chem. Res.
Toxicol. 1992, 5, 713.
37. Coulson, C. A.; Streitwieser, A. Dictionary of ꢀ-Electron
Calculations; W. H Freeman: San Francisco, 1965.
38. Kim, H. S.; LeBreton, P. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 3725.
39. Kim, H. S.; Yu, M.; Jaing, Q.; LeBreton, P. R. J. Am.
Chem. Soc. 1993, 115, 6169.
40. Broch, H.; Hamza, A.; Vasilescu, D. J. Biomol. Struct.
Dyn. 1996, 13, 903.
41. Hamza, A.; Broch, H.; Vasilescu, D. J. Biomol. Struct.
Dyn. 1996, 13, 915.
42. Maxam, A. M.; Gilbert, W. Methods Enzymol. 1980, 65,
499.
32. Zhou, Q. B.; Turnbull, K. D. J. Org. Chem. 1999, 64,
2847.
43. Maliepaard, M.; deMol, N. J.; Tomasz, M.; Gargiulo, D.;
Janssen, L. H. M.; vanDuynhoven, J. P. M.; vanVelzen, E. J. J.;
Verboom, W.; Reinhoudt, D. N. Biochemistry 1997, 36, 9211.
44. Maliepaard, M.; Wolfs, A.; Groot, S. E.; de Mol, N. J.;
Janssen, L. H. M. Br. J. Cancer 1995, 71, 836.
45. Workman, P.; Binger, M.; Kooistra, K. L. Int. J. Radiat.
Oncol. Biol. Phys 1992, 22, 713.
33. Franck, R. W.; Tomasz, M. The Chemistry of Mitomy-
cins. In The Chemistry of Antitumor Agents; Wilman, D. E.,
Ed.; Blackie & Sons; Glasgow, UK, 1990; p 379.
34. Boruah, R. C.; Skibo, E. B. J. Org. Chem. 1995, 60, 2232.
35. Skiles, G. L.; Yost, G. S. Chem. Res. Toxicol. 1996, 9, 291.