DNA Alkylation by leinamycin can be triggered by cyanide and phosphines

Article abstract of DOI:10.1016/S0960-894X(01)00196-2

Previous work has shown that alkylation of DNA by the antitumor agent leinamycin (1) is potentiated by reaction of the antibiotic with thiols. Here, it is shown that other soft nucleophiles such as cyanide and phosphines can also trigger DNA alkylation by leinamycin. Overall, the results suggest that reactions of cyanide and phosphines with leinamycin produce the oxathiolanone intermediate (2), which is known to undergo rearrangement to the DNA-alkylating episulfonium ion 4.

Full text of DOI:10.1016/S0960-894X(01)00196-2

Bioorganic & Medicinal Chemistry Letters 11 (2001) 1511–1515
DNA Alkylation by Leinamycin Can Be Triggered by Cyanide
and Phosphines
Hong Zang, Leonid Breydo, Kaushik Mitra, Jeffrey Dannaldson and Kent S. Gates*
Departments of Chemistry and Biochemistry, University of Missouri-Columbia, Columbia, MO 65211, USA
Received 24 January 2001; accepted 14 March 2001
Abstract—Previous work has shown that alkylation of DNA by the antitumor agent leinamycin (1) is potentiated by reaction of the
antibiotic with thiols. Here, it is shown that other soft nucleophiles such as cyanide and phosphines can also trigger DNA alkylation
by leinamycin. Overall, the results suggest that reactions of cyanide and phosphines with leinamycin produce the oxathiolanone
intermediate (2), which is known to undergo rearrangement to the DNA-alkylating episulfonium ion 4. # 2001 Elsevier Science
Ltd. All rights reserved.
Leinamycin (1) is a potent antitumor antibiotic that
contains a unique 1,2-dithiolan-3-one 1-oxide hetero-
cycle.1,2 Reaction of thiols with this sulfur heterocycle in
leinamycin initiates chemistry that leads to oxidative
DNA damage and DNA alkylation.3ꢀ6 Thiol-triggered
DNA alkylation by leinamycin involves initial conver-
sion of the parent antibiotic to the oxathiolanone form
2,5,7 followed by rearrangement to the episulfonium ion
4 that efficiently modifies double-stranded DNA at the
N7-position of guanine residues (Scheme 1).5 The same
DNA-alkylating episulfonium ion (4) is also produced
by a slower, thiol-independent pathway initiated by
hydrolysis of leinamycin’s sulfur heterocycle (Scheme 1).8,9
Although hydrolytic and thiolytic activation of leina-
mycin occur by different chemical pathways, both pro-
cesses are believed to proceed via the intermediacy of
the same crucial oxathiolanone species (2).
properties of the antibiotic. Here we report that phos-
phines and cyanide are able to trigger DNA alkylation
by the antitumor antibiotic leinamycin. Results of our
chemical model reactions and DNA-damage experi-
ments suggest that reaction of cyanide and phosphines
with leinamycin affords the critical oxathiolanone inter-
mediate (2), which subsequently undergoes rearrangement
to the DNA-alkylating episulfonium ion 4.
The results of previous chemical model reactions sug-
gested that phosphines might be capable of triggering
DNA alkylation by leinamycin. Specifically, we found
that the reaction of triphenylphosphine with 3H-1,2-
benzodithiol-3-one 1-oxide (6) provides a good yield of
the cleft-shaped dimer 8.11,12 In the context of the cur-
rent work, it is important to note that this reaction is
most easily rationalized11 by invoking the involvement
of the 3H-2,1-benzoxathiol-3-one intermediate
7
In order to increase our overall understanding of leina-
mycin’s fundamental chemical properties, we investi-
gated the ability of nucleophiles other than thiol to
trigger DNA alkylation by this antibiotic. So-called
‘soft’ nucleophiles such as thiols, phosphines, cyanide,
and iodide typically display high ‘thiophilicity’. In other
words, these nucleophiles react readily with electrophilic
sulfur centers.10 Therefore, we anticipated that various
soft nucleophiles might react with leinamycin’s electro-
philic sulfur heterocycle in a manner similar to thiols—
that is, in a manner that potentiates the DNA-alkylating
(Scheme 2), which corresponds to the ‘activated form’
of leinamycin (2). Importantly, past work has shown
that 6 provides an accurate model for predicting the
chemical reactivity of leinamycin’s sulfur heterocycle.5,7
As part of the current work chemical model reactions
were performed and the results suggest that cyanide can
trigger DNA alkylation by leinamycin. Treatment of 6
with sodium cyanide (10 equiv) in dry tetrahydrofuran
containing 18-crown-6 (0.07 equiv) affords 2-thiocyana-
tobenzoic acid (10) as the only detectable product (by
TLC). The compound was isolated in 75% yield as its
methyl ester derivative after treatment of the reaction
with diazomethane.13,14 We envision that 10 is formed
*Corresponding author. Tel.: +1-573-882-6763; fax: +1-573-882-
2754; e-mail: gatesk@missouri.edu
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00196-2

1512
H. Zang et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1511–1515
by a mechanism involving initial attack of the soft cya-
nide nucleophile on the soft central sulfur atom of 6 to
afford the sulfenic acid derivative 9, followed by cycli-
zation to yield 3H-2,1-benzoxathiol-3-one (7) and one
equivalent of thiocyanate (ꢀSCN) (Scheme 3). Attack of
a second equivalent of cyanide on 7 is expected to yield
2-thiocyanatobenzoic acid (10). Consistent with the
proposed mechanism, the reaction does not proceed to
completion when less than two equivalents of cyanide
are employed. In addition, we find that one equivalent
of thiocyanate is produced in the reaction (quantita-
tively detected by the characteristic absorbance of
Fe3+–thiocyanate complexes at 450 nm).15 The results
of this model reaction suggest that attack of cyanide ion
on leinamycin will produce the ‘activated’ oxathiola-
none form of the antibiotic (2).
Our first evidence that phosphines and cyanide can, in
fact, trigger DNA damage by leinamycin was obtained
using a plasmid-based assay that can detect DNA alky-
lation. In this assay, supercoiled plasmid DNA that
ha been exposed to an alkylating reagent is subjected
to a ‘plasmid-compatible’ Maxam–Gilbert-type workup
involving warming the alkylated DNA in the presence
of N,N0-dimethylethylenediamine (DMEDA, 100 mM,
1.5 h, 37 ꢁC) or putrescine (100 mM, 6 h, 37 ꢁC). Ther-
mal treatment induces loss of certain alkylated bases
from the DNA (e.g., purines alkylated at N7) to yield
Scheme 1.
Scheme 2.
Scheme 3.

H. Zang et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1511–1515
1513
abasic sites,16 which are subsequently converted to
strand cleavage sites by the reaction with the polyamine
(DMEDA or putrescine).17,18 The number of resulting
DNA-cleavage events can be quantitatively assessed
using agarose gel electrophoresis to monitor the con-
version of supercoiled plasmid (form I) into the open
circular form (form II) that occurs when the supercoiled
plasmid sustains a single-strand break.
to the thiol-triggered process under these conditions
(compare thiol-triggered DNA cleavage shown in Fig. 3 to
the cleavage shown in Figs. 1 and 2). It is noteworthy
that water soluble phosphines such as tris(2-hydro-
xyethyl)phosphine are comparable to triphenylpho-
sphine in their ability to trigger DNA alkylation in this
system (data not shown). The notion that DNA clea-
vage observed in these assays results from alkylation
rather than an oxidative, radical-mediated process is
supported by the finding that addition of oxygen radical
scavengers (e.g., EtOH, 100 mM) or catalase (which
destroys H2O2) has no significant effect on the cleavage
yields in our assays (data not shown).
Using this plasmid-based alkylation assay, we find that
treatment of supercoiled plasmid DNA with leinamycin
(500 nM) and Ph3P (10–250 mM) or potassium cyanide
(0.5–5 mM) for 6 h at 24 ꢁC, followed by the DMEDA
or putrescine workup, leads to significantly more strand
breaks than does treatment of the DNA with leinamycin
alone (Figs. 1 and 2). Cleavage efficiency increases with
increasing concentration of the phosphine or cyanide
triggering agent. Relatively high concentrations of
sodium cyanide are required to obtain efficient cleavage,
presumably because relatively low concentrations of the
nucleophilic cyanide ion exist in neutral aqueous solu-
tion (pKa of HCN ꢂ9.4). Consistent with the lower
thiophilicity of phosphines and cyanide relative to
thiols,10 higher concentrations of these agents are
required to achieve DNA-alkylation yields comparable
Figure 3. DNA cleavage by leinamycin (1) with varying concentra-
tions of 2-mercaptoethanol. Supercoiled plasmid DNA (pBR322, 38
mM in bp) was incubated for 7 h at 24 ꢁC with 1 (0.5 mM) and various
concentrations of of 2-mercaptoethanol in sodium phosphate buffer
(50 mM, pH 7) followed by treatment with DMEDA (100 mM) for 1.5
h at 37 ꢁC. The resulting DNA cleavage was analyzed using agarose
gel electrophoresis. The numbers in parentheses following the descrip-
tion of each lane indicate the mean number of strand breaks per plas-
mid molecule calculated using the equation S=ꢀln fI where fI is the
fraction of plasmid present as form I. The values are corrected for
strand breaks present in the untreated DNA. Lane 1, DNA alone; lane
2, 1 (0.5 mM) (0.1); lane 3, thiol alone (100 mM) (0.0); lane 4, 1 (0.5
mM)+thiol (1 mM) (0.5); lane 5, 1 (0.5 mM)+thiol (5 mM) (1.2); lane 6,
1 (0.5 mM)+thiol (10 mM) (1.3); lane 7, 1 (0.5 mM)+thiol (50 mM)
(1.6); lane 8, 1 (0.5 mM)+thiol (100 mM) (1.7).
Figure 1. DNA cleavage by leinamycin (1) with varying concentra-
tions of triphenylphosphine. Supercoiled plasmid DNA (pBR322, 38
mM in bp) was incubated for 6 h at 24 ꢁC with 1 (0.5 mM) and various
concentrations of Ph3P in sodium phosphate buffer (50 mM, pH 7)
followed by treatment with putrescine (100 mM) for 6 h at 37 ꢁC. The
resulting DNA cleavage was analyzed using agarose gel electrophor-
esis. The numbers in parentheses following the description of each lane
indicate the mean number of strand breaks per plasmid molecule cal-
culated using the equation S=ꢀln fI where fI is the fraction of plasmid
present as form I. The values are corrected for strand breaks present in
the untreated DNA. Lane 1, DNA alone; lane 2, 1 (0.5 mM) (0.1); lane
3, Ph3P alone (10 mM) (0.0); lane 4, Ph3P alone (50 mM) (0.0); lane 5,
Ph3P alone (100 mM) (0.0); lane 6, Ph3P (250 mM) (0.0); lane 7, 1 (0.5
mM)+Ph3P (10 mM) (1.0); lane 8, 1 (0.5 mM)+Ph3P (50 mM) (2.1);
lane 9, 1 (0.5 mM)+Ph3P (100 mM) (4.4); lane 10, 1 (0.5 mM)+Ph3P
(250 mM) (4.4).
Figure 2. DNA cleavage by leinamycin (1) with varying concentra-
tions of cyanide. Supercoiled plasmid DNA (pBR322, 38 mM in bp)
was incubated for 6 h at 24 ꢁC with 1 (0.5 mM) and various con-
centrations of KCN in sodium phosphate buffer (50 mM, pH 7) fol-
lowed by treatment with putrescine (100 mM) for 6 h at 37 ꢁC. The
resulting DNA cleavage was analyzed using agarose gel electrophor-
esis. The numbers in parentheses following the description of each lane
indicate the mean number of strand breaks per plasmid molecule cal-
culated using the equation S=ꢀln fI where fI is the fraction of plasmid
present as form I. The values are corrected for strand breaks present in
the untreated DNA. Lane 1, DNA alone; lane 2, 1 (0.5 mM) (0.1); lane
3, KCN alone (250 mM) (0.0); lane 4, KCN alone (1 mM) (0.0); lane 5,
KCN alone (5 mM) (0.0); lane 6, KCN alone (10 mM) (0.0); lane 7, 1
(0.5 mM)+KCN (250 mM) (0.8); lane 8, 1 (0.5 mM)+KCN (1 mM)
(1.5); lane 9, 1 (0.5 mM)+KCN (5 mM) (1.8); lane 10, 1 (0.5
mM)+KCN (10 mM) (2.6).
Figure 4. Triphenylphosphine-triggered alkylation of DNA by leina-
mycin. In the alkylation reactions, a 50-32P-labeled 145-base pair DNA
fragment was incubated at 37 ꢁC with leinamycin (13 mM) and triphe-
nylphosphine (1 mM) or 2-mercaptoethanol (1 mM) for 2 h in sodium
phosphate buffer (50 mM, pH 7.0), followed by Maxam–Gilbert
workup, denaturing 20% polyacrylamide gel electrophoresis, and
imaging of the gel by exposure to Fuji X-ray film at 4 ꢁC for 60 h.
Lane 1, DNA alone; lane 2, Maxam–Gilbert G rxn; lane 3, leinamycin
(13 mM) alone; lane 4, Ph3P alone (1 mM); lane 5, 2-mercaptoethanol
alone (1 mM); lane 6, leinamycin (13 mM)+Ph3P (1 mM); lane 7, lei-
namycin (13 mM)+2-mercaptoethanol (1 mM).

1514
H. Zang et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1511–1515
To further characterize phosphine and cyanide-trig-
gered DNA damage by leinamycin, we examined the
resulting DNA cleavage using denaturing poly-
acrylamide gel electrophoresis (PAGE). We find that
incubation of a 50-32P-labeled 145-base pair restriction
fragment of duplex DNA with leinamycin (13 mM) and
Ph3P (1 mM), followed by Maxam–Gilbert workup
(0.1 M piperidine, 90 ꢁC, 30 min) affords cleavage at
every guanine residue in the fragment (Fig. 4). Under
these conditions (2 h incubation), phosphine-triggered
DNA alkylation (lane 6, Fig. 4) is significantly more
efficient than background alkylation by leinamycin
alone (lane 3) and is comparable to thiol triggered DNA
alkylation (lane 7).
Our results clearly show that phosphines and cyanide
can trigger DNA alkylation by the antibiotic leinamy-
cin. Leinamycin has recently been shown to undergo
slow conversion to the DNA-alkylating episulfonium
ion (4) in aqueous buffered solution;9 however, under
the conditions of the experiments described here, DNA
alkylation by this pathway is relatively inefficient (lane
2, Fig. 1; lane 2, Fig. 2, lane 3, Fig. 4) though it can
clearly be seen in the experiments where leinamycin is
incubated alone with DNA for longer times at 37 ꢁC
(e.g., lane 2, Fig. 5). In all experiments, the efficiency of
phosphine- and cyanide-triggered DNA alkylation is
well above the background levels of alkylation resulting
from the incubation of DNA with leinamycin alone.
Similarly, treatment of a 50-32P-labeled 19-base pair oli-
gonucleotide duplex with leinamycin (50 mM) and
sodium cyanide (10 mM), followed by Maxam–Gilbert
workup, produces cleavage specifically at guanine resi-
dues. Phosphorimage analysis of the gel reveals that the
yield of cyanide-triggered DNA alkylation is nearly
equivalent (ꢂ90%) to that afforded by the thiol-trig-
gered alkylation process and the sequence-specificity of
guanine alkylation resulting from the cyanide-promoted
activation of leinamycin is nearly identical to that of the
thiol-mediated process (Fig. 5). In this experiment,
DNA alkylation by leinamycin in the absence of any acti-
vating agent (lane 2, Fig. 5) is approximately 15% of that
obtained in the thiol-activated case (lane 4, Fig. 5).
The similarities in the nature of the DNA damage
resulting from activation of leinamycin by phosphine,
cyanide and thiol suggest that all three reactions gen-
erate the same DNA-alkylating intermediate (4, Scheme
1). Studies employing the leinamycin model compound
3H-1,2-benzodithiol-3-one 1-oxide (6, Schemes 2 and 3)
further suggest that attack of cyanide or phosphine on
the sulfur heterocycle of leinamycin converts the anti-
biotic to its ‘activated’ oxathiolanone form 2 which
undergoes rearrangement to the DNA-alkylating epi-
sulfonium ion 4. Consistent with this mechanism, LC/
MS analysis reveals that treatment of leinamycin with
phosphines in aqueous solution affords the character-
istic rearrangment product5 resulting from hydrolysis of
the episulfonium ion (4). Overall, these studies reveal
novel, chemically interesting routes by which the anti-
tumor antibiotic leinamycin can be converted to its
DNA-alkylating form.
Acknowledgements
The authors are grateful for support of this work from
the National Institutes of Health (GM51565 and
CA83925) and we thank Dr. Yutaka Kanda and other
researchers at Kyowa Hakko Kogyo Ltd. for helpful
discussions and for providing samples of leinamycin.
Finally, we thank members of the Gates research group
for critical review of the manuscript.
References and Notes
1. Hara, M.; Takahashi, I.; Yoshida, M.; Kawamoto, I.;
Morimoto, M.; Nakano, H. J. Antibiot. 1989, 42, 333.
2. Hara, M.; Asano, K.; Kawamoto, I.; Takiguchi, T.; Kat-
sumata, S.; Takahashi, K.; Nakano, H. J. Antibiot. 1989, 42,
1768.
3. Gates, K. S. Chem. Res. Toxicol. 2000, 13, 953.
4. Mitra, K.; Kim, W.; Daniels, J. S.; Gates, K. S. J. Am.
Chem. Soc. 1997, 119, 11691.
5. Asai, A.; Hara, M.; Kakita, S.; Kanda, Y.; Yoshida, M.;
Saito, H.; Saitoh, Y. J. Am. Chem. Soc. 1996, 118, 6802.
6. Behroozi, S. J.; Kim, W.; Dannaldson, J.; Gates, K. S.
Biochemistry 1996, 35, 1768.
7. Behroozi, S. J.; Kim, W.; Gates, K. S. J. Org. Chem. 1995,
60, 3964.
Figure 5. Cyanide-triggered DNA alkylation by leinamycin. In the
alkylation reaction, a DNA duplex consisting of a radiolabeled 30-mer
(50-32P-ATA ATT TGT ATA GGG AGA GAA AGT TAA TAA-30)
hybridized to a complementary 19-mer (50-TTA TTA ACT TTC TCT
CCC T-30) was incubated at 37 ꢁC with leinamycin (50 mM) and
potassium cyanide (10 mM) for 22 h in Hepes buffer (20 mM, pH 7.0)
containing herring sperm DNA (80 mM bp), followed by Maxam–
Gilbert workup, denaturing 16% polyacrylamide gel electrophoresis,
and documentation of the gel by phosphorimager analysis. The clearly
separated bands in the lower and middle portion of the gel result from
alkylation and cleavage at the six guanines in the duplex region of the
labeled oligonucleotide. Lane 1, DNA alone; lane 2, leinamycin alone;
lane 3, b-ME alone (0.5 mM); lane 4, leinamycin+b-ME (0.5 mM);
lane 5, leinamycin+KCN (10 mM); lane 6, Maxam–Gilbert G reaction.

H. Zang et al. / Bioorg. Med. Chem. Lett. 11 (2001) 1511–1515
1515
8. Asai, A.; Saito, H.; Saitoh, Y. Bioorg. Med. Chem. 1997, 5,
723.
9. Breydo, L.; Zang, H.; Mitra, K.; Gates, K. S. J. Am. Chem.
Soc. 2001, 123, 2060.
10. Oae, S. Organic Sulfur Chemistry: Structure and Mechanism;
CRC: Boca Raton, 1991.
11. Mitra, K.; Gates, K. S. Tetrahedron Lett. 1995, 36, 1391.
12. Mitra, K.; Pohl, M. E.; MacGillivray, L. R.; Barnes, C. L.;
Gates, K. S. J. Org. Chem. 1998, 62, 9361.
13. To a mixture of 7 (400 mg, 2.17 mmol), 18-crown-6 (40
mg, 0.15 mmol) and NaCN (1.07 g, 21.8 mmol) was added
THF (30 mL) to produce a light yellow suspension. After
stirring for 12 h at 24 ꢁC under dry nitrogen, the reaction was
complete as monitored by TLC. Then HCl (2 M, 15 mL) was
added to the reaction mixture, and the resulting clear solution
was extracted with brine (20 mL). CAUTION: All steps of this
procedure should be performed in a well ventilated fume hood
because highly toxic HCN gas may be produced. The aqueous
layer was extracted with CH2Cl2 (2ꢃ30 mL) and to the com-
bined organic layers was added diazomethane (25 mL of a
0.3 M solution in ether; CAUTION: explosion hazard!). The
resulting solution was dried over Na2SO4 and the solvent
removed under reduced pressure to yield a yellow oil which
was purified by column chromatography on silica gel (6:1
hexane/EtOAc) to produce the methyl ester of compound 10
(315 mg, 75%) as a white solid: mp 67–69 ꢁC, Rf 0.61 (3:1
hexane/EtOAc); 1H NMR (250 MHz, CDCl3) d 8.12 (dd,
J=8, 1.5 Hz, 1H), 7.92 (dd, J=8, 1 Hz, 1H), 7.65 (ddd, J=8
Hz, 8 Hz, and 1.5 Hz, 1H), 7.42 (ddd, J=8 Hz, 8 Hz, and 1
Hz, 1H), 3.97 (s, 3H); 13C NMR (62.5 MHz, CDCl3) d 166.4,
134.1, 131.6, 130.8, 127.8, 127.5, 126.1, 111.5, 52.8; HRMS
(EI) m/z calcd for C9H7NO2S 193.020, found 193.020. Identity
of this compound was further confirmed by comparison
(NMR, TLC) to a sample prepared from anthranilic acid by
the route of Simchen and Wenzelburger.14
14. Simchen, G.; Wenzelburger, J. Chem. Ber. 1970, 103, 413.
15. Vogel, A. Vogel’s Textbook of Quantitative Inorganic
Analysis, 4th ed.; Longman: London, 1978. To a mixture of 7
(51 mg, 0.275 mmol) and KCN (179 mg, 2.75 mmol) was
added DMF (5.5 mL). Solution immediately turned yellow,
and after stirring for 15 min at room temperature the reaction
was judged to be complete by TLC. In order to detect the
presence of thiocyanate, 5 mL aliquots of the reaction mixture
were added to 1 mL of an FeCl3 solution (1 mM, in water)
containing 20% DMSO and 60 mM HCl. Resulting iron–
thiocyanate complexes were detected using UV–vis spectro-
scopy by their absorbance at 450 nm. The yield of thiocyanate
ion was determined by comparing the absorbance to a cali-
bration curve prepared by measuring the absorbance of iron
thiocyanate complexes prepared from known concentrations
of thiocyanate.
16. Epe, B.; Hegler, J. Methods Enzymol. 1994, 234, 122.
17. McHugh, P. J.; Knowland, J. Nucleic Acids Res. 1995, 23,
1664.
18. Povirk, L. F.; Goldberg, I. H. Proc. Natl. Acad. Sci.
U.S.A. 1985, 82, 3182.