Lysine-enediyne conjugates as photochemically triggered DNA double-strand cleavage agents

Article abstract of DOI:10.1039/b417012a

Statistical analysis of DNA-photocleavage by two types of lysine-enediyne conjugates confirms that more double-strand breaks are produced than can be accounted for by coincident single-strand breaks. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b417012a

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Lysine–enediyne conjugates as photochemically triggered DNA double-
strand cleavage agents{
Serguei V. Kovalenko and Igor V. Alabugin*
Received (in Cambridge, UK) 10th November 2004, Accepted 23rd December 2004
First published as an Advance Article on the web 25th January 2005
DOI: 10.1039/b417012a
conjugates 3 and 4 were synthesized in a similar fashion from
3,4-dihydroxybenzoic acid as described in the ESI{.
Statistical analysis of DNA-photocleavage by two types of
lysine–enediyne conjugates confirms that more double-strand
breaks are produced than can be accounted for by coincident
single-strand breaks.
The ability of enediynes to cleave DNA under irradiation was
investigated using conversion of supercoiled plasmid DNA into the
respective relaxed circular and linear forms (Forms II and III). The
relative amounts of the three DNA forms were determined by
densitometric analysis of the gel electrophoresis bands at different
irradiation times.
In recent years, considerable effort has been invested into
development of synthetic reagents that selectively cleave DNA
under irradiation with visible or UV light without the use of metals
or reducing agents.1 The ability of these chemical systems to
generate reactive organic intermediates on demand represents a
promising approach to new antitumor therapeutic strategies.
Enediyne antibiotics are among the most potent natural
anticancer agents. Their biological activity stems from cyclization
of the enediyne moiety to a reactive p-benzyne diradical followed
by simultaneous cleavage of both strands of duplex DNA (double-
stranded cleavage, or ds). Since ds lesions are believed to be more
biologically important than single-strand (ss) breaks,3 numerous
studies have focused on developing new DNA ds photocleavage
agents utilizing the photochemical version of the Bergman
cyclization.2 Despite these efforts, only one of the photoactivated
enediynes reported in the literature shows ds DNA cleavage at mM
concentrations (1026–1025 M)2a and it still remains unclear
whether the photochemical version of the Bergman cyclization
provides true ds DNA cleavage instead of accumulation of
random ss breaks.
As shown in Fig. 1, irradiation of plasmid pB322 in the presence
of conjugate 1 generates linear DNA before all of the supercoiled
DNA is converted to the relaxed circular form (lanes 3–9). Under
conditions where all of the three DNA forms are present at the
same time, one can gain a more thorough insight into the nature of
the cleavage using the statistical test of Povirk.6,7 This test assumes
a Poisson distribution of strand cuts and calculates average
number of ss-(n1) and ds-breaks (n2) per DNA molecule (Table 1).8
In the case of enediyne 1, the n1/n2 ratio decreases before reaching
Scheme 1 C1–C5 cyclization of bis-TFP enediynes.
Our interest in the design of new molecular systems for DNA
cleavage was caused by the discovery of photochemical transfor-
mation of tetrafluoropyridinyl (TFP) substituted enediynes into
indenes.4 Since this reaction is accompained by four formal
H-atom abstractions, we suggested that this chemistry can also be
used for development of potent DNA photocleaving agents.
By taking advantage of the ability of positively charged amino
acids to bind to DNA,5 we designed hybrid molecular systems 1, 3,
which combine bis-TFP enediyne and lysine moieties via two
linkers of different length. For comparison, we also synthesized
bis-Ph enediynes 2 and 4 expected to undergo the photoBergman
cyclization.2f,h Synthesis of enediynes 1 and 2 is outlined in
Scheme 2. Nitration of dibromobenzene yielded 3,4-dibromoni-
trobenzene which was coupled with acetylenes under Sonogashira
conditions. The products were reduced with SnCl2 to afford
anilines which were acylated with Boc2LysOH/POCl3 in pyridine.
Deprotection with HCl provided the requisite lysine–enediyne
conjugates 1 and 2 as diammonium salts. Lysine–enediyne
Scheme 2 (a) HNO3, rt, 75%; (b) Me3SiCCH, PdCl2(PPh3)2, CuI, (i-
Pr)2NH, 78%; (c) PhCCH, PdCl2(PPh3)2, CuI, (i-Pr)2NH, 72 h, rt, 76%;
(d) C5F5N, CsF, DMF, 24 h, rt, 65%; (e) SnCl2, THF, 3 h, rt, 79%; (f) 1.
Boc2Lys(OH), POCl3, Py, 220 uC, 1 h, rt, 2. HCl, CH2Cl2, 41%; (g) SnCl2,
THF, 3 h, rt, 75%; (h) 1. Boc2Lys(OH), POCl3, Py, 220 uC A rt, 1 h; 2.
HCl, CH2Cl2, 39%.
{ Electronic supplementary information (ESI) available: Full experimental
details. See http://www.rsc.org/suppdata/cc/b4/b417012a/
*alabugin@chem.fsu.edu
1444 | Chem. Commun., 2005, 1444–1446
This journal is ß The Royal Society of Chemistry 2005

View Article Online
Table 2 Statistical efficiency of single-strand and double-strand
break formation by enediynes 2–4 as a function of irradiation time
n1/n2 (n2)
Time/min
2
3
4
17
22
27
32
37
42
35 (0.03)
35 (0.04)
35 (0.05)
26 (0.09)
28 (0.09)
29 (0.10)
15 (0.05)
17 (0.05)
12 (0.09)
10 (0.11)
14 (0.13)
14 (0.17)
13 (0.06)
10 (0.10)
13 (0.09)
13 (0.12)
11 (0.14)
10 (0.19)
fluorescence intensity of bound ethidium bromide) of compounds
1–4 are, on average, an order of magnitude smaller than C50 of
spermidine (respectively 0.9 (1), 5.0 (2), 1.3 (3), 1.5 (4) vs. 29 mM
[36 mM in ref. 12a]). This result suggests that non-electrostatic
components contribute significantly to binding of lysine–enediyne
conjugates to DNA. Interestingly, the similarity in binding of
enediynes 3 and 4 correlates well with the respective n1/n2 ratios.
In conclusion, we have unambiguously shown that photoacti-
vated enediynes can cause true non-random ds DNA cleavage.
Further research will concentrate on understanding chemistry
responsible for the cleavage, optimizing DNA binding and
studying its correlation with the cleavage efficiency.
Fig. 1 Photochemical cleavage of pBR322 supercoiled DNA (30 mM) by
1 (20 mM) in phosphate buffer (20 mM, pH 8.0). Lanes 1–9, DNA + 1 + hv
for 2, 7, 13, 16, 22, 25, 30, 35 and 45 min of irradiation (l . 305 nm)
respectively. The relative amounts of the three DNA forms are given by
diamonds (Form I), hollow circles (Form II) and squares (Form III). The
lines are used only to organize the data.
Table 1 Statistical efficiency of single-strand and double-strand
break formation by 1 as a function of irradiation time
Number of
ss-breaks (n1)
and ds-breaks
(n2) per molecule
Relative amounts (%)
The authors are grateful to the National Science Foundation
(CHE-0316598) and to the Material Research and Technology
(MARTECH) Center at Florida State University for partial
support of this research, to the 3M Company for an Untenured
Faculty Award, to Olga Barykina and Charlene Townsend for
their help with the experimental work, to Professor Tim Cross and
Dr Alla Korepanova for the use of the experimental facilities and
Professor Debra Mohler (Emory) for helpful discussions.
Time/min Form I Form II Form III n1
n2
n1/n2
0
2
88.1
75.2
64.0
54.0
52.9
50.0
32.0
38.9
29.5
22.1
11.9
24.8
35.2
44.5
45.5
48.3
63.6
56.2
65.0
68.9
0
0
0.13
0.29
0.44
0.60
0.62
0.68
1.09
0.90
1.16
1.41
0
0
7
0.8
1.5
1.6
1.7
4.4
4.9
5.5
9.0
0.008
0.014
0.016
0.017
0.046
0.052
0.059
0.100
53
44
40
40
24
17
20
14
13
16
22
25
30
35
40
Serguei V. Kovalenko and Igor V. Alabugin*
Department of Chemistry and Biochemistry, Florida State University,
Tallahassee, FL 32306-4390, USA, USA.
E-mail: alabugin@chem.fsu.edu; Fax: +1 850 644 8281;
Tel: +1 850 644 5795
a plateau suggesting that ss and ds events are kinetically
independent and that Form II A Form III conversion occurs at
a slower rate than the initial scission. However, at all times the
range of n1/n2 values (14–53) is significantly smaller than expected
from a completely random process.9
Notes and references
1 B. Armitage, Chem. Rev., 1998, 98, 1171 and references therein.
2 DNA-cleavage studies: (a) J. Kagan, X. Wang, X. Chen, K. Y. Lau,
I. V. Batac, R. W. Tuveson and J. B. Hudson, J. Photochem. Photobiol.,
1993, 21, 1352; (b) R. L. Funk, E. R. R. Young, R. M. Williams,
M. F. Flanagan and T. L. Cecil, J. Am. Chem. Soc., 1996, 118, 3291; (c)
N. Choy, B. Blanko, J. Wen, A. Krishan and K. C. Russell, Org Lett.,
2000, 2, 3761; (d) Intramolecular PET activation: M. Schmittel, G. Viola,
F. Dall’Acqua and G. Morbach, Chem. Commun., 2003, 646; (e)
MLCT-promoted Bergman cyclization: P. J. Benites, R. C. Holmberg,
D. S. Rawat, B. J. Kraft, L. J. Klein, D. G. Peters, H. H. Thorp and
J. M. Zaleski, J. Am. Chem. Soc., 2003, 125, 6434. Fundamental
photochemistry; (f) A. Evenzahav and N. J. Turro, J. Am. Chem. Soc.,
1998, 120, 1835; (g) T. Kaneko, M. Takanashi and M. Hirama, Angew.
Chem., Int. Ed. Engl., 1999, 38, 1267. Peptide cleavage and optimization
of delivery strategies: ; (h) G. B. Jones, J. M. Wright, G. Plourde, II,
A. D. Purohit, J. K. Wyatt, G. Hynd and F. Fouad, J. Am. Chem. Soc.,
2000, 122, 9872; F. S. Fouad, C. F. Crasto, Y. Lin and G. B. Jones,
Tetrahedron Lett., 2004, 45, 7753.
The other lysine–enediyne conjugates 2–4 are also capable of
causing DNA cleavage efficiently: less than 10% of supercoiled
DNA remains after 45 min of radiation. Importantly, in every case
the n1/n2 values cannot be accounted by coincident random ss-
breaks (Tables 2).9,10 Comparable n1/n2 values, in the range of
6–20, have been observed for iron bleomycin.11 The dynamics of
cleavage are interesting—in contrast to 1, the n1/n2 values for
enediynes 2–4 remain relatively constant with time. The n1/n2 ratios
are noticeably smaller for the enediynes 3 and 4 where the DNA-
cleaving moiety is attached to the lysine residue through a longer
linker. These observations suggest that DNA-photocleaver inter-
action plays an important role in the DNA cleavage and that the
longer tether may allow for better alignment of the enediyne for
interaction with opposing DNA strands.
3 (a) L. F. Povirk, Mutat. Res., 1996, 355, 71; (b) D. T. Weaver, Crit. Rev.
Eukaryot. Gene Exp., 1996, 6, 345; (c) A. Sancar, Annu. Rev. Biochem.,
1996, 65, 43; (d) C. Rajani, J. R. Kincaid and D. H. Petering, J. Am.
Chem. Soc., 2004, 126, 3829.
4 Experiment: I. V. Alabugin and S. V. Kovalenko, J. Am. Chem. Soc.,
2002, 124, 9052. Theoretical treatment: I. V. Alabugin and
The interaction of lysine–enediyne conjugates with DNA was
investigated using fluorescence quenching binding assay12 based on
displacement of ethidium bromide by enediynes. The C50 values
(the concentration of conjugate leading to a 50% reduction in
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 1444–1446 | 1445

View Article Online
M. Manoharan, J. Am. Chem. Soc., 2003, 125, 4495 Thermodynamics
of peptide binding to nucleic acids.
5 Thermodynamics of peptide binding to nucleic acids: T. M. Lohman
and D. P. Mascotti, Methods Enzymol., 1992, 212, 400.
6 L. F. Povirk, W. Wu¨bker, W. Ko¨hnlein and F. Hutchinson, Nucleic
Acids Res., 1977, 4, 3573; L. F. Povirk and C. W. Houlgrave,
Biochemistry, 1988, 27, 3850. For an alternative approach based on
multicomponent kinetic fitting, see: J. Goodisman, C. Kirk and
J. C. Dabrowiak, Biophys. Chem., 1997, 69, 249.
7 Experimental conditions: 1 (20 mM, 0–45 min), 2 (20 mM, 0–45 min), 3
(25 mM, 0–45 min), 4 (25 mM, 0–45 min). The solution (100 mL)
containing enediyne, DNA (0.02 mg mL21, 30 mM per bp) in 20 mM
potassium phosphate buffer was incubated for 1 h at 25 uC. After
incubation, 10 samples (10 mL of each) in a Pyrex tubes (OD 0.11 in.)
were sealed tightly with Parafilm. The sample was placed at a
distance of 20 cm from 200 W Hg–Xe lamp (Spectra-Physics, Laser
& Photonics Oriel Instruments with long pass filter with 324 nm cut-on
wavelength) and cooled with the stream of air. The relative quantities
of the supercoiled, nicked, and linear DNA were calculated by the
image analyzer software Total/Lab (Nonlinear Dynamics Ltd., UK).
The amount of supercoiled DNA was multiplied by factor of 1.4
to account for reduced ethidium bromide intercalation into
supercoiled DNA.
8 For random Poisson distribution the average number of dsb per
molecule can be calculated from the equation: n2 5 1/[( fI + fII + fIII)/
fIII 2 1]. The average number of ss breaks can be determined from
the fraction of supercoiled DNA remaining after irradiation
fI 5 exp[2(n1 + n2)].
9 Using typical n1 values obtained from the experiment, one can calculate
the expected n2 from Freifelder–Trubo10 relation n2 5 n1 (2h + 1)/4L,
2
where h is the maximum number of unbroken base pairs between single-
strand breaks in opposite strands that produces a linear form (h 5 16),
L 5 the number of phosphodiester bonds per DNA strand (L 5 4361
for pB322). Under these assumptions, n1/n2 5 1057 (n1 5 0.5), 529
(n1 5 1), 352 (n1 5 1.5).
10 D. Freifelder and B. Trubo, Biopolymers, 1969, 7, 681.
11 L. F. Povirk, W. Wu¨bker and R. J. Steighner, Biochemistry, 1989, 28,
5808; M. J. Absalon, J. W. Kozarich and J. Stubbe, Biochemistry, 1995,
34, 2065; J. Stubbe, J. W. Kozarich, W. Wu and D. E. Vanderwall, Acc.
Chem. Res., 1996, 29, 322; K. Charles and L. F. Povirk, Chem. Res.
Toxicol., 1998, 11, 1580.
12 (a) K. D. Stewart and T. A. Gray, J. Phys. Org. Chem., 1992, 5, 461; (b)
B. F. Cain, B. C. Baguley and W. A. Denny, J. Med. Chem., 1978, 21,
658; (c) H.-P. Hsieh, J. G. Muller and C. J. Burrows, J. Am. Chem. Soc.,
1994, 116, 12077; (d) D. L. Boger, B. E. Fink, S. R. Brunette, W. C. Tse
and M. P. Hedrick, J. Am. Chem. Soc., 2001, 123, 12077.
1446 | Chem. Commun., 2005, 1444–1446
This journal is ß The Royal Society of Chemistry 2005