A clickable psoralen to directly quantify DNA interstrand crosslinking and repair

Article abstract of DOI:10.1016/j.bmc.2016.01.032

DNA interstrand crosslinks (ICLs) represent physical obstacles to advancing replication forks and transcription complexes. A range of ICL-inducing agents have successfully been incorporated into cancer therapeutics. While studies have adopted UVA-activated psoralens as model ICL-inducing agents for investigating ICL repair, direct detection of the lesion has often been tempered by tagging the psoralen scaffold with a relatively large reporter group that may perturb the biological activity of the parent psoralen. Here a minimally-modified psoralen probe was prepared featuring a small alkyne handle suitable for click chemistry. The psoralen probe, designated 8-propargyloxypsoralen (8-POP), can be activated by UVA in vitro to generate ICLs that are susceptible to post-labeling with an azide-tagged fluorescent reporter via a copper-catalyzed reaction. A modified alkaline comet assay demonstrated that UVA-activated 8-POP proficiently generated ICLs in cells. Cellular 8-POP-DNA lesions were amenable to click-mediated ligation to fluorescent reporters in situ, which permitted their detection and quantitation by fluorescence microscopy and flow cytometry. Small molecule DNA repair inhibitors to 8-POP-treated cells attenuated the removal of 8-POP-DNA lesions, validating 8-POP as an appropriate probe for investigating cellular ICL repair. The post-labeling strategy applied in this study is inexpensive, rapid and highly modular in nature with the potential for multiple applications in DNA repair studies.

Full text of DOI:10.1016/j.bmc.2016.01.032

Bioorganic & Medicinal Chemistry 24 (2016) 1071–1078
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
A clickable psoralen to directly quantify DNA interstrand
crosslinking and repair
⇑
Benjamin J. Evison, Marcelo L. Actis, Naoaki Fujii
Department of Chemical Biology and Therapeutics, St Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN 38105, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
DNA interstrand crosslinks (ICLs) represent physical obstacles to advancing replication forks and tran-
scription complexes. A range of ICL-inducing agents have successfully been incorporated into cancer ther-
apeutics. While studies have adopted UVA-activated psoralens as model ICL-inducing agents for
investigating ICL repair, direct detection of the lesion has often been tempered by tagging the psoralen
scaffold with a relatively large reporter group that may perturb the biological activity of the parent pso-
ralen. Here a minimally-modified psoralen probe was prepared featuring a small alkyne handle suitable
for click chemistry. The psoralen probe, designated 8-propargyloxypsoralen (8-POP), can be activated by
UVA in vitro to generate ICLs that are susceptible to post-labeling with an azide-tagged fluorescent repor-
ter via a copper-catalyzed reaction. A modified alkaline comet assay demonstrated that UVA-activated
Received 12 November 2015
Revised 6 January 2016
Accepted 17 January 2016
Available online 18 January 2016
Keywords:
Interstrand DNA crosslink
DNA repair
Click chemistry
Psoralen
8-POP proficiently generated ICLs in cells. Cellular 8-POP–DNA lesions were amenable to click-mediated
Imaging
Chemical screening
ligation to fluorescent reporters in situ, which permitted their detection and quantitation by fluorescence
microscopy and flow cytometry. Small molecule DNA repair inhibitors to 8-POP-treated cells attenuated
the removal of 8-POP–DNA lesions, validating 8-POP as an appropriate probe for investigating cellular ICL
repair. The post-labeling strategy applied in this study is inexpensive, rapid and highly modular in nature
with the potential for multiple applications in DNA repair studies.
Ó 2016 Elsevier Ltd. All rights reserved.
1
. Introduction
DNA interstrand crosslinks (ICLs) are specialized lesions that
represent a greater threat to viability relative to intrastrand cross-
links, covalent lesions that link bases within a single DNA strand.
Given their lethal nature, agents that induce DNA ICLs occupy an
important niche in the repertoire of agents presently available
covalently bridge opposing strands of duplex DNA. Such lesions
are generated by agents that occur either endogenously within
cells (e.g., aldehydes generated by lipid peroxidation) or exoge-
nously via cellular exposure to various environmental mutagens
and certain chemotherapeutic agents. Following their generation
within mammalian cells, ICLs are highly lethal to cell viability as
they resist DNA strand separation, a process typically required
for transcription and DNA replication to proceed. Moreover, these
lesions are challenging for mammalian cells to repair and typically
2
for the treatment of human cancers.
Most ICL-inducing agents generate intrastrand crosslinks and
monofunctional adducts, with ICLs only occupying a small fraction
1,2
1–3
of all adducts,
a feature that renders them difficult to detect in
biological samples. To overcome these limitations, studies have
adopted the use of psoralens as model ICL-inducing agents.
Psoralens are furocoumarins that exhibit a planar tricyclic ring
3
–5
structure that occur naturally in various plants.
Following inter-
calation within duplex DNA, psoralens can be photoactivated by
UVA to generate stable monofunctional adducts and ICLs in both
0
4
Abbreviations: CSK buffer, cytoskeleton buffer; DAPI, 4 ,6-diamidino-2-
naked DNA in vitro and in cellular DNA. The psoralens favor
phenylindole; DMSO, dimethyl sulfoxide; EDTA, ethylenediaminetetraacetic acid;
FBS, fetal bovine albumin; HEPES, 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesul-
fonic acid; HR, homologous recombination; ICL, interstrand crosslink; 8-MOP, 8-
methoxypsoralen; NMR, nuclear magnetic resonance; PARP, poly(ADP-ribose)
polymerase; PBS, phosphate buffered saline; PCR, polymerase chain reaction; 8-
POP, 8-propargyloxypsoralen; TAE, tris-acetate EDTA buffer; TLC, thin layer
chromatography; TCEP, tris(carboxyethyl)phosphine; TRITC, tetramethylrhodamine
isothiocyanate; UVA, ultraviolet A.
0
intercalation at 5 -TpA dinucleotides of duplex DNA and, following
activation by UVA light, covalently bond with a thymine nucleobase
in either strand (to generate a monoadduct) or both strands (to pro-
3
,6
duce an ICL). In comparison to other ICL-inducing agents, the pso-
ralens induce a significantly higher yield of ICLs relative to other
lesion types; up to 40% of 8-methoxypsoralen (8-MOP)-induced
adducts form ICLs following photoactivation by UVA.3 Psoralen-
induced ICLs within cells have successfully been detected via mass
⇑
Corresponding author. Tel.: +1 (901) 595 5854; fax: +1 (901) 595 5715.
E-mail address: naoaki.fujii@stjude.org (N. Fujii).
http://dx.doi.org/10.1016/j.bmc.2016.01.032
968-0896/Ó 2016 Elsevier Ltd. All rights reserved.
0

1
072
B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
spectrometry, high performance liquid chromatography and alka-
line comet assays, although these methods tend to be labor-inten-
sive, time-consuming, technically challenging and often require
using gradient mixtures of (A) water (0.05% TFA) and (B) methanol.
Low-resolution mass spectra (ESI) were collected on a Waters
Micromass ZQ in positive-ion mode. Flash chromatography was
performed on a Biotage Isolera chromatography system using Bio-
tage SNAP KP-SIL pre-packed columns, and the solvent mixture in
brackets was used as eluent. Nuclear magnetic resonance (NMR)
spectra were obtained on a Bruker Avance II NMR spectrometer
1,5–8
the use of specialized equipment and materials.
Direct detec-
tion of cellular 8-MOP–DNA adducts can also be achieved by anti-
bodies raised against 8-MOP–DNA photoadducts, however these
antibodies are no longer available.9 Other efforts have been direc-
ted at chemically tagging psoralens with various reporter groups
including biotin, digoxigenin and fluorescent labels that are amen-
,10
1
at 400 MHz for H NMR spectra. Chemical shifts (ppm) are reported
relative to the solvent peak. Signals are designated as follows: s,
singlet; d, doublet; dd, doublet of doublet; t, triplet; q, quadruplet;
m, multiplet. Coupling constants (J) are shown in Hertz.
1
1,12
able to detection in situ.
While these psoralen conjugates have
13,14
yielded valuable insights in ICL repair studies,
their tags are
rather large and bulky that can perturb the molecular nature of
native psoralens. Indeed, two of psoralen conjugates demonstrated
unusual cellular distribution patterns inconsistent with the native
2.2.1. 8-Hydroxypsoralen
A
two-neck flask was charged with 8-methoxypsoralen
1
2
psoralen, thereby prohibiting their experimental utility. Electro-
poration was required to permit sufficient cellular uptake of a bio-
(200 mg, 0.925 mmol) and dichloromethane (6 mL). The flask was
flushed with nitrogen and cooled to 0 °C. A solution of tribromob-
orane in dichloromethane (1 M, 3.7 mL, 3.7 mmol) was added
dropwise to the stirring reaction mixture. After 3 h, the reaction
mixture was added to water (30 mL) and extracted with ethyl acet-
ate (3 ꢀ 30 mL). The combined organic layers were washed with
saturated brine, dried over anhydrous sodium sulfate, filtered,
and concentrated. The resulting product was used in the following
step without further purification (173 mg, 93% yield). 1H NMR
1
1
tin-labeled psoralen, a condition unnecessary in the case of the
native psoralens.
These issues can be circumvented through post-labeling a min-
imally-modified psoralen probe with a reporter in situ. Perhaps the
best method presently available to achieve this is through the Cu
(
I)-catalyzed reaction between an alkyne- and azide-tagged pair
of molecules, a bioorthogonal conjugation so-called click chem-
1
5,16
istry.
Given their small steric nature, the introduction of either
6
(400 MHz, DMSO-d ) d 10.67 (s, 1H), 8.25–7.98 (m, 2H), 7.46
an alkyne- or azide-handle affords a readily detectable biological
probe without grossly affecting the structure of the parent com-
(s, 1H), 7.05 (s, 1H), 6.41 (d, J = 9.7 Hz, 1H).
1
6
pound. We report herein the preparation and development of a
psoralen probe with a small alkyne handle. The probe, termed 8-
propargyloxypsoralen (8-POP), can be activated by UVA to gener-
ate DNA ICLs and is readily detectable in situ by ligation with an
azide-tagged fluorescent reporter.
2.2.2. 8-Propargyloxypsoralen (8-POP)
Propargyl bromide (114 lL, 1.03 mmol) was added to a solution
of 8-hydroxypsoralen (173 mg, 0.856 mmol) and potassium car-
bonate (473 mg, 3.42 mmol) in acetone (4 mL). The reaction mix-
ture was heated to reflux, stirred for 3 h, then cooled and
concentrated. Water (20 mL) was added and extracted with ethyl
acetate (3 ꢀ 20 mL). The combined organic layers were washed
with saturated brine, dried over anhydrous sodium sulfate, filtered,
and concentrated. The crude product was purified by flash column
chromatography (Biotage Isolera, 25 g SNAP column, eluting with
2
2
. Materials and methods
.1. Materials
All chemicals were sourced from Sigma Chemical Co. (St. Louis,
hexanes/ethyl acetate, 0–60% gradient) to produce a white solid
1
MO) unless stated otherwise. Sodium ascorbate was from Acros
Organics (Geel, Belgium) while hydrogen peroxide (30%) was pur-
chased from Thermo Fisher Scientific (Waltham, MA). Electron
microscopy grade paraformaldehyde solution (16%) was obtained
from Electron Microscopy Sciences (Hatfield, PA). NdeI was pur-
chased from New England Biolabs (Beverly, MA). SYBR green I stain
and Alexa fluor azide 488 reagent (US Patents US20080050731 A1)
was from Life Technologies (Carlsbad, CA). Azide-flour 545 reagent
(40 mg, 19% yield). H NMR (400 MHz, DMSO-d
6
) d 8.25–8.09 (m,
2H), 7.74 (s, 1H), 7.11 (d, J = 2.2 Hz, 1H), 6.45 (d, J = 9.6 Hz, 1H),
5.14 (d, J = 2.4 Hz, 2H), 3.56 (t, J = 2.5 Hz, 1H). MS (ESI) m/z, [M
+H] calculated for C14H O , 241.05; observed 241.00.
9 4
2.3. In vitro crosslinking assay
Supercoiled pEGFP-PCNA plasmid DNA was initially linearized
by digestion with NdeI using standard procedures and then puri-
fied using a QIAquick PCR Purification Kit. Linearized plasmid
(
5-carboxytetramethylrhodamine-azide) was from Sigma Chemical
Co. Vectashield mounting medium was from Vector Laboratories
Burlingame, CA). Propidium iodide was obtained from Roche
(
(40 ng/lL) was subsequently reacted with 31 lM 8-POP or DMSO
Life Sciences (Indianapolis, IA). Micro Bio-Spin 6 chromatography
columns were purchased from Bio-Rad (Hercules, CA). T2AA
vehicle in 35 mM HEPES, pH 7.4 and exposed to UVA light centered
at 365 nm at room temperature. Irradiation was delivered using a
UVP model UVGL-58 Handheld UV Lamp (Upland, CA) from a dis-
tance of 5.08 cm for 0–30 min as indicated. Samples were dena-
tured at 80 °C for 5 min in strand separation buffer (30% DMSO,
1 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol) and
subsequently loaded onto a 0.8% agarose gel in 1 ꢀ TAE buffer.
Control samples that were not thermally denatured were run in
parallel and were representative of double stranded DNA. DNA
was fractionated by electrophoresis for 2–3 h at 80 V and then
(
(S)-4-(4-(2-amino-3-hydroxypropyl)-2,6-diiodophenoxy)phenol
17
hydrochloride) was prepared as previously described. Olaparib
was obtained from Selleck Chemicals (Houston, TX). A QIAquick
PCR Purification Kit was from Qiagen (Valencia, CA). A pEGFP-PCNA
plasmid was originally obtained from Daniel Gerlich18 via Addgene
(
Cambridge, MA).
2
.2. Chemical synthesis
stained with 0.5 lg/mL ethidium bromide. Agarose gels were
destained and then visualized and photographed under UV illumi-
All commercial reagents were used without further purification.
Reactions requiring the exclusion of air were carried out under an
atmosphere of dry nitrogen in oven dried glassware. All reactions
were monitored by thin-layer chromatography (TLC) carried out
on EMD Chemicals silica gel 60-F 254 coated glass plates and
visualized using UV light (254 nm). Analysis by LC–MS was
performed by using an XBridge C18 column run at 1 mL/min and
nation. Quantitation was performed using ImageJ software (NIH).
2.4. In vitro conjugation assay
Supercoiled pEGFP-PCNA plasmid DNA (100 ng/
lL) was initially
incubated with 50 M 8-POP in 35 mM HEPES, pH 7.4 and then
l

B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
1073
exposed to UVA irradiation for 30 min at room temperature as
described above. To each sample, the following was then added:
ice to remove non-chromatin bound protein and then fixed with
4% paraformaldehyde in PBS for 12 min at room temperature. Each
sample was then washed with 3% FBS in PBS and then incubated
2
0
l
M Azide-flour 545, 1 mM copper(II) sulfate and 0–5 mM
sodium ascorbate or TCEP, and the samples incubated for a further
0 min at room temperature. Unreacted fluor azide was removed
with a click reaction mixture (50 lM Azide-flour 545, 1 mM cop-
9
per(II) sulfate, 100 mM sodium ascorbate, 100 mM Tris–HCl pH
8.0) at room temperature for 30 min. Cells were then washed
extensively using 3 ꢀ 15 min 3% FBS in PBS rinses with mild
agitation and then mounted onto glass slides in VectaShield
by gel filtration using Micro Bio-Spin 6 chromatography columns.
The contents of each sample were then fractionated by
electrophoresis through a 0.8% agarose gel in 1ꢀ TAE buffer for
9
0 min at 100 V. Agarose gels were first scanned and visualized
containing 1 lg/mL DAPI. Stained samples were captured and ana-
on an ImageQuant LAS 4000 instrument (GE Healthcare, Piscat-
away, NJ) operating in the green epi-illumination mode (520 nm
peak excitation wavelength and a 575DF20 emission filter). Gels
were subsequently stained with SYBR green I and then visualized
again on the ImageQuant LAS 4000 imager using the blue
epi-illumination mode (460 nm peak excitation wavelength and a
Y515 Di emission filter).
lyzed on a Nikon E800 microscope equipped with a 20ꢀ/0.5 NA
Plan Apo objective and Nikon NIS elements software (Nikon, Tokyo,
Japan).
2.8. Analysis of 8-POP–DNA adduct repair by flow cytometry
HeLa cells (330,000 per dish) were seeded into 6 cm culture
dishes and incubated overnight. The following day, samples were
2
.5. Cell culture
treated with 50
UVA irradiated as indicated above. After irradiation, cells were
released into media containing DMSO vehicle or 40 M of T2AA,
lM 8-POP or DMSO vehicle, PBS washed and
HeLa and HT1080 cells were obtained from American Type
l
Culture Collection (ATCC, Manassas, VA) and maintained at
sub-confluent levels in Dulbecco’s modified Eagle’s medium
olaparib, or aphidicolin for a total of 24 h. At designated time
points, samples were harvested by trypsinization and then incu-
bated in hypotonic swelling buffer (10 mM HEPES pH 7.5,
(
Mediatech, Inc., Manassas, VA) supplemented with 10% fetal
bovine serum (Thermo Scientific, Waltham, MD) at 37°C in a
humidified atmosphere containing 5% CO
2
1.5 mM MgCl , 10 mM KCl) for 10 min on ice. Cells were then
2
.
hypotonically lysed by passing each sample through a 27 gauge
needle with a total of 3–4 repetitions. The integrity of samples
was monitored by visual inspection using light microscopy after
trypan blue staining. Nuclei were subsequently pelleted by cen-
trifugation and then fixed with 4% paraformaldehyde in PBS for
15 min at room temperature. Samples were subjected to a single
wash with 3% FBS in PBS and stored at 4 °C in a buffer of the same
composition until the day of analysis. On the day of analysis, nuclei
2
.6. Alkaline comet assay
HeLa cells (100,000 per well) were seeded into 6-well cluster
plates and allowed to settle overnight. Cells were initially treated
with DMSO vehicle, 50 M 8-MOP or 8-POP for 60 min and then
thoroughly washed with PBS prior to irradiation with UVA light
l
2
(
(
100 J/m ) at 365 nm using a Stratalinker UV Crosslinker 1800
Stratagene, La Jolla, CA). Following UVA exposure, cell samples
were initially incubated in a click reaction buffer (50 lM Alexa
fluor 488 azide, 1 mM copper(II) sulfate, 100 mM sodium ascor-
bate, 100 mM Tris–HCl pH 8.0) for 30 min at room temperature
and then washed extensively using 3 ꢀ 3% FBS in PBS rinses to
were treated with PBS as a control or 200
2 2
lM H O for 15 min at
3
7 °C before harvest by trypsinization. Detached samples were
subsequently resuspended in molten 1% Type VII low gelling tem-
perature agarose and then allowed to set on glass slides pre-coated
with agarose. Cells were then lysed by bathing slides in ice-cold
remove unbound reporter. Samples were then stained with 5 lg/
mL propidium iodide in 3% FBS in PBS and then analyzed using
an LSR Fortessa flow cytometer (BD Biosciences, San Jose, CA)
equipped with a 488 nm laser line which was used for sample exci-
tation. Green and red fluorescence was collected using the FL1 and
FL3 channels, respectively. Ten thousand events were collected for
each sample and gated using forward- versus side-scatter to elim-
inate cell debris and doublets. The data were analyzed using FlowJo
software version 7.6.5 (Tree Star, San Carlos, CA).
lysis buffer (100 mM Na
1
4
2
EDTA, 2.5 M NaCl, 10 mM Tris–HCl (pH
0.5), 1% Triton X-100) for 60 min and then subjected to
ꢀ 15 min washes with ice-cold MilliQ H O. Each slide was then
2
submerged in alkali electrophoresis buffer (300 mM NaOH, 1 mM
Na EDTA) for 60 min and then electrophoresed at 30 V for
0 min at 4 °C. Samples were neutralized by the addition of neu-
2
3
tralization buffer (500 mM Tris–HCl pH 7.5) for 10 min and then
allowed to dry overnight at ambient temperature. Comets were
stained with SYBR green I for 10 min and then destained using
2.9. Analysis of 8-POP–DNA adduct repair by chromatin foci
quantitation
3
ꢀ MilliQ H
E800 microscope equipped with a 20ꢀ/0.5 NA Plan Apo objective
Nikon, Tokyo, Japan) and the level of DNA damage assessed using
2
O washes. Samples were visualized using a Nikon
HT1080 cells were seeded (330,000 per well) onto poly-L-
(
lysine-treated coverslips in 6-well cluster plates and allowed to
attach overnight. Samples were treated with 8-POP, rinsed with
PBS and irradiated with UVA as described above for the in situ
detection of 8-POP adducts by microscopy. Following irradiation,
cells were released into media containing DMSO vehicle or 63–
125 nM gemcitabine for 19 h. At designated time points following
release, samples were processed as detailed for in situ detection of
8-POP lesions by microscopy. Cells were subsequently analyzed on
a TE2000 microscope equipped with a C2 confocal and a 40ꢀ/0.8
NA Plan Apo objective (Nikon, Tokyo, Japan). Images were captured
and processed using Nikon NIS-elements software. The software
was also used to calculate the mean red fluorescence intensity of
each nucleus using the DAPI staining pattern to define to perimeter
of the nucleus. At least 68 nuclei were scored per sample.
19
the scoring system detailed by Collins. At least 100 comets were
scored per slide.
2
.7. In situ detection of 8-POP–DNA adducts in cells by
microscopy
HeLa cells were initially seeded onto coverslips in 6-well cluster
plates at 330,000 cells per well and allowed to attach overnight.
The following day, cells were treated with 50 M 8-POP or DMSO
vehicle for 60 min, washed with PBS and then exposed to UVA irra-
diation as described above. Cell samples were then pre-extracted
with ice-cold CSK buffer (100 mM NaCl, 10 mM HEPES pH 7.8,
l
3
2
mM MgCl , 300 mM sucrose, 0.5% Triton X-100) for 2 min on

1
074
B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
3
3
. Results
.1. A rationale for the design of 8-POP
Upon NMR solution structures of psoralen-induced DNA mono-
adducts and ICLs, it was revealed that the 8-position of each pso-
ralen lesion was located in the minor groove of the duplex and
2
0
was solvent exposed. Accordingly, it was surmised that the 8-
position may provide a suitable site for the introduction of an
alkyne appendage that may be accessible for conjugation with an
azide-tagged reporter (Fig. 1A). To this end, the psoralen 8-MOP
was adopted as a starting material as the compound was readily
amenable to an initial demethylation at the 8-position to yield 8-
hydroxypsoralen. The 8-hydroxypsoralen intermediate was subse-
quently functionalized by a propargyl group to yield 8-POP
(Fig. 1B). 8-POP does not decompose or self-crosslink under UVA
dose being used for crosslinking DNA (Fig. S1).
3
.2. 8-POP generates stable DNA ICLs in vitro following
Figure 2. 8-POP can generate DNA ICLs following activation by UVA light in vitro.
A) Linearized pEGFP-PCNA plasmid DNA was initially reacted with 0 (NT) or 31
-POP and exposed to UVA light for 0–30 min as indicated. Each sample was then
subjected to denaturation and electrophoresis as described in the Section 2.
Controls that were not thermally denatured are denoted by and are
representative of double stranded linearized plasmid DNA. (B) The fraction of
double stranded DNA stabilized by 8-POP represented in (A) was quantitated and is
expressed as a function of time of UVA exposure.
activation by UVA
(
8
lM
It was a reasonable concern that the alkyne handle of 8-POP
may impair DNA binding and subsequent UVA-mediated adduct
formation. To address this concern, an in vitro crosslinking assay
was applied. In the absence of 8-POP (NT, lanes 0–30), linearized
pEGFP-PCNA plasmid DNA migrates more rapidly through agarose
gels relative to a double stranded control (NT, lane C, Fig. 2A).
These results suggest that the fragment was successfully dena-
tured into its single stranded components and that UVA light alone
was insufficient for duplex stabilization. Upon the introduction of
a
C
3.3. 8-POP–DNA adducts can efficiently be labeled with an
azide-tagged fluorescent reporter in vitro via click chemistry
8
-POP (lanes 0–30, Fig. 2A), there was a clear UVA dose-dependent
Next, to establish if 8-POP ICLs and monofunctional DNA
adducts were susceptible to conjugation with a reporter, a super-
coiled plasmid DNA was initially reacted with 8-POP and UVA to
generate 8-POP–DNA adducts. Following ligation with an azide-
tagged fluorophore under reducing agent to generate the requisite
Cu(I) species, each DNA sample was fractionated by electrophore-
sis and then analyzed for successful conjugation to the reporter by
scanning in-gel for red fluorescence. Figure 3 (top panel) exhibits
the staining pattern of the red fluorescent reporter along with
the same gel post-stained with SYBR green (bottom panel) repre-
senting the total DNA content of each sample. Figure 3 (bottom
panel, lanes 2–7) reveals that the application of increasing doses
of sodium ascorbate coincides with degradation of plasmid DNA
increase in the fraction of double stranded DNA, a result that is
consistent with the generation of a duplex-stabilizing ICL that pre-
cludes complete denaturation of linearized plasmid. Quantitation
of these bands revealed that approximately 50% of linearized plas-
mid DNA was stabilized following 30 min of UVA exposure
(Fig. 2B). Notably, 8-POP failed to stabilize double stranded DNA
in the absence of UVA (8-POP, lane 0, Fig. 2A), suggesting that
intercalated non-covalent 8-POP was insufficient to stabilize the
duplex without UVA. Collectively, these results strongly indicate
that 8-POP generated stable covalent DNA ICLs in vitro following
activation by UVA light, much like 8-MOP under similar assay
2
1
conditions.
Figure 1. (A) A schematic representation of the post-labeling strategy adopted in this report. In this scheme, 8-POP, an analog of the clinically-applied 8-MOP, is activated by
UVA to generate covalent 8-POP–DNA adducts. The alkyne-appended 8-POP–DNA lesion may be post-labeled via ligation with an azide-tagged fluorescent reporter
(
3
designated here as N -linker-FT) using a Cu(I)-catalyzed dipolar cycloaddition. The level of 8-POP–DNA lesions can readily be quantified by measuring fluorescence. (B) The
synthetic route used to produce 8-POP.

B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
1075
Figure 4. 8-POP generates DNA ICLs in HeLa cells following UVA exposure. HeLa
cells were initially treated with DMSO vehicle, 50 M 8-MOP or 8-POP as indicated
for 60 min and then exposed to 0 or 100 J/m2 UVA. Following irradiation, samples
were subjected to treatment with 200 M H or PBS as a control for 15 min at
7 °C. Samples were then processed and assayed for their levels of DNA damage
l
l
2 2
O
3
using an alkaline comet assay as detailed in the Section 2. Each column represents
the mean level of DNA damage for each sample as reflected by the visual score in
arbitrary units ± SEM. At least 100 comets were scored per sample.
breakage as reflected by its high visual comet score (sample 4,
Fig. 4) in the absence of UVA. Neither psoralen alone (samples 5
Figure 3. The ligation of 8-POP-modified plasmid DNA with an azide-tagged
fluorescent reporter via click chemistry. Supercoiled pEGFP-PCNA plasmid DNA was
initially reacted with 50 lM 8-POP and UVA for 30 min. The contents of each
reaction were then subjected to conjugation with Azide-flour 545 via Cu(I)-
catalyzed click chemistry using 0–5 mM sodium ascorbate or TCEP. Samples were
fractionated through agarose gels by electrophoresis and the gels scanned for Alexa
fluorescence (top panel). Gels were then counterstained with SYBR green (bottom
panel) and scanned to reveal the total DNA content of each sample. M denotes
molecular weight markers.
and 6, Fig. 4) influenced the level of damage introduced by H
sample 4, Fig. 4), indicating that non-activated, intercalated 8-
MOP or 8-POP was insufficient to crosslink DNA. None of the
UVA-treated controls in the absence of H (samples 7–9, Fig. 4)
2 2
O
(
2 2
O
induced strand breakage relative to their non-irradiated counter-
parts, suggesting that: (a) UVA alone induced no detectable dam-
age and (b) UVA-activated 8-MOP and 8-POP–DNA lesions were
not processed by the cell to generate DNA strand breakage in these
into smaller fragments as indicated by their increased
electrophoretic mobility relative to intact plasmid (bottom panel,
2 2
conditions. While the combination of H O and UVA treatment
induced an exceptionally high level of strand breakage (sample
10, Fig. 4), the introduction of 8-MOP to cells in parallel yielded a
large decrease of ꢁ70% in the level of damage (sample 11, Fig. 4).
Since non-activated, intercalated 8-MOP failed to influence the
2
2,23
lane 1, Fig. 3), a feature that has been noted previously.
The
red fluorescent staining profile (top panel, lanes 2–7, Fig. 3) exhib-
ited essentially the same staining pattern as the DNA content, indi-
cating that the red fluorescent reporter was conjugated to DNA
containing covalent 8-POP–DNA adducts. Ligation of the red fluo-
rescent reporter to 8-POP–DNA adducts did not proceed in the
absence of copper catalyst (top panel, lane 1, Fig. 3). Sodium ascor-
bate was substituted for TCEP in a parallel series of reactions (lanes
2 2
level of damage generated by H O (sample 5, Fig. 4), the reversal
of damage by activated 8-MOP reflects the formation of duplex-
6
stabilizing 8-MOP ICLs. An equimolar dose of UVA-activated
8-POP yielded an even greater, >99% reversal in the amount of
H O -induced damage (sample 12, Fig. 4), indicating that UVA-
2 2
9
–14, Fig. 3) that also yielded an overlapping red fluorescent stain-
activated 8-POP was more proficient at ICL generation than 8-MOP.
ing profile (top panel) with total DNA content (bottom panel). The
application of TCEP generated no detectable degradation of
plasmid DNA (bottom panel, lanes 8–14, Fig. 3) unlike the use of
sodium ascorbate. However, ligation of the fluorescent reporter
did not smoothly proceed at 5 mM TCEP (top panel, lane 14,
Fig. 3). Despite this anomaly, these results collectively suggest that
3.5. 8-POP–DNA lesions can be detected within HeLa cells in situ
Having established that the post-labeling protocol was effective
in cell-free systems, it was necessary to test whether the strategy
could be applied to detect 8-POP–DNA lesions in cells. In a prelim-
inary experiment, sodium ascorbate visualized 8-POP–DNA lesions
in cells better than TCEP did (not shown), although TCEP was supe-
rior for maintaining DNA integrity than ascorbate in vitro (Fig. 3).
Since DNA integrity is not an absolute requirement in cellular
imaging, we used ascorbate afterwards. Cultured HeLa cells were
initially exposed to 8-POP or DMSO vehicle and then irradiated
with or without UVA. Whole cell samples were immediately pre-
extracted to remove non-chromatin bound protein, fixed and then
subjected to a click reaction mixture to enable ligation of the azide-
tagged red fluorescent reporter to 8-POP–DNA adducts in situ.
Samples were co-stained with the nuclear dye DAPI, mounted
and then analyzed by epi-fluorescence microscopy. Representative
images of samples treated with 8-POP are indicated in Figure 5.
Control samples exposed to DMSO vehicle consistently
demonstrated very low levels of red fluorescence (data not shown),
8
-POP–DNA adducts can be successfully post-labeled with a
reporter.
3
.4. Relative to 8-MOP, 8-POP more readily induces DNA ICLs in
HeLa cells following UVA irradiation
A modified alkaline comet assay has previously been applied in
the detection of 8-MOP-induced DNA ICLs in human
6
keratinocytes. Similarly, the assay was deployed to test whether
8
-POP was proficient at ICL induction within cells. HeLa cells were
treated by 8-MOP or 8-POP, UVA irradiated and then exposed to
to generate random single strand breaks. In the absence of
UVA and H , neither 8-MOP nor 8-POP generated a visual comet
score beyond the vehicle-treated control (samples 1–3, Fig. 4).
exposure induced a relatively high level of DNA strand
2 2
H O
2 2
O
2 2
H O

1
076
B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
Figure 5. 8-POP–DNA lesions can be detected within HeLa cells in situ. Represen-
tative microscopy images of HeLa cells that were treated with 50 M 8-POP for
0 min and then irradiated with 0 J/m (top panels) or 100 J/m (bottom panels).
Samples were subsequently pre-extracted, fixed and subjected to post-labeling
with an azide-tagged red fluorescent reporter as described in the Section 2. Samples
were counterstained with DAPI, mounted on glass slides and then visualized by epi-
fluorescence microscopy using DAPI and TRITC filters to define the nucleus and
reveal the red 8-POP adduct staining pattern, respectively.
l
2
2
6
Figure 6. The post-labeling strategy of 8-POP can be applied to measure the cellular
repair of 8-POP–DNA adducts and its inhibition by small molecules. (A and B) HeLa
cells were exposed to 50
UVA. Cells were subsequently released into media containing DMSO vehicle or
M aphidicolin, olaparib or T2AA for 24 h. All samples were then hypotonically
l
M 8-POP for 60 min and then irradiated with 100 J/m2
40 l
lysed, fixed, subjected to post-labeling with an azide-tagged Alexa 488 reporter by
click chemistry, counterstained with propidium iodide and analyzed by flow
cytometry as described in the Section 2. (A) A representative histogram displaying
the relative levels of 8-POP–DNA adducts as reflected by the Alexa 488 signal
immediately following UVA irradiation (‘start’, solid black histogram) and 24 h after
indicating that the fluorescent reporter was not non-specifically
ligated to DNA (or other biomolecules) by the click reaction in
the absence of 8-POP. Similarly, samples exposed to 8-POP but
not UVA irradiation showed comparably low levels of fluorescence
release into media containing DMSO as a control (light gray histogram) or 40 lM
T2AA (medium gray histogram). (B) A quantitative representation of the histogram
displaying the median ICL (Alexa 488) signal as a function of treatment. (C) HT1080
cells were initially treated with 8-POP and UVA as described for (A and B) and then
released into media containing DMSO vehicle or 62.5–125 nM gemcitabine for 19 h.
Cells were subsequently pre-extracted, fixed and post-labeled with an azide-tagged
red fluorescent reporter using click chemistry as detailed in the Section 2. The mean
nuclear intensity of the red fluorescent reporter was quantitated by confocal
microscopy and is plotted for each cells counted for analysis. ‘Start’ indicates the
mean nuclear fluorescence intensity immediately following UVA irradiation. Error
bars indicate the standard deviation. Statistics were analyzed by a Student’s t test
(
top right panel, Fig. 5), evidence that suggests non-covalently
bound 8-POP was largely removed by the pre-extraction proce-
dure. Upon the UVA application to 8-POP-treated cells, there was
a prominent increase in red fluorescent signal (bottom right panel,
Fig. 5), indicating that UVA exposure had activated 8-POP to gener-
ate covalent DNA adducts and that the post-labeling procedure
was effective in situ. Co-staining with the nuclear dye DAPI (bot-
tom left panel, Fig. 5) revealed that 8-POP lesions were predomi-
nantly localized within the nucleus with relatively little detected
in cytoplasmic regions. These results suggest that the post-labeling
strategy is effective for in situ detection of UVA-activated covalent
(
two-tailed, unpaired). Numbers of cells counted for the analysis are indicated at
the top for each data.
intensity relative to the DMSO control (medium gray histogram
vs light grey histogram, Fig. 6A). The inclusion of aphidicolin, an
8
-POP DNA lesions.
2
5
inhibitor of B-family DNA polymerases, or olaparib, a poly(ADP-
2
6
3
.6. The post-labeling strategy can be applied to measure the
ribose) polymerase (PARP) inhibitor, also impaired the reduction
in fluorescence intensity when compared to the DMSO control
(Fig. 6B). Given the reputation of these three agents as DNA repair
inhibitors, it is likely that the loss of fluorescence with time most
likely reflects the repair of 8-POP DNA adducts. The results suggest
that the assay may be exploited to identify or further validate small
molecule inhibitors of DNA repair.
cellular repair of 8-POP–DNA adducts and its inhibition by
small molecules
In an effort to validate the post-labeling strategy for assaying
chemical compounds, the cellular repair of 8-POP–DNA adducts
was evaluated. HeLa cells were initially treated with 8-POP and
then exposed to UVA. Following irradiation, cells were released into
media containing DMSO vehicle or a small molecule compound. At
designated time points, samples were subsequently assayed for
their levels of 8-POP DNA adducts by applying the post-labeling
strategy in conjunction with flow cytometry. Figure 6A and B
demonstrates that 8-POP DNA adducts were readily detectable by
flow cytometry immediately following UVA exposure. Upon their
release into media containing DMSO vehicle and 24 h later, there
was a notable shift in Alexa fluorescence intensity of the cell popu-
lation (Fig. 6A) with the population demonstrating a reduction of
the median fluorescence intensity (Fig. 6B). The decrease in Alexa
fluorescence signal noted here may reflect the cellular repair of 8-
POP DNA adducts with time. The introduction of T2AA, an inhibitor
A
ribonucleotide reductase inhibitor gemcitabine potently
inhibited repair of an ICL-containing reporter plasmid in HT1080
2
7
cells. The development of 8-POP as a potential probe for evaluat-
ing cellular ICL repair provided an opportunity to independently
verify whether gemcitabine inhibits ICL repair within genomic
DNA of this cell line. In addressing this, HT1080 cells were treated
with 8-POP, irradiated with UVA and then released into media con-
taining DMSO vehicle or gemcitabine. At designated time points,
samples were processed, clicked and evaluated for their levels of
8-POP–DNA adducts by fluorescence microscopy. In the absence
of gemcitabine, the mean nuclear fluorescence intensity signifi-
cantly decreased 19 h later (Fig. 6C), a result consistent with the
cellular repair of 8-POP–DNA lesions in this time interval. Under
identical conditions, gemcitabine dose-dependently attenuated
2
4
of ICL repair, to cells attenuated the decrease in fluorescence

B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
1077
the reduction in fluorescence, indicating that gemcitabine inhib-
ited the repair of the 8-POP–DNA lesion at this dose and time.
The introduction of 8-POP as a click-ready tool therefore represents
a valuable new addition to these existing platinum-based agents as
their differing structures are likely to elicit distinct cellular DNA
damage responses.
It is worth highlighting that the in situ detection of 8-POP DNA
lesions by either microscopy (Figs. 5 and 6C) or flow cytometry
4
. Discussion
In the in vitro post-labeling protocol, the application of sodium
(Fig. 6A and B) affords numerous advantages that are contingent
ascorbate to maintain copper in its requisite Cu(I) state was asso-
ciated with a profound degradation of plasmid DNA (Fig. 3). Both
DNA and protein are susceptible to cleavage induced by the com-
bination of ascorbate and Cu(II) ions in a series of redox-based
on the method of analysis. In preparing samples for microscopy,
a pre-extraction procedure using CSK buffer was adopted to facili-
tate the removal of cellular constituents (such as soluble proteins)
while permitting the retention of nuclear chromatin on a solid
2
2,23
reactions.
The substitution of sodium ascorbate for TCEP as a
4
1
slide support. Given that pre-extraction is widely applied in
immunofluorescence studies and that the procedure is readily
compatible with the detection of 8-POP DNA lesions (Fig. 5), it is
entirely conceivable that the two methods can be coupled to mon-
itor the spatiotemporal dynamics of specific DNA damage response
reducing agent generated an increase in the yield of clicked pro-
duct with little detectable damage to plasmid DNA (Fig. 3). Inter-
estingly, there was a notable shift in the mobility of plasmid
DNA from a supercoiled form to an apparently larger, less mobile
species as the TCEP dose increased (lanes 9–13, Fig. 3). Most likely,
this less mobile species is an open circular form of the plasmid that
has lost its superhelicity. It has been established that the transition
of plasmids from supercoiled to open circular states can occur
4
2
(
8
DDR) proteins to sites of 8-POP–DNA damage. Localized tracts of
-POP–DNA lesions can potentially be generated by the application
of a 364 nm laser to drug-treated cells, as highlighted originally by
1
2
Thazhathveetil et al. In the context of detecting the lesion by flow
cytometry (Fig. 6A and B), co-staining samples for DNA content
using a DNA-specific dye will enable ready quantitation of the
lesion within specific stages of the cell cycle. Therefore, by combin-
ing microscopic and flow cytometric methods, the assay could be
useful to determine role of the recruited DDR proteins for ICL
repair in the context of DNA replication.
via reactive oxygen species-mediated ‘nicking’ of the DNA back-
bone2
8–30
and the Cu(II)/TCEP combination may have generated
nicks (Fig. 3). Although it is unclear why TCEP, but not ascorbate,
induces little damage in combination with Cu(II), the use of TCEP
may be prudent in contexts where it is necessary to maintain the
structural integrity of the clicked DNA product.
Extending on these in vitro studies, it was subsequently demon-
strated that 8-POP not only formed UVA-dependent DNA ICLs
within HeLa cells, but the compound strikingly yielded more of
these lesions relative to 8-MOP at equimolar doses (Fig. 4). It is
conceivable that the terminal alkyne group of 8-POP participates
in the crosslinking reaction with DNA following UVA activation.
However there is little, if any, precedent for this in the literature
at large. There is some evidence in favor of UVA-dependent thiol-
In summary, we have developed an alkyne-appended psoralen
probe termed 8-POP that can be activated by UVA light to generate
DNA ICLs in vitro and within cells. 8-POP–DNA lesions were amen-
able to conjugation with azide-tagged fluorescent reporters via
click chemistry, a feature that readily permits their quantitation
by microscopy and flow cytometry. The post-labeling strategy
applied in this study is rapid and highly modular in nature with
the potential for multiple applications in DNA repair studies.
yne covalent coupling of alkyne-tagged compounds with thiol-
bearing cysteine residues,3
1,32
however there are no readily avail-
Author contributions
able reports for an equivalent UVA-dependent direct coupling of
alkyne functional groups with DNA. It is worth noting that, despite
the requirement for UVA in the induction of 8-POP–DNA ICLs, the
alkyne handle remains amenable to conjugation with an azide-
tagged fluorescent reporter (Figs. 3, 5 and 6), indicating that the
alkyne functional group remained intact. Perhaps it is more likely
that the alkyne tag of 8-POP engenders the psoralen scaffold with
other features that favor greater ICL formation relative to 8-MOP
B.J.E. designed the research, performed all experiments except
production of 8-POP, analyzed the data, and wrote the manuscript.
MLA produced 8-POP. N.F. conducted the project, analyzed the
data, and edited the manuscript. All authors have approved the
final version of the manuscript.
Funding sources
(Fig. 4). For instance, initial non-covalent intercalative binding of
the psoralens with DNA is significantly influenced by differences
in the chemical groups that decorate the central tri-cyclic furo-
The American Lebanese Syrian Associated Charities (ALSAC) and
American Cancer Society Research Scholar Grant #RSG CDD-
coumarin ring system.4
,33–35
Although it is beyond the scope of
1
20969 (to N.F.). Microscopy for the comet assay and the detection
the current study, it would be fascinating to establish a mechanis-
tic basis for the propensity of UVA-activated 8-POP to generate a
greater fraction of DNA ICLs relative to 8-MOP and whether this
feature may be exploited in a therapeutic context. In any case, this
feature is especially favorable in the context of the present study as
of 8-POP–DNA lesions was performed at the Cell & Tissue Imaging
Center, supported by ALSAC and NCI P30 CA021765-34.
Acknowledgments
8
-POP is intended as a tool for use in ICL repair.
The development of 8-POP and its application as a tool for ICL
We thank Jennifer Peters for microscopic foci quantification and
Akira Inoue for insightful discussions and advise.
induction and repair within cells now marks the addition of the
compound to an expanding list of clickable DNA- and RNA-directed
agents originally designed for similar purposes. 8-POP now joins
picazoplatin
compound
tools to better understand the basis of the biological activity of
their parent compounds. As analogs of cisplatin, both picazoplatin
and the platinum-acridine hybrid likely form a majority of DNA
monoadducts that are single stranded in nature (e.g., intrastrand
crosslinks) as opposed to the psoralens, a group of compounds
that typically generate a much greater fraction of DNA ICLs.
Supplementary data
3
6,37
and an azide-appended platinum-acridine hybrid
as DNA-reactive, click-ready agents designed as
3
8,39
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2016.01.032.
References and notes
1
.
Clauson, C.; Scharer, O. D.; Niedernhofer, L. Cold Spring Harb. Perspect. Biol.
013, 5, a012732.
2. Deans, A. J.; West, S. C. Nat. Rev. Cancer 2011, 11, 467.
1
3
2
3
,13,40

1
078
B. J. Evison et al. / Bioorg. Med. Chem. 24 (2016) 1071–1078
23. Wang, Y.; Van Ness, B. Nucleic Acids Res. 1989, 17, 6915.
3
.
.
.
.
.
.
.
Dronkert, M. L.; Kanaar, R. Mutat. Res. 2001, 486, 217.
Cimino, G. D. et al Annu. Rev. Biochem. 1985, 54, 1151.
Lai, C. et al Anal. Chem. 2008, 80, 8790.
Wu, J. H. et al Mutagenesis 2009, 24, 173.
Liu, S.; Wang, Y. Anal. Chem. 2013, 85, 6732.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
24. Inoue, A. et al J. Biol. Chem. 2014, 289, 7109.
25. Baranovskiy, A. G. et al Nucleic Acids Res. 2014, 42, 14013.
26. Lee, J. M.; Ledermann, J. A.; Kohn, E. C. Ann. Oncol. 2014, 25, 32.
27. Fujii, N. et al Bioorg. Med. Chem. 2015, 23, 6912.
28. Hoffman, M. et al Chem. Res. Toxicol. 2012, 25, 1609.
29. Li, Y.; Trush, M. A. Carcinogenesis 1993, 14, 1303.
30. Lewis, J. G.; Stewart, W.; Adams, D. O. Cancer Res. 1988, 48, 4762.
31. Lo Conte, M. et al J. Org. Chem. 2010, 75, 4644.
Bevilacqua, P. M.; Edelson, R. L.; Gasparro, F. P. J. Invest. Dermatol. 1991, 97, 151.
Santella, R. M. et al Nucleic Acids Res. 1985, 13, 2533.
0. Yang, X. Y.; DeLeo, V.; Santella, R. M. Cancer Res. 1987, 47, 2451.
1. Tsuchida, K.; Komatsu, K. Cancer Sci. 2008, 99, 2238.
2. Thazhathveetil, A. K. et al Bioconjugate Chem. 2007, 18, 431.
3. Huang, J. et al Mol. Cell. 2013, 52, 434.
32. Minozzi, M. et al J. Org. Chem. 2011, 76, 450.
33. Gupta, M.; Ali, R. J. Biochem. 1984, 95, 1253.
4. Muniandy, P. A. et al J. Biol. Chem. 2009, 284, 27908.
5. Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 2003, 125, 4686.
6. Speers, A. E.; Cravatt, B. F. Chem. Biol. 2004, 11, 535.
7. Punchihewa, C. et al J. Biol. Chem. 2012, 287, 14289.
8. Steigemann, P. et al Cell 2009, 136, 473.
34. Isaacs, S. T.; Wiesehahn, G.; Hallick, L. M. Natl. Cancer Inst. Monogr. 1984, 66, 21.
35. Wiesehahn, G.; Hearst, J. E. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 2703.
36. White, J. D. et al J. Am. Chem. Soc. 2013, 135, 11680.
37. Osborn, M. F. et al ACS Chem. Biol. 2014, 9, 2404.
38. Ding, S. et al Angew. Chem., Int. Ed. 2013, 52, 3350.
39. Qiao, X. et al J. Biol. Inorg. Chem. 2014, 19, 415.
9. Collins, A. R. Mol. Biotechnol. 2004, 26, 249.
0. Spielmann, H. P. et al Biochemistry 1995, 34, 12937.
1. Hartley, J. A.; Souhami, R. L.; Berardini, M. D. J. Chromatogr. 1993, 618, 277.
2. Chiou, S. H. J. Biochem. 1983, 94, 1259.
40. Raschle, M. et al Science 2015, 348, 1253671.
41. Sawasdichai, A. et al J. Vis. Exp. 2010.
42. Polo, S. E.; Jackson, S. P. Genes Dev. 2011, 25, 409.