Genotoxicity screening for N-nitroso compounds. Electrochemical and electrochemiluminescent detection of human enzyme-generated DNA damage from N-nitrosopyrrolidine

Article abstract of DOI:10.1039/b703012f

We report for the first time voltammetric/electrochemiluminescent sensors applied to predict genotoxicity of N-nitroso compounds bioactivated by human cytochrome P450 enzymes. The Royal Society of Chemistry.

Full text of DOI:10.1039/b703012f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Genotoxicity screening for N-nitroso compounds. Electrochemical and
electrochemiluminescent detection of human enzyme-generated DNA
damage from N-nitrosopyrrolidine{
Sadagopan Krishnan,a Eli G. Hvastkovs,a Besnik Bajrami,a Ingela Jansson,b John B. Schenkmanb and
James F. Rusling*ab
Received (in Cambridge, UK) 27th February 2007, Accepted 15th March 2007
First published as an Advance Article on the web 3rd April 2007
DOI: 10.1039/b703012f
We report for the first time voltammetric/electrochemilumines-
cent sensors applied to predict genotoxicity of N-nitroso
compounds bioactivated by human cytochrome P450 enzymes.
Metabolic activation of lipophilic pollutants and drugs in
Scheme 1
mammalian liver by cytochrome (cyt) P450 enzymes can cause
DNA damage in a major genotoxicity pathway.1 Detection of
DNA damage caused by metabolites can be used in toxicity
screening and molecular genotoxicity studies.2,3 We recently
described electrochemical sensors for genotoxicity screening that
combine metabolic enzymes and DNA in thin films.2 The enzyme
reaction may produce reactive metabolites, and relative rates of
reaction with DNA are detected by catalytic voltammetry2a,3 or
mass spectrometry (MS).2b By comparison with nucleobase
formation rates measured by MS, these sensors have been
validated so far for hydrocarbons such as styrene, benzo[a]pyrene,
and simple methylating agents, and can detect y0.05% DNA
damage.2a,3 However, to be widely applicable, the sensors need to
predict genotoxicity for other classes of compounds such as
nitrogen heterocycles, which appear often in drugs and toxic
chemicals. Herein, we report the first application of in vitro sensors
with voltammetric and electrochemiluminescent outputs to predict
genotoxicity caused by metabolites of N-nitrosamines, formed in
the sensors by reactions catalyzed by cyt P450 2E1.
(dG) using LC/MS and UV–vis spectra.10 They identified 13
possible adducts between a-acetoxy N-nitrosopyrrolidine and dG,
demonstrating the complex nature of this reaction. In the present
work, we used a similar reaction scheme in thin films to develop
in vitro toxicity screening sensors for such molecules. However, to
date, in vitro genotoxicity testing for nitrosamines has only
included microsomal and reconstituted enzymatic systems.
Therefore, alternative screening methods addressing desirable
analytical aspects (i.e. cost, throughput, rapid sample handling)
are needed.
Electrochemical methods offer simple, rapid, inexpensive
approaches to detect DNA damage.2a,11 A particularly sensitive
approach is to use ruthenium tris(2,29-bipyridyl) [Ru(bpy)32+] to
electrocatalytically oxidize guanines in DNA.11c We previously
showed that electrode-immobilized ruthenium bipyridyl polymer
[Ru(bpy)2PVP10]2+, denoted as RuPVP, can also be used to
catalytically oxidize guanine in DNA in a similar fashion,
according to Scheme 2, where G 5 guanine.12
Nitrosamines constitute a large class of compounds that are
reported as carcinogenic in more than 30 species.4 Cytochrome
P450s play a key role in bioactivating N-nitrosamines for
genotoxicity, mainly by a-hydroxylation of carbon adjacent to
the heterocyclic nitrogen. N-Nitrosopyrrolidine (NPYR) is a
suspected human carcinogen with well established rodent carcino-
Damage to the nucleobases causes DNA to partly unwind, and
allows better access of the catalyst to guanines. This causes faster
cycling of RuIII/II in eqn 2 and yields an enhanced catalytic
oxidation current detected by square wave voltammetry (SWV).3
Electrochemiluminescence (ECL) at 610 nm is generated from
RuPVP by further oxidation of guanine (eqns 3 and 4), or by
reaction of reduced RuI and RuIII to produce excited RuII*.12
Detected simultaneously, SWV and ECL signals are proportional
to the relative amount of chemically damaged DNA bases in the
sensor films.12 Immobilized RuPVP metallopolymer in the sensor
film removes the need for soluble catalyst, or for sacrificial
reductant as the guanines serve the latter purpose.
genicity.5
Exogenous
and
endogenous
sources
of
N-nitrosopyrrolidine include diet, tobacco smoke and nitrosation
of pyrrolidine.6,7 N-Nitrosopyrrolidine requires metabolic bioacti-
vation to exert carcinogenic effects,8 and cyt P450 2E1 is a major
enzyme that bioactivates N-nitrosopyrrolidine (Scheme 1).9
Adduct formation occurs between the metabolites of N-nitrosopy-
rrolidine and the deoxyguanosines of DNA (see Fig. S1).
Hecht et al. detailed the mechanism of DNA adduct formation
from the reaction of a-acetoxy N-nitrosopyrrolidine, a stable
precursor to a-hydroxy N-nitrosopyrrolidine, and deoxyguanosine
aDepartment of Chemistry (U-60), 55 N. Eagleville Rd., University of
Connecticut, USA. E-mail: james.rusling@uconn.edu
bDepartment of Pharmacology, University of Connecticut, USA
{ Electronic supplementary information (ESI) available: Experimental
details, Figs. S1–S8. See DOI: 10.1039/b703012f
Scheme 2
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 1713–1715 | 1713

View Article Online
In the present work, DNA–RuPVP–enzyme films were
constructed as active components of genotoxicity sensors. Layers
of RuPVP, DNA and enzymes were adsorbed sequentially on
basal plane pyrolytic graphite disks (PG, A 5 0.2 cm2)2,3,12 and
rinsed with water between adsorption steps (see Supplemental
Information). Quartz crystal microbalance (QCM) studies con-
firmed regular and reproducible film formation (ESI Fig. S2). Film
architectures expressed as order of layer deposition for SWV were
DNA/RuPVP/DNA/cyt P450 2E1/DNA or DNA/RuPVP/DNA/
myoglobin/DNA. For ECL, additional bilayers of DNA/RuPVP
were necessary to improve signal/noise.
The peak potential was 1.13 V vs. SCE and peak currents
increased with incubation time. Fig. 1b shows the increase in signal
as peak current ratio (Ip,f/Ip,i), where Ip,i is the peak for sensors not
exposed to NPYR and peroxide, and Ip,f is the peak after the
enzyme reaction. This ratio minimizes deviations due to inter-
electrode variability. NPYR is bioactivated by cyt P450 2E1 to
reactive metabolites. The Ip,f/Ip,i increase in Fig. 1 is due to
guanine–NPYR metabolite adduct formation, which increases the
catalytic oxidation rate as mentioned above. The signal increase
reflects DNA damage similar to that reported previously from
styrene and benzo[a]pyrene metabolites.2a,3,12 Fig. 1b also shows
control peak ratios for sensors exposed to H2O2 only, and for
RuPVP/DNA/Mb sensors exposed to NPYR and H2O2. The
H2O2-only control shown in Fig. 1a and plotted in Fig. 1b
demonstrates that peroxide does not damage DNA or contribute
to the oxidative current under these conditions. The negligible
signal increase when using Mb as a model oxygenase13c in the films
suggests Mb does not bioactivate NPYR under these conditions.
ESI Figs. S6 to S8 demonstrate that voltammetric peaks are due to
RuPVP and increases due to catalytic DNA oxidation.
Cyt P450 2E1 immobilized in similar thin films demonstrated
proper electrochemical behavior characteristic of the native
enzyme (ESI Figs. S3–S4). Peroxide activates cyt P450 enzymes
to give the same metabolites as in vivo, where activation occurs via
NADPH and cyt P450 reductase.13a,b To drive the enzyme
reaction, sensors were exposed to 150 mM N-nitrosopyrrolidine
(NPYR, Sigma) + 1 mM H2O2 in acetate buffer pH 5.5 at 37 uC. A
water rinse of the sensor stopped the reaction. Fig. 1a shows SWVs
obtained for sensors for increasing time of the enzyme reaction.
Figs. 2a and 2b are digitally reconstructed images from ECL
arrays12b that demonstrate the ECL response when the DNA/cyt
P450 2E1 and DNA/Mb films are exposed to increasing times of
damage solution (see Supplementary Information for description
of ECL arrays12b). Each spot in the array contains enzyme, DNA
and RuPVP. The increase in light intensity as the cyt P450 2E1
reaction proceeds results from increasing damage to the DNA.
Results for RuPVP/DNA/Mb confirm that Mb does not
bioactivate NPYR. Fig. 2c shows an ECL ratio plot comparing
the ECL from different enzyme reaction times to that at t 5 0,
Fig. 1 Voltammetric sensor data after exposure to NPYR: (a) SWV of
DNA/RuPVP/DNA/cyt P450 2E1/DNA films before and after incuba-
tions at 37 uC with 150 mM NPYR and 1 mM H2O2 in 10 mM acetate
buffer + 50 mM NaCl, pH 5.5. Controls show incubations with only 1 mM
H2O2 (labeled times in min) (SWV ampl. 25 mV; freq. 15 Hz; step 4 mV).
Damage exposure peaks are offset for clarity; controls are not. (b)
Influence of NPYR/H2O2 incubation time on SWV peak current ratio for
DNA/RuPVP/DNA/cyt P450 2E1/DNA (solid line) and DNA/RuPVP/
DNA/Mb/DNA films. Controls are incubation of DNA/cyt P450 2E1
films with only 1 mM H2O2 (dashed line). Error bars are sd for n 5 3 trials.
Fig. 2 ECL array data after exposure to NPYR: (a) Digitally
reconstructed image demonstrating CCD captured ECL emitted from
RuPVP/DNA/cyt P450 2E1 array and (b) RuPVP/DNA/Mb array
exposed to 150 mM NPYR + 1 mM H2O2 for the denoted amounts of
time (s). (c) Ratio plot demonstrating the ECL signal increase from films
containing cyt P450 2E1 or Mb.
1714 | Chem. Commun., 2007, 1713–1715
This journal is ß The Royal Society of Chemistry 2007

View Article Online
N-nitrosopyrrolidine a-hydroxylation metabolite 2-hydroxytetra-
hydrofuran (2-OH-THF).14 Controls not exposed to cyt P450 2E1
showed an SIR chromatogram peak at 27 min (dashed line,
Fig. 3b). The MS for the 27 min peak showed a peak at m/z 101
Da associated with [M + H]+ of N-nitrosopyrrolidine.
Overall, the genotoxicity sensors detected DNA damage from
the NPYR metabolites from cyt P450 2E1 by SWV and ECL.
Bioactivation of NPYR to reactive metabolites in our sensor films
was effected by cyt P450 2E1 but not Mb, indicating the specificity
of the sensor to enzyme identity. This result agrees with previous
reports that cyt P450 2E1 is a major human liver enzyme involved
in metabolizing NPYR.9
The authors gratefully acknowledge financial support from the
National Institute of Environmental Health Sciences, NIH (grant
No. ES03154).
Fig. 3 Single ion recording capLC-MS chromatogram detecting m/z 70–
100 Da for: (a) 150 mM NPYR + 1 mM H2O2 reacted 10 min with CYP
2E1 immobilized on silica microspheres; (b) 150 mM NPYR + 1 mM H2O2
reacted with control microspheres with no enzyme.
Notes and references
1 (a) Cytochrome P450, ed. J. B. Schenkman and H. Greim, Springer
Verlag, Berlin, 1993; (b) Cytochrome P450: Structure,mechanism, and
biochemistry, ed. P. R. Ortiz de Montellano, Plenum, New York, 3rd
edn, 2005.
2 (a) J. F. Rusling, in Electrochemistry of nucleic acids and proteins, ed.
E. Palecek, F. Scheller and J. Wang, Elsevier, Amsterdam, 2005,
pp. 433–450; (b) M. Tarun and J. F. Rusling, Crit. Rev. Eukaryotic Gene
Expression, 2005, 15, 295–315.
3 (a) L. Zhou, J. Yang, C. Estavillo, J. D. Stuart, J. B. Schenkman and
J. F. Rusling, J. Am. Chem. Soc., 2003, 125, 1431–1436; (b) B. Wang,
I. Jansson, J. B. Schenkman and J. F. Rusling, Anal. Chem., 2005, 77,
1361–1367; (c) B. Wang and J. F. Rusling, Anal. Chem., 2003, 75,
4229–4235.
4 S. S. Hecht, Chem. Res. Toxicol., 1998, 11, 559–603.
5 R. Preussmann and B. W. Stewart, N-Nitroso carcinogens, in Chemical
carcinogenesis, ed. C. E. Searle, ACS monograph 182, American
Chemical Society, Washington, DC, 1984, pp. 643–828.
6 (a) H. Bartsch and B. Spiegelhalder, Eur. J. Cancer Prev., 1996, 5,
11–18; (b) D. Hoffmann and I. Hoffmann, J. Toxicol. Environ. Health,
1997, 50, 307–364.
7 A. R. Tricker, B. Pfundstein, T. Kalble and R. Preussmann,
Carcinogenesis, 1992, 13, 563–568.
8 H. L. Wong, S. E. Murphy, M. Wang and S. S. Hecht, Carcinogenesis,
2003, 24, 291–300.
9 A. M. Flammang, H. V. Gelboin, T. Aoyama, F. J. Gonzalez and
G. D. McCoy, Biochem. Arch., 1993, 9, 197–204.
10 M. Wang, E. J. McIntee, Y. Shi, G. Cheng, P. Upadhyaya, P. W. Villalta
and S. S. Hecht, Chem. Res. Toxicol., 2001, 14, 1435–1445.
11 (a) E. Palecek and M. Fojta, Anal. Chem., 2001, 73, 74A–83A; (b) H. H.
Thorp, Trends Biotechnol., 1998, 16, 117–121; (c) D. H. Johnston, K. C.
Glasgow and H. H. Thorp, J. Am. Chem. Soc., 1995, 117, 8933–8938.
12 (a) L. Dennany, R. J. Forster and J. F. Rusling, J. Am. Chem. Soc.,
2003, 125, 5213–5218; (b) E. G. Hvastkovs, M. So, S. Krishnan,
B. Bajrami, M. Tarun, I. Jansson, J. B. Schenkman and J. F. Rusling,
Anal. Chem., 2007, 79, 1897–1906.
13 (a) F. P. Guengerich, Chem. Res. Toxicol., 2001, 14, 611–650; (b)
P. C. Cirino and F. H. Arnold, Angew. Chem., Int. Ed., 2003, 42,
3299–3301; (c) P. R. Ortiz de Montellano and C. E. Catalano, J. Biol.
Chem., 1985, 260, 9265–9271.
similar to the SWV ratio plots. The ECL ratio increased with
reaction time for the RuPVP/DNA/cyt P450 2E1 films, denoting
increasing amounts of DNA damage from N-nitrosopyrrolidine
metabolites.
Increased ECL signals (Fig. 2c) at smaller reaction times
compared to SWV (cf. Fig. 1b) are attributable to differences in
film composition and detection format. Larger amounts of
RuPVP and DNA in the ECL films force more ruthenium centers
in close proximity to DNA resulting in more ECL. The higher
amount of RuPVP in ECL films was necessary to generate
sufficient light to be imaged by the CCD camera.12b These
additional RuPVP layers were unnecessary for voltammetric
analysis, and only one layer of RuPVP was used to generate a
satisfactory signal to background ratio as seen in Fig. 1a. The
rapidly achieved plateau in Fig. 2c is likely due to increased
amounts of RuPVP and prolonged oxidation conditions compared
to SWV, resulting in detection of the majority of damaged
guanines formed during NPYR/H2O2 exposure. We also observe
plateau responses for electrochemical studies at lengthy NPYR
exposure times (data not shown), as well as for direct DNA
damage studies,12a lending causal evidence.
Verification of active NPYR metabolite produced by cyt P450
2E1 was accomplished by forming analogous films on hydro-
xylated 0.5 mm silica microspheres, running the enzyme reaction,
and analyzing reaction solutions by CapLC-MS (see Supplemental
Information for protocols). MS was acquired by monitoring the
elution of compounds with selected m/z of 70–100 Da range (single
ion recording, SIR). Fig. 3a shows the chromatogram for a SIR-
MS of NPYR reaction solution extract in 10 mM acetate buffer
plus 50 mM NaCl, pH 5.5 (150 mM NPYR + 1 mM H2O2) after
exposure for 10 min to microspheres with immobilized cyt P450
2E1. This shows the [M + H]+ product peak in the chromatogram
at 19 min (solid line). The observed [M + H]+ MS peak at m/z 89
Da for this peak (Fig. S5) is consistent with the major
14 (a) C. B. Chen, G. D. McCoy, S. S. Hecht, D. Hoffmann and
E. L. Wynder, Cancer Res., 1978, 38, 3812–3816; (b) M. Wang,
P. Upadhyaya, T. T. Dinh, L. E. Bonilla and S. S. Hecht, Chem. Res.
Toxicol., 1998, 11, 1567–1573.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 1713–1715 | 1715