Mutation Research 653 (2008) 63–69

Contents lists available at

ScienceDirect

Mutation Research/Genetic Toxicology and

Environmental Mutagenesis

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / g e n t o x

C o m m u n i t y a d d r e s s : w w w . e l s e v i e r . c o m / l o c a t e / m u t r e s

Development of a highthroughput yeast-based assay for detection of
metabolically activated genotoxins

Xuemei Liu

∗

, Jeffrey A. Kramer, Jonathan C. Swaffield, Yi Hu, Guixuan Chai, Alan G.E. Wilson

Drug Metabolism, Pharmacokinetics, and Toxicology, Lexicon Pharmaceuticals, Inc., 8800 Technology Forest Place, The Woodlands, TX 77381, United States

a r t i c l e i n f o

Article history:
Received 24 January 2008
Received in revised form 14 March 2008
Accepted 19 March 2008
Available online 1 April 2008

Keywords:
Yeast
Dual luciferase
Highthroughput
Metabolic activation
Genetic toxicology
Mutagens

a b s t r a c t

The potential genotoxicity of drug candidates is a serious concern during drug development. Therefore, it
is important to assess the potential genotoxicity and mutagenicity of a compound early in the discovery
phase of drug development. AMES Salmonella assay is the most widely used assay for the assessment
of mutagenicity and genotoxicity. However, the AMES assay is not readily adaptable to highthroughput
screening and several strains of Salmonella must be employed to ensure that different types of DNA damage
can be studied. Therefore, an additional robust highthroughput genotoxicity screen would be of significant
value in the early detection and elimination of genotoxicity. The complexity of DNA damage requires
numerous cellular pathways, thus using single model organism to predict genotoxicity in early stage is
challenging. Another critical component of such screens is that they incorporate the capability of metabolic
activation to ensure that no genotoxic metabolites are generated.

We have developed a novel highthroughput reporter assay for DNA repair that detects genotoxicity,

and which incorporates metabolic activation. The assay has a low compound requirement as compared to
Ames, and relies upon two different reporter genes cotransfected into a yeast strain. The gene encoding
Renilla luciferase is fused to the constitutive 3-phosphoglycerate kinase (PGK1) promoter and integrated
into the yeast genome to provide a control for cell numbers. The firefly luciferase gene is fused to the
RAD51 (bacterial RecA homolog) promoter and used to report an increase in DNA repair activity. A dual
luciferase assay is performed by measuring the firefly and Renilla luciferase activities in the same sample.
The result is expressed as the ratio of the two luciferase activities; changes from the base level (control) are
interpreted as induction of the RAD51 promoter and evidence of DNA repair activity in eukaryote cells due
to DNA damage. The yeast dual luciferase reporter has been characterized with and without S-9 activation
using positive and negative control agents. This assay is efficient, requires little time and low amounts of
compound. The assay is compatible with metabolic activation, adaptable to a highthroughput platform,
and yields data that accurately and reproducibly detects DNA damage.

Whereas the normal yeast cell wall, plasma membrane composition and the presence of active trans-

porters can prevent the entry or persistence of some compounds internally in yeast cells, our assay did
show concordance with regulatory mutagenicity assays, many of which require metabolic activation and
are poorly detected by bacterial mutagenicity assays. Although there were false negative results, in our
hands this assay performs as well as or better than other commercially available genetox assays. Fur-
thermore, the RAD51 gene is strongly inducible by homologous intrachromosomal recombination; thus
this assay may provide a means to detect clastogens. The RAD51 promoter fused dual luciferase assay
represents a valuable addition to the armamentarium for the early detection of genotoxic compounds.

© 2008 Elsevier B.V. All rights reserved.

Abbreviations: GFP, green fluorescent protein; DSB, double strand break; EMS,

ethylmethanesulfonate;

2AAF,

2-acetylaminofluorene;

4AAF,

4-acetylamino-

fluorene; MMS, methylmethane sulfonate; 2AA, 2-aminoanthracene; 4NQO1,
4-nitroquinoline-N-oxide; 2NF, 2-nitrofluorene; DMN, dimethylnitrosamine; YNB,
yeast nitrogen base; RLU, relative light units.

∗ Corresponding author. Tel.: [phone]; fax: [phone].

E-mail address:

[e-mail]

(X. Liu).

1. Introduction

Rapid, highthroughput biological assays that measure chem-

ically induced mutations can provide “early warnings” of
mutagenicity and genotoxicity. Among these assays, bacterial
screens are generally robust, economical, and well characterized.
The most widely used systems employ microplate-based bacterial
screens: the Ames test (Ames II) and the SOS response reporters

[1–6]

. The principle drawback of the bacteria-based tests is that

they lack eukaryotic chromosomes; thus they are unable to detect

1383-5718/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:

10.1016/j.mrgentox.2008.03.006

64

X. Liu et al. / Mutation Research 653 (2008) 63–69

clastogenic and aneugenic events

[7]

. Chromosome damage can

lead to genomic rearrangements such as deletions, transloca-
tions, and amplifications

[8,9]

, and bacteria-based tests are not

suitable for revealing such chromosomal aberrations. Typically,
chromosome breakage is scored microscopically

[3]

. However, this

technique remains labor intensive, time consuming, and its appli-
cation for screening in drug discovery is severely limited. Therefore,
it is desirable to have a screening test that reduces the false negative
results of bacterial tests, and which is amenable to highthroughput
methods.

Yeast (Saccharomyces cerevisiae) provide a valuable system for

the analysis of recombination in eukaryotes

[10,11]

. Several publica-

tions have described the development of a yeast-based genotoxicity
test in which the DNA damage-inducible promoter of the RAD54
gene is fused to the green fluorescent protein (GFP)

1

[12–15]

. In

yeast, the RAD54 protein participates in the recombinational repair
of double-strand DNA breaks together with the RAD51, RAD52,
RAD55, and RAD57 proteins. In this process, RAD54 interacts with
RAD51 and stimulates DNA strand exchange, promoted by RAD51
protein

[16,17]

. For several reasons, GFP is an unsuitable reporter for

this study because the GFP signal is invariably contaminated with
endogenous (or media-based) autofluorescence

[13–15]

although

there are methods to mitigate this

[18]

. Moreover, the limited

metabolic capability of yeast requires the use of the autofluorescent
nicotinamide nucleotides (NADPH) and S-9 mix for the activation
of most promutagens.

To overcome autofluorescence limitations, we employed two

luciferase genes as reporters, and this dual luciferase assay pro-
vides advantages for measuring gene expression in yeast. First,
the sensitivity of the luciferase assay allows the reporter to be
integrated into the genome and still provide measurable signal,
alleviating the issue of copy number variation observed with epi-
somal reporters

[19]

. Second, the use of a constitutive promoter to

drive expression of one of the reporters provides an internal con-
trol for cell numbers. In the current study, we have used the RAD51
promoter as an inducible reporter of DNA damage, and the PGK1
promoter as the constitutive internal control. The RAD51 gene is a
homolog of bacterial recA which promotes the pairing of homol-
ogous DNA molecules and strand exchange reactions

[16,20–23]

;

therefore, we surmised that our assay would assist in the detection
of DNA-damaging agents.

We evaluated RAD51-based induction of luminescence using 28

model compounds. The testing compounds included DNA dam-
aging agents, many of which are promutagens, and well-known
reagents employed as controls in the regulatory test. Our data show
that the assay has a high level of concordance with regulatory muta-
genicity assays, and suggests that yeast dual luciferase reporter
system affords a new tool to identify metabolically activated muta-
gens and genotoxins.

2. Materials and methods

2.1. Caution

All the positive controls were handled in accordance with NIH Guidelines for

the Laboratory Use of Chemical Carcinogens

[24]

.

All solvents and chemicals were purchased from Aldrich Chemicals (Milwaukee,

WI) or Sigma (St. Louis, MO) and the purities are greater than 99%. Aroclor1254-
induced rat liver S-9 homogenate was obtained from Molecular Toxicology Inc.
(Boone, NC). The dual luciferase assay kit (No. E1960) was purchased from Promega
Corp. (Madison, WI).

2.2. Yeast strains and growth media

The haploid strain S. cerevisiae YPH499 was obtained from Stratagene (La Jolla,

CA). The strain has not been modified to increase permeability or sensitivity to DNA
damage. Transfected yeast strains were grown on yeast nitrogen base (YNB) with
URA and LEU dropout medium as previously described

[19]

.

2.3. Luciferase reporter plasmids

The Renilla and firefly luciferase gene sequences were amplified by PCR

from pRL-CMV and pGL-3-Basic plasmids (Promega Corp., Madison, WI), respec-
tively. The Renilla primers used were forward: 5



-GGGCAAGCTTCCCTTATGAC-

TTCGAAAGTTTATGATCC-3



and reverse: 5



-GGGCCAGCTGTTATTGTTCATTTTTGAG-

AAC-3



. The firefly primers used were forward: 5



-GGGCAAGCTTCCCTTATGGAA-

GACGCCAAAAACATAAAG-3



and reverse: 5



-GGGCCAGCTGTTACACGGCGATCTTTCC-

GCCC-3



. For both Renilla and firefly genes the forward primer incorporates a unique

HindIII site into the amplified product, whereas a unique PvuII site was introduced
to the end of reverse primers.

The assay strain contains two nuclear, integrative replicating, single copy plas-

mids YIp358 (URA3) and YIp368 (LEU2) ordered from the ATCC (Manassas, VA)

[25]

.

The vectors YIp358 and YIp368 were digested with HindIII and PvuII to remove
lacZ, purified by agarose gel electrophoresis, and ligated with firefly and Renilla
genes generated via PCR to produce the YIp358Firefly and YIp368Renilla vectors,
respectively.

To generate a PGK1-YIp368Renilla promoter fusion construct, the PGK1 pro-

moter was amplified by PCR from yeast (HMS-1) genomic DNA using primers
forward: 5



-GGGCGGATCCGAAGTACCTTCAAAGAATGGGGTCTCATC-3



and reverse:

5



-GGGCAAGCTTTTGTTTTATATTTGTTGTAAAAAGTAG-3



, which incorporated unique

BamHI and HindIII sites, respectively

[19]

. The RAD51-YIp358Firefly reporter

construct was initiated by introducing a KpnI site at -1.3kb of the RAD51 5



NTS, and the HindIII sites at the third codon. The primers used were for-
ward: 5



-AGATTAATGGTACCTATTTTGTGTTGGGGTTGTTTTTGGGACC-3



and reverse:

5



-AGACACATAAGCTTACATATGACGATAACAAATTAGTAGGCC-3



. The final vectors

bear the up-stream noncoding DNA sequence of the S. cerevisiae PGK1 promoter
in front of Renilla luciferase reporter and RAD51 gene fused with firefly luciferase
reporter. All the sequences were verified using the dye terminator cycle sequencing
method on an ABI PRISM.

2.4. Construction of luciferase reporter yeast strain

Yeast transformation was as described elsewhere; a lithium acetate protocol was

employed

[26]

. Double transformants were selected on Ura DO, Leu DO SC-media

plates

[27]

. Individual colonies were picked and streaked twice for clonal purification

before storage at

−80

◦

C. Nonpromoter containing YIp358Firefly and YIp368Renilla

vectors were also transfected and used for background evaluation.

2.5. RAD51 dual luciferase assay

Assays were carried out in 96-well, white microplates (Dynex Technologies,

Chantilly, VA). Before treatment with possible genotoxic chemicals, yeast cells were
grown in YNB-URE-LEU medium at 30

◦

C and 200 rpm orbital shaking for 24 h

to achieve logarithmic growth. Cells were collected and reseeded (75

␮L, 3 × 10

5

cells/mL) into a 96-well, clear plate (Becton Dickinson Labware, Franklin Lakes, NJ).
All the test compounds were dissolved in 15% DMSO to make 30,000, 15,000, 3000,
1500, 300 and 150

␮g/mL. The test compounds were then added to yeast cells (10 ␮L

to make final concentration 2000, 1000, 200, 100, 20, and 10

␮g/mL in the incubation

systems) combined with 65

␮L of 11% S-9 mix to make 5% final concentration of S-9;

or a phosphate buffer according to bacterial Ames assay

[28]

. Since previous results

have shown that DMSO increases yeast DEL recombination

[10]

, the final concen-

tration for DMSO in the medium is 1%. After 18 h treatment as previously described

[29]

, 40

␮L cells were harvested for assessment of dual luciferase activity in 200 ␮L

passive lysis buffer (Promega luciferase kits). Luciferase activity was determined for
25

␮L aliquots of cell lysates in 100 ␮L lysis buffer mixed with 25 ␮L of the luciferase

assay solution, and read on a Tecan Ultra-384 (Tecan UK Ltd., Theale, UK) in a 96-
well plate format. Changes in induction ratio (fold changes) were calculated using
the equation (F = firefly, R = Renilla):

Fold changes

=

F

treated

/R

treated

F

DMSO

/R

DMSO

.

The results are expressed as the mean

± S.D. where n equals six.

2.6. Ames assay

Ames assays were performed as described by Mortelmans and Zeiger

[28]

.

Briefly, the tester strains TA100 and TA98 were combined with S-9 mix or buffer,
test or control article, as well as a trace of histidine and molten agar (Moltox, Inc.,
Boone, NC). After 48 h, the background lawn density was scored, followed by count-
ing the number of revertant colonies. Mutation results are reported as numbers
of revertants per plate. The results are expressed as the mean

± S.D. based on six

dilutions scored in triplicate.

2.7. Data analysis

Data analysis was similar to that described by others

[29]

. For each com-

pound, data from three independent experiments were collected and the mean
and standard deviation was calculated. A compound was considered genotoxic in

X. Liu et al. / Mutation Research 653 (2008) 63–69

65

Fig. 1. Sensitivity, reproducibility, and stability of yeast luciferase assay. Yeast strain
containing the RAD51firefly and PGK1Renilla reporters was grown in YNB-URA-LEU
medium to exponential phase. The cell numbers were quantified by hemocytometer,
and dilutions were prepared such that the indicated number of cells was used to
assay both firefly (A) and Renilla (B) luciferase activities. The data represent the

the luciferase assay if it caused at least a 1.6-fold induction as compared to the
DMSO and the observed increase was concentration dependent. If the test com-
pounds showed cytotoxicity (both firefly and Renilla signals are lower than twofold
of background) or precipitation in more than three concentrations (totally six con-
centrations were screened), further dilutions need to be tested. A compound was
considered negative, if it did not satisfy the above mentioned criteria but led to a
cytotoxic effect or precipitation was found in the medium within the treated con-
centration range. Methylmethane sulfonate (MMS), at concentrations of 250 and
750

␮g/mL served as a positive control mutagen for experiments performed in the

absence of metabolic activation. 2-Aminoanthracene (2AA) at concentrations of 13.3
and 66.7

␮g/mL was used as positive control for studies in the presence of metabolic

activation.

3. Results

3.1. Background

An important criterion for many reporter assays is the back-

ground signal. To evaluate background luminescence resulting from
the spontaneous oxidation of luciferin with components in the
yeast lysate and S-9, we performed both firefly and Renilla luciferase
assays using the double transfected nonpromoter YIp358Firefly
and YIp368Renilla vector strains. The background signals for the
firefly luciferase assay (mean

± S.D.) with and without S-9 were

85

± 49 and 96 ± 47 relative light units (RLU), respectively. Renilla

luciferase assay yielded 61

± 30 RLU without S-9 and 94 ± 31 RLU

with S-9. The background signals for lysis buffer without adding
luciferase substrate were 105

± 42 RLU without S-9 and 96 ± 47 RLU

with S-9. Thus, in both assays the background signal was negligi-
ble.

3.2. Sensitivity, reproducibility, and stability

Other important criteria for reporter assays are sensitivity, lin-

earity, reproducibility, and stability. The luciferase yeast strain was
grown to exponential phase in YNB-URE-LEU medium at 30

◦

C.

The cell numbers were quantified by hemocytometer, serial dilu-
tions were prepared, and between 600 and 50,000 cells were
used per assay for both firefly and Renilla luciferase activities.
Our data showed that both the firefly and Renilla luciferase activ-
ities increased linearly with increasing numbers of cells (

Fig. 1

A

and B), and the correlation of firefly to Renilla luciferase activi-
ties was very strong (

Fig. 1

C). The firefly/Renilla activity ratio was

calculated at each cell concentration (

Fig. 1

D), and the average

mean value was 0.086. The error between replicates was slightly
higher at lower cell concentrations. Therefore, 16,000–25,000 cells
per assay were employed for validation studies. The yeast dual
luciferase assay was also tested with different concentrations of
S-9 (1, 2, 5, and 10%) co-incubation with eight replicates. The fire-
fly luciferase assay signals were 1648

± 234 (0% S9), 1765 ± 210

(1% S9), 1626

± 251 (2% S9), 1480 ± 247 (5% S9), and 1386 ± 125

(10% S9) RLU. Renilla luciferase assay yielded 23,634

± 2034 (0%

S9), 22,374

± 1757 (1% S9), 22,853 ± 1907 (2% S9), 22,199 ± 2410

(5% S9), and 19,586

± 1982 (10% S9) RLU. The concentration range

for S9 is based on those used in the bacterial Ames assay

[28]

. Our data showed that as the S9 concentration increased,

mean of three replicate assays performed at each dilution. (C) Firefly vs. Renilla
luciferase activity is linear in 600–51,200 range of cell concentrations analyzed.
These data are from graphs (A) and (B) and the correlation coefficient is indicated (R).
(D) Firefly and Renilla activity ratio is independent of the number of cells assayed.
The data represent those shown in (A) and (B). The bars indicate standard deviation.
(E) Stability of firefly and Renilla luciferase activities in cell lysates with or without
S-9. Yeast cells were lysed in passive lysis buffer, and the ratios firefly and Renilla
luciferase activity were determined immediately as day 0. The lysates were then
stored in

−20

◦

C and luciferase assays were performed on the same aliquots of the

lysate on days 1 and 7. The data shown represent the means from three independent
experiments, with the bars showing standard deviation.

66

X. Liu et al. / Mutation Research 653 (2008) 63–69

Fig. 2. Effect of MMS and 2AA on induction of firefly luciferase. Yeast cells were
treated with different concentrations of MMS, 2AA, and vehicle for 18 h. (A) MMS.
(B) 2AA + Aroclor1254-induced rat liver S-9. The data represent the means from three
independent experiments, with the bars showing standard deviation.

the cytotoxicity is slightly increased. Although the S-9 concen-
trations might raise concerns about the reproducibility of the
firefly/Renilla ratios, our data demonstrate that those theoreti-
cal concerns do not raise any practical problems with this assay
(

Fig. 1

E).

To determine the stability of yeast lysate for the purposes of

the genotoxicity assay, lysates were examined over time. The dual
luciferase strain was assayed immediately after lysis; and the fire-
fly/Renilla activity ratio at this point was defined as day 0. The
lysates were then stored at

−20

◦

C and luciferase assays were per-

formed on the same lysate aliquots on days 1 and 7. The data (

Fig. 1

E)

demonstrate that the yeast lysates are stable up to 7 days at

−20

◦

C

but variation between replicates increased slightly as a function of
time. Therefore the dual luciferase reporter assay is useful for geno-
toxicity studies with respect to sensitivity, reproducibility, linearity,
and stability.

3.3. Effect of MMS and 2AA on induction of firefly luciferase

The system was optimized by using the alkylating agent MMS,

which mutates DNA and promotes activation of the intra-S-phase
checkpoint response

[30,31]

. We selected 2AA as a model com-

pound for optimizing the metabolic activation conditions using
the Aroclor 1254-induced liver S-9. The mutagenic potencies of
MMS and 2AA were measured by serial dilutions of MMS or 2AA
in the exponentially growing cells, with or without S-9 activa-
tion.

Fig. 2

A and B demonstrate that both MMS and 2AA cause a

dose-dependent stimulation of RAD51 gene expression. Five and
ten percent S-9 caused similar inductions of luminescence with
test concentrations of 2AA, but 10% S-9 treated alone was slightly

Table 1
Twenty-eight compounds tested in yeast dual luciferase assay

Compounds

CAS no.

Salmonella

Carcinogenicity

a

Rat

Mouse

Urethane

51-79-6

+

+

+

Ethylmethane sulfonate

62-50-0

+

+

+

Dimethylnitrosamine

62-75-9

+

+

+

Vinblastin

865-21-4

−

−

b

4-Acetylaminofluorene

28322-02-3

+

−

2-Acetylaminofluorene

53-96-3

+

+

+

Etoposide

33419-42-0

−

Bleomycin

9041-93-4

−

2-Aminoanthracene

613-13-8

+

Cyclophosphamide

50-18-0

+

+

+

Mitomycin C

50-07-7

+

+

Sodium azide

26628-22-8

+

−

4-Nitroquinoline-N-oxide

56-57-5

+

+

+

Methylmethane sulfonate

66-27-3

+

+

Cisplatin

15663-27-1

+

+

Arsenic III oxide

1327-53-3

−

−

Methyl viologen

1910-42-5

+

Nalidixic acid

389-08-2

−

+

−

2-Nitrofluorene

607-57-8

+

+

Taxol

33069-62-4

−

Ara C

147-94-4

−

Chlorambucil

305-03-3

+

+

+

Actinomycin D

50-76-0

−

+

Caffeine

58-08-2

−

−

−

HCl

7647-01-0

−

−

NaOH

1310-73-2

−

−

−

MeOH

67-56-1

−

DMSO

67-68-5

−

−

−

a

Data are collected from Carcinogenic Potency Project (University of California,

Berkeley).

b

Data are not available.

cytotoxic to yeast cells. Thus, 5% S-9 was chosen as the working
concentration.

3.4. Induction by other DNA-damaging agents

In order to evaluate the ability of the dual luciferase assay to

detect mutagens with and without metabolic activation, we exam-
ined the levels of RAD51-based luciferase induction using a variety
of DNA-damaging agents.

Tables 1 and 2

list chemicals that were

tested in the yeast dual luciferase assay, including both cytotoxic
genotoxins and noncytotoxic nongenotoxins. The dual luciferase
assay was performed in parallel, with and without metabolic acti-
vation (Aroclor1254-induced rat liver S-9), using the same reagents
and single cell purified colonies as a source of the tester cells. MMS
was employed as a positive control for the induction of RAD51.
2AA was used as a positive control for the S-9 metabolic activation
luciferase studies. The background luminescence was measured
using 1% DMSO treatment. All of the direct-acting mutagens
including MMS, EMS, sodium azide, 2-nitrofluoren (2-NF), and 4-
nitroquinoline-N-oxide (4NQO1) were shown to be positive in our
system. The cross-linking agents (mitomycin C and cisplatin) and
a reactive oxygen species generator (methyl viologen dichloride)
were also evaluated and showed greater than 1.6-fold RAD51 induc-
tion. The negative controls, caffeine, HCl, NaOH, MeOH, and DMSO,
did not show any RAD51 induction in the dual luciferase assay.

We also tested clastogens that are negative in the Ames assay,

including: etoposide, vinblastin, arsenic III oxide, bleomycin, taxol,
araC, chlorambucil, and actinomycin D. No induction was observed
for taxol, chlorambucil, araC, and actinomycin D but induction was
seen with etoposide, vinblastin, arsenic III oxide, and bleomycin.
Because the dual luciferase assay could detect promutagens, we
selected 2AAF, 4AAF, dimethylnitrosamine, and cyclophosphamide
as model compounds for metabolic activation evaluation. All of

X. Liu et al. / Mutation Research 653 (2008) 63–69

67

Table 2
Comparison between yeast dual reporter assay and published data

a

Compounds

Luciferase Assay

Genotoxicity Reference

b

Concentrations

Lowest Detectable

Maximum

With S-9

Without S-9

(

␮g/mL)

Concentration (

␮g/mL)

Fold Changes

Urethane

25,000-50

12,500 (cytotoxic)

1.8

+

−

[45]

Ethylmethane sulfonate

2500-5

500

1.9

+

+

Alkylation of DNA

[46]

Dimethylnitrosamine

100-0.2

50

2.5

+

−

Require metabolic activation

[47]

Vinblastin

500-1

250 (cytotoxic)

1.9

+

+

[48]

4-Acetylaminofluorene

1250-2.5

313

3.7

+

−

Require metabolic activation

[49]

2-Acetylaminofluorene

1250-2.5

313

3.4

+

−

Require metabolic activation

[50]

Etoposide

1200-2.4

300 (cytotoxic)

1.9

+

+

[51]

Bleomycin

250-0.5

50 (cytotoxic)

2.4

+

+

Cleavage of DNA.

[52]

2-Aminoanthracene

500-1

50

2.9

+

+

Require metabolic activation

[53]

Cyclophosphamide

1000-2

250

2.2

+

−

Require metabolic activation

[54]

Mitomycin C

200-0.4

100 (cytotoxic)

1.9

+

+

Alkylation of DNA

[55]

Sodium azide

5-0.01

2.5 (cytotoxic)

2.2

+

+

[56]

4-Nitroquinoline-N-oxide

30-0.06

0.5 (cytotoxic)

2.6

+

+

[50]

Methylmethane sulfonate

2000-4

196 (cytotoxic)

11

+

+

Alkylation of DNA

[57]

Cisplatin

30-0.06

15 (cytotoxic)

2.2

+

+

[58]

Arsenic III oxide

1250-2.5

125 (cytotoxic)

1.8

+

+

[59]

Methyl viologen

1000-2

100 (cytotoxic)

1.8

+

+

[60]

Nalidixic acid

500-1

25 (cytotoxic)

2.1

+

+

[61]

2-Nitrofluorene

100-0.2

5 (cytotoxic)

2.2

+

+

[62]

Taxol

1000-2

−

−

−

−

[63]

Ara C

1000-2

−

−

−

−

[64]

Chlorambucil

1000-2

−

−

−

−

Alkylation of DNA

[65]

Actinomycin D

250-0.5

−

−

−

−

Inhibition of DNA and RNA synthesis

[66]

Caffeine

500-1

−

−

−

−

Negative

HCl

10%-0.1%

−

−

−

−

Negative

NaOH

400-0.8

−

−

−

−

Negative

MeOH

10%-0.1%

−

−

−

−

Negative

DMSO

1%

−

−

−

−

Negative

a

Cells were treated with compounds for 18 h. Values are expressed by three determinations.

b

Experimental details are described in Section

2

.

c

Data are collected from published literature.

these four compounds exhibited a dose-dependent induction of
RAD51 (

Fig. 3

).

3.5. Comparisons of dual luciferase assay with regulatory tests

We compared the relative concordance of our assay to the reg-

ulatory assays using the published data (

Tables 1 and 2

).

Fig. 3. Effect of 2AAF (closed bar 625

␮g/mL; dotted bar 313 ␮g/mL; hashed

bar 62.5

␮g/mL; open bar 31.3 ␮g/mL), 4AAF (closed bar 625 ␮g/mL; dotted bar

313

␮g/mL; hashed bar 62.5 ␮g/mL; open bar 31.3 ␮g/mL), DMN (closed bar

50

␮g/mL; dotted bar 25 ␮g/mL; hashed bar 12.5 ␮g/mL; open bar 2.5 ␮g/mL),

and cyclophosphamide (closed bar 500

␮g/mL; dotted bar 250 ␮g/mL; hashed bar

50

␮g/mL open bar 25 ␮g/mL) on induction of firefly luciferase with 5% S9. The data

shown represent the means of three replicates, with the bars showing standard
deviation.

4. Discussion

Currently, screening of DNA damaging agents is achieved with

a battery of standardized genetic toxicity assays that have the abil-
ity to monitor both cytotoxic and genotoxic effects

[32,33]

. These

assays provide high precision and predictivity of potential mech-
anisms of genetic toxicity. Negative findings in these assays are
a strong indication that carcinogenicity is unlikely. In contrast,
indications with multiple assays that a sample can produce DNA
damage or subcellular damage is a strong indication that additional
testing may be required. Whilst this is true in relation to the detec-
tion of human carcinogens as positive in these regulatory assays,
it has been shown that the established in vitro mammalian assays
in particular have very poor specificity and thus there is a need
for developing new assays to improve specificity. These regulatory
approaches provide reliable prediction of human carcinogenic haz-
ard, but time and cost factors limit the number of chemicals that
can be evaluated in these systems and most screens are still applied
late in preclinical drug development.

Measurement of altered gene expression to detect DNA dam-

age response in living cells has shown value in preclinical safety
assessment

[34]

. While damage-inducible genes have been iden-

tified in numerous organisms, regulation of the damage response
has been most extensively characterized in Escherichia coli and S.
cerevisiae

[12,34–37]

. A rapid turn-around yeast-based genotoxicity

test has been reported in which RAD54 gene induction is moni-
tored as induction of green fluorescence protein in yeast

[12,13,18]

.

The RAD54 protein performs crucial functions in the repair of DNA
double-strand breaks (DSBs), which are the major genotoxic lesions
induced by radiation and clastogens. Unrepaired DSBs can cause
loss of chromosomes or cell death. If misrepaired, DSBs can give rise
to mutations and chromosomal rearrangements; therefore, DSBs, if

68

X. Liu et al. / Mutation Research 653 (2008) 63–69

unrepaired, can lead on to cellular changes that may contribute
to stages in the carcinogenic process in multicellular organisms

[38–41]

. In S. cerevisiae RAD51, RAD52, RAD54, and RAD55 have been

assigned to one epistasis group, the recombinational repair group.
In response to DSBs, the RAD51 protein binds to single-stranded
DNA and helps to scan double-stranded DNA until a homologous
sequence is found, where it forms a nucleofilament on the single-
stranded DNA and catalyzes homologous stand exchange together
with RAD54

[17,20,42]

. Therefore, it maybe that the RAD51 gene

used in this study is more sensitive than the RAD54 gene because
of its apparently early induction in response to DNA damage.

Reporter genes have been used extensively to monitor

DNA damage-induced responses, including

␤-galactosidase, CAT,

luciferase, and GFP. Among them, GFP may be the most commonly
used. The major drawback of GFP assays is that many cellular
metabolites and crucial cellular extracts in culture media exhibit
intense autofluorescent. Such autofluorescence background can
cause experimental inaccuracy in GFP assays. Additionally, the nor-
malization of GFP assays has routinely been achieved by quantifying
total protein in the yeast lysate, or on the basis of the optical density
(A

600

) of the yeast culture

[12,18,29]

. The use of the colorimetric

assays to measure total protein requires significant time in addi-
tion to the GFP assay, and optical density measurements can yield
highly variable results (compound precipitation and background
absorption interference) depending on the spectrophotometer and
the yeast strain.

The yeast luciferase assay we have described here uses RAD51

as the inducible damaged DNA sensor and the PGK1 promoter as
a constitutive control. Using luminescence to replace fluorescence
represents an advantage over the other yeast-based assays in that
metabolic activation can be assessed in our assay

[29]

. Because

of its promising performance as a sensitive and specific assay for
carcinogens, the yeast dual luciferase assay has great utility as a
screening tool for predicting mutagenic and genotoxic activity and
investigating the mechanism and genetic control of homologous
recombination and genomic instability. Furthermore, this high-
throughput assay allows for rapid testing of compounds. Another
advantage of the yeast dual luciferase assay is that the internal cell
number control provides an improved means to perform quantita-
tive assays of chemomutagenic potential.

In the present study, 28 compounds were tested for induction of

RAD51. These compounds have been studied previously for muta-
genicity activity. An unexpected finding of this study was that taxol,
chlorambucil, araC, and actinomycin D (

Table 2

) did not show any

induction in the yeast dual luciferase assay. The reasons for lack of
agreement with the traditional in vivo regulatory assays might arise
from the complex mechanisms of DNA damage sensing that require
numerous cellular pathways for the final biological consequences.
Not surprisingly, a single model organism or cell line can offer
only modest predictive power for genotoxicity in general. Other
factors include the complex differences between mammalian cells
and yeast such as the presence of a cell wall and different plasma
membrane composition. Finally, many compounds are pumped
out of yeast cells by ABC transporters. Despite these limitations,
other genotoxicity assays using unmodified yeast strains have been
applied to the early stages of drug discovery

[10,29]

. The results of

our assay do compare favorably with other early in vitro screens
for mutagenicity (e.g. Ames). Future studies could be focused on
increasing cell permeability and or inactivating ABC transporters
making the yeast more sensitive to DNA damaging agents. Another
clear direction for future assay development will focus on large-
scale validation. It is difficult to choose the assessment criteria
applied in the new assays. The criteria used in this paper extended
the previously published method

[29]

. Future large-scale validation

could provide more information and improvement for the criteria

applied in this assay making this assay suitable for the detection of
mutagenicity in the early drug discovery screening.

In summary, previous studies have shown that RAD51 expres-

sion is controlled at the level of transcription, with levels varying
in response to DNA damage

[17,20,35,43,44]

. The stable transfor-

mation of YPH499 with dual reporter constructs is, therefore, a
useful addition to the predictive tests for genotoxicity and muta-
genicity, especially because the assay detects a DNA DSB response.
This simple, rapid assay is likely to be a valuable addition as an
early screen for the mutagenicity and genotoxicity of mutagens and
promutagens.

Acknowledgements

We thank Melinda M. Albright, Suma Gopinathan, Joe J. Shaw,

and Lance Ishimoto for useful suggestions during the preparation
of this manuscript.

References

[1] E. Lorge, V. Gervais, N. Becourt-Lhote, C. Maisonneuve, J.L. Delongeas, N. Claude,

Genetic toxicity assessment: employing the best science for human safety eval-
uation. Part IV. A strategy in genotoxicity testing in drug development: some
examples, Toxicol. Sci. 98 (2007) 39–42.

[2] V. Thybaud, M. Aardema, J. Clements, K. Dearfield, S. Galloway, M. Hayashi, D.

Jacobson-Kram, D. Kirkland, J.T. MacGregor, D. Marzin, W. Ohyama, M. Schuler,
H. Suzuki, E. Zeiger, Strategy for genotoxicity testing: hazard identification
and risk assessment in relation to in vitro testing, Mutat. Res. 627 (2007)
41–58.

[3] L. Muller, Y. Kikuchi, G. Probst, L. Schechtman, H. Shimada, T. Sofuni, D. Tweats,

ICH-harmonised guidances on genotoxicity testing of pharmaceuticals: evolu-
tion, reasoning and impact, Mutat. Res. 436 (1999) 195–225.

[4] S. Fluckiger-Isler, M. Baumeister, K. Braun, V. Gervais, N. Hasler-Nguyen, R.

Reimann, J. Van Gompel, H.G. Wunderlich, G. Engelhardt, Assessment of the per-
formance of the Ames II assay: a collaborative study with 19 coded compounds,
Mutat. Res. 558 (2004) 181–197.

[5] L.R. Ptitsyn, G. Horneck, O. Komova, S. Kozubek, E.A. Krasavin, M. Bonev, P. Ret-

tberg, A biosensor for environmental genotoxin screening based on an SOS lux
assay in recombinant Escherichia coli cells, Appl. Environ. Microbiol. 63 (1997)
4377–4384.

[6] P. Quillardet, O. Huisman, R. D’Ari, M. Hofnung, SOS chromotest, a direct assay

of induction of an SOS function in Escherichia coli K-12 to measure genotoxicity,
Proc. Natl. Acad. Sci. U.S.A. 79 (1982) 5971–5975.

[7] C. Chen, R.D. Kolodner, Gross chromosomal rearrangements in Saccharomyces

cerevisiae replication and recombination defective mutants, Nat. Genet. 23
(1999) 81–85.

[8] S. Vamvakas, E.H. Vock, W.K. Lutz, On the role of DNA double-strand breaks in

toxicity and carcinogenesis, Crit. Rev. Toxicol. 27 (1997) 155–174.

[9] K.K. Khanna, S.P. Jackson, DNA double-strand breaks: signaling, repair and the

cancer connection, Nat. Genet. 27 (2001) 247–254.

[10] Z. Kirpnick, M. Homiski, E. Rubitski, M. Repnevskaya, N. Howlett, J. Aubrecht,

R.H. Schiestl, Yeast DEL assay detects clastogens, Mutat. Res. 582 (2005)
116–134.

[11] S.L. Forsburg, The art and design of genetic screens: yeast, Nat. Rev. Genet. 2

(2001) 659–668.

[12] N. Billinton, M.G. Barker, C.E. Michel, A.W. Knight, W.D. Heyer, N.J. God-

dard, P.R. Fielden, R.M. Walmsley, Development of a green fluorescent protein
reporter for a yeast genotoxicity biosensor, Biosens. Bioelectron. 13 (1998) 831–
838.

[13] A.W. Knight, N. Billinton, P.A. Cahill, A. Scott, J.S. Harvey, K.J. Roberts, D.J. Tweats,

P.O. Keenan, R.M. Walmsley, An analysis of results from 305 compounds tested
with the yeast RAD54-GFP genotoxicity assay (GreenScreen GC) including rela-
tive predictivity of regulatory tests and rodent carcinogenesis and performance
with autofluorescent and coloured compounds, Mutagenesis (2007).

[14] A.W. Knight, P.O. Keenan, N.J. Goddard, P.R. Fielden, R.M. Walmsley, A yeast-

based cytotoxicity and genotoxicity assay for environmental monitoring using
novel portable instrumentation, J. Environ. Monit. 6 (2004) 71–79.

[15] R.M. Walmsley, N. Billinton, W.D. Heyer, Green fluorescent protein as a reporter

for the DNA damage-induced gene RAD54 in Saccharomyces cerevisiae, Yeast 13
(1997) 1535–1545.

[16] B.O. Krogh, L.S. Symington, Recombination proteins in yeast, Annu. Rev. Genet.

38 (2004) 233–271.

[17] E.L. Ivanov, J.E. Haber, DNA repair: RAD alert, Curr. Biol. 7 (1997) R492–R495.
[18] N. Billinton, A.W. Knight, Seeing the wood through the trees: a review of

techniques for distinguishing green fluorescent protein from endogenous aut-
ofluorescence, Anal. Biochem. 291 (2001) 175–197.

[19] D.S. McNabb, R. Reed, R.A. Marciniak, Dual luciferase assay system for rapid

assessment of gene expression in Saccharomyces cerevisiae, Eukaryot. Cell 4
(2005) 1539–1549.

X. Liu et al. / Mutation Research 653 (2008) 63–69

69

[20] B. Clever, H. Interthal, J. Schmuckli-Maurer, J. King, M. Sigrist, W.D. Heyer,

Recombinational repair in yeast: functional interactions between Rad51 and
Rad54 proteins, Embo J. 16 (1997) 2535–2544.

[21] M.A. Resnick, P. Martin, The repair of double-strand breaks in the nuclear DNA

of Saccharomyces cerevisiae and its genetic control, Mol. Gen. Genet. 143 (1976)
119–129.

[22] A. Shinohara, S. Gasior, T. Ogawa, N. Kleckner, D.K. Bishop, Saccharomyces cere-

visiae recA homologues RAD51 and DMC1 have both distinct and overlapping
roles in meiotic recombination, Genes Cells 2 (1997) 615–629.

[23] M.U. Keller-Seitz, U. Certa, C. Sengstag, F.E. Wurgler, M. Sun, M. Fasullo, Tran-

scriptional response of yeast to aflatoxin B1: recombinational repair involving
RAD51 and RAD1, Mol. Biol. Cell 15 (2004) 4321–4336.

[24] U.S. Government Printing Office, NIH Guidelines for the Laboratory Use of

Chemical Carcinogens, U.S. Government Printing Office, Washington, DC, 1981.

[25] A.M. Myers, A. Tzagoloff, D.M. Kinney, C.J. Lusty, Yeast shuttle and integrative

vectors with multiple cloning sites suitable for construction of lacZ fusions,
Gene 45 (1986) 299–310.

[26] R.D. Gietz, R.A. Woods, Transformation of yeast by lithium acetate/single-

stranded carrier DNA/polyethylene glycol method, Methods Enzymol. 350
(2002) 87–96.

[27] F. Sherman, Getting started with yeast, Methods Enzymol. 350 (2002) 3–41.
[28] K. Mortelmans, E. Zeiger, The Ames Salmonella/microsome mutagenicity assay,

Mutat. Res. 455 (2000) 29–60.

[29] P.A. Cahill, A.W. Knight, N. Billinton, M.G. Barker, L. Walsh, P.O. Keenan, C.V.

Williams, D.J. Tweats, R.M. Walmsley, The GreenScreen genotoxicity assay: a
screening validation programme, Mutagenesis 19 (2004) 105–119.

[30] M. Chang, M. Bellaoui, C. Boone, G.W. Brown, A genome-wide screen for methyl

methanesulfonate-sensitive mutants reveals genes required for S phase pro-
gression in the presence of DNA damage, Proc. Natl. Acad. Sci. U.S.A. 99 (2002)
16934–16939.

[31] P.J. Lupardus, T. Byun, M.C. Yee, M. Hekmat-Nejad, K.A. Cimprich, A requirement

for replication in activation of the ATR-dependent DNA damage checkpoint,
Genes Dev. 16 (2002) 2327–2332.

[32] D. Kirkland, M. Aardema, L. Henderson, L. Muller, Evaluation of the ability of a

battery of three in vitro genotoxicity tests to discriminate rodent carcinogens
and non-carcinogens. Part I. Sensitivity, specificity and relative predictivity,
Mutat. Res. 584 (2005) 1–256.

[33] D. Kirkland, M. Aardema, L. Muller, H. Makoto, Evaluation of the ability of a bat-

tery of three in vitro genotoxicity tests to discriminate rodent carcinogens and
non-carcinogens. Part II. Further analysis of mammalian cell results, relative
predictivity and tumour profiles, Mutat. Res. 608 (2006) 29–42.

[34] E.L. Perkins, J.F. Sterling, V.I. Hashem, M.A. Resnick, Yeast and human genes that

affect the Escherichia coli SOS response, Proc. Natl. Acad. Sci. U.S.A. 96 (1999)
2204–2209.

[35] D.A. Bressan, B.K. Baxter, J.H. Petrini, The Mre11-Rad50-Xrs2 protein complex

facilitates homologous recombination-based double-strand break repair in Sac-
charomyces cerevisiae, Mol. Cell Biol. 19 (1999) 7681–7687.

[36] R. Arai, Y. Makita, Y. Oda, T. Nagamune, Construction of green fluorescent pro-

tein reporter genes for genotoxicity test (SOS/umu-test) and improvement of
mutagen-sensitivity, J. Biosci. Bioeng. 92 (2001) 301–304.

[37] R.J. Mitchell, M.B. Gu, An Escherichia coli biosensor capable of detecting both

genotoxic and oxidative damage, Appl. Microbiol. Biotechnol. 64 (2004)46–52.

[38] S.P. Jackson, Colworth Medal lecture. Detection, repair and signalling of DNA

double-strand breaks, Biochem. Soc. Trans. 27 (1999) 1–13.

[39] J. Thacker, The nature of mutants induced by ionising radiation in cultured

hamster cells. Part III. Molecular characterization of HPRT-deficient mutants
induced by gamma-rays or alpha-particles showing that the majority have
deletions of all or part of the hprt gene, Mutat. Res. 160 (1986) 267–275.

[40] P.A. Jeggo, Identification of genes involved in repair of DNA double-strand

breaks in mammalian cells, Radiat. Res. 150 (1998) S80–S91.

[41] R. Kanaar, J.H. Hoeijmakers, D.C. van Gent, Molecular mechanisms of DNA dou-

ble strand break repair, Trends Cell. Biol. 8 (1998) 483–489.

[42] J.H. New, T. Sugiyama, E. Zaitseva, S.C. Kowalczykowski, Rad52 protein stim-

ulates DNA strand exchange by Rad51 and replication protein A, Nature 391
(1998) 407–410.

[43] Z. Dudasova, A. Dudas, A. Alemayehu, D. Vlasakova, E. Markova, M. Chovanec,

V. Vlckova, J. Brozmanova, Disruption of the RAD51 gene sensitizes S. cerevisiae

cells to the toxic and mutagenic effects of hydrogen peroxide, Folia Microbiol.
(Praha) 49 (2004) 259–264.

[44] L.H. Jensen, M. Dejligbjerg, L.T. Hansen, M. Grauslund, P.B. Jensen, M.

Sehested, Characterisation of cytotoxicity and DNA damage induced by
the topoisomerase II-directed bisdioxopiperazine anti-cancer agent ICRF-187
(dexrazoxane) in yeast and mammalian cells, BMC Pharmacol. 4 (2004) 31.

[45] O.H. Iversen, Urethan (ethyl carbamate) alone is carcinogenic for mouse skin,

Carcinogenesis 5 (1984) 911–915.

[46] G. Spagnuolo, C. Mauro, A. Leonardi, M. Santillo, R. Paterno, H. Schweikl, E.V.

Avvedimento, S. Rengo, NF-kappaB protection against apoptosis induced by
HEMA, J. Dent. Res. 83 (2004) 837–842.

[47] S. Sato, M. Taketomi, T. Morita, Simplified mouse peripheral reticulocyte

micronucleus test with dimethylnitrosamine, Mutat. Res. 278 (1992) 103–107.

[48] G.C. Jagetia, P. Jyothi, Enhancement of micronuclei frequency by vindesine in

mouse bone marrow, Mutat. Res. 388 (1997) 1–5.

[49] H. Tinwell, P.A. Lefevre, J. Ashby, Relative activities of methyl methane-

sulphonate (MMS) as a genotoxin, clastogen and gene mutagen to the liver and
bone marrow of MutaMouse mice, Environ. Mol. Mutagen. 32 (1998) 163–172.

[50] I. Valentin-Severin, V. Thybaud, A.M. Le Bon, J.C. Lhuguenot, M.C. Chagnon, The

autoradiographic test for unscheduled DNA synthesis: a sensitive assay for the
detection of DNA repair in the HepG2 cell line, Mutat. Res. 559 (2004) 211–
217.

[51] A.M. Hooker, R. Horne, A.A. Morley, P.J. Sykes, Dose-dependent increase or

decrease of somatic intrachromosomal recombination produced by etoposide,
Mutat. Res. 500 (2002) 117–124.

[52] I.M. Gavin, S.M. Melnik, N.P. Yurina, M.I. Khabarova, S.G. Bavykin, Zero-length

protein-nucleic acid crosslinking by radical-generating coordination com-
plexes as a probe for analysis of protein-DNA interactions in vitro and in vivo,
Anal. Biochem. 263 (1998) 26–30.

[53] J.A. Bond, A.P. Li, Rat nasal tissue activation of benzo(a)pyrene and 2-

aminoanthracene to mutagens in Salmonella typhimurium, Environ. Mutagen.
5 (1983) 311–318.

[54] K.L. Witt, J.B. Bishop, Mutagenicity of anticancer drugs in mammalian germ

cells, Mutat. Res. 355 (1996) 209–234.

[55] S.M. Galloway, J.E. Miller, M.J. Armstrong, C.L. Bean, T.R. Skopek, W.W. Nichols,

DNA synthesis inhibition as an indirect mechanism of chromosome aberra-
tions: comparison of DNA-reactive and non-DNA-reactive clastogens, Mutat.
Res. 400 (1998) 169–186.

[56] O. Olsen, X. Wang, D. von, Wettstein, Sodium azide mutagenesis: preferential

generation of A.T–>G.C transitions in the barley Ant18 gene, Proc. Natl. Acad.
Sci U.S.A. 90 (1993) 8043–8047.

[57] G.S. Lee, K.S. Blonsky, D.L. Van On, E.A. Savage, A.R. Morgan, R.C. von, Borstel,

Base alterations in yeast induced by alkylating agents with differing Swain-
Scott substrate constants, J. Mol. Biol. 223 (1992) 617–626.

[58] D. Burnouf, M. Duane, R.P. Fuchs, Spectrum of cisplatin-induced mutations in

Escherichia coli, Proc. Natl. Acad. Sci. U.S.A. 84 (1987) 3758–3762.

[59] A. Hartwig, U.D. Groblinghoff, D. Beyersmann, A.T. Natarajan, R. Filon, L.H.

Mullenders, Interaction of arsenic(III) with nucleotide excision repair in UV-
irradiated human fibroblasts, Carcinogenesis 18 (1997) 399–405.

[60] J.M. Kang, N.M. Iovine, M.J. Blaser, A paradigm for direct stress-induced muta-

tion in prokaryotes, FASEB J. 20 (2006) 2476–2485.

[61] J.E. Ambler, R.J. Pinney, Positive R plasmid mutator effect on chromosomal

mutation to nalidixic acid resistance in nalidixic acid-exposed cultures of
Escherichia coli, J. Antimicrob. Chemother. 35 (1995) 603–609.

[62] X.S. Cui, L.C. Eriksson, L. Moller, Formation and persistence of DNA adducts

during and after a long-term administration of 2-nitrofluorene, Mutat. Res.
442 (1999) 9–18.

[63] J.B. Mailhes, M.J. Carabatsos, D. Young, S.N. London, M. Bell, D.F. Albertini, Taxol-

induced meiotic maturation delay, spindle defects, and aneuploidy in mouse
oocytes and zygotes, Mutat. Res. 423 (1999) 79–90.

[64] W.F. Benedict, M. Karon, Chromatid breakage: cytosine arabinoside-induced

lesions inhibited by ultraviolet irradiation, Science 171 (1971) 680–682.

[65] G.C. Kundu, J.R. Schullek, I.B. Wilson, The alkylating properties of chlorambucil,

Pharmacol. Biochem. Behav. 49 (1994) 621–624.

[66] S. Pincus, D. Rekosh, The inhibition of both initiation and elongation of aden-

ovirus DNA replication by actinomycin D, Biochem. Biophys. Res. Commun. 121
(1984) 1033–1041.