An in vitro method for detecting genetic toxicity based on inhibition of RNA synthesis by DNA lesions

Previously, we demonstrated that our in vivo method for visualizing transcription in mammalian cells can detect UV- and/or chemically
(e.g., camptothecin, etoposide, 4NQO, cisplatin) induced damage in genomic DNA by
inhibiting RNA polymerase during transcription elongation 20], 22], 23]. Here, we modified this in vivo method to establish a new in vitro method (Fig. 1) using T7 RNAP and DNA templates containing a T7 promoter. RNA polymerase synthesizes
new transcripts from a DNA template, which can be detected by RT-PCR (Fig. 1a). However, RNA polymerase stops at chemically induced DNA lesions, and no products
are detected by RT-PCR (Fig. 1b). First, we tested the detection of UV-induced DNA lesions (e.g., CPD and 6–4 pp
that trigger NER) using conventional PCR. Under these experimental conditions, approximately
12 DNA lesions were expected in plasmid DNA 26]. Although 6–4 pp is more frequently induced than CPD, both lesions are thought to
inhibit mammalian RNA polymerase II transcription almost completely 14], 27]. However, conventional PCR (Fig. 2a) could obscure the difference between UV-irradiated and non-irradiated DNA (Fig. 2b), as Taq DNA polymerase might synthesize DNA from the undamaged region (301 bp) in
the UV-irradiated DNA template and/or beyond UV lesions such as CPD or 6–4 pp in the
PCR cycles 28]. In contrast, T7 RNAP covers the DNA sequence between the T7 promoter and the PCR-detecting
region and shows inefficient RNA primer extension from arrested RNAs in a single reaction
29]. This approach seems to be more suitable for detecting DNA lesions than conventional
PCR as it reveals the effects of DNA lesions on transcription.

Fig. 1. Experimental design. a In the absence of DNA damage, RNA polymerase generates RNA transcripts from DNA templates
(normal transcription). After purifying RNA, RT-PCR is performed, and the PCR products are analyzed. b If chemical substances damage DNA, the resulting lesions inhibit RNA synthesis, as
RNA polymerase cannot synthesize transcripts from damaged templates, and PCR products
will not be detected

Fig. 2. Analysis of PCR products from UV-irradiated DNA templates. a Scheme for detecting DNA damage using PCR products from UV-irradiated (450 J/m
²
) DNA templates. b Agarose gel (1 %) demonstrating PCR products (301 bp) from UV-irradiated DNA templates.
M: size marker. Odd lanes: PCR products from mock DNA templates. Even lanes: PCR products
from UV-irradiated DNA templates. DNA was amplified for 9 (lanes 1 and 2), 12 (lanes 3 and 4), 15 (lanes 5 and 6), 18 (lanes 7 and 8), or 21 (lanes 9 and 10) PCR cycles

When we tested this method to detect transcription inhibition by DNA lesions in vitro (Fig. 3a), we identified newly synthesized RNA transcripts (Fig. 3b) and cDNA products from those transcripts using RT-PCR (Fig. 3c), but we could not detect any PCR products without using reverse transcriptase (Fig. 3c, lanes 11 and 12). As shown in Fig. 3c, the amounts of PCR products (301 bp) from non-irradiated DNA (UV -) were higher
than those from UV-irradiated DNA (UV +) until 15 PCR cycles (lanes 5 and 6), indicating
that UV-induced DNA lesions blocked RNA synthesis by T7 RNAP. However, after 18 PCR
cycles, no difference between RT-PCR products from non-irradiated and UV-irradiated
templates could be detected (Fig. 3c, lanes 7–10). These results indicate that RT-PCR using agarose gel electrophoresis
for detection (Fig. 3a) might not be suitable for determining the effects of UV-induced DNA damage on transcription
because it is necessary to optimize the number of PCR cycles.

Fig. 3. Analysis of PCR products generated by RT-PCR from RNA transcripts of UV-irradiated
DNA templates. a Scheme for detecting transcription inhibition using RNA transcripts from UV-irradiated
(450 J/m
²
) DNA templates. b Agarose gel (1 %) demonstrating RNA transcripts from UV-irradiated DNA templates.
M: size marker. Lane 1: RNA transcripts from mock DNA templates. Lane 2: RNA transcripts
from UV-irradiated DNA templates. c Agarose gel (1 %) demonstrating RT-PCR products (301 bp) from UV-irradiated DNA templates.
M: size marker. Odd lanes: RT-PCR products from mock DNA templates. Even lanes: RT-PCR
products from UV-irradiated DNA templates. DNA was amplified for 9 (lanes 1 and 2), 12 (lanes 3 and 4), 15 (lanes 5 and 6), 18 (lanes 7 and 8), or 21 (lanes 9 and 10) PCR cycles. Lanes 11 and 12: no RT reaction

Therefore, we tested the utility of qPCR for detecting RNA transcription inhibition
by UV-induced DNA lesions (Fig. 4a). qPCR is both powerful and sensitive and is used for a broad range of applications.
Combined with reverse transcription, it can quantify RNA in cells or tissues. To adapt
the new qPCR method, we first assessed primer sets using melting curve analysis (Fig. 4b) and confirmed that these primers generated one PCR product under our experimental
conditions. This primer design is a crucial step because inefficient or non-specific
primer annealing will negatively affect the quality and reliability of the assay.
The amplification plot of the qPCR analysis showed a delay in the accumulation of
qPCR products to later cycles, indicating fewer RNA transcripts from UV-irradiated
templates (Fig. 4c). Normalized to non-irradiated templates, the amount of RNA transcripts from UV-irradiated
templates was markedly decreased by 0.052-fold (Fig. 4d). As expected, qPCR improved detection of the effects of DNA damage on RNA transcription.
Without reverse transcriptase, no specific qPCR products were detected (Fig. 4c and d), confirming the origin of these products from T7 RNAP transcription of UV-irradiated
DNA templates. Considering the melting curves shown in Fig. 4c and d, however, we cannot rule out the possibility that PCR products derived from DNA remaining
after RNA purification. As qPCR proved suitable for detecting damaged DNA templates,
we tried to directly apply this method to UV-irradiated templates (Fig. 5). However, while the amplification plot showed slight differences, the method was
not sensitive enough to significantly detect damage of UV-irradiated DNA templates
(Fig. 5c and d).

Fig. 4. RT-qPCR analysis of transcripts from UV-irradiated DNA templates. a Scheme for detecting transcription inhibition from UV-irradiated (450 J/m
²
) DNA templates. b Melting curve of RT-qPCR products from RNA transcripts of UV-irradiated DNA templates.
Each reaction was run in triplicate. c Amplification plot of RT-qPCR analysis of RNA transcripts of UV-irradiated DNA templates.
Each reaction was run in triplicate. d Relative fold change of transcripts from mock (set as 1.0) and UV-irradiated DNA
templates. Data show the mean of three samples?±?standard deviation (SD)

Fig. 5. qPCR analysis of UV-irradiated DNA templates. a Scheme for detecting DNA damage using qPCR products from UV-irradiated (450 J/m
²
) DNA templates. b Melting curve of qPCR products from UV-irradiated DNA templates. Each reaction was
run in triplicate. c Amplification plot of qPCR analysis of UV-irradiated DNA templates. Each reaction
was run in triplicate. d Relative fold change of PCR products from mock (set as 1.0) and UV-irradiated DNA
templates. Data show the mean of three experiments?±?standard deviation (SD)

Next, we used cisplatin to directly induce DNA adducts 30], 31] as intrastrand or interstrand crosslinks and monoadducts, which interfere with replication
and transcription. Although these adducts are mainly eliminated by NER 32], they are thought to mediate the cytotoxic activity of cisplatin in tumor cells.
Using cisplatin-treated DNA samples, we investigated the generation of PCR products
from damaged templates. Unlike with UV-damaged templates, conventional PCR (Fig. 6a) revealed differences between cisplatin-treated and untreated DNA (Fig. 6b). These results suggest that cisplatin DNA adducts efficiently block DNA synthesis
by Taq DNA polymerase and/or prevent primer annealing. Consistent with this observation,
we obtained similar results using qPCR to detect cisplatin-induced DNA adducts, revealing
a clear difference between damaged and non-damaged DNA (Fig. 7).

Fig. 6. Analysis of PCR products generated from cisplatin-treated DNA templates. a Scheme for detecting DNA damage using PCR products from cisplatin-treated DNA templates
(drug/nucleotide ratio?=?0.005). b Agarose gel (1 %) demonstrating PCR products (301 bp) from cisplatin-treated DNA
templates. M: size marker. Odd lanes: PCR products from mock DNA templates. Even lanes:
PCR products from UV-irradiated DNA templates. DNA was amplified for 9 (lanes 1 and 2), 12 (lanes 3 and 4), 15 (lanes 5 and 6), 18 (lanes 7 and 8), or 21 (lanes 9 and 10) PCR cycles

Fig. 7. qPCR analysis of cisplatin-treated DNA templates. a Scheme for detecting DNA damage using PCR products from cisplatin-treated DNA templates
(drug/nucleotide ratio?=?0.005). b Melting curve of qPCR products from cisplatin-treated DNA templates. Each reaction
was run in triplicate. c Amplification plot of qPCR analysis of cisplatin-treated DNA templates. Each reaction
was run in triplicate. d Relative fold change of PCR products from mock (set as 1.0) and cisplatin-treated
DNA templates. Data show the mean of three experiments?±?standard deviation (SD)

Previous biochemical studies demonstrated that T7 RNAP stalls at cisplatin DNA adducts
33], 34]. Accordingly, when we tested the effects of cisplatin on DNA templates via transcription
and RT-PCR (Fig. 8a), we detected newly synthesized RNA transcripts from untreated DNA templates (Fig. 8b, lane 1) but not from cisplatin-treated templates (Fig. 8b, lane 2) and detected cDNA from untreated DNA templates (Fig. 8c, lanes 3, 5, 7, and 9) but not cisplatin-treated templates (Fig. 8c, lanes 4, 6, 8, and 10). No PCR products were detected in the absence of RT (Fig. 8c, lanes 11 and 12). These results indicate that cisplatin DNA adducts inhibited T7
RNAP transcription initiation and/or elongation. qPCR melting curve analysis (Fig. 9a) indicated that the primer set generated one PCR product after RNA transcription
from cisplatin-treated DNA templates (Fig. 9b), suggesting that the PCR product is specific. The amplification plot (Fig. 9c) and the relative fold change (Fig. 9d) showed little T7 RNAP transcription of cisplatin-treated DNA templates, and insignificant
amounts of PCR products were generated without RT, indicating that this method can
detect the inhibition of RNA synthesis by cisplatin DNA damage.

Fig. 8. Analysis of PCR products generated by RT-PCR from RNA transcripts of cisplatin-treated
DNA templates. a Scheme for detecting transcription inhibition using RNA transcripts from cisplatin-treated
DNA templates (drug/nucleotide ratio?=?0.005). b Agarose gel (1 %) demonstrating RNA transcripts from cisplatin-treated DNA templates.
M: size marker. Lane 1: RNA transcripts from mock DNA templates. Lane 2: RNA transcripts
from cisplatin-treated DNA templates. c Agarose gel (1 %) demonstrating RT-PCR products (301 bp) from cisplatin-treated DNA
templates. M: size marker. Odd lanes: RT-PCR products from mock DNA templates. Even
lanes: RT-PCR products from cisplatin-treated DNA templates. DNA was amplified for
9 (lanes 1 and 2), 12 (lanes 3 and 4), 15 (lanes 5 and 6), 18 (lanes 7 and 8), or 21 (lanes 9 and 10) PCR cycles. Lanes 11 and 12: no RT reaction

Fig. 9. RT-qPCR analysis of transcripts from cisplatin-treated DNA templates. a Scheme for detecting transcription inhibition from cisplatin-treated DNA templates
(drug/nucleotide ratio?=?0.005). b Melting curve of RT-qPCR products from RNA transcripts of cisplatin-treated DNA templates.
Each reaction was run in triplicate. c Amplification plot of RT-qPCR analysis of RNA transcripts of cisplatin-treated DNA
templates. Each reaction was run in triplicate. d Relative fold change of transcripts from mock (set as 1.0) and cisplatin-treated
DNA templates. Data show the mean of three samples?±?standard deviation (SD)