Evaluation of the sensor for detection of 8-oxoguanine in plasmid DNA

We next sought to evaluate our sensor for detecting 8-oxoG in the context of a much larger DNA target that would better mimic genomic DNA. Towards this goal we chose to utilize a plasmid DNA target containing multiple sites of methylated cytosine. To generate a plasmid target, a high level of CpG methylation must first be achieved since localization of the sensor to DNA is mediated by a domain that recognizes methylation. Because typical laboratory strains of bacteria lack an endogenous methylase specific for generating 5-methylcytosine in CpG dinucleotides, we exogenously methylated a 7667 bp pETDuet-derived plasmid using M.SssI CpG methyltransferase, creating pETm. The degree of methylation was assessed by exposure to a methylation-sensitive restriction enzyme, BstUI, which cleaves only non-methylated 5'-CGCG sites. Only the non-methylated plasmid was digested by BstUI, while the methylated plasmid was fully protected from digestion (Figure 4A, compare lanes 2 and 3). Next, the methylated plasmid was exposed to oxidizing conditions consisting of 1 mM H₂O₂ and 30 μM CuCl₂. We produced the optimized split-protein biosensors, CLuciferase-OGG1 (18 AA linker) and MBD1-NLuciferase, in a cell-free translation system, followed by addition of oxidized or non-oxidized pETm. Upon incubation of 8 ng of oxidized pETm with the split-luciferase constructs, a 6-fold change in luminescent signal over untreated DNA was achieved (Figure 4B). This gain in signal in an oxidation- and methylation-dependent fashion, confirmed that the 8-oxoguanine sensor may potentially be used for detection of DNA oxidation in a genomic target with extensive methylation at CpG sites.

Open in a separate window

Figure 4

Detection of oxidation in plasmid DNA using the 8-oxoG sensor. (A) Exogenously methylated plasmid DNA (pETm) was compared to the non-methylated plasmid (pET) by restriction enzyme analysis using BstUI, which cleaves non-methylated 5'-CGCG sites. (B) The CLuciferase-OGG1 and MBD1-NLuciferase sensors were incubated with 8 ng of pETm that was exposed to oxidizing conditions (Ox.) of 30 μM CuCl₂ and 1 mM H₂O₂ or non-oxidizing conditions of 30 μM CuCl₂ only (No Ox.).

Application of the sensor for detection of 8-oxoguanine in mammalian DNA

We next turned toward confirming the broader utility of our sensor by evaluating its ability to directly detect the oxidation-dependent modification of natively methylated mammalian DNA. We isolated genomic DNA from HeLa cultures and then exposed it to 1 mM H₂O₂ in the presence of 30 μM CuCl₂ to generate 8-oxoG lesions. In the presence of our split proteins, the oxidized HeLa DNA (50 ng) generated a 27-fold signal induction over the non-treated equivalent, thus confirming the utility of our sensor for detection of lesions in mammalian genomic DNA (Figure 5A). Importantly, evaluation of oxidized HeLa DNA using a split-luciferase based methylation sensor did not result in an observable decrease in luminescence, indicating that oxidation at guanines in mCpG sites does not significantly affect MBD1 binding (Supplementary Information).⁵⁴

Open in a separate window

Figure 5

Detection of oxidation in HeLa DNA using the 8-oxoG sensor. (A) The CLuciferase-OGG1 and MBD1-NLuciferase sensors were incubated with 50 ng of DNA isolated from HeLa cells that was exposed to oxidizing conditions (Ox.) of 30 μM CuCl₂ and 1 mM H₂O₂ for 30 min or non-oxidizing conditions of 30 μM CuCl₂ only (No Ox.). (B) The sensor pair was incubated with 50 ng of HeLa DNA that was treated for 30 min with 1 mM H₂O₂ in the presence of 15, 30, or 60 μM CuCl₂. (C) The sensor pair was incubated with 50 ng of HeLa DNA that was treated with 1 mM H₂O₂ and 60 μM CuCl₂ for 10, 20 or 30 min.

In order to probe oxidation-dependent changes in signal, we next evaluated the extent of DNA damage induced by varying the CuCl₂ concentration (15, 30, or 60 μM) present during treatment of HeLa DNA with 1 mM H₂O₂ for 30 min. The damaged HeLa DNA was exposed to our translated split-protein sensors. A CuCl₂ dose-dependent increase in luminescence was observed, wherein the 15, 30, and 60 μM CuCl₂ treated oxidized samples provided 13-, 27-, and 32-fold signal over the non-oxidized equivalents (Figure 5B). The effect of varying the H₂O₂ concentration (0.5 mM, 1 mM, and 1.5 mM) was also investigated, although this did not have a great effect on sensor response at a given concentration of CuCl₂ (Supplementary Information). Finally, the exposure time for the 1 mM H₂O₂ and 60 μM CuCl₂ treatment was varied (t = 10, 20, and 30 min) to identify the final optimized conditions. Equivalent signals were observed for the 10 and 20 min treatments (~37-fold), while a decrease in signal to 31-fold was observed as the time of oxidation was increased to 30 min, possibly indicating the formation of alternate oxidation products (for example, further oxidation of 8-oxoguanine⁵⁵ or increased strand breakage⁵¹) that are not recognized by OGG1 (Figure 5C). At times longer than 2 h, the 8-oxoG sensor no longer bound the oxidized target with greater than 2-fold preference over the non-oxidized equivalent (Supplementary Information). Gel electrophoresis of the oxidized DNA revealed a direct correlation between the time of oxidation and increased degradation of the DNA target, possibly preventing OGG1 binding (Supplementary Information). Finally, we chose to use an independent enzyme-linked immunosorbent assay (ELISA), which was more sensitive than LC-MS, for quantifying 8-oxoguanine from hydrolyzed DNA. We directly confirmed the presence of 8-oxoG lesions using ELISA of hydrolyzed nucleosides from both oxidized and non-oxidized HeLa DNA. By comparison to the ELISA standard curve utilizing pure 8-oxoguanosine, it was determined that DNA exposed to oxidizing conditions of 1 mM H₂O₂ and 60 μM CuCl₂ for 10 min yields ~ 1200 8-oxoG lesions in 10⁶ DNA bases (Supplementary Information), which is consistent with results from previous studies using liquid chromatography with electrochemical detection (LC-ECD).⁵⁶ The equivalent non-oxidized samples did not yield any 8-oxoguanosine detectable in the ELISA. Considering the proportion of 8-oxoG lesions generated using our optimized oxidation conditions, the 50 ng DNA sample evaluated with our split-protein biosensors contains ~ 200 fmol of 8-oxoG. Further confirmation of 8-oxoG in the oxidized HeLa DNA by LC-MS was not feasible due to limitations in instrument sensitivity of ~ 30 pmol 8-oxoguanosine. These results illustrate the ability of our sensors to report on the relative degree of DNA damage generated by modulating the damage-inducing conditions.

Initial design and validation of a split-luciferase biosensor for UV photodamage

To further investigate the generality of our DNA damage sensor platform and to expand the repertoire of detectable lesions, we next designed a biosensor for detection of UV-induced photoproducts including cyclobutane pyrimidine dimers (CPDs) and 6–4 pyrimidine-pyrimidone photoproducts (6-4PPs). If left unrepaired these bulky DNA adducts can cause base misincorporation or stalled DNA replication.⁵⁷ DNA photoproducts are reported to be initially recognized in the global genome nucleotide excision repair (GG-NER) pathway by the DDB1-DDB2 complex,^58–60 which binds 6-4PPs (K_d ~ 63 pM), trans, syn-CPDs (K_d ~ 77 pM), and Dewar thymine dimers (K_d ~ 210 pM),⁶¹ as well as other lesions that cause distortions or flexibility in the DNA.^62–64 The DDB2 monomer has also recently been reported to be capable of independently binding photoproducts in DNA,^65,66 indicating its potential utility in our split-protein reassembly system. For this detection system, we chose to use the Danio rerio DDB2 ortholog of which a high resolution crystal structure has recently been solved.²⁷ The domain utilized herein has 52% identity and 73% similarity to the Homo sapiens protein, and the overlaid structures have a Cα-backbone RMSD of 1.1 Å.

Thus for the design of a UV-induced photoproduct(s) sensor, we attached DDB2 (residues 94–457) to the C-terminus of CLuciferase (residues 398–550) through an 18 AA linker and utilized MBD1-NLuciferase for localization to methylated target DNA (Figure 6A). We initially evaluated our sensors using a 23 bp double-stranded methylated oligonucleotide target 5'-GCGTAmCGTACGCCCACGCCACCG (mC represents 5-methylcytosine), which was exposed to 2 h of UV radiation from a germicidal lamp (peak output at 254 nm). The expected modifications induced by exposure to UVC radiation have been previously described, with 5'-TT and 5'-TC dinucleotides being more photoreactive than 5'-CT and 5'-CC.⁶⁷ In our oligonucleotide target only 5'-CC dinucleotides are present, which, in contrast to the other dipyrimidine sequences, yield the 6–4 photoproduct (~50 % of products) as the predominant lesion over the cyclobutane dimer and the Dewar valence isomer.⁶⁷

Open in a separate window

Figure 6

Split-luciferase biosensor for the detection of UV damage in DNA. (A) Cartoon representation showing CFluc-DDB2 and MBD1-NFluc bound to a DNA target containing a 6-4 pyrimidine-pyrimidone photoproduct adjacent to a methylated CpG dinucleotide. (B) The sensor pair was incubated with 7.1 nM of methylated or non-methylated oligonucleotides (inset, site of methylation shown in red) that were exposed to 2 h of UV radiation (254 nm) or not exposed to UV radiation (No UV). Results are the average of two independent trials, with luminescence readings presented as relative to the No UV control. (C) Two sets of UV damage sensors, CFluc-DDB2 with MBD1-NFluc and CFluc-MBD1 with DDB2-NFluc, were compared to evaluate the effect of fusion site on luciferase reassembly. (D) Each set of biosensors shown in (C) was reassembled in the presence of 7.1 nM of methylated oligonucleotide exposed to 2 h of UV radiation.

Therefore, we expect a high yield of cytosine 6–4 adducts in the oligonucleotide target. We generated our split proteins, CLuciferase-DDB2 and MBD1-NLuciferase, in a cell-free translation system, followed by addition of the UV-irradiated oligonucleotide. Incubation with 7.1 nM UV-induced target resulted in a 45-fold enhancement of reassembled luciferase activity, establishing a DDB2-specific interaction with irradiated DNA (Figure 6B). Additionally, signal was not observed for the non-methylated oligonucleotides, confirming the requirement for methylation in the target for MBD1-mediated sensor recognition.

We next pursued optimization of the sensor architecture by altering the sites of attachment of the DDB2 and MBD1 domains, creating CLuciferase-MBD1 and DDB2-NLuciferase (Figure 6C). We produced the two complementary sets of biosensors in a cell-free translation system, and then added 7.1 nM of UV or non-UV irradiated oligonucleotides. Despite similar relative signals for the two pairs (27-fold vs. 22-fold), we found that the original biosensor pair, CLuciferase-DDB2 with MBD1-NLuciferase, conferred a significantly higher absolute signal (42-fold) than the newly generated CLuciferase-MBD1 and DDB2-NLuciferase pair (Figure 6D). Therefore, the site of attachment of the DDB2 domain has a significant effect on reassembly efficiency in this case as compared to OGG1, and the original sensor geometry was utilized in subsequent experiments.

Detection of UV damage in plasmid DNA

We next sought to evaluate our sensors for detection of UV damage in exogenously methylated plasmid DNA. To generate UV-induced lesions, we exposed the exogenously methylated plasmid target, pETm, to 2 h UVC light from a germicidal lamp (peak output at 254 nm). To directly confirm the presence of UV-photoproducts in pETm, the UV-treated and untreated plasmids were incubated with MseI, which cleaves the sequence 5'-TTAA between the adjacent thymines (Supplementary Information), while the presence of thymine dimers prevents MseI endonuclease activity. Cleavage was only observed with the untreated plasmid, while the UV-treated plasmid was protected from nuclease cleavage by the formation of pyrimidine dimers (Figure 7A, compare lanes 2 and 4). In order to test the ability of our sensors to detect this UV-induced damage, we expressed the split proteins, CLuciferase-DDB2 and MBD1-NLuciferase, in a cell-free translation system, followed by addition of the UV-irradiated pETm target. We observed a 200-fold increase in signal when compared to non-irradiated target (8 ng), indicating detection of a significant gain in pyrimidine dimers (Figure 7B). Detection of UV damage in HeLa DNA. We next assessed our UV damage sensor in the context of mammalian DNA. We isolated natively methylated genomic DNA from HeLa cells, followed by a 2 h exposure to UVC light. When 50 ng of the UV-treated HeLa DNA was added to the sensors, we observed a ~100-fold increase in signal when compared to an equivalent amount of non-irradiated target (Figure 8A). It has been previously shown that 5-methylcytosine has a reactivity similar to unmodified cytosine for forming cyclobutane pyrimidine dimers upon exposure to UVC (254 nm) radiation.⁶⁸ Therefore, sequences such as 5'-TmCG and 5'-CmCG may form cyclobutane dimers. However, incubation of the HeLa DNA targets with our methylation-specific biosensor revealed no change in luminescent signal upon UV-irradiation, indicating that MBD1 binding is not significantly compromised by UV treatment of the DNA target (Supplementary Information). Given the sensitivity of the DDB2 domain, a titration of UV-treated HeLa DNA revealed a linear correlation within the range tested (2.5–12.5 ng), with a detection limit of at least 2.5 ng, which was detectable over the average background signal plus three standard deviations (99% confidence level) (Figure 8B).

Open in a separate window

Figure 7

Detection of UV photoproducts in plasmid DNA using the UV damage sensor. (A) Methylated plasmid DNA, pETm, was treated with UVC light (h_ν) for 2 h or not exposed to UV light (No h_ν). Irradiated and non-irradiated pETm were incubated with the endonuclease MseI, which is inhibited by the formation of thymine dimers in its recognition sequence (5'-TTAA). (B) The CLuciferase-DDB2 and MBD1-NLuciferase sensor system was incubated with 8 ng of pETm treated with 2 h of UV radiation or no UV.

Open in a separate window

Figure 8

Detection of UV photoproducts in mammalian DNA using the UV damage sensor. (A) The CLuciferase-DDB2 and MBD1-NLuciferase sensor system was incubated with 50 ng HeLa DNA that was treated with 2 h UVC with a peak output at 254 nm or No UV. (B) The sensor pair was incubated with 12.5, 8.3, 5.6, 3.7, or 2.5 ng of HeLa DNA treated as described in (A), yielding a linear response. (C) (inset) UVC irradiation of HeLa cells results in the formation of DNA photoproducts, such as the 6–4 photoproduct (6–4 PP). The sensor pair was incubated with 50 ng DNA isolated from HeLa cells treated with 10 min of UV exposure or No UV.

Detection of UV damage induced in mammalian cells

Following establishment of our UV damage sensors for the detection of UV-induced photoproducts in exogenously damaged DNA, we proceeded to evaluate our split proteins for detection of DNA damage initiated in live cells. Because endogenous damage may result in cellular toxicity that could prevent further analysis of DNA samples, we first sought to determine an appropriate length of UV exposure time during which cell viability was not greatly compromised. We exposed cells to UV radiation for varying lengths of time and evaluated cell survival using the formazan-based MTT assay. Only ~70% of cells remained viable after a 2 h exposure, which reflects typically observed toxicity associated with longer or more intense UV exposure (Supplementary Information).^69,70 However, a 10 min exposure resulted in greater than 95% viability. Therefore, to determine if our sensors were capable of detecting UV damage in DNA directly isolated from cultures, we exposed HeLa cells to UV radiation for only 10 min, followed by immediate harvesting of genomic DNA to limit repair of DNA lesions. Isolated DNA was quantified, and 50 ng was incubated with the UV damage sensor, resulting in a 25-fold signal over background levels of damage (Figure 8C). To further validate the ability of our sensor to detect DNA damage induced in cell populations, various UV exposure times, ranging from 15 min to 1 min, were investigated (Supplementary Information). Thus, these results demonstrate that this split-protein sensor can be utilized for the direct detection of UV lesions in genomic DNA induced in a cellular context using a method that does not require extensive or destructive sample processing following isolation of genomic DNA.

Cell culture

HeLa cells were maintained at 37 °C and 5% CO₂ in 90% DMEM/F12 1:1 media (Lonza) supplemented with 10% FBS (Lonza), penicillin-streptomycin (Mediatech), and ampotericin B (JR Scientific). To obtain mammalian genomic DNA, HeLa DNA was isolated from 10⁶ trypsinized cells using the Wizard® Genomic DNA Purification Kit (Promega) according to the manufacturer’s instructions.

Cloning

The luciferase halves utilized in all biosensors were split as previously described.^33,39 DNA encoding the damage detection domains, Homo sapiens OGG1 (residues 12–325) and Danio rerio DDB2 (residues 94–457), was amplified from existing plasmids obtained from Open Biosystems (I.M.A.G.E. clone IDs: 3350168 and 7402966, respectively). The resulting PCR products were introduced into a pcDNA3.1(+) vector (Invitrogen) containing CLuciferase (residues 398–505) followed by an 18 AA linker, creating pcDNA3.1(+)-CLuciferase-OGG1 and pcDNA3.1(+)-CLuciferase-DDB2. Site-directed mutagenesis using the QuikChange II Site-Directed Mutagenesis Kit (Stratagene) was carried out on pcDNA3.1(+)-CLuciferase-OGG1 to generate the corresponding plasmid containing the catalytically inactive OGG1(K249Q) mutant. Cloning of the other sensor architectures is described in the Supplementary Information. The pcDNA3.1(+)-MBD1-NLuciferase construct was generated by amplifying MBD1-NLuciferase (MBD1 synthesized by GenScript) from an existing plasmid and inserting it into the pcDNA3.1(+) vector. All sequences were confirmed using dideoxynucleotide sequencing.

Oligonucleotide target preparation

All oligonucleotide targets were obtained HPLC purified from Integrated DNA Technologies and are listed in Table S1 (Supplementary Information). All designed DNA targets were annealed in 1× annealing buffer (10 mM Tris-HCl, pH 7.9, 150 mM NaCl, 10 mM MgCl₂, 1mM DTT) using the following procedure: heating to 95 °C for 7 min, cooling to 56 °C at a rate of 0.1 °C sec⁻¹, equilibrating at 56 °C for 5 min, cooling to 25 °C at a rate of 0.1 °C sec⁻¹ and finally equilibrating at 25 °C for 10 min using a Labnet Multi Gene II thermocycler.

Generation of DNA oxidation IN VITRO

A pETDuet-derived plasmid (7667 bp) was fully methylated, generating pETm, by incubation with S-adenosylmethionine and M.SssI CpG methyltransferase (New England Biolabs) according to the manufacturer’s protocol. The extent of protection was determined by exposure to the methylation sensitive endonuclease BstUI. To introduce oxidation, 708 nM annealed oligonucleotide, 8 ng/μL pETm, or 50 ng/μL HeLa DNA was treated with 1 mM H₂O₂ in the presence of 30 μM CuCl₂ for 30 min at RT, followed by quenching with 1 mM EDTA. To assess the effect of Cu²⁺ concentration on induction of base damage, 50 ng/μL HeLa DNA was treated with 1 mM H₂O₂ in the presence of 15, 30, or 60 μM CuCl₂ for 30 min, followed by quenching with 1 mM EDTA. To assess the effect of oxidation time on induction of base damage, 50 ng/μL HeLa DNA was treated with 60 μM CuCl₂ and 1 mM H₂O₂ for 10, 20, or 30 min, followed by quenching with 1 mM EDTA. In all cases control reactions were carried out with CuCl₂ in the absence of H₂O₂, followed by addition of 1 mM EDTA. The DNA was analyzed without further purification.

Detection of DNA oxidation

Preparation of mRNA is described in the Supplementary Information. The mRNAs encoding CLuciferase-OGG1(K249Q) and MBD1-NLuciferase were translated in the Flexi Rabbit Reticulocyte Lysate System (Promega), consisting of 25 μL reactions prepared according to the manufacturer’s instructions. A typical reaction was performed at 30 °C for 1.5 h and contained 0.2 pmol of each mRNA transcript. Following translation, 1.25 μL of 708 nM annealed oligonucleotide, 8 ng/μL pETm, or 50 ng/μL HeLa DNA target was added to 23.75 μL of the translation and binding was allowed to occur for 1 h at 4 °C. Activity was monitored as a luminescent signal produced upon addition of luminescence reagent, where 20 μL of each translation was added to 80 μL of reagent. Luminescent readings were acquired 1 min after mixing using a Turner TD-20e Luminometer with a 10 s integration time. Results are presented as the relative average of two independent experiments.

Optimization of OGG1 sensor architecture

The mRNAs encoding the three sensor pairs (CLuciferase-18 AA-OGG1 + MBD1-NLuciferase, CLuciferase-MBD1 + OGG1-15 AA-NLuciferase, and CLuciferase-33AA-OGG1 + MBD1-NLuciferase) were translated in the Flexi Rabbit Reticulocyte Lysate System (Promega), consisting of 25 μL reactions prepared according to the manufacturer’s instructions. A typical reaction was performed at 30 °C for 1.5 h and contained 0.2 pmol of a pair of mRNA transcripts. Following translation, 1.25 μL of 708 nM oxidized oligonucleotide (23 bp) was added to 23.75 μL of the translation, and binding was allowed to occur for 1 h at 4 °C. Activity was monitored as a luminescent signal produced upon addition of luminescence reagent, where 20 μL of each translation was added to 80 μL of reagent. Luminescent readings were acquired 1 min after mixing using a Turner TD-20e Luminometer with a 10 s integration time. Results are presented as the average readings for two independent experiments.

Evaluation of target length dependence on split-luciferase reassembly

The mRNAs encoding the CLuciferase-Zif268 and MBD1-NLuciferase were translated in the Flexi Rabbit Reticulocyte Lysate System (Promega), consisting of 25 μL reactions prepared according to the manufacturer’s instructions. A typical reaction was performed at 30 °C for 1.5 h and contained 10 μM ZnCl₂, 1 pmol of each mRNA transcript, and 1.25 μL of 708 nM methylated oligonucleotide with 0, 1, 2, 3, or 10 bp spacing between the methylation site and the zinc finger binding site. Activity was monitored as a luminescent signal produced upon addition of luminescence reagent, where 20 μL of each translation was added to 80 μL of reagent. Luminescent readings were acquired 1 min after mixing using a Turner TD-20e Luminometer with a 10 s integration time. Results are presented as the relative average of two independent experiments.

Induction of UV damage

For UV damage induction, 708 nM annealed oligonucleotide, 8 ng/μL pETm, or 362 ng/μL HeLa DNA was introduced into a quartz cuvette and exposed for 2 h to a germicidal UV lamp (LightSources, Inc.) with a UV fluency of 240 μW/cm² at 254 nm according to the manufacturer. The DNA was collected and analyzed without further preparation. To confirm the presence of thymine dimer formation in the UV-irradiated pETm, a nuclease protection assay was performed. Plasmid was incubated with the MseI restriction endonuclease, which cleaves 5'-TTAA sequences between adjacent thymines, followed by agarose gel electrophoresis. To generate UV-induced lesions in vivo, HeLa cells were plated in complete medium at 10⁶ cells per 60 mm dish 24 h prior to treatment. Cells were exposed for 10 min, followed by immediate trypsinization and isolation of genomic DNA using the Wizard® Genomic DNA Purification Kit according to the manufacturer’s instructions. To determine cellular toxicity associated with excessive UV damage, 2.4 x 10⁴ cells per well in a 96 well plate were assayed with a methylthiazolyl tetrazolium (MTT)-based in vitro Toxicology Assay Kit (Sigma) after UV exposure for 2 h, 1 h, 10 min, or no UV. MTT formazan absorbance was measured at 570 nm using a Synergy^TM 2 microplate reader (BioTek® Instruments, Inc.), and results are presented as the average of two independent trials.

Detection of UV damage

The mRNAs encoding the split-proteins CLuciferase-DDB2 and MBD1-NLuciferase were translated in the Flexi Rabbit Reticulocyte Lysate System (Promega), consisting of 25 μL reactions prepared according to the manufacturer’s instructions. A typical reaction was performed at 30 °C for 1.5 h and contained 0.2 pmol of each mRNA transcript. Following translation, 1.25 μL of 708 nM UV-treated oligonucleotide, 8 ng/μL UV-treated pETm, 50 ng/μL UV-treated HeLa DNA, or 50 ng/μL HeLa DNA isolated from UV-treated cells was added to 23.75 μL of the translation and binding was allowed to occur for 1 h at 4 °C. To determine the lowest detectable amount of UV-treated HeLa DNA, 1.25 μL of 12.5, 8.3, 5.6, 3.7, or 2.5 ng/μL dilutions of treated or untreated DNA was added to 23.75 μL of the translation. In all cases activity was monitored as a luminescent signal produced upon addition of luminescence reagent, where 20 μL of each translation was added to 80 μL of reagent. Luminescent readings were acquired 1 min after mixing using a Turner TD-20e Luminometer with a 10 s integration time. Results are presented as the relative average of two independent experiments.

We thank D. Piwnica-Worms for the luciferase constructs. This research was supported by R21CA143661 and R01GM077403 (NIH). J. L. F. thanks the Philanthropic Education Organization (P.E.O.) Chapter BB and the University of Arizona TRIF Imaging Fellowship for financial support. Molecular structures were rendered using PyMol; DeLano, W. L. (www.pymol.org).