DOI: 10.1002/cbic.201500184
Full Papers
In Vitro Fluorogenic Real-Time Assay of the Repair of Oxidative DNA Damage Sarah K. Edwards,[a] Toshikazu Ono,[a, b] Shenliang Wang,[a] Wei Jiang,[a] Raphael M. Franzini,[a] Jong Wha Jung,[a, c] Ke Min Chan,[a] and Eric T. Kool*[a] The repair of oxidative damage to DNA is essential to avoid mutations that lead to cancer. Oxidized DNA bases, such as 8oxoguanine, are a main source of these mutations, and the enzyme 8-oxoguanine glycosylase 1 (OGG1) is the chief human enzyme that excises 8-oxoguanine from DNA. The activity of OGG1 has been linked to human inflammation responses and to cancer, and researchers are beginning to search for inhibitors of the enzyme. However, measuring the activity of the enzyme typically requires laborious gel-based measurements of radiolabeled DNAs. Here we report the design and properties of fluorogenic probes that directly report on the activity of OGG1 (and its bacterial homologue Fpg) in real time as the
oxidized base is excised. The probes are short, modified DNA oligomers containing fluorescent DNA bases and are designed to utilize 8-oxoguanine itself as a fluorescence quencher. Screening of combinations of fluorophores and 8-oxoguanine revealed two fluorophores, pyrene and tCo, that are strongly quenched by the damaged base. We tested 42 potential probes containing these fluorophores: the optimum probe, OGR1, yields a 60-fold light-up signal in vitro with OGG1 and Fpg. It can report on oxidative repair activity in mammalian cell lysate and with bacterial cells overexpressing a repair enzyme. Such probes might prove useful in quantifying enzyme activity and performing competitive inhibition assays.
Introduction 8-Oxoguanine is an abundant form of oxidative DNA damage that is generated by reactive oxygen species resulting from metabolism, ionizing radiation, and chemicals.[1–5] 8-Oxoguanine in DNA can adopt a syn conformation and pair with adenine during replication,[6, 7] ultimately leading to G-to-T mutations.[8–10] This form of nucleobase damage might be the chief molecular source of oxidation-induced mutagenesis in cells.[11] As protection against the mutagenic potential of 8-oxoguanine, cells have evolved multiple ways to repair this lesion. One of the chief eukaryotic 8-oxoguanine repair enzymes is 8oxoguanine glycosylase 1 (OGG1), which preferentially removes 8-oxoguanine opposite cytosine in double-stranded DNA.[12, 13] OGG1 acts as a glycosylase to excise the 8-oxoguanine base. OGG1 can then act as an AP lyase to performing a b-elimina[a] S. K. Edwards, Dr. T. Ono, Dr. S. Wang, Dr. W. Jiang, Dr. R. M. Franzini, Dr. J. W. Jung, K. M. Chan, Prof. E. T. Kool Department of Chemistry, Stanford University Stanford, CA 94305 (USA) E-mail: [email protected] [b] Dr. T. Ono Present Address: Department of Chemistry and Biochemistry Graduate School of Engineering Center for Molecular Systems (CMS), Kyushu University 744 Motooka, Nishi-ku, Fukuoka, 819-0395 (Japan) [c] Dr. J. W. Jung Present Address: College of Pharmacy Research Institute of Pharmaceutical Sciences Kyungpook National University Daegu, 702-701 (Republic of Korea) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201500184.
ChemBioChem 2015, 16, 1637 – 1646
tion reaction on the 3’-side of the abasic site to generate an a,b-unsaturated aldehyde at the 3’ terminus.[12, 13] However, OGG1’s lyase activity proceeds slowly and is believed not to be relevant in vivo; it has been proposed that the endonuclease APE1 instead cleaves the DNA backbone following 8-oxoguanine excision by OGG1.[14–18] Formamidopyrimidine-DNA glycosylase (Fpg), the functional homologue of OGG1 in Escherichia coli, excises the 8-oxoguanine base and then performs b and d elimination reactions at the abasic site to generate a single nucleotide gap flanked by 3’- and 5’-phosphates.[19–23] Both OGG1 and Fpg prefer double-stranded DNA substrates, but Fpg can also act on single-stranded DNA.[24, 25] E. coli, Saccharomyces cerevisiae, and mice engineered to be deficient in Fpg or OGG1 display increased levels of genomic 8-oxoguanine and increased mutations, thus highlighting the importance of these enzymes in maintaining genome integrity.[26–30] OGG1 mutants with decreased or abolished 8-oxoguanine repair activity have been found in many cancer specimens,[31–39] and lower OGG1 activity has been associated with increased cancer risk.[40–43] In addition to its prominent association with genetic mutations leading to malignancies, altered OGG1 activity has been associated with many other important diseases. Several OGG1 mutants with decreased or abolished activity have been discovered in Alzheimer’s patients.[44, 45] Two OGG1 polymorphisms have been associated with rheumatoid arthritis,[46] and decreased OGG1 activity has been linked with decreased lung inflammation in an allergy model.[47] Finally, altered levels of OGG1 have been associated with type 2 diabetes.[48, 49]
1637
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers A simple method to measure OGG1 activity would help further elucidate these multiple links between OGG1 activity and disease. Such a method would also aid screening for OGG1 inhibitors or activators, which would be useful as drugs and research tools, and assist determining disease prognosis from patient samples. Currently, quick and easy study of OGG1 activity is hindered by the time-consuming and indirect nature of available methods (e.g., PAGE, comet assay, GC/MS, ELISA, HPLC, and qRT-PCR). For example, the strand-nicking assay (one of the most common methods used to measure OGG1 activity) involves radiolabeling an 8-oxoguanine-containing oligonucleotide, incubating it with the sample of interest, and analyzing the sample at various time points by gel electrophoresis and radioactivity measurement. This method requires at least a day and only measures OGG1 activity at discrete, predetermined times. More recently developed approaches using fluorescently labeled oligonucleotides[50–53] give a fluorescent readout of OGG1 activity, but these probes rely on backbone cleavage (which is done slowly by OGG1, and is believed to be performed by a separate enzyme in human cells)[14–17] and subsequent DNA dehybridization to generate a fluorescence signal. Thus, these FRET-based probes do not directly measure the excision of 8-oxoguanine but instead rely on downstream events following the excision. Other DNA-cleaving and DNA-unwinding activities (by various nucleases and helicases) can yield false signals with this design. A split-luciferase sensor that detects 8-oxoguanine in DNA could potentially be used to identify competitive inhibitors of OGG1 or to indirectly measure OGG1 activity,[54] but this sensor measures the binding of OGG1 to 8-oxoguanine rather than OGG1 enzymatic activity. Given the need for a more direct and simple measurement of 8-oxoguanine excision activity, we aimed to develop “lightup” probes that directly measure excision of this oxidized base in real time. By identifying and applying a fluorophore that is strongly quenched by 8-oxoguanine, we have developed a probe that gives a 60-fold fluorescence increase upon reaction with OGG1 or Fpg. Such probes enable direct and specific measurements of OGG1 and Fpg activity in vitro and are expected to be useful in screening polymorphisms, mutations, post-translational modifications, and competitive inhibitors of 8-oxoguanine repair enzyme activity.
Results and Discussion Probe design Our probe design relies on a fluorophore that is quenched specifically by 8-oxoguanine, but becomes emissive following excision of the damaged base. Most desirable would be a fluorophore that is not strongly quenched by native (undamaged) nucleobases, so that a DNA enzyme-binding sequence can be constructed while retaining efficient fluorescence emission. Although the “8-oxo-G clamp” is known to be quenched by 8-oxoguanine,[55–57] this nucleobase analogue pairs with and surrounds the damaged base, and thus might prevent recognition or excision by repair enzymes. To screen for fluorophores that are quenched by 8-oxoguanine, we assembled dinucleotides ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
consisting of a fluorophore deoxyriboside (a nucleoside with a fluorophore replacing the nucleobase) and either 8-oxoguanine or an abasic site mimic (representing the product following 8-oxoguanine excision). (Structures of the dye monomers and a representative dinucleotide are shown in Figure S1 in the Supporting Information.) By comparing the fluorescence emission intensities of the 8-oxoguanine-containing dinucleotides with those of the abasic controls, we found that the fluorophores pyrene, 2-aminopurine, ethenoadenine, pyrrolo-dC, and tCo were quenched two- to 30-fold by 8-oxoguanine, with pyrene and tCo most strongly quenched (Figure 1). Pyrene has been reported to be quenched strongly by thymine and cytosine, and moderately by guanine.[58] Similarly to a previous report of quenching of tCo by guanine, tCo was more strongly quenched by 8-oxoguanine on the 5’-side of the fluorophore than on its 3’-side.[59] tCo is reported not to be strongly quenched by the other natural nucleobases.[59] In contrast to these fluorophores, phenylethynylpyrene, perylene, and QB dye were not quenched (or were poorly quenched) by 8-oxoguanine (Figure 1). Having identified at least two dyes that are strongly quenched by 8-oxoguanine in DNA, we synthesized 42 candidate probes as putative enzymatic substrates that contain 8oxoguanine and one or more of these fluorophores in place of natural nucleobases (Figure S2). Oligonucleotide design was guided by the desire to keep probes short, in order to aid synthesis yield and enhance the likelihood of cellular uptake. Some probes were single-stranded; others were double-stranded with a hairpin design. For the latter, we chose a highly stabilizing GAA loop[60] to enhance thermal stability in short sequences. The probes were prepared by standard phosphoramidite chemistry on an automated DNA synthesizer, purified by HPLC or PAGE, and characterized by MALDI-TOF MS to confirm incorporation of both the damaged nucleobase and the fluorescent deoxyriboside. The probes were screened for increases in fluorescence emission after incubation with recombinant human OGG1 or Fpg (signal enhancements for all probes with the enzymes in Figure S3). We observed some trends with pyrene-containing probes. Firstly, short double-stranded (hairpin) probes were much better substrates than single-stranded probes (compare OGR217 vs. OGR18-32 in Figure S3). Secondly, a fluorophore on the 3’-side of 8-oxoguanine appeared to be more favorable for enzymatic reaction than on the 5’-side (OGR19 vs. OGR21). To several of the pyrene-based probes, we added a second fluorophore as a FRET or exciplex partner, as a possible strategy for yielding longer-wavelength emission. In these cases, quenching by 8-oxoguanine was less efficient than for the single-fluorophore designs, giving high initial fluorescence signals. Many of the pyrene-based probes exhibited low turnover; this might may be a consequence of the bulky, hydrophobic fluorophore, or the affinity of OGG1 for the abasic site included as the pyrene pairing partner.[14, 61] We also observed trends with tCo-containing probes. Adding a second 8-oxoguanine to the other side of tCo did not diminish the initial signal but did decrease the final signal (OGR1 vs. OGR36). MALDI-TOF MS analysis (data not shown)
1638
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 1. Fluorescence quenching (emission spectra) of various fluorophores by 8-oxoguanine. DNA dinucleotides consisted of a fluorescent nucleoside adjac) or an abasic site mimic (c c). Fluorescent bases: A) 2-aminopurine, B) ethenoadenine, C) perylene, D) pyrene, E) pyrrolo-dC, cent to 8-oxoguanine (c F) QB, G) and H) tCo, and I) V. Dinucleotides in F and G have the fluorophore on the 3’-side of 8-oxoguanine or the abasic site mimic. In all other panels the fluorophore is on the 5’-side of 8-oxoguanine or the abasic site mimic. ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1639
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers indicated that only one 8-oxoguanine was excised, presumably rendering the probe a poor substrate for excision of the second 8-oxoguanine and thus keeping the probe in a semiquenched state. Increasing the stem length or removing the loop did not increase probe performance, thus indicating that the hairpin structure did not inhibit the enzyme (OGR1 vs. OGR37, OGR41, and OGR42).
fied the probe following cleavage of the DNA backbone with phosphates at the new 3’ and 5’ termini. These observations are consistent with the known mechanisms of these enzymes (Figure 3). Thus the analysis confirms enzymatic excision of the fluorescence-quenching damaged base as the mechanism of probe response. Performance with bacterial cells overexpressing a repair enzyme
In vitro characterization Of the 42 probes tested, the best-performing was a short hairpin that contains the fluorescent cytosine analog tCo opposite guanine and flanked on the 5’-side by 8-oxoguanine and on the 3’-side by guanine (OGR1, Scheme 1). The hairpin structure was stable (Tm > 90 8C; Figure S4), and the probe yielded a 60fold fluorescence increase upon reaction with Fpg (Figure 2 A.) The quantum yield of the probe (emission maximum 460 nm) was 0.008 before the reaction and 0.170 after completion. The probe reacted more quickly with Fpg than hOGG1 (reaching 90 % of final fluorescence after 10 min and 3 h, respectively, for a 1:6 molar ratio of enzyme to substrate; Figure 2 B). From the initial rates of increase as a function of concentration, we determined apparent KM values of 773 120 nm (hOGG1) and 14.5 2.9 nm (Fpg), and kcat of 0.07 and 0.2 min¢1, respectively. KM has been reported as 3.4 nm (hOGG1) and 6 nm (Fpg) and kcat as 300 min¢1 (hOGG1) and 3 min¢1 (Fpg), with conventional double-stranded DNA substrates.[24, 45] Analysis of the hOGG1 reaction products by PAGE and MALDI-TOF MS identified the probe with 8-oxoguanine excised and the probe following cleavage of the DNA backbone with a phosphate at the new 5’ terminus and an abasic site at the new 3’ terminus. Analysis of the Fpg reaction products identi-
Although assays with purified enzyme have general utility, the ability to directly evaluate enzyme activity in bacterial cells (without purification), would save steps in some applications. In a preliminary experiment, we used pOGR1 (a nuclease-protected variant of probe OGR1 with 2’-O-methyl groups at the termini; Figure S5) with E. coli overexpressing Fpg. Probe pOGR1 incubated with these cells in the presence of 0.25 % Triton-X 100 to permeabilize the cells showed an increase in fluorescence after just 1 hour (controls: cells not overexpressing Fpg, or without pOGR1; Figure 4). The results show that the probe can be used to assess enzymatic activity directly in bacterial cell suspensions. Performance in mammalian cell extracts To test whether OGG1 activity can be detected in mammalian cell extracts, we incubated pOGR1 with extract from WT and Ogg1¢/¢ MEF cells[47] (Figure 5). The WT cell extract yielded higher fluorescence (~ 2.5 times greater than in the deleted cell line), and the reproducibility was high, with small errors relative to the signal. The difference between the two cell lines was reproducible beyond the error limits after 8 h, and became more readily apparent at longer times. We conclude that the probe can detect OGG1 activity at native levels in mammalian cell extracts. Utility in detecting inhibitors of OGG1
Scheme 1. A) Quenching and generation of fluorescence signals with base excision. 8-Oxoguanine is represented as “oG” and the fluorophore is represented as “F”. The fluorophore is quenched by 8-oxoguanine but becomes emissive following excision of 8-oxoguanine. B) Short hairpin probe OGR1 (the best-performing probe in this study) undergoes base excision by OGG1. 8-oxoguanine is represented as G*; tCo (1,3-diaza-2-oxophenoxazine) is represented as C*.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1640
Next we tested whether OGR1 can be used in reporter assays for competitive inhibitors of OGG1. A previous study identified the unnatural nucleobase 3Cl-4-F-indole in double-stranded DNA as a strong binder of OGG1.[62] We synthesized a short stable hairpin containing 3-Cl-4F-indole (sequence in Figure S6) as a test inhibitor. From the initial rate of fluorescence increase, we determined apparent KM of OGR1 with hOGG1 to be 0.77 0.12 mm in the absence and 3.71 0.52 mm in the presence
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers window would suffice. At 0.8 mm probe in a reaction volume of 20 mL, one micromole of probe (the scale of a standard DNA synthesis) would be sufficient for approximately 60 000 assays.
Conclusion
Figure 2. A) Fluorescence emission of 4 mm probe OGR1 in Fpg reaction buffer at 37 8C without Fpg (gray) and after 25 min with Fpg (black). B) Time course of 4 mm probe OGR1 emission at 465 nm following addition of 658 nm Fpg (~) or hOGG1 (*) at 37 8C (representative curves).
of the inhibitor (Figure 6). We determined the KI of the inhibitor to be 0.53 mm. Thus, simple fluorescence measurement with probe OGR1 can be used to identify and characterize competitive inhibitors of OGG1. A FRET-based probe has been used to screen for inhibitors of the glycosylase/AP lyase activities of several base-excisionrepair enzymes.[63] To demonstrate the suitability of probe OGR1 in a high-throughput screen, we tested the probe at a lower concentration (0.8 mm probe and 150 nm hOGG1) in a 384-well plate format, and quantified the Z’ factor and signal-to-noise ratio (Figure 7). Z’ factor > 0.4, signal-to-noise ratio > 2, and coefficient of variance < 20 % are acceptable for high-throughput screening.[64] In our assay, the signal-to-noise ratio passed the cutoff after only 10 min, and the Z’ factor passed the cutoff after 25 min, thus showing that long incubation times are not needed to assess enzyme activity with the probe (coefficient of variance < 20 % at all times). Although greater signal intensity was achieved with OGR1 after longer reaction times, it is not necessary for the reaction to go to completion; comparison of the initial rates of fluorescence increase or of the fluorescence obtained after a specified time ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
We have generated a fluorescence turn-on probe for 8-oxoguanine excision. It yields a 60-fold increase in fluorescence upon incubation with human OGG1 or bacterial Fpg. In this novel design, 8-oxoguanine acts as a quencher, thereby providing a direct, real-time fluorescence signal that depends on its excision from DNA. Because of this direct measurement of enzymatic bond cleavage, the probe offers advantages over FRET-based probes, which require downstream events such as DNA backbone cleavage and dehybridization to yield a fluorescent signal, as well as other common indirect and time-consuming methods to measure 8-oxoguanine excision. Compared with a previously reported FRET sensor for OGG1,[51] our probe displays higher fluorescence fold-change, and its small size (relative to previous FRET-based probes) makes it easier and less costly to synthesize. Furthermore, the highly stable doublestranded hairpin design eliminates the need to carefully anneal the probe and obviates issues of strand stoichiometry. We expect that this and related probes will be useful for screening the effects of polymorphisms, mutations, post-translational modifications, and competitive inhibitors on 8-oxoguanine repair enzymes. In our molecular strategy for direct measurement of DNA base excision by fluorescence, the nucleobase to be excised acts as a fluorescence quencher. Specific coupling of the damaged base with a photophysically matched fluorescent nucleobase gives a probe that yields a light-up signal in real time only when the damaged base is removed. This strategy was previously tested with uracil deoxyglycosidase, and the uracil (which arises from hydrolytic damage of cytosine) was combined with a fluorescent pyrene nucleoside to report removal of uracil.[65, 66] The current experiments suggest that the application of such a fluorescence-quenching strategy might be more general for measuring base excision repair. Here we have shown that it can be employed for 8-oxoguanine, by taking advantage of previously unknown quenching of tCo by the damaged base. Fpg reacted approximately 18 times faster than human OGG1 with probe OGR1 under our conditions (Figure 2 B). Whereas OGG1 is known to have a strong preference for double-stranded DNA substrates with 8-oxoguanine paired opposite a cytosine,[25] Fpg also accepts single-stranded substrates.[24] The higher activity of Fpg on the new probe relative to OGG1 might reflect Fpg’s broader substrate requirement. The fact that OGR1 is a significantly less efficient substrate than natural DNA for OGG1 suggests that the extra steric bulk of the fluorescent cytosine derivative tCo interacts unfavorably in the OGG1 active site. However, the probe showed useful activity in vitro and in cell lysates. In the cell-based tests, the signal was slower to develop than with pure enzyme in vitro; we attribute this to inherently low enzyme activity in these systems and to significant cellular background fluorescence at
1641
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers pyrrolo-dC, 2-aminopurine, ethenoadenine, Dss, tCo phosphoramidites, 1 mmol columns of dG-CPG, dA-CPG, Universal Support III, and DNA synthesis reagents were purchased from Glen Research (Sterling, VA). 5’-DMT 3’-cyanoethyl a-pyrene phosphoramidite,[67] 5’-DMT 3’-cyanoethyl a-perylene phosphoramidite,[68] 5’-DMT 3’-cyanoethyl aphenylethynylpyrene phosphoramidite,[69] 5’-DMT 3’-cyanoethyl bN,N-dimethyl-2-(quinolin-2-yl)benzo[d]oxazol-6-amine phosphoramidite (QB),[70] and 5’-DMT 3’-cyanoethyl b-3-chloro-4-fluoro-indole phosphoramidite[62] were prepared according to published methods. All oligonucleotides excluding OGR37–40 were synthesized on a model 394 DNA/RNA synthesizer (Applied Biosystems) with a DMToff program. The coupling of each monomer used standard 3’-to-5’ cyanoethyl phosphoramidite chemistry with 999 s coupling times for pyrene, perylene, and phenylethynylpyrene, 180 s for tCo, and 240 s for other bases. Oligonucleotides OGR37–40 were synthesized by the Peptide and Nucleic Acid Facility (Stanford University, CA). Oligonucleotides containing 8-oxoguanine were deprotected in concentrated ammonium Figure 3. PAGE analysis of 4 mm probe OGR1 5’-labeled with [32P]phosphate and reacted with 658 nm A) hOGG1 or B) Fpg at 37 8C for the indicated times. Piperidine (200 mm) was added and incubated (95 8C, 5 min) to cleave the hydroxide with 2-mercaptoethanol DNA backbone at abasic sites as indicated to identify the 8-oxoguanine excision product and the backbone cleav(0.25 m) for 17 h at 55 8C. Dinucleoage product. Below: MALDI-TOF MS analysis of OGR1 following overnight reaction with C) hOGG1 or D) Fpg. tides containing pyrene, perylene, or phenylethynylpyrene (but not 8oxoguanine) were deprotected in potassium carbonate (0.05 m in methanol) for 24 h at room temperature. All other oligonucleotides this wavelength. Nevertheless, assays over 6–12 h gave reprowere deprotected in ammonium hydroxide for 17 h at room temducible signals of this enzyme activity. perature. The deprotected oligonucleotides were dried in a SpeedTaken together, the results show that the OGR1 probe perVac (Savant).
forms well in in vitro assays of enzyme activity with purified enzyme and with cell lysates, as well as for the measurement of activity in bacteria, which is expected to be useful for measurement of inhibition and of the effects of mutations on the enzymes. The probe gives a sufficiently high signal and low enough variation for use in high-throughput screens for competitive OGG1 inhibitors. A single DNA synthesis (1 mmol scale) would allow approximately 60 000 assays in a 384-well plate format. The fact that all components of probe OGR1 are commercially available makes it widely accessible for researchers.
Experimental Section
Oligonucleotides were purified by either HPLC or PAGE. For purification by HPLC, samples were purified on an HPLC device (Shimadzu) with a Jupiter C5 column (Phenomenex, Torrance, CA) or a Prosphere C18 column (Grace, Columbia, MD); acetonitrile and triethylammonium acetate (0.1 m, pH 7) were the eluents. For purification by PAGE, crude oligonucleotide was loaded onto a 20 % denaturing polyacrylamide gel (acrylamide/bis (19:1, 20 %), urea (7.5 m), tris borate EDTA buffer). The gel was cooled with a fan during electrophoresis. The gel was visualized by UV absorbance, and the darkest band was excised and extracted in NaCl (0.2 m) and 2-mercaptoethanol (0.25 m) overnight at 4 8C with shaking. The sample was dialyzed at 4 8C, filtered, dried in a SpeedVac, and resuspended in water.
Oligonucleotide preparation: Adenine, thymine, guanine, cytosine, 2’-O-methylcytosine, 8-oxoguanine, spacer (tetrahydrofuran),
Probe concentration was determined by fluorophore absorbance: pyrene e342 = 39 000 m¢1 cm¢1,[67] perylene e440 = 39 200 m¢1 cm¢1,[68]
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1642
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 6. Michaelis–Menten curves (n = 1) of 0.11 mm hOGG1 with varying concentrations of probe OGR1 and no inhibitor (&) or 2 mm inhibitor (*) in hOGG1 reaction buffer at 37 8C. The curves were fitted to the Michaelis– Menten equation with Origin software (c c). Figure 4. Reporting on 8-oxoguanine excision with bacterial cells expressing a repair enzyme. A) E. coli overexpressing Fpg after 17 h with 4 mm nucleaseprotected probe pOGR1 and 0.25 % Triton X-100 (c c); overexpressing Fpg after 17 h with no probe and 0.25 % Triton X-100 (c c); basal Fpg expression after 17 h with 4 mm probe and 0.25 % Triton X-100 (a a). B) E. coli overexpressing Fpg with 4 mm probe and 0.25 % Triton X-100 after 1 h (c c), 4 h (a a), or 17 h (c c).
Figure 5. Reporting 8-oxoguanine excision in mammalian cell extracts. Probe pOGR1 (2 mm) was incubated with 6.7 ng mL¢1 WT (~) or Ogg1¢/¢ (*) MEF cell extracts at 37 8C. The fluorescence at 460 nm (lex = 355 nm) was measured at the indicated time points. Background fluorescence (without probe) was subtracted. Error bars show standard deviation (n = 5).
2-aminopurine e260 = 1000 m¢1 cm¢1,[71] ethenoadenine e260 = 4800 m¢1 cm¢1,[72] pyrrolo-dC e345 = 3700 m¢1 cm¢1,[73] Dss e380 = 31 000 m¢1 cm¢1,[74] tCo e360 = 9000 m¢1 cm¢1,[59] phenylethynylpyrene e363 = 67 000 m¢1 cm¢1,[69] 3-Cl-4-F-indole e260 = 13 700 m¢1 cm¢1.[62] The QB extinction coefficient was calculated to be 22 400 m¢1 cm¢1 according the Beer–Lambert Law by measuring the absorbances of known concentrations of QB diol in methanol. For oligonucleotides with multiple nucleobases or fluorophores absorbing at the specified extinction-coefficient wavelength, the extinction coefficient was calculated as the sum of the extinction coefficients of each nucleobase or fluorophore in the oligonucleotide. Oligonucleotide identities were confirmed by MALDI-TOF MS performed at the Stanford Peptide and Nucleic Acid Facility. See
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
Figure 7. Probe OGR1 in 384-well assay. A) 0.8 mm probe OGR1 was incubated with (~) or without (*) 0.15 mm hOGG1 in 20 mL of buffer. The fluorescence at 460 nm (lex = 355 nm) was measured. Error bars represent standard deviation with (n = 5) or without (n = 3) the probe. B) Z’ factor at each timepoint (a: cutoff value of 0.4). C) Signal-to-noise ratio at each timepoint (a: cutoff value of 2).
1643
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Table S1 for the calculated and found masses. Samples were stored at ¢20 8C. Fluorescence analysis of in vitro enzymatic activity: Fpg reaction buffer (NEBuffer 1), hOGG1 reaction buffer (NEBuffer 4), and BSA (10 mg mL¢1) were purchased from New England Biolabs. The emission spectra were obtained with a Jobin Yvon Spex Fluorolog 3 spectrophotometer (Horiba, Kyoto, Japan), by exciting the sample at the fluorophore excitation maximum (pyrene 342 nm, perylene 440 nm, 2-aminopurine 303 nm, ethenoadenine 276 nm, pyrrolo-dC 347 nm, Dss 380 nm, tCo 360 nm, phenylethynylpyrene 363 nm, QB 400 nm) and measuring emission starting at 10 nm above the excitation wavelength (step size 1 nm). Aqueous solutions of 1 Õ Fpg or hOGG1 reaction buffer and BSA (100 mg mL¢1) were prepared in a quartz cuvette. The fluorescence emission spectrum was measured before addition of probe to determine buffer fluorescence. Fluorescence emission spectra were recorded immediately after addition of probe (4 mm) and after 10 min of incubation at 37 8C (to confirm stable fluorescence upon heating). For OGR1 studies, 658 nm of Fpg or hOGG1 (New England Biolabs) was added. For probe screening, Fpg (833 nm) or hOGG1 (658 nm) was added. The fluorescent emission spectrum was measured at various timepoints following addition of enzyme. The sample was maintained at 37 8C throughout. PAGE analysis of in vitro enzymatic activity: Probes were labeled at the 5’-end with 32P by using T4 polynucleotide kinase (Invitrogen) and [g-32P]ATP (PerkinElmer), and purified by ethanol precipitation as previously described.[75] BSA (100 mg mL¢1), unlabeled probe (4 mm), labeled probe (0.05 mm), and Fpg or hOGG1 (658 nm) were mixed in Fpg or hOGG1 reaction buffer and incubated at 37 8C. Aliquots (10 mL) were removed at the indicated timepoints. Non-piperidine-treated aliquots were mixed with loading buffer (10 mL; EDTA (25 mm), formamide (95 %), Bromophenol Blue (0.1 %), xylene cyanol (0.1 %)). Samples to be reacted with piperidine were mixed with piperidine/water (1:10, 2 mL) for 5 min at 90 8C and then mixed with the loading buffer. The samples were frozen for up to 24 h before PAGE analysis. Samples (10 mL) were loaded on a 20 % denaturing analytical PAGE gel (20 % acrylamide/ bis solution (19:1), urea (7.5 m), tris borate EDTA buffer) and electrophoresed. The gel was exposed overnight on a phosphor screen (Molecular Dynamics) and imaged on a Typhoon 9410 gel imager (GE Healthcare). MALDI-TOF MS analysis of enzymatic activity: Samples were prepared as for the fluorescence assays. An aliquot was submitted to the Stanford Peptide and Nucleic Acid Facility for zip-tip clean up and MALDI-TOF analysis. The samples were analyzed in HPA matrix in linear negative mode. Quantum yield measurements: Quinine sulfate (fluorescence quantum yield standard)[76] was diluted in H2SO4 (0.1 m), and the probes were diluted in water to an absorbance of 0.04–0.05 at 350 nm. For the reacted probes, the absorbance of protein in buffer was subtracted from the absorbance of the sample. The absorbance was measured at 350 nm on a Varian Cary 100 Bio UV/Vis spectrophotometer. The fluorescence emission spectrum was measured a Jobin Yvon-Spex Fluorolog 3 spectrophotometer (lex = 350 nm, lem = 360–680 nm, step size 1 nm). Quantum yield was calculated according to Fsample = (Fstandard)(Isample/Istandard)(Astandard/ Asample)(n2sample/n2standard), where F is quantum yield, I is the fluorescence emission over 360–680 nm, A is the absorbance at 350 nm, and n is the refractive index of the solvent for either quinine sulfate (standard) or probe (sample). ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
E. coli methods: In order to test OGR1 in bacterial cell solutions, we added 2’-O-methyl monomers for protection at the 5’ and 3’ ends of the probe thus generating probe pOGR1. The quenched quantum yield was 0.004–0.005, comparable to the quantum yield of probe OGR1 (0.008). As a control we synthesized an unquenched probe (G-OGR1; containing guanine instead of 8-oxoguanine), which has a quantum yield of 0.08. E. coli One Shot BL21(DE3) cells (Invitrogen/Life Technologies) were transformed with plasmid pKK223.3-WT Fpg (a gift from Prof. Sheila David, University of California–Davis)[77] by following the Invitrogen protocol. A single colony was cultured overnight at 30 8C in lysogeny broth (LB) containing ampicillin (50 mg mL¢1). The culture was diluted in LB containing ampicillin (50 mg mL¢1) to OD600 = 0.1 and grown at 37 8C to OD600 = 0.4. Isopropyl-1-thio-b-galactopyranoside (IPTG, 1 mm) was added, and the culture was grown for 3 or 24 h at 30 8C. The cells were washed in PBS, centrifuged (5000 g, 5 min, 4 8C), and resuspended in Fpg reaction buffer (to OD600 = 0.6). Triton X-100 (0.25 %) and probe (4 mm) were added. Samples were incubated with shaking at 37 8C for 1, 4, or 17 h, then diluted (1:10) in water, and fluorescence emission was measured on a Jobin Yvon-Spex Fluorolog 3 spectrophotometer (lex = 360 nm). Mammalian cell extracts: Ogg1¢/¢ MEF cells (a gift from Dr. Istvan Boldogh, University of Texas Medical Branch at Galveston)[47] and WT MEF cells were grown in DMEM/F12 (3:1) (HyClone) supplemented with fetal bovine serum (10 %), penicillin (100 U mL¢1), and streptomycin (100 mg mL¢1) in a humidified incubator at 37 8C with 5 % CO2. Cell extracts were obtained by the protocol by Folco et al.[78] with a few changes: the cells were collected by trypsinization at room temperature instead of scraping; Roche complete mini EDTA-free tablets were used instead of PMSF; the cells were lysed by passing 10–15 times through a 21 gauge needle and 5– 10 times through a 16 gauge needle instead of through a Dounce homogenizer. Total protein concentration was determined by a Bradford assay. Equal amounts (6.7 ng mL¢1) of WT and Ogg1¢/¢ MEF cell extract were incubated with probe pOGR1 (2 mm) and MgCl2 (10 mm) at 37 8C in 80 mL reaction volumes in a black 384-well plate. Fluorescence was measured at 460 nm on a Fluoroskan Ascent FL fluorescence plate reader (Thermo Scientific; lex = 355 nm). Melting temperature measurements: Samples containing probe OGR1 (4 mm) in Fpg or hOGG1 reaction buffer reaction buffer were prepared in stoppered 1 cm path-length quartz cells. A Varian Cary 100 UV/Vis spectrophotometer equipped with a thermoprogrammer was used to measure absorbance (260 nm) while the temperature was changed from 25 to 99 8C and from 99 to 25 8C (1 8C min¢1). Because of the high melting temperature of OGR1 and the lack of an absorbance plateau in the high temperature range, MeltWin software could not be used to determine the melting temperatures; melting temperatures were determined by visual inspection of the melting curves and by approximation of the first derivative of the melting curve. Measurement of kinetic parameters: Michaelis–Menten curves for hOGG1 with and without inhibitor (3-Cl-4-F-indole-containing oligonucleotides) were generated by preincubating hOGG1 (0.11 mm) with/without inhibitor (2 mm) in hOGG1 reaction buffer with BSA (100 mg mL¢1) at 37 8C for 5 min. OGR1 (0.5–10 mm) was then added at 37 8C, and the initial velocity was evaluated (rate of fluorescence increase (linear) at 460 nm over the first 10 min). Michaelis–Menten curves for Fpg were generated by incubating OGR1 (0.03–0.3 mm) with Fpg (6 nm) in Fpg reaction buffer with
1644
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers BSA (100 mg mL¢1) at 37 8C and evaluating the initial velocity at 37 8C (rate of fluorescence increase (linear) at 460 nm over the first 6 min). The curves were fitted in Origin (OriginLab, Northampton, MA) by using the Hill fit with n = 1 (yielding the Michaelis–Menten equation) to determine KM and Vmax. The dissociation constant, KI, of the hOGG1 inhibitor was calculated from apparent KM : KM, inhibitor = KM, no inhibitor(1 + ([I]/KI)).[79] To calculate kcat (kcat = Vmax/[enzyme]), we made a calibration curve to convert fluorescence units to amount of product by incubating known concentrations of probe with excess Fpg and measuring the fluorescence increase after the fluorescence signal plateaued. Vmax was converted from fluorescence units per s to mm product per s from the calibration curve. 384-well assay: OGR1 (0.8 mm) and hOGG1 (150 nm) were incubated in hOGG1 reaction buffer with BSA (100 mg mL¢1) at 37 8C in 20 mL reaction volumes in a black 384-well plate. Fluorescence at 460 nm was measured on a Thermo Fluoroskan Ascent FL fluorescence plate reader (lex = 355 nm). The signal-to-noise ratio was calculated by dividing the signal of OGR1 with hOGG1 by the signal of OGR1 without hOGG1. The coefficient of variance was calculated by dividing the standard deviation by the average signal intensity at that timepoint. Z’ factor was calculated from: 1¢
3 SDsignal þ 3 SDbaseline meansignal ¢meanbaseline
ð1Þ
Acknowledgements We thank Philip Yuen and Prof. Sheila David (University of California–Davis) for providing the plasmid encoding Fpg. We thank Dr. Istvan Boldogh (University of Texas Medical Branch at Galveston) for providing Ogg1¢/¢ MEF cells. This work was supported by the U.S. National Institutes of Health (GM067201 and GM106067), and the National Science Foundation (graduate fellowship to S.K.E.). Funding for the open access charge was provided by Stanford University. Keywords: DNA damage · fluorescent probes · Fpg · OGG1 · oxoguanine [1] H. Kasai, S. Nishimura, Nucleic Acids Res. 1984, 12, 2137 – 2145. [2] B. N. Ames, L. S. Gold, Mutat. Res. 1991, 250, 3 – 16. [3] H. Kasai, S. Nishimura, Y. Kurokawa, Y. Hayashi, Carcinogenesis 1987, 8, 1959 – 1961. [4] H. Kasai, S. Nishimura, Environ. Health Perspect. 1986, 67, 111 – 116. [5] E. S. Fiala, C. C. Conaway, J. E. Mathis, Cancer Res. 1989, 49, 5518 – 5522. [6] M. Kouchakdjian, V. Bodepudi, S. Shibutani, M. Eisenberg, F. Johnson, A. P. Grollman, D. J. Patel, Biochemistry 1991, 30, 1403 – 1412. [7] K. E. McAuley-Hecht, G. A. Leonard, N. J. Gibson, J. B. Thomson, W. P. Watson, W. N. Hunter, T. Brown, Biochemistry 1994, 33, 10266 – 10270. [8] K. C. Cheng, D. S. Cahill, H. Kasai, S. Nishimura, L. A. Loeb, J. Biol. Chem. 1992, 267, 166 – 172. [9] S. Shibutani, M. Takeshita, A. P. Grollman, Nature 1991, 349, 431 – 434. [10] M. L. Wood, M. Dizdaroglu, E. Gajewski, J. M. Essigmann, Biochemistry 1990, 29, 7024 – 7032. [11] R. De Bont, N. van Larebeke, Mutagenesis 2004, 19, 169 – 185. [12] D. O. Zharkov, T. A. Rosenquist, S. E. Gerchman, A. P. Grollman, J. Biol. Chem. 2000, 275, 28607 – 28617. [13] P.-M. Girard, N. Guibourt, S. Boiteux, Nucleic Acids Res. 1997, 25, 3204 – 3211.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
[14] J. W. Hill, T. K. Hazra, T. Izumi, S. Mitra, Nucleic Acids Res. 2001, 29, 430 – 438. [15] A. E. Vidal, I. D. Hickson, S. Boiteux, J. P. Radicella, Nucleic Acids Res. 2001, 29, 1285 – 1292. [16] T. Saitoh, K. Shinmura, S. Yamaguchi, M. Tani, S. Seki, H. Murakami, Y. Nojima, J. Yokota, Mutat. Res. 2001, 486, 31 – 40. [17] V. S. Sidorenko, G. A. Nevinsky, D. O. Zharkov, DNA Repair 2007, 6, 317 – 328. [18] B. Dalhus, M. Forsbring, I. H. Helle, E. S. Vik, R. J. Forstrøm, P. H. Backe, I. Alseth, M. Bjørs, Structure 2011, 19, 117 – 127. [19] B. Sun, K. A. Latham, M. L. Dodson, R. S. Lloyd, J. Biol. Chem. 1995, 270, 19501 – 19508. [20] V. Bailly, W. G. Verly, T. O’Connor, J. Laval, Biochem. J. 1989, 262, 581 – 589. [21] T. R. O’Connor, J. Laval, Proc. Natl. Acad. Sci. USA 1989, 86, 5222 – 5226. [22] R. J. Graves, I. Felzenszwalb, J. Laval, T. R. O’Connor, J. Biol. Chem. 1992, 267, 14429 – 14435. [23] M. Bhagwat, J. A. Gerlt, Biochemistry 1996, 35, 659 – 665. [24] A. A. Ishchenko, N. V. Bulychev, G. A. Maksakova, F. Johnson, G. A. Nevinsky, IUBMB Life 1999, 48, 613 – 618. [25] H. Dou, S. Mitra, T. K. Hazra, J. Biol. Chem. 2003, 278, 49679 – 49684. [26] O. Minowa, T. Arai, M. Hirano, Y. Monden, S. Nakai, M. Fukuda, M. Itoh, H. Takano, Y. Hippou, H. Aburatani, K.-i. Masamura, T. Nohmi, S. Nishimura, T. Noda, Proc. Natl. Acad. Sci. USA 2000, 97, 4156 – 4161. [27] A. Klungland, I. Rosewell, S. Hollenbach, E. Larsen, G. Daly, B. Epe, E. Seeberg, T. Lindahl, D. E. Barnes, Proc. Natl. Acad. Sci. USA 1999, 96, 13300 – 13305. [28] M. Cabrera, Y. Nghiem, J. H. Miller, J. Bacteriol. 1988, 170, 5405 – 5407. [29] D. Thomas, A. D. Scot, R. Barbey, M. Padula, S. Boiteux, Mol. Gen. Genet. 1997, 254, 171 – 178. [30] K. Sakumi, Y. Tominaga, M. Furuichi, P. Xu, T. Tsuzuki, M. Sekiguchi, Y. Nakabeppu, Cancer Res. 2003, 63, 902 – 905. [31] E. Speina, K. D. Arczewska, D. Gackowski, M. Zielin´ska, A. Siomek, J. Kowalewski, R. Olin´ski, B. Tudek, J. T. Kus´mierek, J. Natl. Cancer Inst. 2005, 97, 384 – 395. [32] S. Chevillard, J. P. Radicella, C. Levalois, J. Lebeau, M.-F. Poupon, S. Oudard, B. Dutrillaux, S. Boiteux, Oncogene 1998, 16, 3083 – 3086. [33] J.-W. Hyun, J.-Y. Choi, H.-H. Zeng, Y.-S. Lee, H.-S. Kim, S.-H. Yoon, M.-H. Chung, Oncogene 2000, 19, 4476 – 4479. [34] J.-W. Hyun, Y.-C. Jung, H.-S. Kim, E.-Y. Choi, J.-E. Kim, B.-H. Yoon, S.-H. Yoon, Y.-S. Lee, J. Choi, H.-J. You, M.-H. Chung, Mol. Cancer Res. 2003, 1, 290 – 299. [35] M. Audebert, J. P. Radicella, M. Dizdaroglu, Nucleic Acids Res. 2000, 28, 2672 – 2678. [36] M. Audebert, S. Chevillard, C. Levalois, G. Gyapay, A. Vieillefond, J. Klijanienko, P. Vielh, A. K. El Naggar, S. Oudard, S. Boiteux, J. P. Radicella, Cancer Res. 2000, 60, 4740 – 4744. [37] I.-J. Kim, J.-L. Ku, H. C. Kang, J.-H. Park, K.-A. Yoon, Y. Shin, H.-W. Park, S. G. Jang, S.-K. Lim, S. Y. Han, Y.-K. Shin, M. R. Lee, S.-Y. Jeong, H.-R. Shin, J.-S. Lee, W.-H. Kim, J.-G. Park, Hum. Genet. 2004, 115, 498 – 503. [38] P. Karihtala, S. Kauppila, U. Puistola, A. Jukkola-Vuorinen, Br. J. Cancer 2012, 106, 344 – 347. [39] S. D. Bruner, D. P. G. Norman, G. L. Verdine, Nature 2000, 403, 859 – 866. [40] D. Gackowski, E. Speina, M. Zielinska, J. Kowalewski, R. Rozalski, A. Siomek, T. Paciorek, B. Tudek, R. Olinski, Cancer Res. 2003, 63, 4899 – 4902. [41] T. Paz-Elizur, M. Krupsky, S. Blumenstein, D. Elinger, E. Schechtman, Z. Livneh, J. Natl. Cancer Inst. 2003, 95, 1312 – 1319. [42] T. Paz-Elizur, R. Ben-Yosef, D. Elinger, A. Vexler, M. Krupsky, A. Berrebi, A. Shani, E. Schechtman, L. Freedman, Z. Livneh, Cancer Res. 2006, 66, 11683 – 11689. [43] R. Gangwar, D. Ahirwar, A. Mandhani, R. D. Mittal, Mutat. Res. 2009, 680, 56 – 63. [44] G. Mao, X. Pan, B.-B. Zhu, Y. Zhang, F. Yuan, J. Huang, M. A. Lovell, M. P. Lee, W. R. Markesbery, G.-M. Li, L. Gu, Nucleic Acids Res. 2007, 35, 2759 – 2766. [45] V. S. Sidorenko, A. P. Grollman, P. Jaruga, M. Dizdaroglu, D. O. Zharkov, FEBS J. 2009, 276, 5149 – 5162. [46] S.-Y. Chen, L. Wan, C.-M. Huang, Y.-C. Huang, J. J.-C. Sheu, Y.-J. Lin, S.-P. Liu, Y.-C. Lan, C.-H. Lai, C.-W. Lin, C.-H. Tsai, F.-J. Tsai, Rheumatol. Int. 2012, 32, 1165 – 1169.
1645
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers [47] A. Bacsi, L. Aguilera-Aguirre, B. Szczesny, Z. Radak, T. K. Hazra, S. Sur, X. Ba, I. Boldogh, DNA Repair 2013, 12, 18 – 26. [48] B. Tyrberg, K. A. Anachkov, S. A. Dib, J. Wang-Rodriguez, K.-H. Yoon, F. Levine, BMC Endocr. Disord. 2002, 2, 2. [49] S. Simone, Y. Gorin, C. Velagapudi, H. E. Abboud, S. L. Habib, Diabetes 2008, 57, 2626 – 2636. [50] A. Maksimenko, A. A. Ishchenko, G. Sanz, J. Laval, R. H. Elder, M. K. Saparbaev, Biochem. Biophys. Res. Commun. 2004, 319, 240 – 246. [51] Y. Kim, I. S. Hong, Bull. Korean Chem. Soc. 2009, 30, 2149 – 2151. [52] L. Mirbahai, R. M. Kershaw, R. M. Green, R. E. Hayden, R. A. Meldrum, N. J. Hodges, DNA Repair 2010, 9, 144 – 152. [53] T. T. Nikiforov, S. Roman, Anal. Biochem. 2015, 477, 69 – 77. [54] J. L. Furman, P.-W. Mok, A. H. Badran, I. Ghosh, J. Am. Chem. Soc. 2011, 133, 12518 – 12527. [55] O. Nakagawa, S. Ono, Z. Li, A. Tsujimoto, S. Sasaki, Angew. Chem. Int. Ed. 2007, 46, 4500 – 4503; Angew. Chem. 2007, 119, 4584 – 4587. [56] T. Nasr, Z. Li, O. Nakagawa, Y. Taniguchi, S. Ono, S. Sasaki, Bioorg. Med. Chem. Lett. 2009, 19, 727 – 730. [57] Y. Taniguchi, R. Kawaguchi, S. Sasaki, J. Am. Chem. Soc. 2011, 133, 7272 – 7275. [58] J. N. Wilson, Y. Cho, S. Tan, A. Cuppoletti, E. T. Kool, ChemBioChem 2008, 9, 279 – 285. [59] P. Sandin, K. Bçrjesson, H. Li, J. Mrtensson, T. Brown, L. M. Wilhelmsson, B. Albinsson, Nucleic Acids Res. 2008, 36, 157 – 167. [60] S. Yoshizawa, T. Ueda, Y. Ishido, K. Miura, K. Watanabe, I. Hirao, Nucleic Acids Res. 1994, 22, 2217 – 2221. [61] T. J. Matray, E. T. Kool, J. Am. Chem. Soc. 1998, 120, 6191 – 6192. [62] P. L. McKibbin, A. Kobori, Y. Taniguchi, E. T. Kool, S. S. David, J. Am. Chem. Soc. 2012, 134, 1653 – 1661. [63] A. C. Jacobs, M. J. Calkins, A. Jadhav, D. Dorjsuren, D. Maloney, A. Simeonov, P. Jaruga, M. Dizdaroglu, A. K. McCullough, R. S. Lloyd, PLoS One 2013, 8, e81667. [64] M. Entzeroth, H. Flotow, P. Condron, Curr. Protoc. Pharmacol. 2009, Unit 9.4.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
[65] T. Ono, S. Wang, C.-K. Koo, L. Engstrom, S. S. David, E. T. Kool, Angew. Chem. Int. Ed. 2012, 51, 1689 – 1692; Angew. Chem. 2012, 124, 1721 – 1724. [66] T. Ono, S. K. Edwards, S. Wang, W. Jiang, E. T. Kool, Nucleic Acids Res. 2013, 41, e127. [67] R. X.-F. Ren, N. C. Chaudhuri, P. L. Paris, S. Rumney IV, E. T. Kool, J. Am. Chem. Soc. 1996, 118, 7671 – 7678. [68] J. Gao, S. Watanabe, E. T. Kool, J. Am. Chem. Soc. 2004, 126, 12748 – 12749. [69] S. Wang, J. Guo, T. Ono, E. T. Kool, Angew. Chem. Int. Ed. 2012, 51, 7176 – 7180; Angew. Chem. 2012, 124, 7288 – 7292. [70] L. H. Yuen, R. M. Franzini, S. Wang, P. Crisalli, V. Singh, W. Jiang, E. T. Kool, Angew. Chem. Int. Ed. 2014, 53, 5361 – 5365; Angew. Chem. 2014, 126, 5465 – 5469. [71] K. Evans, D. Xu, Y. Kim, T. M. Nordlund, J. Fluoresc. 1992, 2, 209 – 216. [72] J. A. Secrist III, J. R. Barrio, N. J. Leonard, G. Weber, Biochemistry 1972, 11, 3499 – 3506. [73] R. A. Tinsley, N. G. Walter, RNA 2006, 12, 522 – 529. [74] M. Kimoto, T. Mitsui, S. Yokoyama, I. Hirao, J. Am. Chem. Soc. 2010, 132, 4988 – 4989. [75] J. Sambrook, D. Russel, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, 2001, p. 10 – 17. [76] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, Berlin, 2006, p. 54. [77] M. D. Leipold, J. G. Muller, C. J. Burrows, S. S. David, Biochemistry 2000, 39, 14984 – 14992. [78] E. G. Folco, H. Lei, J. L. Hsu, R. Reed, J. Vis. Exp. 2012, 64, 4140. [79] J. A. Todhunter in Methods in Enzymology, Vol. 63: Enzyme Kinetics and Mechanism Part A: Initial Rate and Inhibitor Methods (Ed.: D Purlich), Elsevier, Amsterdam, 1979, pp. 383 – 411. Manuscript received: April 10, 2015 Final article published: June 12, 2015
1646
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
In Vitro Fluorogenic Real-Time Assay of the Repair of Oxidative DNA Damage Sarah K. Edwards,[a] Toshikazu Ono,[a, b] Shenliang Wang,[a] Wei Jiang,[a] Raphael M. Franzini,[a] Jong Wha Jung,[a, c] Ke Min Chan,[a] and Eric T. Kool*[a] The repair of oxidative damage to DNA is essential to avoid mutations that lead to cancer. Oxidized DNA bases, such as 8oxoguanine, are a main source of these mutations, and the enzyme 8-oxoguanine glycosylase 1 (OGG1) is the chief human enzyme that excises 8-oxoguanine from DNA. The activity of OGG1 has been linked to human inflammation responses and to cancer, and researchers are beginning to search for inhibitors of the enzyme. However, measuring the activity of the enzyme typically requires laborious gel-based measurements of radiolabeled DNAs. Here we report the design and properties of fluorogenic probes that directly report on the activity of OGG1 (and its bacterial homologue Fpg) in real time as the
oxidized base is excised. The probes are short, modified DNA oligomers containing fluorescent DNA bases and are designed to utilize 8-oxoguanine itself as a fluorescence quencher. Screening of combinations of fluorophores and 8-oxoguanine revealed two fluorophores, pyrene and tCo, that are strongly quenched by the damaged base. We tested 42 potential probes containing these fluorophores: the optimum probe, OGR1, yields a 60-fold light-up signal in vitro with OGG1 and Fpg. It can report on oxidative repair activity in mammalian cell lysate and with bacterial cells overexpressing a repair enzyme. Such probes might prove useful in quantifying enzyme activity and performing competitive inhibition assays.
Introduction 8-Oxoguanine is an abundant form of oxidative DNA damage that is generated by reactive oxygen species resulting from metabolism, ionizing radiation, and chemicals.[1–5] 8-Oxoguanine in DNA can adopt a syn conformation and pair with adenine during replication,[6, 7] ultimately leading to G-to-T mutations.[8–10] This form of nucleobase damage might be the chief molecular source of oxidation-induced mutagenesis in cells.[11] As protection against the mutagenic potential of 8-oxoguanine, cells have evolved multiple ways to repair this lesion. One of the chief eukaryotic 8-oxoguanine repair enzymes is 8oxoguanine glycosylase 1 (OGG1), which preferentially removes 8-oxoguanine opposite cytosine in double-stranded DNA.[12, 13] OGG1 acts as a glycosylase to excise the 8-oxoguanine base. OGG1 can then act as an AP lyase to performing a b-elimina[a] S. K. Edwards, Dr. T. Ono, Dr. S. Wang, Dr. W. Jiang, Dr. R. M. Franzini, Dr. J. W. Jung, K. M. Chan, Prof. E. T. Kool Department of Chemistry, Stanford University Stanford, CA 94305 (USA) E-mail: [email protected] [b] Dr. T. Ono Present Address: Department of Chemistry and Biochemistry Graduate School of Engineering Center for Molecular Systems (CMS), Kyushu University 744 Motooka, Nishi-ku, Fukuoka, 819-0395 (Japan) [c] Dr. J. W. Jung Present Address: College of Pharmacy Research Institute of Pharmaceutical Sciences Kyungpook National University Daegu, 702-701 (Republic of Korea) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201500184.
ChemBioChem 2015, 16, 1637 – 1646
tion reaction on the 3’-side of the abasic site to generate an a,b-unsaturated aldehyde at the 3’ terminus.[12, 13] However, OGG1’s lyase activity proceeds slowly and is believed not to be relevant in vivo; it has been proposed that the endonuclease APE1 instead cleaves the DNA backbone following 8-oxoguanine excision by OGG1.[14–18] Formamidopyrimidine-DNA glycosylase (Fpg), the functional homologue of OGG1 in Escherichia coli, excises the 8-oxoguanine base and then performs b and d elimination reactions at the abasic site to generate a single nucleotide gap flanked by 3’- and 5’-phosphates.[19–23] Both OGG1 and Fpg prefer double-stranded DNA substrates, but Fpg can also act on single-stranded DNA.[24, 25] E. coli, Saccharomyces cerevisiae, and mice engineered to be deficient in Fpg or OGG1 display increased levels of genomic 8-oxoguanine and increased mutations, thus highlighting the importance of these enzymes in maintaining genome integrity.[26–30] OGG1 mutants with decreased or abolished 8-oxoguanine repair activity have been found in many cancer specimens,[31–39] and lower OGG1 activity has been associated with increased cancer risk.[40–43] In addition to its prominent association with genetic mutations leading to malignancies, altered OGG1 activity has been associated with many other important diseases. Several OGG1 mutants with decreased or abolished activity have been discovered in Alzheimer’s patients.[44, 45] Two OGG1 polymorphisms have been associated with rheumatoid arthritis,[46] and decreased OGG1 activity has been linked with decreased lung inflammation in an allergy model.[47] Finally, altered levels of OGG1 have been associated with type 2 diabetes.[48, 49]
1637
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers A simple method to measure OGG1 activity would help further elucidate these multiple links between OGG1 activity and disease. Such a method would also aid screening for OGG1 inhibitors or activators, which would be useful as drugs and research tools, and assist determining disease prognosis from patient samples. Currently, quick and easy study of OGG1 activity is hindered by the time-consuming and indirect nature of available methods (e.g., PAGE, comet assay, GC/MS, ELISA, HPLC, and qRT-PCR). For example, the strand-nicking assay (one of the most common methods used to measure OGG1 activity) involves radiolabeling an 8-oxoguanine-containing oligonucleotide, incubating it with the sample of interest, and analyzing the sample at various time points by gel electrophoresis and radioactivity measurement. This method requires at least a day and only measures OGG1 activity at discrete, predetermined times. More recently developed approaches using fluorescently labeled oligonucleotides[50–53] give a fluorescent readout of OGG1 activity, but these probes rely on backbone cleavage (which is done slowly by OGG1, and is believed to be performed by a separate enzyme in human cells)[14–17] and subsequent DNA dehybridization to generate a fluorescence signal. Thus, these FRET-based probes do not directly measure the excision of 8-oxoguanine but instead rely on downstream events following the excision. Other DNA-cleaving and DNA-unwinding activities (by various nucleases and helicases) can yield false signals with this design. A split-luciferase sensor that detects 8-oxoguanine in DNA could potentially be used to identify competitive inhibitors of OGG1 or to indirectly measure OGG1 activity,[54] but this sensor measures the binding of OGG1 to 8-oxoguanine rather than OGG1 enzymatic activity. Given the need for a more direct and simple measurement of 8-oxoguanine excision activity, we aimed to develop “lightup” probes that directly measure excision of this oxidized base in real time. By identifying and applying a fluorophore that is strongly quenched by 8-oxoguanine, we have developed a probe that gives a 60-fold fluorescence increase upon reaction with OGG1 or Fpg. Such probes enable direct and specific measurements of OGG1 and Fpg activity in vitro and are expected to be useful in screening polymorphisms, mutations, post-translational modifications, and competitive inhibitors of 8-oxoguanine repair enzyme activity.
Results and Discussion Probe design Our probe design relies on a fluorophore that is quenched specifically by 8-oxoguanine, but becomes emissive following excision of the damaged base. Most desirable would be a fluorophore that is not strongly quenched by native (undamaged) nucleobases, so that a DNA enzyme-binding sequence can be constructed while retaining efficient fluorescence emission. Although the “8-oxo-G clamp” is known to be quenched by 8-oxoguanine,[55–57] this nucleobase analogue pairs with and surrounds the damaged base, and thus might prevent recognition or excision by repair enzymes. To screen for fluorophores that are quenched by 8-oxoguanine, we assembled dinucleotides ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
consisting of a fluorophore deoxyriboside (a nucleoside with a fluorophore replacing the nucleobase) and either 8-oxoguanine or an abasic site mimic (representing the product following 8-oxoguanine excision). (Structures of the dye monomers and a representative dinucleotide are shown in Figure S1 in the Supporting Information.) By comparing the fluorescence emission intensities of the 8-oxoguanine-containing dinucleotides with those of the abasic controls, we found that the fluorophores pyrene, 2-aminopurine, ethenoadenine, pyrrolo-dC, and tCo were quenched two- to 30-fold by 8-oxoguanine, with pyrene and tCo most strongly quenched (Figure 1). Pyrene has been reported to be quenched strongly by thymine and cytosine, and moderately by guanine.[58] Similarly to a previous report of quenching of tCo by guanine, tCo was more strongly quenched by 8-oxoguanine on the 5’-side of the fluorophore than on its 3’-side.[59] tCo is reported not to be strongly quenched by the other natural nucleobases.[59] In contrast to these fluorophores, phenylethynylpyrene, perylene, and QB dye were not quenched (or were poorly quenched) by 8-oxoguanine (Figure 1). Having identified at least two dyes that are strongly quenched by 8-oxoguanine in DNA, we synthesized 42 candidate probes as putative enzymatic substrates that contain 8oxoguanine and one or more of these fluorophores in place of natural nucleobases (Figure S2). Oligonucleotide design was guided by the desire to keep probes short, in order to aid synthesis yield and enhance the likelihood of cellular uptake. Some probes were single-stranded; others were double-stranded with a hairpin design. For the latter, we chose a highly stabilizing GAA loop[60] to enhance thermal stability in short sequences. The probes were prepared by standard phosphoramidite chemistry on an automated DNA synthesizer, purified by HPLC or PAGE, and characterized by MALDI-TOF MS to confirm incorporation of both the damaged nucleobase and the fluorescent deoxyriboside. The probes were screened for increases in fluorescence emission after incubation with recombinant human OGG1 or Fpg (signal enhancements for all probes with the enzymes in Figure S3). We observed some trends with pyrene-containing probes. Firstly, short double-stranded (hairpin) probes were much better substrates than single-stranded probes (compare OGR217 vs. OGR18-32 in Figure S3). Secondly, a fluorophore on the 3’-side of 8-oxoguanine appeared to be more favorable for enzymatic reaction than on the 5’-side (OGR19 vs. OGR21). To several of the pyrene-based probes, we added a second fluorophore as a FRET or exciplex partner, as a possible strategy for yielding longer-wavelength emission. In these cases, quenching by 8-oxoguanine was less efficient than for the single-fluorophore designs, giving high initial fluorescence signals. Many of the pyrene-based probes exhibited low turnover; this might may be a consequence of the bulky, hydrophobic fluorophore, or the affinity of OGG1 for the abasic site included as the pyrene pairing partner.[14, 61] We also observed trends with tCo-containing probes. Adding a second 8-oxoguanine to the other side of tCo did not diminish the initial signal but did decrease the final signal (OGR1 vs. OGR36). MALDI-TOF MS analysis (data not shown)
1638
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 1. Fluorescence quenching (emission spectra) of various fluorophores by 8-oxoguanine. DNA dinucleotides consisted of a fluorescent nucleoside adjac) or an abasic site mimic (c c). Fluorescent bases: A) 2-aminopurine, B) ethenoadenine, C) perylene, D) pyrene, E) pyrrolo-dC, cent to 8-oxoguanine (c F) QB, G) and H) tCo, and I) V. Dinucleotides in F and G have the fluorophore on the 3’-side of 8-oxoguanine or the abasic site mimic. In all other panels the fluorophore is on the 5’-side of 8-oxoguanine or the abasic site mimic. ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1639
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers indicated that only one 8-oxoguanine was excised, presumably rendering the probe a poor substrate for excision of the second 8-oxoguanine and thus keeping the probe in a semiquenched state. Increasing the stem length or removing the loop did not increase probe performance, thus indicating that the hairpin structure did not inhibit the enzyme (OGR1 vs. OGR37, OGR41, and OGR42).
fied the probe following cleavage of the DNA backbone with phosphates at the new 3’ and 5’ termini. These observations are consistent with the known mechanisms of these enzymes (Figure 3). Thus the analysis confirms enzymatic excision of the fluorescence-quenching damaged base as the mechanism of probe response. Performance with bacterial cells overexpressing a repair enzyme
In vitro characterization Of the 42 probes tested, the best-performing was a short hairpin that contains the fluorescent cytosine analog tCo opposite guanine and flanked on the 5’-side by 8-oxoguanine and on the 3’-side by guanine (OGR1, Scheme 1). The hairpin structure was stable (Tm > 90 8C; Figure S4), and the probe yielded a 60fold fluorescence increase upon reaction with Fpg (Figure 2 A.) The quantum yield of the probe (emission maximum 460 nm) was 0.008 before the reaction and 0.170 after completion. The probe reacted more quickly with Fpg than hOGG1 (reaching 90 % of final fluorescence after 10 min and 3 h, respectively, for a 1:6 molar ratio of enzyme to substrate; Figure 2 B). From the initial rates of increase as a function of concentration, we determined apparent KM values of 773 120 nm (hOGG1) and 14.5 2.9 nm (Fpg), and kcat of 0.07 and 0.2 min¢1, respectively. KM has been reported as 3.4 nm (hOGG1) and 6 nm (Fpg) and kcat as 300 min¢1 (hOGG1) and 3 min¢1 (Fpg), with conventional double-stranded DNA substrates.[24, 45] Analysis of the hOGG1 reaction products by PAGE and MALDI-TOF MS identified the probe with 8-oxoguanine excised and the probe following cleavage of the DNA backbone with a phosphate at the new 5’ terminus and an abasic site at the new 3’ terminus. Analysis of the Fpg reaction products identi-
Although assays with purified enzyme have general utility, the ability to directly evaluate enzyme activity in bacterial cells (without purification), would save steps in some applications. In a preliminary experiment, we used pOGR1 (a nuclease-protected variant of probe OGR1 with 2’-O-methyl groups at the termini; Figure S5) with E. coli overexpressing Fpg. Probe pOGR1 incubated with these cells in the presence of 0.25 % Triton-X 100 to permeabilize the cells showed an increase in fluorescence after just 1 hour (controls: cells not overexpressing Fpg, or without pOGR1; Figure 4). The results show that the probe can be used to assess enzymatic activity directly in bacterial cell suspensions. Performance in mammalian cell extracts To test whether OGG1 activity can be detected in mammalian cell extracts, we incubated pOGR1 with extract from WT and Ogg1¢/¢ MEF cells[47] (Figure 5). The WT cell extract yielded higher fluorescence (~ 2.5 times greater than in the deleted cell line), and the reproducibility was high, with small errors relative to the signal. The difference between the two cell lines was reproducible beyond the error limits after 8 h, and became more readily apparent at longer times. We conclude that the probe can detect OGG1 activity at native levels in mammalian cell extracts. Utility in detecting inhibitors of OGG1
Scheme 1. A) Quenching and generation of fluorescence signals with base excision. 8-Oxoguanine is represented as “oG” and the fluorophore is represented as “F”. The fluorophore is quenched by 8-oxoguanine but becomes emissive following excision of 8-oxoguanine. B) Short hairpin probe OGR1 (the best-performing probe in this study) undergoes base excision by OGG1. 8-oxoguanine is represented as G*; tCo (1,3-diaza-2-oxophenoxazine) is represented as C*.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1640
Next we tested whether OGR1 can be used in reporter assays for competitive inhibitors of OGG1. A previous study identified the unnatural nucleobase 3Cl-4-F-indole in double-stranded DNA as a strong binder of OGG1.[62] We synthesized a short stable hairpin containing 3-Cl-4F-indole (sequence in Figure S6) as a test inhibitor. From the initial rate of fluorescence increase, we determined apparent KM of OGR1 with hOGG1 to be 0.77 0.12 mm in the absence and 3.71 0.52 mm in the presence
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers window would suffice. At 0.8 mm probe in a reaction volume of 20 mL, one micromole of probe (the scale of a standard DNA synthesis) would be sufficient for approximately 60 000 assays.
Conclusion
Figure 2. A) Fluorescence emission of 4 mm probe OGR1 in Fpg reaction buffer at 37 8C without Fpg (gray) and after 25 min with Fpg (black). B) Time course of 4 mm probe OGR1 emission at 465 nm following addition of 658 nm Fpg (~) or hOGG1 (*) at 37 8C (representative curves).
of the inhibitor (Figure 6). We determined the KI of the inhibitor to be 0.53 mm. Thus, simple fluorescence measurement with probe OGR1 can be used to identify and characterize competitive inhibitors of OGG1. A FRET-based probe has been used to screen for inhibitors of the glycosylase/AP lyase activities of several base-excisionrepair enzymes.[63] To demonstrate the suitability of probe OGR1 in a high-throughput screen, we tested the probe at a lower concentration (0.8 mm probe and 150 nm hOGG1) in a 384-well plate format, and quantified the Z’ factor and signal-to-noise ratio (Figure 7). Z’ factor > 0.4, signal-to-noise ratio > 2, and coefficient of variance < 20 % are acceptable for high-throughput screening.[64] In our assay, the signal-to-noise ratio passed the cutoff after only 10 min, and the Z’ factor passed the cutoff after 25 min, thus showing that long incubation times are not needed to assess enzyme activity with the probe (coefficient of variance < 20 % at all times). Although greater signal intensity was achieved with OGR1 after longer reaction times, it is not necessary for the reaction to go to completion; comparison of the initial rates of fluorescence increase or of the fluorescence obtained after a specified time ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
We have generated a fluorescence turn-on probe for 8-oxoguanine excision. It yields a 60-fold increase in fluorescence upon incubation with human OGG1 or bacterial Fpg. In this novel design, 8-oxoguanine acts as a quencher, thereby providing a direct, real-time fluorescence signal that depends on its excision from DNA. Because of this direct measurement of enzymatic bond cleavage, the probe offers advantages over FRET-based probes, which require downstream events such as DNA backbone cleavage and dehybridization to yield a fluorescent signal, as well as other common indirect and time-consuming methods to measure 8-oxoguanine excision. Compared with a previously reported FRET sensor for OGG1,[51] our probe displays higher fluorescence fold-change, and its small size (relative to previous FRET-based probes) makes it easier and less costly to synthesize. Furthermore, the highly stable doublestranded hairpin design eliminates the need to carefully anneal the probe and obviates issues of strand stoichiometry. We expect that this and related probes will be useful for screening the effects of polymorphisms, mutations, post-translational modifications, and competitive inhibitors on 8-oxoguanine repair enzymes. In our molecular strategy for direct measurement of DNA base excision by fluorescence, the nucleobase to be excised acts as a fluorescence quencher. Specific coupling of the damaged base with a photophysically matched fluorescent nucleobase gives a probe that yields a light-up signal in real time only when the damaged base is removed. This strategy was previously tested with uracil deoxyglycosidase, and the uracil (which arises from hydrolytic damage of cytosine) was combined with a fluorescent pyrene nucleoside to report removal of uracil.[65, 66] The current experiments suggest that the application of such a fluorescence-quenching strategy might be more general for measuring base excision repair. Here we have shown that it can be employed for 8-oxoguanine, by taking advantage of previously unknown quenching of tCo by the damaged base. Fpg reacted approximately 18 times faster than human OGG1 with probe OGR1 under our conditions (Figure 2 B). Whereas OGG1 is known to have a strong preference for double-stranded DNA substrates with 8-oxoguanine paired opposite a cytosine,[25] Fpg also accepts single-stranded substrates.[24] The higher activity of Fpg on the new probe relative to OGG1 might reflect Fpg’s broader substrate requirement. The fact that OGR1 is a significantly less efficient substrate than natural DNA for OGG1 suggests that the extra steric bulk of the fluorescent cytosine derivative tCo interacts unfavorably in the OGG1 active site. However, the probe showed useful activity in vitro and in cell lysates. In the cell-based tests, the signal was slower to develop than with pure enzyme in vitro; we attribute this to inherently low enzyme activity in these systems and to significant cellular background fluorescence at
1641
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers pyrrolo-dC, 2-aminopurine, ethenoadenine, Dss, tCo phosphoramidites, 1 mmol columns of dG-CPG, dA-CPG, Universal Support III, and DNA synthesis reagents were purchased from Glen Research (Sterling, VA). 5’-DMT 3’-cyanoethyl a-pyrene phosphoramidite,[67] 5’-DMT 3’-cyanoethyl a-perylene phosphoramidite,[68] 5’-DMT 3’-cyanoethyl aphenylethynylpyrene phosphoramidite,[69] 5’-DMT 3’-cyanoethyl bN,N-dimethyl-2-(quinolin-2-yl)benzo[d]oxazol-6-amine phosphoramidite (QB),[70] and 5’-DMT 3’-cyanoethyl b-3-chloro-4-fluoro-indole phosphoramidite[62] were prepared according to published methods. All oligonucleotides excluding OGR37–40 were synthesized on a model 394 DNA/RNA synthesizer (Applied Biosystems) with a DMToff program. The coupling of each monomer used standard 3’-to-5’ cyanoethyl phosphoramidite chemistry with 999 s coupling times for pyrene, perylene, and phenylethynylpyrene, 180 s for tCo, and 240 s for other bases. Oligonucleotides OGR37–40 were synthesized by the Peptide and Nucleic Acid Facility (Stanford University, CA). Oligonucleotides containing 8-oxoguanine were deprotected in concentrated ammonium Figure 3. PAGE analysis of 4 mm probe OGR1 5’-labeled with [32P]phosphate and reacted with 658 nm A) hOGG1 or B) Fpg at 37 8C for the indicated times. Piperidine (200 mm) was added and incubated (95 8C, 5 min) to cleave the hydroxide with 2-mercaptoethanol DNA backbone at abasic sites as indicated to identify the 8-oxoguanine excision product and the backbone cleav(0.25 m) for 17 h at 55 8C. Dinucleoage product. Below: MALDI-TOF MS analysis of OGR1 following overnight reaction with C) hOGG1 or D) Fpg. tides containing pyrene, perylene, or phenylethynylpyrene (but not 8oxoguanine) were deprotected in potassium carbonate (0.05 m in methanol) for 24 h at room temperature. All other oligonucleotides this wavelength. Nevertheless, assays over 6–12 h gave reprowere deprotected in ammonium hydroxide for 17 h at room temducible signals of this enzyme activity. perature. The deprotected oligonucleotides were dried in a SpeedTaken together, the results show that the OGR1 probe perVac (Savant).
forms well in in vitro assays of enzyme activity with purified enzyme and with cell lysates, as well as for the measurement of activity in bacteria, which is expected to be useful for measurement of inhibition and of the effects of mutations on the enzymes. The probe gives a sufficiently high signal and low enough variation for use in high-throughput screens for competitive OGG1 inhibitors. A single DNA synthesis (1 mmol scale) would allow approximately 60 000 assays in a 384-well plate format. The fact that all components of probe OGR1 are commercially available makes it widely accessible for researchers.
Experimental Section
Oligonucleotides were purified by either HPLC or PAGE. For purification by HPLC, samples were purified on an HPLC device (Shimadzu) with a Jupiter C5 column (Phenomenex, Torrance, CA) or a Prosphere C18 column (Grace, Columbia, MD); acetonitrile and triethylammonium acetate (0.1 m, pH 7) were the eluents. For purification by PAGE, crude oligonucleotide was loaded onto a 20 % denaturing polyacrylamide gel (acrylamide/bis (19:1, 20 %), urea (7.5 m), tris borate EDTA buffer). The gel was cooled with a fan during electrophoresis. The gel was visualized by UV absorbance, and the darkest band was excised and extracted in NaCl (0.2 m) and 2-mercaptoethanol (0.25 m) overnight at 4 8C with shaking. The sample was dialyzed at 4 8C, filtered, dried in a SpeedVac, and resuspended in water.
Oligonucleotide preparation: Adenine, thymine, guanine, cytosine, 2’-O-methylcytosine, 8-oxoguanine, spacer (tetrahydrofuran),
Probe concentration was determined by fluorophore absorbance: pyrene e342 = 39 000 m¢1 cm¢1,[67] perylene e440 = 39 200 m¢1 cm¢1,[68]
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
1642
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Figure 6. Michaelis–Menten curves (n = 1) of 0.11 mm hOGG1 with varying concentrations of probe OGR1 and no inhibitor (&) or 2 mm inhibitor (*) in hOGG1 reaction buffer at 37 8C. The curves were fitted to the Michaelis– Menten equation with Origin software (c c). Figure 4. Reporting on 8-oxoguanine excision with bacterial cells expressing a repair enzyme. A) E. coli overexpressing Fpg after 17 h with 4 mm nucleaseprotected probe pOGR1 and 0.25 % Triton X-100 (c c); overexpressing Fpg after 17 h with no probe and 0.25 % Triton X-100 (c c); basal Fpg expression after 17 h with 4 mm probe and 0.25 % Triton X-100 (a a). B) E. coli overexpressing Fpg with 4 mm probe and 0.25 % Triton X-100 after 1 h (c c), 4 h (a a), or 17 h (c c).
Figure 5. Reporting 8-oxoguanine excision in mammalian cell extracts. Probe pOGR1 (2 mm) was incubated with 6.7 ng mL¢1 WT (~) or Ogg1¢/¢ (*) MEF cell extracts at 37 8C. The fluorescence at 460 nm (lex = 355 nm) was measured at the indicated time points. Background fluorescence (without probe) was subtracted. Error bars show standard deviation (n = 5).
2-aminopurine e260 = 1000 m¢1 cm¢1,[71] ethenoadenine e260 = 4800 m¢1 cm¢1,[72] pyrrolo-dC e345 = 3700 m¢1 cm¢1,[73] Dss e380 = 31 000 m¢1 cm¢1,[74] tCo e360 = 9000 m¢1 cm¢1,[59] phenylethynylpyrene e363 = 67 000 m¢1 cm¢1,[69] 3-Cl-4-F-indole e260 = 13 700 m¢1 cm¢1.[62] The QB extinction coefficient was calculated to be 22 400 m¢1 cm¢1 according the Beer–Lambert Law by measuring the absorbances of known concentrations of QB diol in methanol. For oligonucleotides with multiple nucleobases or fluorophores absorbing at the specified extinction-coefficient wavelength, the extinction coefficient was calculated as the sum of the extinction coefficients of each nucleobase or fluorophore in the oligonucleotide. Oligonucleotide identities were confirmed by MALDI-TOF MS performed at the Stanford Peptide and Nucleic Acid Facility. See
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
Figure 7. Probe OGR1 in 384-well assay. A) 0.8 mm probe OGR1 was incubated with (~) or without (*) 0.15 mm hOGG1 in 20 mL of buffer. The fluorescence at 460 nm (lex = 355 nm) was measured. Error bars represent standard deviation with (n = 5) or without (n = 3) the probe. B) Z’ factor at each timepoint (a: cutoff value of 0.4). C) Signal-to-noise ratio at each timepoint (a: cutoff value of 2).
1643
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Table S1 for the calculated and found masses. Samples were stored at ¢20 8C. Fluorescence analysis of in vitro enzymatic activity: Fpg reaction buffer (NEBuffer 1), hOGG1 reaction buffer (NEBuffer 4), and BSA (10 mg mL¢1) were purchased from New England Biolabs. The emission spectra were obtained with a Jobin Yvon Spex Fluorolog 3 spectrophotometer (Horiba, Kyoto, Japan), by exciting the sample at the fluorophore excitation maximum (pyrene 342 nm, perylene 440 nm, 2-aminopurine 303 nm, ethenoadenine 276 nm, pyrrolo-dC 347 nm, Dss 380 nm, tCo 360 nm, phenylethynylpyrene 363 nm, QB 400 nm) and measuring emission starting at 10 nm above the excitation wavelength (step size 1 nm). Aqueous solutions of 1 Õ Fpg or hOGG1 reaction buffer and BSA (100 mg mL¢1) were prepared in a quartz cuvette. The fluorescence emission spectrum was measured before addition of probe to determine buffer fluorescence. Fluorescence emission spectra were recorded immediately after addition of probe (4 mm) and after 10 min of incubation at 37 8C (to confirm stable fluorescence upon heating). For OGR1 studies, 658 nm of Fpg or hOGG1 (New England Biolabs) was added. For probe screening, Fpg (833 nm) or hOGG1 (658 nm) was added. The fluorescent emission spectrum was measured at various timepoints following addition of enzyme. The sample was maintained at 37 8C throughout. PAGE analysis of in vitro enzymatic activity: Probes were labeled at the 5’-end with 32P by using T4 polynucleotide kinase (Invitrogen) and [g-32P]ATP (PerkinElmer), and purified by ethanol precipitation as previously described.[75] BSA (100 mg mL¢1), unlabeled probe (4 mm), labeled probe (0.05 mm), and Fpg or hOGG1 (658 nm) were mixed in Fpg or hOGG1 reaction buffer and incubated at 37 8C. Aliquots (10 mL) were removed at the indicated timepoints. Non-piperidine-treated aliquots were mixed with loading buffer (10 mL; EDTA (25 mm), formamide (95 %), Bromophenol Blue (0.1 %), xylene cyanol (0.1 %)). Samples to be reacted with piperidine were mixed with piperidine/water (1:10, 2 mL) for 5 min at 90 8C and then mixed with the loading buffer. The samples were frozen for up to 24 h before PAGE analysis. Samples (10 mL) were loaded on a 20 % denaturing analytical PAGE gel (20 % acrylamide/ bis solution (19:1), urea (7.5 m), tris borate EDTA buffer) and electrophoresed. The gel was exposed overnight on a phosphor screen (Molecular Dynamics) and imaged on a Typhoon 9410 gel imager (GE Healthcare). MALDI-TOF MS analysis of enzymatic activity: Samples were prepared as for the fluorescence assays. An aliquot was submitted to the Stanford Peptide and Nucleic Acid Facility for zip-tip clean up and MALDI-TOF analysis. The samples were analyzed in HPA matrix in linear negative mode. Quantum yield measurements: Quinine sulfate (fluorescence quantum yield standard)[76] was diluted in H2SO4 (0.1 m), and the probes were diluted in water to an absorbance of 0.04–0.05 at 350 nm. For the reacted probes, the absorbance of protein in buffer was subtracted from the absorbance of the sample. The absorbance was measured at 350 nm on a Varian Cary 100 Bio UV/Vis spectrophotometer. The fluorescence emission spectrum was measured a Jobin Yvon-Spex Fluorolog 3 spectrophotometer (lex = 350 nm, lem = 360–680 nm, step size 1 nm). Quantum yield was calculated according to Fsample = (Fstandard)(Isample/Istandard)(Astandard/ Asample)(n2sample/n2standard), where F is quantum yield, I is the fluorescence emission over 360–680 nm, A is the absorbance at 350 nm, and n is the refractive index of the solvent for either quinine sulfate (standard) or probe (sample). ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
E. coli methods: In order to test OGR1 in bacterial cell solutions, we added 2’-O-methyl monomers for protection at the 5’ and 3’ ends of the probe thus generating probe pOGR1. The quenched quantum yield was 0.004–0.005, comparable to the quantum yield of probe OGR1 (0.008). As a control we synthesized an unquenched probe (G-OGR1; containing guanine instead of 8-oxoguanine), which has a quantum yield of 0.08. E. coli One Shot BL21(DE3) cells (Invitrogen/Life Technologies) were transformed with plasmid pKK223.3-WT Fpg (a gift from Prof. Sheila David, University of California–Davis)[77] by following the Invitrogen protocol. A single colony was cultured overnight at 30 8C in lysogeny broth (LB) containing ampicillin (50 mg mL¢1). The culture was diluted in LB containing ampicillin (50 mg mL¢1) to OD600 = 0.1 and grown at 37 8C to OD600 = 0.4. Isopropyl-1-thio-b-galactopyranoside (IPTG, 1 mm) was added, and the culture was grown for 3 or 24 h at 30 8C. The cells were washed in PBS, centrifuged (5000 g, 5 min, 4 8C), and resuspended in Fpg reaction buffer (to OD600 = 0.6). Triton X-100 (0.25 %) and probe (4 mm) were added. Samples were incubated with shaking at 37 8C for 1, 4, or 17 h, then diluted (1:10) in water, and fluorescence emission was measured on a Jobin Yvon-Spex Fluorolog 3 spectrophotometer (lex = 360 nm). Mammalian cell extracts: Ogg1¢/¢ MEF cells (a gift from Dr. Istvan Boldogh, University of Texas Medical Branch at Galveston)[47] and WT MEF cells were grown in DMEM/F12 (3:1) (HyClone) supplemented with fetal bovine serum (10 %), penicillin (100 U mL¢1), and streptomycin (100 mg mL¢1) in a humidified incubator at 37 8C with 5 % CO2. Cell extracts were obtained by the protocol by Folco et al.[78] with a few changes: the cells were collected by trypsinization at room temperature instead of scraping; Roche complete mini EDTA-free tablets were used instead of PMSF; the cells were lysed by passing 10–15 times through a 21 gauge needle and 5– 10 times through a 16 gauge needle instead of through a Dounce homogenizer. Total protein concentration was determined by a Bradford assay. Equal amounts (6.7 ng mL¢1) of WT and Ogg1¢/¢ MEF cell extract were incubated with probe pOGR1 (2 mm) and MgCl2 (10 mm) at 37 8C in 80 mL reaction volumes in a black 384-well plate. Fluorescence was measured at 460 nm on a Fluoroskan Ascent FL fluorescence plate reader (Thermo Scientific; lex = 355 nm). Melting temperature measurements: Samples containing probe OGR1 (4 mm) in Fpg or hOGG1 reaction buffer reaction buffer were prepared in stoppered 1 cm path-length quartz cells. A Varian Cary 100 UV/Vis spectrophotometer equipped with a thermoprogrammer was used to measure absorbance (260 nm) while the temperature was changed from 25 to 99 8C and from 99 to 25 8C (1 8C min¢1). Because of the high melting temperature of OGR1 and the lack of an absorbance plateau in the high temperature range, MeltWin software could not be used to determine the melting temperatures; melting temperatures were determined by visual inspection of the melting curves and by approximation of the first derivative of the melting curve. Measurement of kinetic parameters: Michaelis–Menten curves for hOGG1 with and without inhibitor (3-Cl-4-F-indole-containing oligonucleotides) were generated by preincubating hOGG1 (0.11 mm) with/without inhibitor (2 mm) in hOGG1 reaction buffer with BSA (100 mg mL¢1) at 37 8C for 5 min. OGR1 (0.5–10 mm) was then added at 37 8C, and the initial velocity was evaluated (rate of fluorescence increase (linear) at 460 nm over the first 10 min). Michaelis–Menten curves for Fpg were generated by incubating OGR1 (0.03–0.3 mm) with Fpg (6 nm) in Fpg reaction buffer with
1644
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers BSA (100 mg mL¢1) at 37 8C and evaluating the initial velocity at 37 8C (rate of fluorescence increase (linear) at 460 nm over the first 6 min). The curves were fitted in Origin (OriginLab, Northampton, MA) by using the Hill fit with n = 1 (yielding the Michaelis–Menten equation) to determine KM and Vmax. The dissociation constant, KI, of the hOGG1 inhibitor was calculated from apparent KM : KM, inhibitor = KM, no inhibitor(1 + ([I]/KI)).[79] To calculate kcat (kcat = Vmax/[enzyme]), we made a calibration curve to convert fluorescence units to amount of product by incubating known concentrations of probe with excess Fpg and measuring the fluorescence increase after the fluorescence signal plateaued. Vmax was converted from fluorescence units per s to mm product per s from the calibration curve. 384-well assay: OGR1 (0.8 mm) and hOGG1 (150 nm) were incubated in hOGG1 reaction buffer with BSA (100 mg mL¢1) at 37 8C in 20 mL reaction volumes in a black 384-well plate. Fluorescence at 460 nm was measured on a Thermo Fluoroskan Ascent FL fluorescence plate reader (lex = 355 nm). The signal-to-noise ratio was calculated by dividing the signal of OGR1 with hOGG1 by the signal of OGR1 without hOGG1. The coefficient of variance was calculated by dividing the standard deviation by the average signal intensity at that timepoint. Z’ factor was calculated from: 1¢
3 SDsignal þ 3 SDbaseline meansignal ¢meanbaseline
ð1Þ
Acknowledgements We thank Philip Yuen and Prof. Sheila David (University of California–Davis) for providing the plasmid encoding Fpg. We thank Dr. Istvan Boldogh (University of Texas Medical Branch at Galveston) for providing Ogg1¢/¢ MEF cells. This work was supported by the U.S. National Institutes of Health (GM067201 and GM106067), and the National Science Foundation (graduate fellowship to S.K.E.). Funding for the open access charge was provided by Stanford University. Keywords: DNA damage · fluorescent probes · Fpg · OGG1 · oxoguanine [1] H. Kasai, S. Nishimura, Nucleic Acids Res. 1984, 12, 2137 – 2145. [2] B. N. Ames, L. S. Gold, Mutat. Res. 1991, 250, 3 – 16. [3] H. Kasai, S. Nishimura, Y. Kurokawa, Y. Hayashi, Carcinogenesis 1987, 8, 1959 – 1961. [4] H. Kasai, S. Nishimura, Environ. Health Perspect. 1986, 67, 111 – 116. [5] E. S. Fiala, C. C. Conaway, J. E. Mathis, Cancer Res. 1989, 49, 5518 – 5522. [6] M. Kouchakdjian, V. Bodepudi, S. Shibutani, M. Eisenberg, F. Johnson, A. P. Grollman, D. J. Patel, Biochemistry 1991, 30, 1403 – 1412. [7] K. E. McAuley-Hecht, G. A. Leonard, N. J. Gibson, J. B. Thomson, W. P. Watson, W. N. Hunter, T. Brown, Biochemistry 1994, 33, 10266 – 10270. [8] K. C. Cheng, D. S. Cahill, H. Kasai, S. Nishimura, L. A. Loeb, J. Biol. Chem. 1992, 267, 166 – 172. [9] S. Shibutani, M. Takeshita, A. P. Grollman, Nature 1991, 349, 431 – 434. [10] M. L. Wood, M. Dizdaroglu, E. Gajewski, J. M. Essigmann, Biochemistry 1990, 29, 7024 – 7032. [11] R. De Bont, N. van Larebeke, Mutagenesis 2004, 19, 169 – 185. [12] D. O. Zharkov, T. A. Rosenquist, S. E. Gerchman, A. P. Grollman, J. Biol. Chem. 2000, 275, 28607 – 28617. [13] P.-M. Girard, N. Guibourt, S. Boiteux, Nucleic Acids Res. 1997, 25, 3204 – 3211.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
[14] J. W. Hill, T. K. Hazra, T. Izumi, S. Mitra, Nucleic Acids Res. 2001, 29, 430 – 438. [15] A. E. Vidal, I. D. Hickson, S. Boiteux, J. P. Radicella, Nucleic Acids Res. 2001, 29, 1285 – 1292. [16] T. Saitoh, K. Shinmura, S. Yamaguchi, M. Tani, S. Seki, H. Murakami, Y. Nojima, J. Yokota, Mutat. Res. 2001, 486, 31 – 40. [17] V. S. Sidorenko, G. A. Nevinsky, D. O. Zharkov, DNA Repair 2007, 6, 317 – 328. [18] B. Dalhus, M. Forsbring, I. H. Helle, E. S. Vik, R. J. Forstrøm, P. H. Backe, I. Alseth, M. Bjørs, Structure 2011, 19, 117 – 127. [19] B. Sun, K. A. Latham, M. L. Dodson, R. S. Lloyd, J. Biol. Chem. 1995, 270, 19501 – 19508. [20] V. Bailly, W. G. Verly, T. O’Connor, J. Laval, Biochem. J. 1989, 262, 581 – 589. [21] T. R. O’Connor, J. Laval, Proc. Natl. Acad. Sci. USA 1989, 86, 5222 – 5226. [22] R. J. Graves, I. Felzenszwalb, J. Laval, T. R. O’Connor, J. Biol. Chem. 1992, 267, 14429 – 14435. [23] M. Bhagwat, J. A. Gerlt, Biochemistry 1996, 35, 659 – 665. [24] A. A. Ishchenko, N. V. Bulychev, G. A. Maksakova, F. Johnson, G. A. Nevinsky, IUBMB Life 1999, 48, 613 – 618. [25] H. Dou, S. Mitra, T. K. Hazra, J. Biol. Chem. 2003, 278, 49679 – 49684. [26] O. Minowa, T. Arai, M. Hirano, Y. Monden, S. Nakai, M. Fukuda, M. Itoh, H. Takano, Y. Hippou, H. Aburatani, K.-i. Masamura, T. Nohmi, S. Nishimura, T. Noda, Proc. Natl. Acad. Sci. USA 2000, 97, 4156 – 4161. [27] A. Klungland, I. Rosewell, S. Hollenbach, E. Larsen, G. Daly, B. Epe, E. Seeberg, T. Lindahl, D. E. Barnes, Proc. Natl. Acad. Sci. USA 1999, 96, 13300 – 13305. [28] M. Cabrera, Y. Nghiem, J. H. Miller, J. Bacteriol. 1988, 170, 5405 – 5407. [29] D. Thomas, A. D. Scot, R. Barbey, M. Padula, S. Boiteux, Mol. Gen. Genet. 1997, 254, 171 – 178. [30] K. Sakumi, Y. Tominaga, M. Furuichi, P. Xu, T. Tsuzuki, M. Sekiguchi, Y. Nakabeppu, Cancer Res. 2003, 63, 902 – 905. [31] E. Speina, K. D. Arczewska, D. Gackowski, M. Zielin´ska, A. Siomek, J. Kowalewski, R. Olin´ski, B. Tudek, J. T. Kus´mierek, J. Natl. Cancer Inst. 2005, 97, 384 – 395. [32] S. Chevillard, J. P. Radicella, C. Levalois, J. Lebeau, M.-F. Poupon, S. Oudard, B. Dutrillaux, S. Boiteux, Oncogene 1998, 16, 3083 – 3086. [33] J.-W. Hyun, J.-Y. Choi, H.-H. Zeng, Y.-S. Lee, H.-S. Kim, S.-H. Yoon, M.-H. Chung, Oncogene 2000, 19, 4476 – 4479. [34] J.-W. Hyun, Y.-C. Jung, H.-S. Kim, E.-Y. Choi, J.-E. Kim, B.-H. Yoon, S.-H. Yoon, Y.-S. Lee, J. Choi, H.-J. You, M.-H. Chung, Mol. Cancer Res. 2003, 1, 290 – 299. [35] M. Audebert, J. P. Radicella, M. Dizdaroglu, Nucleic Acids Res. 2000, 28, 2672 – 2678. [36] M. Audebert, S. Chevillard, C. Levalois, G. Gyapay, A. Vieillefond, J. Klijanienko, P. Vielh, A. K. El Naggar, S. Oudard, S. Boiteux, J. P. Radicella, Cancer Res. 2000, 60, 4740 – 4744. [37] I.-J. Kim, J.-L. Ku, H. C. Kang, J.-H. Park, K.-A. Yoon, Y. Shin, H.-W. Park, S. G. Jang, S.-K. Lim, S. Y. Han, Y.-K. Shin, M. R. Lee, S.-Y. Jeong, H.-R. Shin, J.-S. Lee, W.-H. Kim, J.-G. Park, Hum. Genet. 2004, 115, 498 – 503. [38] P. Karihtala, S. Kauppila, U. Puistola, A. Jukkola-Vuorinen, Br. J. Cancer 2012, 106, 344 – 347. [39] S. D. Bruner, D. P. G. Norman, G. L. Verdine, Nature 2000, 403, 859 – 866. [40] D. Gackowski, E. Speina, M. Zielinska, J. Kowalewski, R. Rozalski, A. Siomek, T. Paciorek, B. Tudek, R. Olinski, Cancer Res. 2003, 63, 4899 – 4902. [41] T. Paz-Elizur, M. Krupsky, S. Blumenstein, D. Elinger, E. Schechtman, Z. Livneh, J. Natl. Cancer Inst. 2003, 95, 1312 – 1319. [42] T. Paz-Elizur, R. Ben-Yosef, D. Elinger, A. Vexler, M. Krupsky, A. Berrebi, A. Shani, E. Schechtman, L. Freedman, Z. Livneh, Cancer Res. 2006, 66, 11683 – 11689. [43] R. Gangwar, D. Ahirwar, A. Mandhani, R. D. Mittal, Mutat. Res. 2009, 680, 56 – 63. [44] G. Mao, X. Pan, B.-B. Zhu, Y. Zhang, F. Yuan, J. Huang, M. A. Lovell, M. P. Lee, W. R. Markesbery, G.-M. Li, L. Gu, Nucleic Acids Res. 2007, 35, 2759 – 2766. [45] V. S. Sidorenko, A. P. Grollman, P. Jaruga, M. Dizdaroglu, D. O. Zharkov, FEBS J. 2009, 276, 5149 – 5162. [46] S.-Y. Chen, L. Wan, C.-M. Huang, Y.-C. Huang, J. J.-C. Sheu, Y.-J. Lin, S.-P. Liu, Y.-C. Lan, C.-H. Lai, C.-W. Lin, C.-H. Tsai, F.-J. Tsai, Rheumatol. Int. 2012, 32, 1165 – 1169.
1645
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers [47] A. Bacsi, L. Aguilera-Aguirre, B. Szczesny, Z. Radak, T. K. Hazra, S. Sur, X. Ba, I. Boldogh, DNA Repair 2013, 12, 18 – 26. [48] B. Tyrberg, K. A. Anachkov, S. A. Dib, J. Wang-Rodriguez, K.-H. Yoon, F. Levine, BMC Endocr. Disord. 2002, 2, 2. [49] S. Simone, Y. Gorin, C. Velagapudi, H. E. Abboud, S. L. Habib, Diabetes 2008, 57, 2626 – 2636. [50] A. Maksimenko, A. A. Ishchenko, G. Sanz, J. Laval, R. H. Elder, M. K. Saparbaev, Biochem. Biophys. Res. Commun. 2004, 319, 240 – 246. [51] Y. Kim, I. S. Hong, Bull. Korean Chem. Soc. 2009, 30, 2149 – 2151. [52] L. Mirbahai, R. M. Kershaw, R. M. Green, R. E. Hayden, R. A. Meldrum, N. J. Hodges, DNA Repair 2010, 9, 144 – 152. [53] T. T. Nikiforov, S. Roman, Anal. Biochem. 2015, 477, 69 – 77. [54] J. L. Furman, P.-W. Mok, A. H. Badran, I. Ghosh, J. Am. Chem. Soc. 2011, 133, 12518 – 12527. [55] O. Nakagawa, S. Ono, Z. Li, A. Tsujimoto, S. Sasaki, Angew. Chem. Int. Ed. 2007, 46, 4500 – 4503; Angew. Chem. 2007, 119, 4584 – 4587. [56] T. Nasr, Z. Li, O. Nakagawa, Y. Taniguchi, S. Ono, S. Sasaki, Bioorg. Med. Chem. Lett. 2009, 19, 727 – 730. [57] Y. Taniguchi, R. Kawaguchi, S. Sasaki, J. Am. Chem. Soc. 2011, 133, 7272 – 7275. [58] J. N. Wilson, Y. Cho, S. Tan, A. Cuppoletti, E. T. Kool, ChemBioChem 2008, 9, 279 – 285. [59] P. Sandin, K. Bçrjesson, H. Li, J. Mrtensson, T. Brown, L. M. Wilhelmsson, B. Albinsson, Nucleic Acids Res. 2008, 36, 157 – 167. [60] S. Yoshizawa, T. Ueda, Y. Ishido, K. Miura, K. Watanabe, I. Hirao, Nucleic Acids Res. 1994, 22, 2217 – 2221. [61] T. J. Matray, E. T. Kool, J. Am. Chem. Soc. 1998, 120, 6191 – 6192. [62] P. L. McKibbin, A. Kobori, Y. Taniguchi, E. T. Kool, S. S. David, J. Am. Chem. Soc. 2012, 134, 1653 – 1661. [63] A. C. Jacobs, M. J. Calkins, A. Jadhav, D. Dorjsuren, D. Maloney, A. Simeonov, P. Jaruga, M. Dizdaroglu, A. K. McCullough, R. S. Lloyd, PLoS One 2013, 8, e81667. [64] M. Entzeroth, H. Flotow, P. Condron, Curr. Protoc. Pharmacol. 2009, Unit 9.4.
ChemBioChem 2015, 16, 1637 – 1646
www.chembiochem.org
[65] T. Ono, S. Wang, C.-K. Koo, L. Engstrom, S. S. David, E. T. Kool, Angew. Chem. Int. Ed. 2012, 51, 1689 – 1692; Angew. Chem. 2012, 124, 1721 – 1724. [66] T. Ono, S. K. Edwards, S. Wang, W. Jiang, E. T. Kool, Nucleic Acids Res. 2013, 41, e127. [67] R. X.-F. Ren, N. C. Chaudhuri, P. L. Paris, S. Rumney IV, E. T. Kool, J. Am. Chem. Soc. 1996, 118, 7671 – 7678. [68] J. Gao, S. Watanabe, E. T. Kool, J. Am. Chem. Soc. 2004, 126, 12748 – 12749. [69] S. Wang, J. Guo, T. Ono, E. T. Kool, Angew. Chem. Int. Ed. 2012, 51, 7176 – 7180; Angew. Chem. 2012, 124, 7288 – 7292. [70] L. H. Yuen, R. M. Franzini, S. Wang, P. Crisalli, V. Singh, W. Jiang, E. T. Kool, Angew. Chem. Int. Ed. 2014, 53, 5361 – 5365; Angew. Chem. 2014, 126, 5465 – 5469. [71] K. Evans, D. Xu, Y. Kim, T. M. Nordlund, J. Fluoresc. 1992, 2, 209 – 216. [72] J. A. Secrist III, J. R. Barrio, N. J. Leonard, G. Weber, Biochemistry 1972, 11, 3499 – 3506. [73] R. A. Tinsley, N. G. Walter, RNA 2006, 12, 522 – 529. [74] M. Kimoto, T. Mitsui, S. Yokoyama, I. Hirao, J. Am. Chem. Soc. 2010, 132, 4988 – 4989. [75] J. Sambrook, D. Russel, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, 2001, p. 10 – 17. [76] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Springer, Berlin, 2006, p. 54. [77] M. D. Leipold, J. G. Muller, C. J. Burrows, S. S. David, Biochemistry 2000, 39, 14984 – 14992. [78] E. G. Folco, H. Lei, J. L. Hsu, R. Reed, J. Vis. Exp. 2012, 64, 4140. [79] J. A. Todhunter in Methods in Enzymology, Vol. 63: Enzyme Kinetics and Mechanism Part A: Initial Rate and Inhibitor Methods (Ed.: D Purlich), Elsevier, Amsterdam, 1979, pp. 383 – 411. Manuscript received: April 10, 2015 Final article published: June 12, 2015
1646
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
