- Email: [email protected]
A homogeneous, high-throughput fluorescence resonance energy transfer-based DNA polymerase assay
ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 347 (2005) 254–261 www.elsevier.com/locate/yabio
A homogeneous, high-throughput fluorescence resonance energy transfer-based DNA polymerase assay Adam Shapiro a,*, Olga Rivin a, Ning Gao a, Laurel Hajec b a
b
Department of Biochemistry, Infection Discovery, AstraZeneca R&D Boston, Waltham, MA 02451, USA Department of Molecular Sciences, Infection Discovery, AstraZeneca R&D Boston, Waltham, MA 02451, USA Received 11 July 2005 Available online 13 October 2005
Abstract A homogeneous, fluorescence resonance energy transfer (FRET)-based DNA polymerase assay that is suitable for high-throughput screening for inhibitors, and can also be used for steady-state kinetic investigations, is described. The activity, kinetic mechanism, and processivity of the isolated a subunit of DNA polymerase III, the product of the dnaE gene, from the gram-negative pathogen Haemophilus influenzae were investigated using the FRET assay. 2005 Elsevier Inc. All rights reserved. Keywords: DNA polymerase; Fluorescence resonance energy transfer; FRET; DnaE; Pol III; High-throughput
DNA polymerases are required for DNA replication and repair in bacterial and eukaryotic cells. The catalytic a subunit of the replicative DNA polymerase III (pol III),1 for example, is an essential enzyme in bacteria [1–3] and, therefore, is suitable as an antimicrobial drug target. High-throughput screening of large libraries of drug-like molecules is frequently used in the pharmaceutical industry to identify target enzyme inhibitors as lead compounds for drug discovery. Traditional methods for measuring activity of DNA polymerases in vitro involve incorporation of radiolabeled deoxyribonucleoside triphosphate (dNTPs) into DNA and separation of the radioactive DNA product from the unused substrate. Such methods are not optimal for automated high-throughput screening for inhibitors *
Corresponding author. Fax: +1 781 839 4500. E-mail address: [email protected] (A. Shapiro). 1 Abbreviations used: pol III, polymerase III; dNTP, deoxyribonucleoside triphosphate; FRET, fluorescence resonance energy transfer; PCR, polymerase chain reaction; LB, Luria–Bertani; IPTG, isopropyl-b-Dthiogalactoside; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; LC–MS, liquid chromatography–mass spectrometry; HPLC, high-performance liquid chromatography; FAM, carboxyfluorescein; TAMRA, carboxytetramethylrhodamine; GC, guanine–cytosine; CGE, capillary gel electrophoresis. 0003-2697/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2005.09.023
due to the need for a separation step and the use of radioactivity. Other methods suitable for high-throughput screening for DNA polymerase inhibitors have been described in the literature. Earnshaw and Pope [4] described a scintillation proximity assay in which radiolabeled nucleotides were added by the enzyme to a biotinylated DNA substrate that was immobilized on a solid scintillant. Seville and coworkers [5] described an assay in which DNA polymerase activity was measured by the fluorescence intensity of a dye, the quantum yield of which increased when bound to double-stranded DNA. An assay for DNA polymerase activity described by Griep [6] used the recovery of the intrinsic ultraviolet fluorescence of single-stranded DNA binding protein on displacement from the newly synthesized double-stranded DNA. This method, although advantageous by virtue of providing a continuous signal, would not be suitable for high-throughput screening due to interference from UV absorbance and fluorescence of the test compounds. This article describes a novel, homogeneous fluorescence resonance energy transfer (FRET) assay for DNA polymerase activity that measures template-directed addition of natural dNTPs to the 3 0 -OH of a synthetic DNA primer.
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
The FRET assay has several advantages. It is fully compatible with automated high-throughput screening and is also useful for steady-state kinetic analysis and mechanistic studies. It requires no radioactivity or scintillant. By measuring the ratio of fluorescence intensities at two wavelengths, it achieves high measurement precision. Because the DNA substrate is covalently labeled with high-quantum yield fluorescent dyes, measurements can readily be made at low-nanomolar DNA concentrations and the samples from the assay can be reused for other analyses, such as capillary gel electrophoresis, to investigate the products. The activity, kinetics, and processivity of the isolated a subunit of DNA pol III, the product of the dnaE gene, from the gram-negative pathogen Haemophilus influenzae were investigated using the FRET assay. Materials and methods Reagents Highly homogeneous, fluorescently labeled DNA oligonucleotides were synthesized and purified by TriLink Biotechnologies. The primer and template were mixed together at 100 lM in 50 mM MOPS–HCl (pH 7.5) and 100 mM KCl and then annealed by heating briefly at 100 C and cooling slowly to room temperature. dNTPs were purchased from Sigma. DnaE proteins were cloned, expressed, and purified in-house as described below. Overexpression and purification of H. influenzae DnaE The dnaE gene was cloned from chromosomal DNA of wild-type H. influenzae strain ATCC 51907 using the polymerase chain reaction (PCR). Amplification was performed using High Fidelity PCR Master (Roche Applied Science) and the following primers: 5 0 -GGATTTCATATGTCAT CCCAACCTCGCTTCATCC-3 0 and 5 0 -CGAATTCGT CTTATTCAAACTCTAATTCCACC-3 0 (NdeI and EcoRI sites underlined). The PCR product was purified using the QuickStep 2 PCR Purification Kit (Edge Biosystems) and was digested with NdeI and EcoRI. The resulting fragment was then purified and ligated into NdeI- and EcoRI-digested expression vector pET30a (Novagen), producing the plasmid pET30a dnaE. The DNA sequence of the cloned dnaE was confirmed by sequencing on an ABI PRISM 3100 DNA sequencer (Applied Biosystems) using the Big Dye Terminator Cycle Sequencing Kit (Applied Biosystems). For protein overproduction, the plasmid was transformed into Escherichia coli BL21(DE3) (Novagen) and plated on Luria–Bertani (LB) containing 25 lg/ml of kanamycin at 37 C overnight. A single colony of BL21(DE3)/ pET30a–dnaE was inoculated into a 70-ml culture of Terrific Broth containing 25 lg/ml of kanamycin and grown overnight at 25 C. Samples (6 · 10 ml) of the overnight culture were added to 6 · 1 L of Terrific Broth containing 25 lg/ml of kanamycin and grown at 25 C with aeration to mid-logarithmic phase (OD600 = 0.5). Isopropyl-b-D-thi-
255
ogalactoside (IPTG) was added to a final concentration of 0.5 mM. After 3 h of induction, the cells were harvested by centrifugation at 5000g for 10 min at 4 C. Cell paste was stored at 20 C. Protein expression and solubility were checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). Purification of H. influenzae DnaE Frozen cell paste from E. coli cells expressing H. influenzae DnaE was suspended in 50 ml of lysis buffer (50 mM Tris–HCl [pH 7.5], 1 mM ethylenediamine tetraacetic acid [EDTA], 5 mM dithiothreitol, 10% [v/v] glycerol, 1 mM phenylmethylsulfonyl fluoride, and one protease inhibitor cocktail tablet) (Roche Molecular Biochemical). Cells were disrupted by being passed twice through a French press operated at 18,000 psi. The crude extract was centrifuged at 20,000 rpm in a Beckman 45Ti rotor for 30 min at 4 C. The supernatant was loaded at a flow rate of 1.5 ml/min onto a 20-ml Q Sepharose HP (HR 16/10) column (Pharmacia) preequilibrated with buffer A (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, and 10% [v/v] glycerol). The column was washed with buffer A, and the protein was eluted by a linear gradient from 0 to 1 M NaCl in buffer A. Fractions containing DnaE were pooled, and 3 M (NH4)2SO4 in buffer A was added to a final concentration of 0.8 M. The sample was applied at a flow rate of 1.0 ml/min to an 8-ml phenyl-Sepharose HP (HR 16/10) column (Pharmacia) preequilibrated with buffer B [buffer A containing 1 M (NH4)2SO4]. The column was washed with buffer B, and the protein was eluted by a linear gradient from 1 to 0 M (NH4)2SO4 in buffer A. Fractions containing DnaE were pooled (37.5 ml) and diluted with 100 ml of buffer A. The diluted sample was loaded at a flow rate of 1.0 ml/min to an 8-ml Heparin Sepharose CL-6B (HR 16/10) column (Pharmacia) preequilibrated with buffer A. The column was washed with buffer A, and the protein was eluted by a gradient from 0 to 1 M NaCl in buffer A. All chromatography steps were performed at 4 C. Fractions containing DnaE were pooled. Solid (NH4)2SO4 (0.4 g/ml) was added to precipitate all of the proteins and was mixed on ice for 1 h. The sample was centrifuged at 10,000 rpm for 30 min at 4 C in a Beckman JA12 rotor, and the pellet was dissolved in 5 ml of buffer A. The 5-ml sample was applied at a flow rate of 1.0 ml/min to a 320-ml Sephacryl S-300 (HR 26/60) column (Pharmacia) preequilibrated with buffer C (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, 10% [v/v] glycerol, and 150 mM NaCl). The fractions containing DnaE were pooled and dialyzed against 1 L of storage buffer (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, and 20% [v/v] glycerol). The protein was characterized by SDS–PAGE analysis and analytical liquid chromatography–mass spectrometry (LC–MS). The determined mass of the protein indicated that the N-terminal methionine of the DnaE predicted from the DNA sequence was not present (expected mass = 129,658 Da, observed
256
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
mass = 129,665 Da). Aliquots of the protein were flash-frozen and stored at 80 C. FRET assay In the FRET assay, annealed DNA substrate was mixed with DnaE protein and equimolar dATP and dGTP in a black, flat-bottomed, 96-well polystyrene assay plate (Greiner). Reactions were initiated with MgCl2 at a final concentration of 8 mM. The total reaction volume was 100 ll. The reaction buffer contained 50 mM MOPS–HCl (pH 7.5), 20% (v/v) glycerol, 10 mM dithiothreitol, 1 mM EDTA, 4 mM n-octyl-b-D-glucopyranoside, 0.002% (w/v) Brij-35, and 100 nM bovine serum albumin. Reactions were performed at room temperature. To terminate the reactions and separate unelongated primer from template strands, 100 ll of 8 M urea, 1 M Tris base, and 50 mM EDTA were added and mixed thoroughly. After a minimum of 15 min, the fluorescence in each well was read twice in a Tecan Ultra plate reader with a 485-nm excitation filter and 535- and 595-nm emission filters. HPLC assay A high-performance liquid chromatography (HPLC) assay was used to measure the consumption of denaturable DNA substrate in the FRET assay samples. The HPLC column was a 7.8-mM · 15-cm QC-PAC GFC 200 size exclusion column (Toso–Haas). The mobile phase contained 50 mM Tris–HCl (pH 7.0), 0.3 M NaCl, and 6 M urea. The reason for using 6 M urea at neutral pH as the denaturing mobile phase in the HPLC assay rather than 4 M urea with 0.5 M Tris base, the final denaturant condition in the FRET assay, is that the silica packing of the HPLC column is unstable at basic pH. The two denaturants are equally capable of denaturing the DNA substrate (data not shown). The flow rate was 1 ml/min at room temperature. A Waters HPLC system with a model 474 fluorescence detector was used. Fluorescence of the FAM-labeled primer was measured at an excitation wavelength of 485 nm and an emission wavelength of 535 nm. The area of the unelongated, carboxyfluorescein (FAM)-labeled primer peak was integrated. Undenatured primer–template complex was not detected due to essentially complete quenching by the proximity of the carboxytetramethylrhodamine (TAMRA) acceptor. Capillary gel electrophoresis The distribution of elongated primer lengths was measured in FRET assay samples by capillary gel electrophoresis and laser-induced fluorescence detection of FAM. Samples were mixed with an equal volume of deionized formamide (Applied Biosystems) to ensure complete denaturation and then were analyzed with a Beckman P/ACE MDQ capillary electrophoresis system. The capillary was
50 cm long with an internal diameter of 100 lm. The gel matrix was 100-R gel for single-stranded DNA. The running buffer was 44% Tris base, 56% boric acid, and 7 M urea. The running temperature was 30 C. The voltage was 13 kV. The detector was a Beckman Coulter 488-nm laser module with measurement at 520 nm emission. Results and discussion The basis for the DNA polymerase FRET assay is shown in Fig. 1A. A synthetic DNA primer carrying a 5 0 fluorescent donor or acceptor is annealed to a synthetic DNA template carrying a complementary 3 0 fluorescent acceptor or donor. The primer may carry the donor, and the template may carry the acceptor (or vice versa). The donor and acceptor are chosen so that energy transfer between them in the annealed primer–template pair substantially reduces the fluorescence intensity of the donor. In the work reported here, the donor was FAM and the acceptor was TAMRA. The length and sequence of the primer were chosen so that it can be readily separated from the template by a mild chemical denaturant. In the current case, denaturation was achieved in 4 M urea and 0.5 M Tris base, conditions that did not cause quenching of the fluorophores. Elongation of the primer by DNA polymerase activity in the presence of Mg2+ and an appropriate mixture of dNTPs results in a longer primer that is resistant to denaturation by the basic urea. Several design features of the DNA substrate (Fig. 1B) deserve mention. First, the template sequence was chosen to avoid hairpin formation that might interfere with DNA polymerase activity. Second, the template and primer complementary sequence was chosen to ensure a uniform binding location of the primer to the template. Guanine– cytosine (GC) base pairs were placed at the 5 0 and 3 0 ends of the complementary sequence to anchor the primer. Third, for reasons of economy, the length of the elongation region of the template was kept to the minimum necessary to ensure that elongation of the primer by DNA polymerase would generate a nondenaturable product. For the current work, the elongation region of the template consists of TC repeats, so that only dATP and dGTP are required in the assay. Because some aspects of the design of the DNA substrate were chosen out of convenience, it should be possible to design different DNA substrates such as those that require the incorporation of all four dNTPs. When changing the substrate, it would be prudent to test that the denaturation step is complete. This can be done by measuring the quenching of FAM–primer fluorescence by the addition of equimolar TAMRA–template under the conditions of the quenched assay. Quenching should be minimal if denaturation is complete. In this study, the DNA polymerase used was the a subunit of DNA pol III, the product of the dnaE gene. The protein is referred to herein as DnaE. In both gram-positive and gram-negative bacteria, DnaE is essential for rep-
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
257
Fig. 1. (A) Basis of the DNA polymerase FRET assay. In the presence of DNA polymerase (DnaE), Mg2+, and dNTPs, template-directed extension of the primer 3 0 end occurs. The substrate DNA is denaturable by the chemical denaturant, a basic urea solution, separating the FRET donor (D) and acceptor (A). When separated, energy transfer between the donor and acceptor does not occur, so that the donor is fluorescent (green D) and the acceptor is not (black A). The product DNA, in contrast, is not denaturable by the basic urea solution, so that the FRET donor and acceptor remain in proximity. Energy transfer occurs as a result, so that when the solution is exposed to light of a wavelength that excites the donor, the donor is not fluorescent (black D) and the acceptor is (red A). (B) DNA substrate used for the DNA polymerase FRET assay. Extension of the primer was performed using an equimolar mixture of dATP and dGTP. The melting temperature of the double-stranded part of the DNA substrate was calculated to be 34 C based on the GC content. The inclusion of three unpaired nucleotides at the 3 0 end of the primer strand is to avoid potential steric clashing of the two FRET probes. The necessity of this spacer has not been determined.
lication of the bacterial chromosome. Although the pol III holoenzyme contains 10 different polypeptides in E. coli, the isolated a subunit is capable of low-processivity DNA polymerization in vitro [7]. DnaE-catalyzed elongation of the primer resulted in an increase of FRET between the fluorescence donor (FAM) and acceptor (TAMRA) due to the resistance of the elongated product to denaturation by basic urea. The effect was observed as a decrease in FAM fluorescence at 535 nm and little change in TAMRA fluorescence at 595 nm when FAM was excited with 485 nm light (Fig. 2A). Monitoring the 535 nm FAM fluorescence is sufficient to measure polymerase activity. The fluorescence decrease is proportional to the fraction of substrate converted to nondenaturable product. A refinement of the measurement, however, is to measure the change in the ratio of 595 to 535 nm fluorescence when exciting FAM fluorescence with 485 nm light (Fig. 2B). The ratio measurement provides higher precision (Table 1) by correcting for slight differences in path length and meniscus geometry between assay plate wells. For example, when screening for DNA polymerase inhibitory compounds in multiwell assay plates, alterations of the liquid meniscus geometry, caused by effects on the surface tension by the test compounds, have less of an effect on the fluorescence intensity ratio measurement than on the single-wavelength intensity measurement. When using the two-wavelength fluorescence ratio measurement, the change in the ratio (DF595/F535) is a hyperbolic function of the fraction of substrate converted to nondenaturable product (fraction elongated, Fig. 2B). See Appendix A for a derivation of the equation for this function. The equation of this hyperbolic function can be mea-
sured experimentally by using HPLC size exclusion chromatography to measure substrate consumption in samples on which fluorescence measurements have been made. The HPLC separation was performed in 6 M urea at neutral pH, which has similar primer length-dependent ability to denature the DNA as the basic 4 M urea solution in the FRET assay. The parameters of the hyperbolic function depend on the efficiency of FRET between the donor and the acceptor in the substrate and optical characteristics of the assay sample and plate reader. Once the parameters of the hyperbolic function have been determined, DF595/ F535 can be converted to the fraction of substrate converted to nondenaturable product. The conversion is useful when using the FRET assay for steady-state kinetic analysis. It is not required, however, for percentage inhibition or IC50 measurements of inhibitory compounds because the deviation of the hyperbolic function from linearity was negligible when the fraction of substrate consumed was less than approximately 30%, that is, within the linear range of the assay. Fig. 2C shows the range over which the fraction elongated was proportional to the concentration of DnaE. The FRET assay was used for steady-state kinetic analysis of H. influenzae DnaE. The concentrations of DNA substrate and equimolar mixture of dATP and dGTP were varied, and the fraction of substrate converted to nondenaturable product was measured. For a fixed reaction time, the concentration of enzyme was varied to ensure that initial reaction rates pertained to all samples. The data were fit to an Ordered Bi model (Fig. 3), with the DNA template being the first-binding substrate. That the DNA is the first substrate to bind is certain because dNTP base recognition occurs by base pairing to the template. The Km values mea-
258
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
sured for H. influenzae DnaE with the FRET assay are similar to those measured for E. coli DnaE by Kim and McHenry [7] using a gap-filling assay: 22 lM for dNTPs and approximately 670 nM for DNA 3 0 -OH groups compared with 43 lM for dNTPs and 141 nM for DNA in the FRET assay. The gap-filling assay uses natural DNA that has been treated with DNase to introduce singlestranded gaps. Interestingly, it was not necessary to consider processivity of the DnaE enzyme to obtain good agreement to the
Table 1 Comparison of measurement precision in the DNA polymerase FRET assay using the change in single-wavelength fluorescence intensity at 535 nm versus the change in the ratio of fluorescence intensities at 595 and 535 nm DnaE (nM)
0 20
F535
DF535
Mean
SD
Mean
SD
CV (%)
Z0
34,242 24,767
685 664
— 9475
— 954
— 10.1
— 0.57
Mean
SD
Mean
SD
CV (%)
Z0
0.318 0.429
0.001 0.004
— 0.111
— 0.004
— 3.6
— 0.86
F595/F535 0 20
DF595/F535
Note. The data are from Fig. 2A. n = 8. SD, standard deviation; CV, coefficient of variation. Z 0 is a figure of merit for screening assays where Z0 ¼ 1
3 SDmax þ 3 SDmin ; j F max F min j
where Fmax and Fmin are the signals measured in the presence and absence of DnaE, respectively.
data by the kinetic model. McClure and Chow [8] described a kinetic model for processive DNA polymerization in which processivity affected only product terms in the denominator of the ordered Bi–Bi rate equation. Because only initial rate conditions were used in the current investigation, product terms were not used. Therefore, it was not necessary to account for processivity. The fact that the DNA primer was labeled with FAM allowed the distribution of primer product lengths of DnaE to be determined using capillary gel electrophoresis (CGE) combined with laser-induced fluorescence detection, the same method used for DNA sequencing. This analytical technique allowed single-base resolution and quantitation
b Fig. 2. (A) Effect of DNA polymerase activity on the 535- and 595-nm fluorescence intensities of the labeled DNA when using an excitation wavelength of 485 nm to excite FAM. The DnaE concentration was varied from 0 to 300 pM. The DNA concentration was 80 nM. The reaction time was 20 min. Data are the means and standard deviations of eight replicates. The concentrations of dATP and dGTP both were 30 lM. The 535-nm emission of the FAM donor decreased in proportion to the fraction of substrate elongated to product by DnaE. The 595-nm emission, which combines emission from the FAM donor and the TAMRA acceptor, remained essentially constant because the FAM emission decreased by approximately the same extent as the TAMRA emission increased. (B) Effect of DnaE-catalyzed elongation of the FAM-labeled primer into a nondenaturable product (fraction elongated) on the change in the ratio of 595 to 535 nm fluorescence of the DNA when excited with 485 nm light (DF595/F535 or DR). The fluorescence data are from the same experiment shown in panel A and are the ratios of eight replicates. The fraction of substrate elongated in each sample was measured by HPLC (see Materials and Methods). The data were graphed and fit to the aFE following equation: DR ¼ 1bFE ; where a = 0.282 and b = 0.915 using the SigmaPlot program (Systat Software) and FE represents fraction elongated. (C) Effect of DnaE enzyme concentration on the fraction of DNA substrate in which the primer was elongated, rendering it nondenaturable. The data are from the same experiment shown in (A) and are the means and standard deviations of eight replicates. Fluorescence ratio changes were converted to fraction elongated values using the following equation: DR FE ¼ aþbDR ; where a = 0.282 and b = 0.915.
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
259
Fig. 3. Steady-state kinetic analysis of H. influenzae DnaE-catalyzed primer elongation in the FRET assay. The DNA concentrations were varied from 20 to 300 nM. The concentration of dNTPs, which refers to the concentration of each nucleotide in an equimolar mixture of dATP and dGTP, was varied from 0 to 250 lM. The DnaE concentration was varied from 5 to 57 pM to obtain initial rates for all samples, such that the fraction of DNA elongated did not exceed 0.3. The reaction time was 30 min. The means and standard deviations of triplicate data were fit and graphed, using the Grafit program V max ½A½B . The kinetic parameters resulting from the global (Erithacus Software), to an Ordered Bi mechanism with DNA binding first: V ¼ K ia K mB þK mB ½AþK mA ½Bþ½A½B fit (±SE) were Vmax = 3.9 ± 0.2 lmol/min/mg, KmA(DNA) = 141 ± 12 nM, KmB(dNTPs) = 43 ± 4 lM, and Kia(DNA) = 251 ± 12 nM.
of primer product lengths from the FRET assay samples (Fig. 4). By adding the peak areas in the electropherograms and comparing them with the fraction elongated measurements from the FRET assay, it was found that the addition of 4 or 5 nucleotides to the primer, on average, was required to render the product nondenaturable under assay conditions. The processivity of the DnaE enzyme, which lacks the sliding clamp subunit of the DNA pol III holoenzyme, was clearly lower than that of the holoenzyme, one molecule of which is capable of continuously replicating the entire leading strand of the bacterial chromosome, given that fully elongated primers were not observed under conditions in which only a small amount of elongation occurred (Fig. 4A). The alternating heights of the product peaks, with the peaks terminating in A having greater area than the peaks terminating in G, indicate that DnaE was more likely to release the DNA after the addition of dATP and before the addition of dGTP than after the addition of dGTP and before the addition of dATP. Numerical simulations (not shown) revealed that the probability of release of the DNA by DnaE was approximately 0.8 after the addition of A and approximately 0.2 after the addition of G. The paucity of primers to which 14 or more nucleotides were added, even at very high DnaE concentrations (Fig. 4D), shows that the enzyme had difficulty in adding nucleotides to the primer within 4 nucleotides of the end of the template, the singlestranded elongation region of which was 17 nucleotides long. Detailed investigations of the determinants of processivity by DnaE or other DNA polymerases could be undertaken by varying the sequence and length of the elongation region of the template oligonucleotide and the composition of the dNTP mixture. The utility of the FRET assay described here for highly processive DNA polymerases, such as pol III holoenzyme, has not been tested and
may be limited by the shortness of the template relative to the large size of the polymerase complex. Longer templates could readily be substituted without affecting the assay because the products could still be distinguished from the substrates by their resistance to denaturation. Care should be taken to avoid designing templates that form stable hairpins that could interfere with elongation. A requirement for a longer primer would be more problematic because the longer primer might result in a substrate that is resistant to denaturation. This potential difficulty could be alleviated by using a low GC primer and also by increasing the strength of the denaturing conditions, for example, by increasing the final urea concentration. As with any fluorescence-based assay, fluorescent or strongly absorbing compounds in the screening library will interfere in the measurement of inhibition. Fortunately, interference is readily detectable by comparing the fluorescence intensity in the well with the expected fluorescence intensity in the absence of enzyme activity. In cases where the interference is relatively small (e.g., <2-fold), it is possible to correct for the effect mathematically. When the interference is relatively large, the data must be rejected. In the case of the FRET assay for DNA polymerase activity, optical interference in screening could be minimized by using fluorescent probes that have longer wavelength excitation and emission than does the FAM–TAMRA pair. This is due to the fact that the occurrence of strong absorbance and fluorescence of the lead- and drug-like compounds in high-throughput screening libraries decreases substantially as the wavelength increases. Conclusion The FRET assay for DNA polymerase activity described in this article offers several useful features. It is a low-cost,
260
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
nonradioactive homogeneous assay with a high Z 0 , making it suitable for high-throughput automated screening of large compound libraries to search for inhibitors of DNA poly-
merases as drug targets. The assay can also be used for enzymological studies of DNA polymerases. The synthetic DNA substrate can be varied to investigate effects of DNA
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
261
b Fig. 4. CGE analysis of the lengths of elongated FAM-labeled primers following reaction of the DNA with various concentrations of DnaE and complete denaturation by formamide. Elongation of the primer by one base, beginning with A and then alternating G and A, results in the peaks eluting at approximately equally spaced intervals after the substrate peak. A maximum of 17 nucleotides can be added to the primer because that is the number of nucleotides in the single-stranded elongation region of the template. The numbering beneath the electropherograms refers to the number of nucleotides added for each peak. Odd-numbered peaks represent primers terminating in A. Even-numbered peaks represent primers terminating in G. Peaks eluting earlier than the substrate (0) primer are due to degradation of the substrate. A single large peak was observed in the absence of DnaE (not shown). The DnaE concentrations were 5 pM (A), 20 pM (B), 60 pM (C), and 200 pM (D). The fractions of substrate elongated to nondenaturable product (FE) were 0.007 (A), 0.089 (B), 0.328 (C), and 0.771 (D). Each value of FE is the mean of four measurements.
sequence and length on polymerase activity. The FRET assay readout can be used for kinetic measurements, and the same samples can subsequently be analyzed by capillary gel electrophoresis to examine processivity.
Then D
c1 fðpÞ F 595 ¼ . F 535 c2 þ c3 fðpÞ
Let
Acknowledgments
a¼
c1 F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ ¼ c2 F 2535ðsÞ
Thanks are due to Laura Paganessi for providing assistance with cloning the H. influenzae dnaE gene and to Jim Whiteaker for growing E. coli cells expressing the DnaE protein.
and
Appendix A. Derivation of Eq. (1)
Then
F 535 ¼F 535ðsÞ fs þ F 535ðpÞ fp
D
F 595 ¼F 595ðsÞ fs þ F 595ðpÞ fp ; where F535 is the total fluorescence intensity at 535 nm, F595 is the total fluorescence intensity at 595 nm, F535(s) is the fluorescence intensity at 535 nm due to the denaturable substrate, F535(p) is the fluorescence intensity at 535 nm due to the nondenaturable product, F595(s) is the fluorescence intensity at 595 nm due to the denaturable substrate, F595(p) is the fluorescence intensity at 595 nm due to the nondenaturable product, fs is the fraction of denaturable substrate remaining, and fp is the fraction of substrate converted to nondenaturable product (FE): F 595 F 595ðsÞ fs þ F 595ðpÞ fðpÞ ¼ and F 535 F 535ðsÞ fs þ F 535ðpÞ fðpÞ F 595ðsÞ fs þ F 595ðpÞ fðpÞ F 595ðsÞ F 595 D ¼ DR ¼ ; F 535 F 535ðsÞ fs þ F 535ðpÞ fðpÞ F 535ðsÞ fs þ fp ¼ 1 D
F 595 F 535
f s ¼ 1 fp F 595ðsÞ 1 fp þ F 595ðpÞ fðpÞ F 595ðsÞ ¼ . F 535ðsÞ 1 fp þ F 535ðpÞ fðpÞ F 535ðsÞ and
This equation rearranges to F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ fðpÞ F 595 h i ¼ . D F 535 F 2 þ F 535ðsÞ F 535ðpÞ F2 fðpÞ 535ðsÞ
Let c1 ¼F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ ; c2 ¼F 2535ðsÞ ; c3 ¼F 535ðsÞ F 535ðpÞ F 2535ðsÞ .
535ðsÞ
2
b¼
F 535ðpÞ c3 F 535ðsÞ F 535ðsÞ F 535ðpÞ ¼ ¼1 . 2 c2 F 535ðsÞ F 535ðsÞ
a fðpÞ F 595 ¼ F 535 1 b fðpÞ
ð1Þ
and fðpÞ ¼ FE ¼
D FF 595 535 aþb
D FF 595 535
;
where DR ¼ D
F 595 . F 535
ð2Þ
Note that b is FRET efficiency from the donor to the acceptor References [1] M.M. Welch, C.S. McHenry, Cloning and identification of the product of the dnaE gene of Escherichia coli, J. Bacteriol. 152 (1982) 351–356. [2] R. Inoue, C. Kaito, M. Tanabe, K. Kamura, N. Akimitsu, K. Sekimizu, Genetic identification of two distinct DNA polymerases, DnaE and PolC, that are essential for chromosomal DNA replication in Staphylococcus aureus, Mol. Genet. Genom. 266 (2001) 564– 571. [3] E. Dervyn, C. Suski, R. Daniel, C. Bruand, J. Chapuis, J. Errington, L. Janniere, S.D. Ehrlich, Two essential DNA polymerases at the bacterial replication fork, Science 294 (2001) 1716– 1719. [4] D.L. Earnshaw, A.J. Pope, FlashPlate scintillation proximity assays for characterization and screening of DNA polymerases, primase, and helicase activities, J. Biomol. Screen. 6 (2001) 39–46. [5] M. Seville, A.B. West, M.G. Cull, C.S. McHenry, Fluorometric assay for DNA polymerases and reverse transcriptase, BioTechniques 21 (1996) 664–672. [6] M.A. Griep, Fluorescence recovery assay: a continuous assay for processive DNA polymerases applied specifically to DNA polymerase III holoenzyme, Anal. Biochem. 232 (1995) 180–189. [7] D.R. Kim, C.S. McHenry, In vivo assembly of overproduced DNA polymerase III, J. Biol. Chem. 271 (1996) 20681–20689. [8] W.R. McClure, Y. Chow, The kinetics and processivity of nucleic acid polymerases, Methods Enzymol. 64 (1980) 277–297.
A homogeneous, high-throughput fluorescence resonance energy transfer-based DNA polymerase assay Adam Shapiro a,*, Olga Rivin a, Ning Gao a, Laurel Hajec b a
b
Department of Biochemistry, Infection Discovery, AstraZeneca R&D Boston, Waltham, MA 02451, USA Department of Molecular Sciences, Infection Discovery, AstraZeneca R&D Boston, Waltham, MA 02451, USA Received 11 July 2005 Available online 13 October 2005
Abstract A homogeneous, fluorescence resonance energy transfer (FRET)-based DNA polymerase assay that is suitable for high-throughput screening for inhibitors, and can also be used for steady-state kinetic investigations, is described. The activity, kinetic mechanism, and processivity of the isolated a subunit of DNA polymerase III, the product of the dnaE gene, from the gram-negative pathogen Haemophilus influenzae were investigated using the FRET assay. 2005 Elsevier Inc. All rights reserved. Keywords: DNA polymerase; Fluorescence resonance energy transfer; FRET; DnaE; Pol III; High-throughput
DNA polymerases are required for DNA replication and repair in bacterial and eukaryotic cells. The catalytic a subunit of the replicative DNA polymerase III (pol III),1 for example, is an essential enzyme in bacteria [1–3] and, therefore, is suitable as an antimicrobial drug target. High-throughput screening of large libraries of drug-like molecules is frequently used in the pharmaceutical industry to identify target enzyme inhibitors as lead compounds for drug discovery. Traditional methods for measuring activity of DNA polymerases in vitro involve incorporation of radiolabeled deoxyribonucleoside triphosphate (dNTPs) into DNA and separation of the radioactive DNA product from the unused substrate. Such methods are not optimal for automated high-throughput screening for inhibitors *
Corresponding author. Fax: +1 781 839 4500. E-mail address: [email protected] (A. Shapiro). 1 Abbreviations used: pol III, polymerase III; dNTP, deoxyribonucleoside triphosphate; FRET, fluorescence resonance energy transfer; PCR, polymerase chain reaction; LB, Luria–Bertani; IPTG, isopropyl-b-Dthiogalactoside; SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gel electrophoresis; LC–MS, liquid chromatography–mass spectrometry; HPLC, high-performance liquid chromatography; FAM, carboxyfluorescein; TAMRA, carboxytetramethylrhodamine; GC, guanine–cytosine; CGE, capillary gel electrophoresis. 0003-2697/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2005.09.023
due to the need for a separation step and the use of radioactivity. Other methods suitable for high-throughput screening for DNA polymerase inhibitors have been described in the literature. Earnshaw and Pope [4] described a scintillation proximity assay in which radiolabeled nucleotides were added by the enzyme to a biotinylated DNA substrate that was immobilized on a solid scintillant. Seville and coworkers [5] described an assay in which DNA polymerase activity was measured by the fluorescence intensity of a dye, the quantum yield of which increased when bound to double-stranded DNA. An assay for DNA polymerase activity described by Griep [6] used the recovery of the intrinsic ultraviolet fluorescence of single-stranded DNA binding protein on displacement from the newly synthesized double-stranded DNA. This method, although advantageous by virtue of providing a continuous signal, would not be suitable for high-throughput screening due to interference from UV absorbance and fluorescence of the test compounds. This article describes a novel, homogeneous fluorescence resonance energy transfer (FRET) assay for DNA polymerase activity that measures template-directed addition of natural dNTPs to the 3 0 -OH of a synthetic DNA primer.
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
The FRET assay has several advantages. It is fully compatible with automated high-throughput screening and is also useful for steady-state kinetic analysis and mechanistic studies. It requires no radioactivity or scintillant. By measuring the ratio of fluorescence intensities at two wavelengths, it achieves high measurement precision. Because the DNA substrate is covalently labeled with high-quantum yield fluorescent dyes, measurements can readily be made at low-nanomolar DNA concentrations and the samples from the assay can be reused for other analyses, such as capillary gel electrophoresis, to investigate the products. The activity, kinetics, and processivity of the isolated a subunit of DNA pol III, the product of the dnaE gene, from the gram-negative pathogen Haemophilus influenzae were investigated using the FRET assay. Materials and methods Reagents Highly homogeneous, fluorescently labeled DNA oligonucleotides were synthesized and purified by TriLink Biotechnologies. The primer and template were mixed together at 100 lM in 50 mM MOPS–HCl (pH 7.5) and 100 mM KCl and then annealed by heating briefly at 100 C and cooling slowly to room temperature. dNTPs were purchased from Sigma. DnaE proteins were cloned, expressed, and purified in-house as described below. Overexpression and purification of H. influenzae DnaE The dnaE gene was cloned from chromosomal DNA of wild-type H. influenzae strain ATCC 51907 using the polymerase chain reaction (PCR). Amplification was performed using High Fidelity PCR Master (Roche Applied Science) and the following primers: 5 0 -GGATTTCATATGTCAT CCCAACCTCGCTTCATCC-3 0 and 5 0 -CGAATTCGT CTTATTCAAACTCTAATTCCACC-3 0 (NdeI and EcoRI sites underlined). The PCR product was purified using the QuickStep 2 PCR Purification Kit (Edge Biosystems) and was digested with NdeI and EcoRI. The resulting fragment was then purified and ligated into NdeI- and EcoRI-digested expression vector pET30a (Novagen), producing the plasmid pET30a dnaE. The DNA sequence of the cloned dnaE was confirmed by sequencing on an ABI PRISM 3100 DNA sequencer (Applied Biosystems) using the Big Dye Terminator Cycle Sequencing Kit (Applied Biosystems). For protein overproduction, the plasmid was transformed into Escherichia coli BL21(DE3) (Novagen) and plated on Luria–Bertani (LB) containing 25 lg/ml of kanamycin at 37 C overnight. A single colony of BL21(DE3)/ pET30a–dnaE was inoculated into a 70-ml culture of Terrific Broth containing 25 lg/ml of kanamycin and grown overnight at 25 C. Samples (6 · 10 ml) of the overnight culture were added to 6 · 1 L of Terrific Broth containing 25 lg/ml of kanamycin and grown at 25 C with aeration to mid-logarithmic phase (OD600 = 0.5). Isopropyl-b-D-thi-
255
ogalactoside (IPTG) was added to a final concentration of 0.5 mM. After 3 h of induction, the cells were harvested by centrifugation at 5000g for 10 min at 4 C. Cell paste was stored at 20 C. Protein expression and solubility were checked by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). Purification of H. influenzae DnaE Frozen cell paste from E. coli cells expressing H. influenzae DnaE was suspended in 50 ml of lysis buffer (50 mM Tris–HCl [pH 7.5], 1 mM ethylenediamine tetraacetic acid [EDTA], 5 mM dithiothreitol, 10% [v/v] glycerol, 1 mM phenylmethylsulfonyl fluoride, and one protease inhibitor cocktail tablet) (Roche Molecular Biochemical). Cells were disrupted by being passed twice through a French press operated at 18,000 psi. The crude extract was centrifuged at 20,000 rpm in a Beckman 45Ti rotor for 30 min at 4 C. The supernatant was loaded at a flow rate of 1.5 ml/min onto a 20-ml Q Sepharose HP (HR 16/10) column (Pharmacia) preequilibrated with buffer A (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, and 10% [v/v] glycerol). The column was washed with buffer A, and the protein was eluted by a linear gradient from 0 to 1 M NaCl in buffer A. Fractions containing DnaE were pooled, and 3 M (NH4)2SO4 in buffer A was added to a final concentration of 0.8 M. The sample was applied at a flow rate of 1.0 ml/min to an 8-ml phenyl-Sepharose HP (HR 16/10) column (Pharmacia) preequilibrated with buffer B [buffer A containing 1 M (NH4)2SO4]. The column was washed with buffer B, and the protein was eluted by a linear gradient from 1 to 0 M (NH4)2SO4 in buffer A. Fractions containing DnaE were pooled (37.5 ml) and diluted with 100 ml of buffer A. The diluted sample was loaded at a flow rate of 1.0 ml/min to an 8-ml Heparin Sepharose CL-6B (HR 16/10) column (Pharmacia) preequilibrated with buffer A. The column was washed with buffer A, and the protein was eluted by a gradient from 0 to 1 M NaCl in buffer A. All chromatography steps were performed at 4 C. Fractions containing DnaE were pooled. Solid (NH4)2SO4 (0.4 g/ml) was added to precipitate all of the proteins and was mixed on ice for 1 h. The sample was centrifuged at 10,000 rpm for 30 min at 4 C in a Beckman JA12 rotor, and the pellet was dissolved in 5 ml of buffer A. The 5-ml sample was applied at a flow rate of 1.0 ml/min to a 320-ml Sephacryl S-300 (HR 26/60) column (Pharmacia) preequilibrated with buffer C (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, 10% [v/v] glycerol, and 150 mM NaCl). The fractions containing DnaE were pooled and dialyzed against 1 L of storage buffer (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 5 mM dithiothreitol, and 20% [v/v] glycerol). The protein was characterized by SDS–PAGE analysis and analytical liquid chromatography–mass spectrometry (LC–MS). The determined mass of the protein indicated that the N-terminal methionine of the DnaE predicted from the DNA sequence was not present (expected mass = 129,658 Da, observed
256
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
mass = 129,665 Da). Aliquots of the protein were flash-frozen and stored at 80 C. FRET assay In the FRET assay, annealed DNA substrate was mixed with DnaE protein and equimolar dATP and dGTP in a black, flat-bottomed, 96-well polystyrene assay plate (Greiner). Reactions were initiated with MgCl2 at a final concentration of 8 mM. The total reaction volume was 100 ll. The reaction buffer contained 50 mM MOPS–HCl (pH 7.5), 20% (v/v) glycerol, 10 mM dithiothreitol, 1 mM EDTA, 4 mM n-octyl-b-D-glucopyranoside, 0.002% (w/v) Brij-35, and 100 nM bovine serum albumin. Reactions were performed at room temperature. To terminate the reactions and separate unelongated primer from template strands, 100 ll of 8 M urea, 1 M Tris base, and 50 mM EDTA were added and mixed thoroughly. After a minimum of 15 min, the fluorescence in each well was read twice in a Tecan Ultra plate reader with a 485-nm excitation filter and 535- and 595-nm emission filters. HPLC assay A high-performance liquid chromatography (HPLC) assay was used to measure the consumption of denaturable DNA substrate in the FRET assay samples. The HPLC column was a 7.8-mM · 15-cm QC-PAC GFC 200 size exclusion column (Toso–Haas). The mobile phase contained 50 mM Tris–HCl (pH 7.0), 0.3 M NaCl, and 6 M urea. The reason for using 6 M urea at neutral pH as the denaturing mobile phase in the HPLC assay rather than 4 M urea with 0.5 M Tris base, the final denaturant condition in the FRET assay, is that the silica packing of the HPLC column is unstable at basic pH. The two denaturants are equally capable of denaturing the DNA substrate (data not shown). The flow rate was 1 ml/min at room temperature. A Waters HPLC system with a model 474 fluorescence detector was used. Fluorescence of the FAM-labeled primer was measured at an excitation wavelength of 485 nm and an emission wavelength of 535 nm. The area of the unelongated, carboxyfluorescein (FAM)-labeled primer peak was integrated. Undenatured primer–template complex was not detected due to essentially complete quenching by the proximity of the carboxytetramethylrhodamine (TAMRA) acceptor. Capillary gel electrophoresis The distribution of elongated primer lengths was measured in FRET assay samples by capillary gel electrophoresis and laser-induced fluorescence detection of FAM. Samples were mixed with an equal volume of deionized formamide (Applied Biosystems) to ensure complete denaturation and then were analyzed with a Beckman P/ACE MDQ capillary electrophoresis system. The capillary was
50 cm long with an internal diameter of 100 lm. The gel matrix was 100-R gel for single-stranded DNA. The running buffer was 44% Tris base, 56% boric acid, and 7 M urea. The running temperature was 30 C. The voltage was 13 kV. The detector was a Beckman Coulter 488-nm laser module with measurement at 520 nm emission. Results and discussion The basis for the DNA polymerase FRET assay is shown in Fig. 1A. A synthetic DNA primer carrying a 5 0 fluorescent donor or acceptor is annealed to a synthetic DNA template carrying a complementary 3 0 fluorescent acceptor or donor. The primer may carry the donor, and the template may carry the acceptor (or vice versa). The donor and acceptor are chosen so that energy transfer between them in the annealed primer–template pair substantially reduces the fluorescence intensity of the donor. In the work reported here, the donor was FAM and the acceptor was TAMRA. The length and sequence of the primer were chosen so that it can be readily separated from the template by a mild chemical denaturant. In the current case, denaturation was achieved in 4 M urea and 0.5 M Tris base, conditions that did not cause quenching of the fluorophores. Elongation of the primer by DNA polymerase activity in the presence of Mg2+ and an appropriate mixture of dNTPs results in a longer primer that is resistant to denaturation by the basic urea. Several design features of the DNA substrate (Fig. 1B) deserve mention. First, the template sequence was chosen to avoid hairpin formation that might interfere with DNA polymerase activity. Second, the template and primer complementary sequence was chosen to ensure a uniform binding location of the primer to the template. Guanine– cytosine (GC) base pairs were placed at the 5 0 and 3 0 ends of the complementary sequence to anchor the primer. Third, for reasons of economy, the length of the elongation region of the template was kept to the minimum necessary to ensure that elongation of the primer by DNA polymerase would generate a nondenaturable product. For the current work, the elongation region of the template consists of TC repeats, so that only dATP and dGTP are required in the assay. Because some aspects of the design of the DNA substrate were chosen out of convenience, it should be possible to design different DNA substrates such as those that require the incorporation of all four dNTPs. When changing the substrate, it would be prudent to test that the denaturation step is complete. This can be done by measuring the quenching of FAM–primer fluorescence by the addition of equimolar TAMRA–template under the conditions of the quenched assay. Quenching should be minimal if denaturation is complete. In this study, the DNA polymerase used was the a subunit of DNA pol III, the product of the dnaE gene. The protein is referred to herein as DnaE. In both gram-positive and gram-negative bacteria, DnaE is essential for rep-
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
257
Fig. 1. (A) Basis of the DNA polymerase FRET assay. In the presence of DNA polymerase (DnaE), Mg2+, and dNTPs, template-directed extension of the primer 3 0 end occurs. The substrate DNA is denaturable by the chemical denaturant, a basic urea solution, separating the FRET donor (D) and acceptor (A). When separated, energy transfer between the donor and acceptor does not occur, so that the donor is fluorescent (green D) and the acceptor is not (black A). The product DNA, in contrast, is not denaturable by the basic urea solution, so that the FRET donor and acceptor remain in proximity. Energy transfer occurs as a result, so that when the solution is exposed to light of a wavelength that excites the donor, the donor is not fluorescent (black D) and the acceptor is (red A). (B) DNA substrate used for the DNA polymerase FRET assay. Extension of the primer was performed using an equimolar mixture of dATP and dGTP. The melting temperature of the double-stranded part of the DNA substrate was calculated to be 34 C based on the GC content. The inclusion of three unpaired nucleotides at the 3 0 end of the primer strand is to avoid potential steric clashing of the two FRET probes. The necessity of this spacer has not been determined.
lication of the bacterial chromosome. Although the pol III holoenzyme contains 10 different polypeptides in E. coli, the isolated a subunit is capable of low-processivity DNA polymerization in vitro [7]. DnaE-catalyzed elongation of the primer resulted in an increase of FRET between the fluorescence donor (FAM) and acceptor (TAMRA) due to the resistance of the elongated product to denaturation by basic urea. The effect was observed as a decrease in FAM fluorescence at 535 nm and little change in TAMRA fluorescence at 595 nm when FAM was excited with 485 nm light (Fig. 2A). Monitoring the 535 nm FAM fluorescence is sufficient to measure polymerase activity. The fluorescence decrease is proportional to the fraction of substrate converted to nondenaturable product. A refinement of the measurement, however, is to measure the change in the ratio of 595 to 535 nm fluorescence when exciting FAM fluorescence with 485 nm light (Fig. 2B). The ratio measurement provides higher precision (Table 1) by correcting for slight differences in path length and meniscus geometry between assay plate wells. For example, when screening for DNA polymerase inhibitory compounds in multiwell assay plates, alterations of the liquid meniscus geometry, caused by effects on the surface tension by the test compounds, have less of an effect on the fluorescence intensity ratio measurement than on the single-wavelength intensity measurement. When using the two-wavelength fluorescence ratio measurement, the change in the ratio (DF595/F535) is a hyperbolic function of the fraction of substrate converted to nondenaturable product (fraction elongated, Fig. 2B). See Appendix A for a derivation of the equation for this function. The equation of this hyperbolic function can be mea-
sured experimentally by using HPLC size exclusion chromatography to measure substrate consumption in samples on which fluorescence measurements have been made. The HPLC separation was performed in 6 M urea at neutral pH, which has similar primer length-dependent ability to denature the DNA as the basic 4 M urea solution in the FRET assay. The parameters of the hyperbolic function depend on the efficiency of FRET between the donor and the acceptor in the substrate and optical characteristics of the assay sample and plate reader. Once the parameters of the hyperbolic function have been determined, DF595/ F535 can be converted to the fraction of substrate converted to nondenaturable product. The conversion is useful when using the FRET assay for steady-state kinetic analysis. It is not required, however, for percentage inhibition or IC50 measurements of inhibitory compounds because the deviation of the hyperbolic function from linearity was negligible when the fraction of substrate consumed was less than approximately 30%, that is, within the linear range of the assay. Fig. 2C shows the range over which the fraction elongated was proportional to the concentration of DnaE. The FRET assay was used for steady-state kinetic analysis of H. influenzae DnaE. The concentrations of DNA substrate and equimolar mixture of dATP and dGTP were varied, and the fraction of substrate converted to nondenaturable product was measured. For a fixed reaction time, the concentration of enzyme was varied to ensure that initial reaction rates pertained to all samples. The data were fit to an Ordered Bi model (Fig. 3), with the DNA template being the first-binding substrate. That the DNA is the first substrate to bind is certain because dNTP base recognition occurs by base pairing to the template. The Km values mea-
258
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
sured for H. influenzae DnaE with the FRET assay are similar to those measured for E. coli DnaE by Kim and McHenry [7] using a gap-filling assay: 22 lM for dNTPs and approximately 670 nM for DNA 3 0 -OH groups compared with 43 lM for dNTPs and 141 nM for DNA in the FRET assay. The gap-filling assay uses natural DNA that has been treated with DNase to introduce singlestranded gaps. Interestingly, it was not necessary to consider processivity of the DnaE enzyme to obtain good agreement to the
Table 1 Comparison of measurement precision in the DNA polymerase FRET assay using the change in single-wavelength fluorescence intensity at 535 nm versus the change in the ratio of fluorescence intensities at 595 and 535 nm DnaE (nM)
0 20
F535
DF535
Mean
SD
Mean
SD
CV (%)
Z0
34,242 24,767
685 664
— 9475
— 954
— 10.1
— 0.57
Mean
SD
Mean
SD
CV (%)
Z0
0.318 0.429
0.001 0.004
— 0.111
— 0.004
— 3.6
— 0.86
F595/F535 0 20
DF595/F535
Note. The data are from Fig. 2A. n = 8. SD, standard deviation; CV, coefficient of variation. Z 0 is a figure of merit for screening assays where Z0 ¼ 1
3 SDmax þ 3 SDmin ; j F max F min j
where Fmax and Fmin are the signals measured in the presence and absence of DnaE, respectively.
data by the kinetic model. McClure and Chow [8] described a kinetic model for processive DNA polymerization in which processivity affected only product terms in the denominator of the ordered Bi–Bi rate equation. Because only initial rate conditions were used in the current investigation, product terms were not used. Therefore, it was not necessary to account for processivity. The fact that the DNA primer was labeled with FAM allowed the distribution of primer product lengths of DnaE to be determined using capillary gel electrophoresis (CGE) combined with laser-induced fluorescence detection, the same method used for DNA sequencing. This analytical technique allowed single-base resolution and quantitation
b Fig. 2. (A) Effect of DNA polymerase activity on the 535- and 595-nm fluorescence intensities of the labeled DNA when using an excitation wavelength of 485 nm to excite FAM. The DnaE concentration was varied from 0 to 300 pM. The DNA concentration was 80 nM. The reaction time was 20 min. Data are the means and standard deviations of eight replicates. The concentrations of dATP and dGTP both were 30 lM. The 535-nm emission of the FAM donor decreased in proportion to the fraction of substrate elongated to product by DnaE. The 595-nm emission, which combines emission from the FAM donor and the TAMRA acceptor, remained essentially constant because the FAM emission decreased by approximately the same extent as the TAMRA emission increased. (B) Effect of DnaE-catalyzed elongation of the FAM-labeled primer into a nondenaturable product (fraction elongated) on the change in the ratio of 595 to 535 nm fluorescence of the DNA when excited with 485 nm light (DF595/F535 or DR). The fluorescence data are from the same experiment shown in panel A and are the ratios of eight replicates. The fraction of substrate elongated in each sample was measured by HPLC (see Materials and Methods). The data were graphed and fit to the aFE following equation: DR ¼ 1bFE ; where a = 0.282 and b = 0.915 using the SigmaPlot program (Systat Software) and FE represents fraction elongated. (C) Effect of DnaE enzyme concentration on the fraction of DNA substrate in which the primer was elongated, rendering it nondenaturable. The data are from the same experiment shown in (A) and are the means and standard deviations of eight replicates. Fluorescence ratio changes were converted to fraction elongated values using the following equation: DR FE ¼ aþbDR ; where a = 0.282 and b = 0.915.
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
259
Fig. 3. Steady-state kinetic analysis of H. influenzae DnaE-catalyzed primer elongation in the FRET assay. The DNA concentrations were varied from 20 to 300 nM. The concentration of dNTPs, which refers to the concentration of each nucleotide in an equimolar mixture of dATP and dGTP, was varied from 0 to 250 lM. The DnaE concentration was varied from 5 to 57 pM to obtain initial rates for all samples, such that the fraction of DNA elongated did not exceed 0.3. The reaction time was 30 min. The means and standard deviations of triplicate data were fit and graphed, using the Grafit program V max ½A½B . The kinetic parameters resulting from the global (Erithacus Software), to an Ordered Bi mechanism with DNA binding first: V ¼ K ia K mB þK mB ½AþK mA ½Bþ½A½B fit (±SE) were Vmax = 3.9 ± 0.2 lmol/min/mg, KmA(DNA) = 141 ± 12 nM, KmB(dNTPs) = 43 ± 4 lM, and Kia(DNA) = 251 ± 12 nM.
of primer product lengths from the FRET assay samples (Fig. 4). By adding the peak areas in the electropherograms and comparing them with the fraction elongated measurements from the FRET assay, it was found that the addition of 4 or 5 nucleotides to the primer, on average, was required to render the product nondenaturable under assay conditions. The processivity of the DnaE enzyme, which lacks the sliding clamp subunit of the DNA pol III holoenzyme, was clearly lower than that of the holoenzyme, one molecule of which is capable of continuously replicating the entire leading strand of the bacterial chromosome, given that fully elongated primers were not observed under conditions in which only a small amount of elongation occurred (Fig. 4A). The alternating heights of the product peaks, with the peaks terminating in A having greater area than the peaks terminating in G, indicate that DnaE was more likely to release the DNA after the addition of dATP and before the addition of dGTP than after the addition of dGTP and before the addition of dATP. Numerical simulations (not shown) revealed that the probability of release of the DNA by DnaE was approximately 0.8 after the addition of A and approximately 0.2 after the addition of G. The paucity of primers to which 14 or more nucleotides were added, even at very high DnaE concentrations (Fig. 4D), shows that the enzyme had difficulty in adding nucleotides to the primer within 4 nucleotides of the end of the template, the singlestranded elongation region of which was 17 nucleotides long. Detailed investigations of the determinants of processivity by DnaE or other DNA polymerases could be undertaken by varying the sequence and length of the elongation region of the template oligonucleotide and the composition of the dNTP mixture. The utility of the FRET assay described here for highly processive DNA polymerases, such as pol III holoenzyme, has not been tested and
may be limited by the shortness of the template relative to the large size of the polymerase complex. Longer templates could readily be substituted without affecting the assay because the products could still be distinguished from the substrates by their resistance to denaturation. Care should be taken to avoid designing templates that form stable hairpins that could interfere with elongation. A requirement for a longer primer would be more problematic because the longer primer might result in a substrate that is resistant to denaturation. This potential difficulty could be alleviated by using a low GC primer and also by increasing the strength of the denaturing conditions, for example, by increasing the final urea concentration. As with any fluorescence-based assay, fluorescent or strongly absorbing compounds in the screening library will interfere in the measurement of inhibition. Fortunately, interference is readily detectable by comparing the fluorescence intensity in the well with the expected fluorescence intensity in the absence of enzyme activity. In cases where the interference is relatively small (e.g., <2-fold), it is possible to correct for the effect mathematically. When the interference is relatively large, the data must be rejected. In the case of the FRET assay for DNA polymerase activity, optical interference in screening could be minimized by using fluorescent probes that have longer wavelength excitation and emission than does the FAM–TAMRA pair. This is due to the fact that the occurrence of strong absorbance and fluorescence of the lead- and drug-like compounds in high-throughput screening libraries decreases substantially as the wavelength increases. Conclusion The FRET assay for DNA polymerase activity described in this article offers several useful features. It is a low-cost,
260
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
nonradioactive homogeneous assay with a high Z 0 , making it suitable for high-throughput automated screening of large compound libraries to search for inhibitors of DNA poly-
merases as drug targets. The assay can also be used for enzymological studies of DNA polymerases. The synthetic DNA substrate can be varied to investigate effects of DNA
FRET-based DNA polymerase assay / A. Shapiro et al. / Anal. Biochem. 347 (2005) 254–261
261
b Fig. 4. CGE analysis of the lengths of elongated FAM-labeled primers following reaction of the DNA with various concentrations of DnaE and complete denaturation by formamide. Elongation of the primer by one base, beginning with A and then alternating G and A, results in the peaks eluting at approximately equally spaced intervals after the substrate peak. A maximum of 17 nucleotides can be added to the primer because that is the number of nucleotides in the single-stranded elongation region of the template. The numbering beneath the electropherograms refers to the number of nucleotides added for each peak. Odd-numbered peaks represent primers terminating in A. Even-numbered peaks represent primers terminating in G. Peaks eluting earlier than the substrate (0) primer are due to degradation of the substrate. A single large peak was observed in the absence of DnaE (not shown). The DnaE concentrations were 5 pM (A), 20 pM (B), 60 pM (C), and 200 pM (D). The fractions of substrate elongated to nondenaturable product (FE) were 0.007 (A), 0.089 (B), 0.328 (C), and 0.771 (D). Each value of FE is the mean of four measurements.
sequence and length on polymerase activity. The FRET assay readout can be used for kinetic measurements, and the same samples can subsequently be analyzed by capillary gel electrophoresis to examine processivity.
Then D
c1 fðpÞ F 595 ¼ . F 535 c2 þ c3 fðpÞ
Let
Acknowledgments
a¼
c1 F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ ¼ c2 F 2535ðsÞ
Thanks are due to Laura Paganessi for providing assistance with cloning the H. influenzae dnaE gene and to Jim Whiteaker for growing E. coli cells expressing the DnaE protein.
and
Appendix A. Derivation of Eq. (1)
Then
F 535 ¼F 535ðsÞ fs þ F 535ðpÞ fp
D
F 595 ¼F 595ðsÞ fs þ F 595ðpÞ fp ; where F535 is the total fluorescence intensity at 535 nm, F595 is the total fluorescence intensity at 595 nm, F535(s) is the fluorescence intensity at 535 nm due to the denaturable substrate, F535(p) is the fluorescence intensity at 535 nm due to the nondenaturable product, F595(s) is the fluorescence intensity at 595 nm due to the denaturable substrate, F595(p) is the fluorescence intensity at 595 nm due to the nondenaturable product, fs is the fraction of denaturable substrate remaining, and fp is the fraction of substrate converted to nondenaturable product (FE): F 595 F 595ðsÞ fs þ F 595ðpÞ fðpÞ ¼ and F 535 F 535ðsÞ fs þ F 535ðpÞ fðpÞ F 595ðsÞ fs þ F 595ðpÞ fðpÞ F 595ðsÞ F 595 D ¼ DR ¼ ; F 535 F 535ðsÞ fs þ F 535ðpÞ fðpÞ F 535ðsÞ fs þ fp ¼ 1 D
F 595 F 535
f s ¼ 1 fp F 595ðsÞ 1 fp þ F 595ðpÞ fðpÞ F 595ðsÞ ¼ . F 535ðsÞ 1 fp þ F 535ðpÞ fðpÞ F 535ðsÞ and
This equation rearranges to F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ fðpÞ F 595 h i ¼ . D F 535 F 2 þ F 535ðsÞ F 535ðpÞ F2 fðpÞ 535ðsÞ
Let c1 ¼F 535ðsÞ F 595ðpÞ F 535ðpÞ F 595ðsÞ ; c2 ¼F 2535ðsÞ ; c3 ¼F 535ðsÞ F 535ðpÞ F 2535ðsÞ .
535ðsÞ
2
b¼
F 535ðpÞ c3 F 535ðsÞ F 535ðsÞ F 535ðpÞ ¼ ¼1 . 2 c2 F 535ðsÞ F 535ðsÞ
a fðpÞ F 595 ¼ F 535 1 b fðpÞ
ð1Þ
and fðpÞ ¼ FE ¼
D FF 595 535 aþb
D FF 595 535
;
where DR ¼ D
F 595 . F 535
ð2Þ
Note that b is FRET efficiency from the donor to the acceptor References [1] M.M. Welch, C.S. McHenry, Cloning and identification of the product of the dnaE gene of Escherichia coli, J. Bacteriol. 152 (1982) 351–356. [2] R. Inoue, C. Kaito, M. Tanabe, K. Kamura, N. Akimitsu, K. Sekimizu, Genetic identification of two distinct DNA polymerases, DnaE and PolC, that are essential for chromosomal DNA replication in Staphylococcus aureus, Mol. Genet. Genom. 266 (2001) 564– 571. [3] E. Dervyn, C. Suski, R. Daniel, C. Bruand, J. Chapuis, J. Errington, L. Janniere, S.D. Ehrlich, Two essential DNA polymerases at the bacterial replication fork, Science 294 (2001) 1716– 1719. [4] D.L. Earnshaw, A.J. Pope, FlashPlate scintillation proximity assays for characterization and screening of DNA polymerases, primase, and helicase activities, J. Biomol. Screen. 6 (2001) 39–46. [5] M. Seville, A.B. West, M.G. Cull, C.S. McHenry, Fluorometric assay for DNA polymerases and reverse transcriptase, BioTechniques 21 (1996) 664–672. [6] M.A. Griep, Fluorescence recovery assay: a continuous assay for processive DNA polymerases applied specifically to DNA polymerase III holoenzyme, Anal. Biochem. 232 (1995) 180–189. [7] D.R. Kim, C.S. McHenry, In vivo assembly of overproduced DNA polymerase III, J. Biol. Chem. 271 (1996) 20681–20689. [8] W.R. McClure, Y. Chow, The kinetics and processivity of nucleic acid polymerases, Methods Enzymol. 64 (1980) 277–297.