Quantitative Microplate Assay for Real-Time Nuclease Kinetics

Utilizing the phenomenon of nucleases exposing oligonucleotide phosphate backbones to phosphatases we present a novel quantitative method for kinetics of nuclease catalysis. Inorganic phosphate released from nuclease products by phosphatases could be quantified in real-time by a fluorescent sensor of inorganic phosphate. Two different nucleases were employed, showing the versatility of this assay for multiple turnover label-free nuclease studies.


Introduction
Nucleases are indispensable to all organisms. They are essential for the DNA replication machinery, removing primers, proofreading etc. Nucleases play integral roles in recombination and repair of the genomic DNA. Many cell defense mechanisms include nucleases; the RISC complex degrading specific RNAs, restriction endonucleases cleaving foreign double-stranded DNA (dsDNA) and the CRISPR-Cas systems storing information about invading viruses in the genome of many bacteria and archaea [1][2][3].
Nucleases have been an important addition to the molecular biology toolbox for decades. Especially important has been the applicability of type II restriction endonucleases on recombinant techniques, allowing scientists to 'cut-and-paste' different dsDNA substrates together [4]. Other examples of common uses of nucleases are biosensors [5], removal of specific nucleic acid species [6] and also genomic editing (ex. CRISPR-Cas) [7].
There is a plethora of different nucleases acting on different substrates with different modes of action. A multitude of assays for determining enzymatic activities of nucleases have been developed; mostly by incorporating fluorescent-or radioactive labels in the substrate oligonucleotides [8,9]. Major drawbacks can be seen in these assays such as: the inability to perform the nuclease reaction under multiple turnover conditions [10], requiring labeling of substrate and/or enzyme, low sensitivity and discontinuity. Assays without modifications generally require high concentrations of substrate and enzyme as the read-outs have low sensitivity, such as running samples in gels or measuring absorbance [11]. Many assays are discontinuous which can cause faulty assumptions as a lot of the reaction data is missed between time-points and there is increased labor compared to continuous assays [12,13]. Nuclease assays often measure either the amount of substrate left or they measure the cleavage of substrate as a yes/no parameter even though many nucleases cleave each substrate molecule several times, causing significant information loss [14].
Nucleosides are chained together by phosphodiesters as a backbone to form oligonucleotides. When a nuclease cleaves an oligonucleotide the phosphodiester backbone connecting two nucleosides is transformed to either a 5' or 3' phosphomonoester or 2', 3'-cyclic phosphodiester bound to one of the products. Nuclease cleavage of the phosphodiester backbone exposes the phosphate for phosphatases to act upon, hydrolyzing the phosphate into inorganic phosphate (P i ). Using this knowledge an assay is presented here where various nucleases can be used together with natural, label-free substrates. Briefly, the assay principle is as follows: Nuclease cleaves oligonucleotide substrate, producing products with exposed phosphates, these are in turn released by a phosphatase as P i which thereafter is quantified in real-time by a fluorescent phosphate sensor. This reaction can be measured continuously using nanomolar concentrations of substrate.
Products formed by nucleases are mainly oligonucleotides or nucleosides with 5' or 3' monophosphates. Alkaline phosphatases have long been known to efficiently cleave as 5' and 3' phosphomonoester substrates [15]. FastAP is an improved alkaline phosphatase capable of quickly cleaving phosphates from DNA, RNA and nucleotides and it is active in a wide variety of buffers. However, alkaline phosphatases are incapable of cleaving 2', 3'-cyclic phosphodiesters which are formed by several ribonucleases and also RNA-cleaving ribozymes and deoxyribozymes (DNAzymes) [16]. T4 polynucleotide kinase (T4PNK) has the ability to cleave 2', 3'cyclic phosphodiesters, an ability which has been naturally evolved to thwart the innate immune response of the bacterium which the T4 phage infects; thus, T4PNK can replace alkaline phosphatase when the nuclease reaction form 2',3'-cyclic phosphodiester products [17]. P i concentrations have been measured by discontinuous low sensitivity assays such as reacting P i with molybdate to form colored complexes [18]. A modern assay utilized phosphate binding protein of E. coli coupled to 7-diethylamino-3-[N-(2-maleimidoethyl)carbamoyl]coumarin (MDCC-PBP) as a biosensor for P i allowing real-time P i concentration determination with high sensitivity as P i binding produces a 6-to 8-fold increase in fluorescence [19].
To show the versatility of this assay two different nucleases are used. Exonucleases are represented by exonuclease III from E. coli (ExoIII) while endonucleases are represented by a '10-23' DNAzyme [20]. The choice of these two enzymes also show the applicability of this assay for both protein-and nucleic acid-based nucleases.

Materials and Methods
All oligonucleotides were synthesized and purified by ThermoFisher (Table 1). DNA oligonucleotides were ordered with desalting and RNA oligonucleotides as HPLC purified. Oligonucleotides were diluted in RNase-and DNase-free ultrapure water to 10 μM working stocks. The water used was produced by purifying ddH 2 O using a Milli-Q Advantage A10 Water Purification System equipped with a BioPak Polisher (Merck Millipore). The purified water had undetectable levels of P i . Tris amino Ultra-pure (Angus Chemical Company). Sodium chloride (Sigma Aldrich). Magnesium chloride hexahydrate was from Riedel-de Haën. Hydrochloric acid for setting Tris buffer pH (Scharlau). MDCC-PBP (commercial name Phosphate Sensor), T4 polynucleotide kinase, FastAP and Exonuclease III (ThermoFisher). MDCC-PBP was kept as 5 μM stocks at -80°C. Purine nucleoside phosphorylase (PNPase) and 7-methylguanosine (7-MEG) (Sigma) were kept as stocks at -80°C of 500 U/ml and 30 mM respectively. Reagents with high P i background levels were pretreated with 0.001 U/ml PNPase and 200 μM 7-MEG at RT until P i background was less than 100 nM. 96 well half area black plates with clear flat bottom and NBS™-coating (Corning) were chosen as assay plates to minimize binding of reagents to the plate. The plates were highly contaminated by P i , this was solved by rinsing the plates with P i -free water and centrifugation upside down to remove the water. After centrifugation each plate was covered with a plastic adhesive seal (MidSci). Flexstation II multiplate reader with 8-channel microfluidics was from Molecular Devices.

Assay-and compound plate setups
ExoIII assay plate wells contained 1.5 nM (0.04 U/μl) ExoIII and 1 μM MDCC-PBP in 1X ExoIII buffer (66 mM Tris-HCl pH 7.5, 0.66 mM MgCl 2 ). Compound plate contained varying concentrations of substrate together with FastAP (final concentration 0.0004 U/μl) in 1X ExoIII buffer. The plates were incubated inside the Flexstation II for 30 min at 37°C prior to reaction initiation for the plate to reach the desired temperature.
For the DNAzyme reactions the assay plate was set up containing 0.3 U/μl T4PNK, 2 μM RNA substrate and 1 μM MDCC-PBP in 1X DNAzyme buffer (40 mM Tris-HCl pH 7.5, 20 mM MgCl 2 , 50 mM NaCl). The compound plate contained DNAzyme of varying concentrations in 1X DNAzyme buffer. The plates were left to incubate inside the Flexstation for 15 min at RT for equilibration.
The different setups is only due to minimizing the loss of expensive material as the compound plate demands a larger volume, leading to loss of material, for efficient pipetting. To initiate the reactions 10 μl (1/5 th of the final reaction volume) from each compound plate well was added to each respective well in the assay plate through the Flexstation fluidics module. Reactions were measured every 5 s over the course of the experiment. ExoIII reactions were measured for 20 minutes after initiation while DNAzyme reactions were measured for 30 minutes as this reaction was slower. Flexstation II settings: λ Ex = 426 nm, λ Em = 466 nm, λ cutoff = 435 nm, reads/well = 30, Photomultiplier tube voltage = High.

Data analysis
All experiments were repeated three times and each time-point show error bars representing standard error of the mean between the three experiments. Each experiment was normalized to a control reaction containing no substrate and then normalized to the initial fluorescence to remove differences in background fluorescence between each reaction. The fluorescence was then changed to P i concentration through normalization to a standard curve made for the MDCC-PBP in each respective reaction buffer (S5 and S6 Figs). The slope of the linear part of each curve was used as initial velocity (v 0 ). The Michaelis constant (K M ) is defined in eq (1) by  [20] 5'-GCA CCC AGG CTA GCT ACA ACG ACT CTC TC-3' I-DzSJ [10] 5'-GCA CCC AGG CTA CCT ACA ACG ACT CTC TC-3' DzSJ RNA substrate [20] 5'-GGA GAG AGA | UGG GUG CG-3'

ExoIII substrate
Sense strand Quantitative Microplate Assay for Real-Time Nuclease Kinetics the rate constants for enzyme-substrate binding and dissociation (k on and k off respectively) together with the turnover number (k cat ) or the apparent turnover number (k app cat ) which is used for ExoIII. The reverse reaction of ligation is disregarded due to the fact that no evidence has been found for the occurrence of ligation in ExoIII reactions and the reverse reaction rate constant of 10-23 DNAzymes is much smaller than k cat . K M and maximal velocity (V max ) were determined by curve-fitting in GraphPad Prism (GraphPad software) using eq (2). k cat was calculated by dividing V max with initial enzyme concentration [21].

Results and Discussion
General nuclease assays depend on labels and/or separating products by size. The assay presented here uses the phenomenon that when nucleases cleave oligonucleotide backbone phosphodiesters are exposed either in the form of nucleoside monophosphates and/or shorter oligonucleotides with 5' or 3' phosphomonoesters or 2', 3'-cyclic phosphodiesters. The exposed phosphates are in turn released from the product by an alkaline phosphatase or T4PNK, which are present in excess over substrate to prevent the dephosphorylation from being rate-limiting, resulting in the formation of P i , which in turn binds to MDCC-PBP thereby inducing a conformational change. This MDCC-PBP conformational change causes an increase in fluorescence of the environment-sensitive fluorophore MDCC. Downstream reactions of coupled enzyme assay can't be allowed to be rate-limiting. To keep the nuclease reaction as rate-limiting, and thus represent the measured kinetic rates, an excess of phosphatase and MDCC-PBP over substrate was used.

P i background
Contamination of reagents and water by P i is common and poses a problem for phosphatesensing assays. To remove any background P i from reagents they were treated with PNPase and 7-MEG [22]. PNPase catalyzes the simultaneous cleavage and phosphorylation of nucleosides (eg. 7-MEG), resulting in a nucleobase and a ribose-1-phosphate as products. However, addition of PNPase could potentially introduce a negative effect on the readout of the nuclease assay results where the perceived increase in P i concentration would be lower than expected. An experiment was set up to determine the maximal allowed concentration of PNPase in the nuclease reactions. The concentration deemed to be high enough to remove background P i from reagents at RT overnight but not high enough to significantly interfere with the assay was 0.0002 U/ml PNPase which has a reaction rate of 0.002 nM s -1 at 37°C (S1 Fig).

ExoIII kinetics
ExoIII was used to evaluate the applicability of the assay for exonucleases [23]. ExoIII binds to blunt or recessed 3' ends of dsDNA, regardless if the 3' is phosphorylated or not, cleaving nucleotides from the 3' end yielding deoxyribonucleoside-5'-monophosphates (dNMPs). Usually described as a processive exonuclease, ExoIII becomes more nonprocessive at higher temperatures (such as 37°C which is used here) [24]. FastAP, present at an excess over substrate, dephosphorylates the dNMPs for direct quantification by MDCC-PBP. Fig 1 depicts a kinetic model of the cleavage of dsDNA by ExoIII and subsequent P i -release by FastAP for MDCC-PBP P i quantification [25]. P i association to MDCC-PBP was determined as >200 nM s -1 in this setup (S2 Fig) and with a previously published association rate constant of 1.36 Á 10 8 M -1 s -1 was considered negligible to the overall rate of the phosphatase-coupled reaction [19]. k on and k off were not determined in this work. Since ExoIII is mostly a nonprocessive exonuclease at 37°C the enzyme will either move along the substrate without dissociating or dissociate and associate before cleavage of another nucleotide. As nucleotides are cleaved and released one at a time as dNMPs each event will have a different turnover numbers (k cat (n) ) since ExoIII shows some sequence specificity [26]. Thus, an average of all k cat (n) is measured and presented as k app cat . The experiment was repeated three times and as seen in  cat is determined by the amount of dNMPs produced in the reaction, which to our knowledge have never been measured for ExoIII before. K M is determined using the concentration of 3'-ends as this serves as the binding site of the enzyme and the value determined (140.9 nM) is similar to previously published values (120 nM) [27]. The influence of FastAP dephosphorylation reaction rate on the overall reaction was negligible, determined in S3 Fig with a reaction rate more than 7 times the V max of ExoIII.

DNAzyme kinetics
To show the versatility of this assay a '10-23' DNAzyme was employed; ExoIII being a proteinbased exonuclease forming 5'-phosphorylated products compared to DNAzyme which endonucleolytically cleave their target once, forming 2', 3'-cyclic phosphorylated products. DNAzymes have been used for gene knockdown, biosensors, molecular machines and many other applications. With the wide range of application for DNAzymes there is need for quick and robust kinetic assays. Several methods have previously been reported; they usually have drawbacks such as discontinuity, substrate has to be modified, can't measure multiple turnovers and/or are not high-throughput. By using the assay presented here, multiple turnover reactions of a '10-23' DNAzyme (eg. DzSJ) could be monitored in real-time.
DNAzyme cleavage reaction yields a 2', 3'-cyclic phosphate on one of the products, a form of organic phosphate that can't be hydrolyzed by alkaline phosphatase. Instead an excess of T4PNK over substrate was used to dephosphorylate the product. Even though T4PNK is defined as a kinase it also has 3' phosphatase activity with the ability to dephosphorylate 2', 3'cyclic phosphate ends. In Fig 3 the DNAzyme cleavage of an RNA substrate strand is shown, producing a cyclic phosphate, which is subsequently released as P i by T4PNK for quantification by MDCC-PBP.
The extremely low K M of DzSJ (<1 nM) [20] is slightly lower than the limit of detection for this assay, hence the Michaelis-Menten equation could not be applied in this case. Instead different concentrations of DzSJ were used to show the linear increase of velocity with enzyme concentration, as can be seen in Fig 4. DzSJ is capable of multiple turnover catalysis which is shown in Fig 4A as    concentrations higher than the DzSJ concentration. Each initial velocity was determined by curve fitting and plotted against enzyme concentration in Fig 4B to show the linearity of velocity increase as the enzyme concentration is increased. With an R 2 -value of 0.98 the assay gives an excellent response to the enzyme increase and validates the versatility and robustness of the assay. A G-C mutation of the 6 th nucleotide in the catalytic loop of DzSJ served as a control for the experiment, showing no increase in fluorescence over the 2000 s time-course ( Fig 4A). As T4PNK might influence the reaction rate of the overall reaction the dephosphorylation rate was determined in S4 Fig as being more than 4 times the highest rate measured for the DzSJ catalyzed cleavage. The perfect linearity of enzyme concentration dependence of the DzSJ cleavage reaction shown in Fig 4B also indicates T4PNK did not limit the overall reaction when high concentration of enzyme was used.

Conclusions
The real-time microplate assay presented in this study proved to be highly versatile in the selection of nuclease, with low variability between experimental repeats. Protein, non-protein, exoand endo-nucleases are assayed equally well with this assay. Comparison and exact determination of kinetics of nuclease catalysis can be easily performed with this assay, as well as mutation analysis and inhibitor screenings. As the need for labels is removed this assay will minimize the work for optimizing modifications of substrates and the cost of synthesis.