Analytical Biochemistry 446 (2014) 53–58
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
A new, sensitive ecto-50 -nucleotidase assay for compound screening Marianne Freundlieb a, Herbert Zimmermann b, Christa E. Müller a,⇑ a b
PharmaCenter Bonn, Pharmaceutical Sciences Bonn (PSB), Pharmaceutical Institute, Pharmaceutical Chemistry I, University of Bonn, D-53121 Bonn, Germany Institute of Cell Biology and Neuroscience, Molecular and Cellular Neurobiology, Goethe University, 60438 Frankfurt am Main, Germany
a r t i c l e
i n f o
Article history: Received 22 July 2013 Received in revised form 5 October 2013 Accepted 8 October 2013 Available online 19 October 2013 Keywords: Cancer therapy CD73 Ecto-50 -nucleotidase assay Lanthanum chloride precipitation Michaelis–Menten analysis Radiometric enzyme assay
a b s t r a c t Ecto-50 -nucleotidase (eN) is a membrane-bound enzyme that hydrolyzes extracellular nucleoside-50 monophosphates yielding the respective nucleoside and phosphate. Increased levels of eN expression have been observed in many cancer cells. By increasing extracellular adenosine concentrations, they contribute to their proliferative, angiogenic, metastatic, and immunosuppressive effects. Therefore, eN is of considerable interest as a novel drug target for the treatment of cancer as well as of inflammatory diseases. In this study, we developed, optimized, and applied a highly sensitive radiometric assay using [3H]adenosine-50 -monophosphate (AMP) as a substrate. The reaction product [3H]adenosine was separated from [3H]AMP by precipitation of the latter with lanthanum chloride and subsequent filtration through glass fiber filters. Conditions were optimized to reproducibly collect the [3H]adenosine-containing filtrate used for quantitative determination. Validation of the assay yielded a mean Z0 factor of 0.73, which demonstrates its suitability for high-throughput screening. The new assay shows a limit of detection that is at least 30-fold lower than those of common colorimetric methods (e.g., optimized malachite green assay and capillary electrophoresis-based assay procedures), and it is also superior to a recently developed luciferase-based assay. Ó 2013 Elsevier Inc. All rights reserved.
Ecto-50 -nucleotidase (eN,1 CD73, EC 3.1.3.5) is a membranebound enzyme that catalyzes the dephosphorylation of extracellular nucleoside-50 -monophosphates to the respective nucleosides. Adenosine-50 -monophosphate (AMP) is its main substrate, with a Km value of 1 to 50 lM [1–3]. Recently, an X-ray structure of the human dimeric eN was obtained showing that dimerization results from the interaction of two C-terminal domains. The N-terminal domain coordinates two zinc ions in its active site that are essential for the hydrolytic reaction, whereas the C-terminal domain provides the substrate binding pocket. On binding of a substrate, the N-terminal domain rotates to the closed form where the substrate binding pocket and the metal ions form the active site [4]. The enzymatic product adenosine activates G protein-coupled receptors, termed P1 or adenosine receptors, which are subdivided into four distinct subtypes: A1, A2A, A2B, and A3 [5]. eN is often directly colocalized with adenosine receptors and contributes to their activation [6]. Therefore, eN is of considerable interest as a novel drug target. Phosphorylated adenosine derivatives have been designed as eN⇑ Corresponding author. Fax: +49 228 73 2567. E-mail address: [email protected] (C.E. Müller). Abbreviations used: eN, ecto-50 -nucleotidase; AMP, adenosine-50 -monophosphate; ATP, adenosine-50 -triphosphate; ADP, adenosine-50 -diphosphate; AOPCP, a,b-methylene-ADP; HTS, high-throughput screening; LOD, limit of detection; HPLC, highperformance liquid chromatography; Tris, tris(hydroxymethyl)aminomethane; DMSO, dimethyl sulfoxide; Hepes, 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid; cpm, counts per minute. 1
0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ab.2013.10.012
activated prodrugs of A2A adenosine receptor agonists for the site-specific treatment of inflammatory diseases without hypotensive side effects [6,7]. Inhibitors of eN may be useful for the treatment of various diseases that are associated with increased adenosine levels, for example, cancer [8]. Many cancer cells show an increased expression level of eN, leading to elevated extracellular adenosine concentrations. Adenosine stimulates tumor growth, proliferation, angiogenesis, and metastasis via the activation of A2A and A2B receptors, and it suppresses immune responses via A2A receptors [9–16]. Moreover, eN functions as a cosignaling molecule on T-lymphocytes and as an adhesion molecule that is required for lymphocyte binding and thereby directly contributes to the immunosuppressive effect [11,17]. Experiments with eN-deficient mice indicated that targeting the enzymatic activity of tumor eN may be an important new approach for cancer immunotherapy [18,19]. So far, only very few inhibitors of eN have been described [2–4,20–24]. Adenosine-50 -triphosphate (ATP) and adenosine-50 diphosphate (ADP), as well as ADP analogs, can act as competitive inhibitors of eN. The ADP analog AOPCP (1) is the most active inhibitor described so far, with a Ki value in the nanomolar range [3,20]. Very few moderately potent, non-nucleotide-derived inhibitors have been described, including anthraquinones [21], methylxanthines [2], flavonoids such as quercetin [4,22,23], and sulfonamides [24]. To identify and optimize new eN inhibitors, a fast and sensitive assay method that is suitable for high-throughput screening (HTS)
54
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
is required. Several methods have been described for assaying eN as well as intracellular 50 -nucleotidases. These include colorimetric, luminescent, chromatographic/spectroscopic, and radiometric procedures [20,25–45]. A major drawback of the currently available methods is their low sensitivity, requiring high substrate concentrations. As a consequence, only relatively potent and/or well water-soluble inhibitors can be identified. Some methods are very sophisticated and/or time-consuming and, thus, might not be suitable for HTS. Recently, a luciferase-based assay was developed to detect eN activity [25]. Luciferase is frequently being used to determine ATP concentrations; however, in the described eN assay, the authors made use of the inhibiting effect of the enzymatic reaction by AMP. eN-mediated decreases in AMP concentrations were determined by increases in the luminescent luciferase signal [25]. Thus, the assay determines concentrations of the eN reaction product adenosine indirectly by measuring a decrease in AMP concentrations. For optimal assay conditions, high AMP substrate concentrations (300 lM), well above the Km value, are required. This results in low sensitivity; that is, relatively high concentrations of test compounds need to be employed to observe an effect. It limits the detectability of moderately potent test compounds with limited solubility in screening campaigns. Furthermore, ATP, an inhibitor of eN (Ki = 8.9 lM) [20], needs to be present. The most frequently used colorimetric method is the malachite green assay, which detects phosphate produced by the hydrolysis of AMP [26– 28]. However, because phosphate is ubiquitously present, a high background is typically observed, and it is difficult to achieve a limit of detection (LOD) below 1 lM phosphate. Only after careful optimization of the assay conditions have we been able to achieve an LOD of 0.2 lM phosphate. Especially, the effects of antibodies as eN inhibitors are difficult to measure because their solutions typically contain phosphate [25]. Moreover, many colored compounds may interfere with the assay. Further assay procedures are based on high-performance liquid chromatography (HPLC) or capillary electrophoresis (CE) separation coupled with ultraviolet (UV) detection [20,29,30]. These methods are also not ideally suited for HTS because they are time-consuming. Two different coupled enzyme reaction methods have been described [31–35]. These assays are complex and vulnerable to false-positive results because it is not clear whether the enzyme of interest or the auxiliary enzyme is affected by an inhibitory test compound [36]. Assays using radiolabeled substrate have also been described applying different methods for the separation of substrate from product, including HPLC, thin layer chromatography (TLC), anion exchange or column chromatography, precipitation methods followed by centrifugation, or determination of residual substrate concentrations after filtration [37–45]. However, all of the described methods have drawbacks limiting their suitability for the screening of large libraries. Therefore, we developed and optimized a new procedure based on radioactive [3H]AMP as a substrate, quantifying formed [3H]adenosine, which is simple, highly sensitive, reproducible, and suitable for HTS.
from BIOLOG Life Science Institute (Bremen, Germany). Tris(hydroxymethyl)aminomethane (Tris), dimethyl sulfoxide (DMSO), 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (Hepes), lanthanum chloride, and sodium chloride were obtained from Carl Roth (Karlsruhe, Germany). [2,8-3H]AMP (solution in ethanol/water, 1:1, 22.9 Ci/mmol, 1.0 mCi/ml, 849 GBq/mmol) was obtained from Hartmann Analytic (Braunschweig, Germany). [2-3H]adenosine (aqueous solution containing 0.1% ethanol, 23.0 Ci/mmol, 1.0 mCi/ml, 851 GBq/mmol) was obtained from Amersham Biosciences (Freiburg, Germany). Recombinant expression of rat ecto-50 -nucleotidase Catalytically active recombinant soluble glutathione S-transferase/eN fusion protein was expressed in insect cells using the baculovirus system and purified by affinity chromatography using agarose-coupled glutathione as described previously [46]. The active sites of rat and human enzyme are nearly identical, with only one minor difference (a Phe [human]/Tyr [rat] exchange) [47]. Preparation of standard solutions Nucleotides and nucleosides were dissolved in deionized water. For the radiometric assay, aqueous solutions with final concentrations of 1 mg/ml and 0.1 mg/ml were obtained. The standard solutions were further diluted in assay buffer as required for the calibration curves and enzyme assays. Radiometric ecto-50 -nucleotidase assay Reactions were carried out in assay buffer consisting of 25 mM Tris, 140 mM sodium chloride, and 50 mM sodium phosphate with a pH value adjusted to 7.4. [3H]AMP diluted to a specific activity of 3.7 109 Bq/mmol (100 mCi/mmol) was used as a substrate. Substrate solution (10 ll) was added to 80 ll of assay buffer (final concentration in assay 5 lM), and the mixture was preincubated for approximately 5 min at 37 °C. The reaction was then initiated by adding 10 ll of enzyme solution (in enzyme buffer containing 40 mM Hepes and 4 mM iodoacetamide, pH 7.4) to each test tube (final concentration in assay 0.3 ng/ll). The mixture was incubated for 25 min at 37 °C. To stop the reaction, 500 ll of ice-cooled precipitation buffer containing 100 mM aqueous lanthanum chloride and 100 mM aqueous sodium acetate solution (pH 4.0) was added. After approximately 30 min of cooling on ice, the precipitation was completed and the reaction mixture was filtered through GF/B glass fiber filters using a Brandel cell harvester (M-24, Brandel, Gaitherburg, MD, USA) equipped with an individual collection box for collecting the filtrates containing the separated product. After washing three times with 1 ml each of the precipitation buffer, the filtrates were poured into scintillation vials containing 6 ml of the scintillation cocktail ULTIMA Gold XR and were quantified by scintillation counting (TRICARB 2900 TR, Packard/PerkinElmer). Calibration curve
Materials and methods Reagents and chemicals Adenosine, calcium chloride dihydrate, sodium acetate, magnesium chloride, ammonium molybdate tetrahydrate (Bio Ultra), and malachite green oxalate were obtained from Sigma–Aldrich (Steinheim, Germany). AMP was obtained from Merck (Darmstadt, Germany). Disodium adenosine 50 -diphosphate monohydrate was obtained from Acros Organics (Nidderau, Germany). Adenosine 50 -O-(a,b-methylene)diphosphate (AMPCP, AOPCP) was obtained
To determine the sensitivity of the test system, a calibration curve for [3H]adenosine using a fixed concentration of [3H]AMP was determined. The detected counts per minute (cpm) was plotted versus the adenosine concentrations. Six different concentrations of adenosine ranging from 0.01 to 3 lM were used in combination with a fixed concentration of 5 lM AMP. The experiments were performed as described above for the assay, but without the addition of enzyme (three separate experiments, each in triplicate). The results were fitted to a linear regression curve using the GraphPad Prism 5 program (GraphPad Software, La Jolla, CA,
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
USA). As an indicator for the sensitivity of this assay, the LOD was calculated using the following formula:
LOD ¼
3r ; slope
where r is the standard deviation of the blank samples and ‘‘slope’’ is the slope of the corresponding calibration curve. Optimization of assay parameters For determination of the optimal enzyme concentration, an enzyme titration was conducted. Seven different enzyme concentrations between 0.032 and 2.0 lg/ml were incubated with a substrate concentration of 5 lM [3H]AMP for 15 min at 37 °C. To stop the enzymatic reaction, 500 ll of the precipitation buffer was added. Separation of the product was carried out by filtration as described above for the enzyme assay. To determine the optimal period of incubation resulting in a level of substrate conversion between 10 and 20%, the enzymatic reaction was carried out as described above but was terminated after different periods of time between 0 and 60 min of incubation. The substrate concentration was 5 lM [3H]AMP, and the enzyme concentration was 0.3 lg/ ml, based on the results of the enzyme titration conducted before. Three separate experiments were performed, each in triplicate. Determination of apparent Km value For the determination of the Michaelis–Menten constant (Km), seven different substrate concentrations of [3H]AMP between 1 and 1000 lM (final concentrations) were used. The enzymatic reaction was carried out using a final concentration of 0.3 lg/ml eN following the method described above. Three separate assays were performed, each in triplicate. Km and Vmax values were calculated by nonlinear fitting of the Michaelis–Menten equation (see below) using GraphPad Prism 5 software:
Km ¼
V max v ½S
v
:
Assay validation The validation of the newly developed assay was performed by measuring a series of positive and negative controls. In the negative controls, 10 ll of substrate solution was added to 90 ll of assay buffer, resulting in a final concentration of 5 lM. In the positive controls, 10 ll of substrate solution (final concentration 5 lM) and 10 ll of eN solution (final concentration 0.3 lg/ml protein) were added to 80 ll of assay buffer. The assay was performed as described above. Three separate assays were performed. For each of the positive and negative controls, 24 wells were analyzed. Assay performance was quantified using the Z0 factor calculated according to the formula of Zhang and coworkers [11] as shown below:
3 rþ þ 3 r ; Z 0 ¼ 1 lþ þ l where r+ and r– are the standard deviations of the positive and the negative control and l+ and l– are the mean values of the positive and negative controls, respectively. Inhibition assay For further validation of the radiometric assay, two standard inhibitors of eN, ADP and AOPCP, were tested under the established assay conditions as described above. Briefly, 10 ll of the aqueous inhibitor solution was added to 70 ll of assay buffer and 10 ll of
55
substrate solution (final substrate concentration 5 lM). To start the enzymatic reaction, 10 ll of the enzyme solution was added (final concentration 0.3 lg/ml). To obtain full concentration–response curves, 11 inhibitor concentrations ranging from 0.001 to 100 lM were tested. Three separate inhibition assays were performed, each in triplicate. Concentration–response curves were fitted using the sigmoidal dose–response curve of GraphPad Prism 5 software from which IC50 values were deduced. Ki values were calculated using the formula below:
Ki ¼
IC50 : ð1 þ ½S=K m Þ
Screening For screening of a small chemical library, compound solutions in DMSO (1 mM, 1 ll) were added to 79 ll of buffer, 10 ll of substrate solution, and 10 ll of enzyme solution, resulting in a final concentration in the assay of 10 lM of the test compound, 1% DMSO, 5 lM substrate, and 0.3 ng/ll enzyme. On each filter, the following controls were included: one containing only DMSO (no inhibitor) and one in the absence of enzyme. We showed in previous experiments that in the presence of the standard inhibitor AOPCP, the same cpm values were obtained as without enzyme. The percentage inhibition was calculated using the formula below:
% inhibition ¼
cpmnegative control cpmcompound
cpmnegative control cpmpositive control
:
Results and discussion Development and optimization of radiometric ecto-50 -nucleotidase assay AMP is the main substrate of eN. For the development of a new assay procedure, we used [3H]AMP that contained a tritium atom either in the 2-position only or in both the 2- and 8-positions. The radioactive substrate is converted to the radiolabeled product [3H]adenosine by eN (see Fig. 1). A method needed to be developed for the separation of the product [3H]adenosine (present in low amounts) from the substrate [3H]AMP (present in excess). The idea was to separate the labeled product from the radioactive substrate by precipitation of unreacted [3H]AMP using lanthanum chloride, which leads to the formation of a poorly soluble salt. An analogous procedure had previously been used for thymidine kinase assays [48]. However, in this reaction, the precipitated nucleotide was the product, which could be collected by filtration on glass fiber filters. In the case of eN, the product is [3H]adenosine, which would remain in solution. Because the lanthanum–AMP salt still shows some solubility in water, coprecipitation with sodium phosphate was investigated, yielding a lanthanum phosphate–AMP complex that possesses negligible solubility. For lanthanum chloride to achieve complete precipitation of [3H]AMP, the ratio of substrate to sodium phosphate and the pH value of the solution are important parameters that needed to be carefully optimized. We found that rather high amounts of both lanthanum chloride and sodium phosphate are required and that the precipitation needs to be performed in an acidic milieu of pH 4.0 rather than at neutral pH or an alkaline pH value of 9.0. The enzymatic reactions were performed in 25 mM Tris–HCl buffer containing 140 mM sodium chloride and 50 mM sodium phosphate with a pH value adjusted to 7.4 at 37 °C. We did not add magnesium or calcium cations because these tend to form precipitates with phosphate. The addition of these divalent cations
56
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
Fig.1. Hydrolysis of AMP by ecto-50 -nucleotidase (eN).
to the buffer was not required because the cations already bound to the added enzyme appeared to be sufficient for full enzymatic activity. Sodium phosphate was required as a buffer component to facilitate precipitation after the addition of lanthanum chloride. Although phosphate had been described to have an inhibitory effect on 50 -nucleotidases, we could observe high enzymatic activity in the applied buffer. Simulating a typical substrate–product ratio of the enzymatic reaction by analyzing a mixture of 0.5 lM [3H]adenosine and 4.5 lM [3H]AMP, a suitable measurement window was obtained in preliminary experiments (3-fold over background, which was measured with [3H]AMP only). Furthermore, we determined the DMSO sensitivity of the assay because many test compounds are only moderately or even poorly soluble in water. DMSO was tolerated up to a concentration of 2% without any effect.
Fig.2. Enzyme titration to determine the optimal enzyme concentration. The reaction was carried out with a substrate concentration of 5 lM AMP. Data points represent means ± standard errors from three different experiments, each run in triplicate. The LOD was calculated to be 0.013 lg/ml ecto-50 -nucleotidase.
Calibration of test system To determine the LOD as a parameter for the sensitivity of the test system, a calibration curve for adenosine in the presence of AMP at a final concentration of 5.00 lM was obtained by plotting the radioactivity versus the adenosine concentrations. The linear regression was calculated, and the following regression equation was obtained: y = 2512x + 996. A correlation coefficient of 0.9997 (n = 3) was obtained. The LOD for adenosine was found to be 0.028 ± 0.002 lM; thus, it was 30-fold lower than that for the recently developed CE method [20] and also more than 30-fold lower than the LOD for phosphate in colorimetric malachite green assays, which is described in the literature to be approximately 1 lM [26,28]. Optimization of assay parameters To optimize the assay parameters, we initially determined a suitable substrate concentration. To achieve appropriate assay conditions for testing competitive as well as uncompetitive inhibitors, the substrate concentrations should range between 0.1 Km and 10 Km. Because a high sensitivity is desired, a concentration of 5 lM AMP was chosen, which is slightly lower than the Km value of 25 lM determined for rat eN [46]. Another very important parameter to improve the assay sensitivity and robustness is the enzyme concentration, which is directly proportional to the initial velocity. If the enzyme concentration is very high and the substrate is quickly depleted, product formation might not be linear. To define the optimal enzyme concentration, an enzyme titration was performed at the selected substrate concentration of 5 lM. A linear increase of the detected signal in relation to an increase in enzyme concentration was observed (Fig. 2). For the determination of the optimal enzyme concentration, the LOD was calculated following the formula listed in Materials and Methods. The LOD corresponds to the minimal concentration of enzyme that can be reliably differentiated from the background. The optimal enzyme concentration is selected to be at least 20-fold higher than the LOD. The result of the enzyme titration (Fig. 2) yielded the regression equation y = 9.75x + 0.0464, with a correlation coefficient of 0.9978 (n = 3). The LOD was found to be 0.013 lg/ml; therefore, an enzyme concentration of 0.3 lg/ml was chosen as the final assay concentration. The third parameter
determined for assay optimization was the incubation time. To achieve an optimal substrate conversion rate between 10 and 20%, a time course analysis was performed. The range of 10 to 20% of product formation was obtained between 15 and 40 min of incubation; therefore, we chose an incubation time of 25 min for further experiments. Thus, the optimized parameters were as follows: 5 lM AMP as a substrate, 0.3 lg/ml enzyme, and 25 min of incubation time. Determination of kinetic parameters of ecto-50 -nucleotidase Because the kinetic parameters can vary between test systems, the assay-specific Km value was determined using the optimized parameters. The kinetic parameters of the purified rat eN for AMP were calculated using the Michaelis–Menten equation (Fig. 3). The determined Km value was 59 lM and, thus, was roughly in the same range as those determined with other assay methods (1–50 lM) [2,20,46] and an enzyme turnover number of 1.1 s1 was calculated.
Fig.3. Michaelis–Menten plot for the determination of the apparent Km value of AMP using the newly developed radiometric method. Data points represent means ± standard errors from three separate experiments, each run in triplicate. The Km value was calculated to be 59 ± 10 lM.
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
57
Method validation The Z0 factor is a widely used statistical parameter to evaluate robustness and quality of HTS assays. To investigate whether the new radiometric assay was suitable for HTS, a pilot Z0 factor study was conducted prior to applying the assay for extensive compound screening (Fig. 4). We determined a high Z0 factor of 0.73. This result demonstrates a very high assay robustness of the applied radiometric method and its excellent suitability for HTS campaigns. Enzyme inhibition assays
Fig.4. Z0 factor study to determine the quality and robustness of the assay.
For further validation of the radiometric assay, the inhibitory effects of two known competitive inhibitors, AOPCP and ADP, were determined using the optimized conditions. Concentration–inhibition curves are shown in Fig. 5. An IC50 value of 0.236 lM for AOPCP and an IC50 value of 4.66 lM for ADP were determined. These values are well in agreement with published data [3,20,42]. Screening of a compound library As a next step, we screened a small compound library consisting of 321 synthetic, drug-like, mostly non-ionic heterocyclic compounds. The compounds were screened at a concentration of 10 lM. Hit compounds were defined as those inhibiting at least 70% of the enzymatic reaction. Applying this high threshold, we identified 12 hit compounds (4%) as inhibitors. One compound had shown an increase in the signal in initial screening; however, it could not be confirmed in subsequent assays. The results are shown in Fig. 6. These results show that the newly developed and optimized assay procedure is suitable for library screening.
Fig.5. Concentration–inhibition curves for the eN inhibitors AOPCP and ADP. Data points represent means ± standard errors from three different experiments, each run in triplicate. The IC50 values were determined to be 0.236 ± 0.006 lM for AOPCP and 4.66 ± 0.13 lM for ADP. The corresponding Ki values were calculated as 0.197 ± 0.005 lM for AOPCP and 3.88 ± 0.11 lM for ADP.
Conclusions eN has recently gained considerable attention as a novel potential drug target. Therefore, a radiometric eN assay using recombinant purified rat eN was developed to provide a sensitive test system suitable for HTS. The rat eN is believed to be predictive
Fig.6. Graphical presentation of the screening results of a small library of 321 drug-like compounds tested at a concentration of 10 lM in the developed eN assay. Each dot represents one compound. Compounds above the red line showing more than 70% inhibition were defined as hit compounds (12 compounds, 4% of library). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.)
58
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
for the human eN because the active sites of both enzymes are nearly identical, differing only in a single amino acid: a tyrosine in the rat, which is homologously exchanged for a phenylalanine in the human enzyme. With the newly developed radiometric method presented here, we were able to develop a new sensitive and fast assay for the analysis of eN activity and to identify and characterize inhibitors of eN. Advantages of the new assay include a very low LOD of 0.03 lM for adenosine. This is much lower than those of all previously developed eN assays and allows screening around or below the Km value. This is particularly useful for identifying hit compounds that possess low water solubility and for fragment-based screening approaches. Moreover, colored compounds and inorganic phosphate do not interfere with the assay, in contrast to the standard malachite green assay. The optimized assay is accurate, robust, convenient, inexpensive, and fast. The reaction product is directly detected and quantified. The developed radiometric eN assay will be highly useful for identifying and characterizing novel eN inhibitors that have the potential to become novel anti-cancer and anti-inflammatory drugs. Acknowledgments We thank Nicole Florin, Peter Ripphausen, Inês Ferreira, and Karin Pelka for preliminary experiments. H.Z. acknowledges support from grants from the Cluster of Excellence (EXC 115, Frankfurt, Germany) and the Gutenberg Research College (GRC) University Mainz. References [1] H. Zimmermann, M. Zebisch, N. Sträter, Cellular function and molecular structure of ecto-nucleotidases, Purinergic Signal. 8 (2012) 437–502. [2] H. Zimmermann, 50 -Nucleotidase: molecular structure and functional aspects, Biochem. J. 285 (1992) 345–365. [3] S.A. Hunsucker, B.S. Mitchell, J. Spychala, The 50 -nucleotidases as regulators of nucleotide and drug metabolism, Pharmacol. Ther. 107 (2005) 1–30. [4] K. Knapp, M. Zebisch, J. Pippel, A. El-Tayeb, C.E. Müller, N. Sträter, Crystal structure of the human ecto-50 -nucleotidase (CD73): insights into the regulation of purinergic signaling, Structure 20 (2012) 2161–2173. [5] G.G. Yegutkin, Nucleotide- and nucleoside-converting ectoenzymes: important modulators of purinergic signalling cascade, Biochim. Biophys. Acta 1783 (2008) 673–694. [6] U. Flögel, S. Burghoff, P.L.E.M. van Lent, S. Temme, L. Galbarz, Z. Ding, A. ElTayeb, S. Huels, F. Bönner, N. Borg, C. Jacoby, C.E. Müller, W.W. van den Berg, J. Schrader, Selective activation of adenosine A2A receptors on immune cells by a CD73-dependent prodrug suppresses joint inflammation in experimental rheumatoid arthritis, Sci. Transl. Med. 4 (2012) 146ra108. [7] A. El-Tayeb, J. Iqbal, A. Behrenswerth, M. Romio, M. Schneider, H. Zimmermann, J. Schrader, C.E. Müller, Nucleoside-50 -monophosphates as prodrugs of adenosine A2A receptor agonists activated by ecto-50 nucleotidase, J. Med. Chem. 52 (2009) 7669–7677. [8] S.P. Colgan, H.K. Eltzschig, T. Eckle, L.F. Thompson, Physiological roles for ecto50 -nucleotidase (CD73), Purinergic Signal. 2 (2006) 351–360. [9] B. Zhang, CD73: A novel target for cancer immunotherapy, Cancer Res. 70 (2010) 6407–6411. [10] J. Stagg, U. Dicisekera, N. McLaughlin, J. Sharkey, S. Pommey, D. Denoyer, K.M. Dwyer, M.J. Smyth, Anti-CD73 antibody therapy inhibits breast tumor growth and metastasis, Proc. Natl. Acad. Sci. USA 107 (2010) 1547–1552. [11] J.H. Zhang, T.D.Y. Chung, K.R. Oldenburg, A simple statistical parameter for use in evaluation and validation of high throughput screening assays, J. Biomol. Screen. 4 (1999) 67–73. [12] X. Zhi, Y. Wang, X. Zhou, J. Yu, R. Jian, S. Tang, L. Yin, P. Zhou, RNAi-mediated CD73 suppression induces apoptosis and cell-cycle arrest in human breast cancer cells, Cancer Sci. 101 (2010) 2561–2569. [13] S. Gessi, S. Merighi, V. Sacchetto, C. Simioni, P.A. Borea, Adenosine receptors and cancer, Biochim. Biophys. Acta 2011 (1808) 1400–1412. [14] J. Spychala, Tumor-promoting functions of adenosine, Pharmacol. Ther. 87 (2000) 161–173. [15] J. Stagg, M.J. Smyth, Extracellular adenosine triphosphate and adenosine in cancer, Oncogene 29 (2010) 5345–5358. [16] V. Kumar, A. Sharma, Adenosine: an endogenous modulator of innate immune system with therapeutic potential, Eur. J. Pharmacol. 616 (2009) 7–15. [17] R. Resta, L.F. Thompson, T cell signalling through CD73, Cell. Signal. 9 (1997) 131–139. [18] G.G. Yegutkin, F. Marttila-Ichihara, M. Karikoski, J. Niemelä, J.P. Laurila, K. Elima, S. Jalkanen, M. Salmi, Altered purinergic signaling in CD73-deficient mice inhibits tumor progression, Eur. J. Immunol. 41 (2011) 1231–1241.
[19] J. Stagg, U. Divisekera, H. Duret, T. Sparwasser, M.W.L. Teng, P.K. Darcy, M.J. Smyth, CD73-deficient mice have increased antitumor immunity and are resistant to experimental metastasis, Cancer Res. 71 (2011) 2892–2900. [20] J. Iqbal, D. Jirovsky, S. Lee, H. Zimmermann, C.E. Müller, Capillary electrophoresis-based nanoscale assays for monitoring ecto-50 -nucleotidase activity and inhibition in preparations of recombinant enzyme and melanoma cell membranes, Anal. Biochem. 373 (2008) 129–140. [21] Y. Baqi, S.Y. Lee, J. Iqbal, A. Lehr, A.B. Scheiff, H. Zimmermann, J. Bajorath, C.E. Müller, Development of potent and selective inhibitors of ecto-50 -nucleotidase based on athraquinone scaffold, J. Med. Chem. 53 (2010) 2076–2086. [22] E. Braganhol, A.S.K. Tamajusuku, A. Bernardi, M.R. Wink, A.M.O. Battastini, Ecto-50 -nucleotidase/CD73 inhibition by quercetin in the human U138MG glioma cell line, Biochim. Biophys. Acta 1770 (2007) 1352–1359. [23] M. Kavutcu, M.F. Melzig, In vitro effects of selected flavonoids on the 50 nucleotidase activity, Pharmazie 54 (1999) 457–459. [24] P. Ripphausen, M. Freundlieb, A. Brunschweiger, H. Zimmermann, C.E. Müller, J. Bajorath, Virtual screening identifies novel sulfonamide inhibitors of ecto-50 nucleotidase, J. Med. Chem. 55 (2012) 6576–6581. [25] K.F. Sachenmeier, C. Hay, E. Brand, L. Clarke, K. Rosenthal, S. Guillard, S. Rust, R. Minter, R. Hollingsworth, Development of a novel ecto-nucleotidase assay suitable for high-throughput screening, J. Biomol. Screen. 17 (2012) 993–998. [26] A.A. Baykov, O.A. Evtushenko, S.M. Avaeva, A malachite green procedure for orthophosphate determination and its use in alkaline phosphatase-based enzyme immunoassay, Anal. Biochem. 171 (1988) 266–270. [27] P.A. Lanzetta, L.J. Alvarez, P.S. Reinach, O.A. Candia, An improved assay for nanomolar amounts of inorganic phosphate, Anal. Biochem. 100 (1979) 95–97. [28] E.B. Cogan, G.B. Birrell, O.H. Griffith, A robotics-based automated assay for inorganic and organic phosphates, Anal. Biochem. 271 (1999) 29–35. [29] A. Amici, M. Emanuelli, N. Raffaelli, S. Ruggieri, G. Magni, One-minute highperformance liquid chromatography assay for 50 -nucleotidase using a 20-mm reverse-phase column, Anal. Biochem. 216 (1994) 171–175. [30] J.N. Stolk, R.A. De Abreu, A.M.T. Boerbooms, D.G. de Koning, R. de Graaf, P.J.S.M. Kerstens, L.B.A. van de Putte, Purine enzyme activities in peripheral blood mononuclear cells: comparison of a new non-radiochemical high performance liquid chromatography procedure and a radiochemical thin-layer chromatography procedure, J. Chromatogr. 666 (1995) 33–43. [31] S. Perez, N. Courtis, D. Kokkinopoulos, M. Papamichail, C.M. Tsiapalis, T. Trangas, A colorimetric assay for the determination of 50 -nucleotidase activity, J. Immunol. Methods 101 (1987) 73–78. [32] C.P. Price, P.G. Hill, H.G. Sammons, A comparison of two colorimetric procedures for the assay of 50 -nucleotidase, Clin. Chim. Acta 33 (1971) 260– 263. [33] V.G. Bethune, M. Fleisher, M.K. Schwarz, Automated method for determination of serum 50 -nucleotidase activity, Clin. Chem. 18 (1972) 1524–1528. [34] F. Heinz, R. Pilz, S. Reckel, J.R. Kalden, R. Haeckel, A new spectrophotometric method for the determination of 50 -nucleotidase, J. Clin. Chem. Clin. Biochem. 18 (1980) 781–788. [35] A.H. Chalmers, C. Hare, A semi-automated method for 50 -ecto-nucleotidase measurement in lymphocytes, Immunol. Cell Biol. 68 (1990) 75–79. [36] N.J. Kruger, Errors and artifacts in coupled spectrophotometric assays of enzyme activity, Phytochemistry 38 (1995) 1065–1071. [37] P.H. Ellims, L. Bailey, M.B. Van der Weyden, An improved method for the determination of human erythrocyte pyrimidin 50 -nucleotidase activity, Clin. Chim. Acta 88 (1978) 99–103. [38] E.P. Garvey, G.T. Lowen, M.R. Almond, Nucleotide and nucleoside analogues as inhibitors of cytosolic 50 -nucleotidase I from heart, Biochemistry 37 (1998) 9043–9051. [39] G.G. Yegutkin, T. Henttinen, S. Jalkanen, Extracellular ATP formation on vascular endothelial cells is mediated by ecto-nucleotide kinase activities via phosphotransfer reactions, FASEB J. 15 (2001) 251–260. [40] A.A. Suran, A simple microradioisotopic assay for 50 -nucleotidase activity, Anal. Biochem. 55 (1973) 593–600. [41] S.K. Chatterjee, M. Bhattacharya, J.J. Barlow, A simple, specific radiometric assay for 50 -nucleotidase, Anal. Biochem. 95 (1979) 497–506. [42] M.K. Gentry, R.A. Olsson, A simple, specific, radioisotopic assay for 50 nucleotidase, Anal. Biochem. 64 (1975) 624–627. [43] C. Tucker-Pian, B. Bakay, W.L. Nyhan, 50 -Nucleotidase: solubilization, radiochemical analysis, and electrophoresis, Biochem. Genet. 17 (1979) 995–1005. [44] T.S. Shenoy, A.J. Cliffod, Adenine nucleotide metabolism in relation to purine enzymes in liver, erythrocytes, and cultured fibroblasts, Biochim. Biophys. Acta 411 (1975) 133–143. [45] D.R. Janero, C. Yarwood, J.K. Thakkar, Application of solid-phase extraction on anion-exchange cartridges to quantify 50 -nucleotidase activity, J. Chromatogr. 573 (1992) 207–218. [46] J. Servos, H. Reilander, H. Zimmermann, Catalytically active soluble ecto-50 nucleotidase purified after heterologous expression as a tool for drug screening, Drug Dev. Res. 45 (1998) 269–276. [47] N. Furtmann, J. Bajorath, Evaluation of molecular model-based discovery of ecto-50 -nucleotidase inhibitors on the basis of X-ray structures, Bioorg. Med. Chem. 21 (2013) 6616–6622. [48] S.T. Gammon, M. Bernstein, D.P. Schuster, D. Piwnica-Worms, A method for quantification of nucleotides and nucleotide analogues in thymidine kinase assays using lanthanum phosphate coprecipitation, Anal. Biochem. 369 (2007) 80–86.
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
A new, sensitive ecto-50 -nucleotidase assay for compound screening Marianne Freundlieb a, Herbert Zimmermann b, Christa E. Müller a,⇑ a b
PharmaCenter Bonn, Pharmaceutical Sciences Bonn (PSB), Pharmaceutical Institute, Pharmaceutical Chemistry I, University of Bonn, D-53121 Bonn, Germany Institute of Cell Biology and Neuroscience, Molecular and Cellular Neurobiology, Goethe University, 60438 Frankfurt am Main, Germany
a r t i c l e
i n f o
Article history: Received 22 July 2013 Received in revised form 5 October 2013 Accepted 8 October 2013 Available online 19 October 2013 Keywords: Cancer therapy CD73 Ecto-50 -nucleotidase assay Lanthanum chloride precipitation Michaelis–Menten analysis Radiometric enzyme assay
a b s t r a c t Ecto-50 -nucleotidase (eN) is a membrane-bound enzyme that hydrolyzes extracellular nucleoside-50 monophosphates yielding the respective nucleoside and phosphate. Increased levels of eN expression have been observed in many cancer cells. By increasing extracellular adenosine concentrations, they contribute to their proliferative, angiogenic, metastatic, and immunosuppressive effects. Therefore, eN is of considerable interest as a novel drug target for the treatment of cancer as well as of inflammatory diseases. In this study, we developed, optimized, and applied a highly sensitive radiometric assay using [3H]adenosine-50 -monophosphate (AMP) as a substrate. The reaction product [3H]adenosine was separated from [3H]AMP by precipitation of the latter with lanthanum chloride and subsequent filtration through glass fiber filters. Conditions were optimized to reproducibly collect the [3H]adenosine-containing filtrate used for quantitative determination. Validation of the assay yielded a mean Z0 factor of 0.73, which demonstrates its suitability for high-throughput screening. The new assay shows a limit of detection that is at least 30-fold lower than those of common colorimetric methods (e.g., optimized malachite green assay and capillary electrophoresis-based assay procedures), and it is also superior to a recently developed luciferase-based assay. Ó 2013 Elsevier Inc. All rights reserved.
Ecto-50 -nucleotidase (eN,1 CD73, EC 3.1.3.5) is a membranebound enzyme that catalyzes the dephosphorylation of extracellular nucleoside-50 -monophosphates to the respective nucleosides. Adenosine-50 -monophosphate (AMP) is its main substrate, with a Km value of 1 to 50 lM [1–3]. Recently, an X-ray structure of the human dimeric eN was obtained showing that dimerization results from the interaction of two C-terminal domains. The N-terminal domain coordinates two zinc ions in its active site that are essential for the hydrolytic reaction, whereas the C-terminal domain provides the substrate binding pocket. On binding of a substrate, the N-terminal domain rotates to the closed form where the substrate binding pocket and the metal ions form the active site [4]. The enzymatic product adenosine activates G protein-coupled receptors, termed P1 or adenosine receptors, which are subdivided into four distinct subtypes: A1, A2A, A2B, and A3 [5]. eN is often directly colocalized with adenosine receptors and contributes to their activation [6]. Therefore, eN is of considerable interest as a novel drug target. Phosphorylated adenosine derivatives have been designed as eN⇑ Corresponding author. Fax: +49 228 73 2567. E-mail address: [email protected] (C.E. Müller). Abbreviations used: eN, ecto-50 -nucleotidase; AMP, adenosine-50 -monophosphate; ATP, adenosine-50 -triphosphate; ADP, adenosine-50 -diphosphate; AOPCP, a,b-methylene-ADP; HTS, high-throughput screening; LOD, limit of detection; HPLC, highperformance liquid chromatography; Tris, tris(hydroxymethyl)aminomethane; DMSO, dimethyl sulfoxide; Hepes, 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid; cpm, counts per minute. 1
0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ab.2013.10.012
activated prodrugs of A2A adenosine receptor agonists for the site-specific treatment of inflammatory diseases without hypotensive side effects [6,7]. Inhibitors of eN may be useful for the treatment of various diseases that are associated with increased adenosine levels, for example, cancer [8]. Many cancer cells show an increased expression level of eN, leading to elevated extracellular adenosine concentrations. Adenosine stimulates tumor growth, proliferation, angiogenesis, and metastasis via the activation of A2A and A2B receptors, and it suppresses immune responses via A2A receptors [9–16]. Moreover, eN functions as a cosignaling molecule on T-lymphocytes and as an adhesion molecule that is required for lymphocyte binding and thereby directly contributes to the immunosuppressive effect [11,17]. Experiments with eN-deficient mice indicated that targeting the enzymatic activity of tumor eN may be an important new approach for cancer immunotherapy [18,19]. So far, only very few inhibitors of eN have been described [2–4,20–24]. Adenosine-50 -triphosphate (ATP) and adenosine-50 diphosphate (ADP), as well as ADP analogs, can act as competitive inhibitors of eN. The ADP analog AOPCP (1) is the most active inhibitor described so far, with a Ki value in the nanomolar range [3,20]. Very few moderately potent, non-nucleotide-derived inhibitors have been described, including anthraquinones [21], methylxanthines [2], flavonoids such as quercetin [4,22,23], and sulfonamides [24]. To identify and optimize new eN inhibitors, a fast and sensitive assay method that is suitable for high-throughput screening (HTS)
54
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
is required. Several methods have been described for assaying eN as well as intracellular 50 -nucleotidases. These include colorimetric, luminescent, chromatographic/spectroscopic, and radiometric procedures [20,25–45]. A major drawback of the currently available methods is their low sensitivity, requiring high substrate concentrations. As a consequence, only relatively potent and/or well water-soluble inhibitors can be identified. Some methods are very sophisticated and/or time-consuming and, thus, might not be suitable for HTS. Recently, a luciferase-based assay was developed to detect eN activity [25]. Luciferase is frequently being used to determine ATP concentrations; however, in the described eN assay, the authors made use of the inhibiting effect of the enzymatic reaction by AMP. eN-mediated decreases in AMP concentrations were determined by increases in the luminescent luciferase signal [25]. Thus, the assay determines concentrations of the eN reaction product adenosine indirectly by measuring a decrease in AMP concentrations. For optimal assay conditions, high AMP substrate concentrations (300 lM), well above the Km value, are required. This results in low sensitivity; that is, relatively high concentrations of test compounds need to be employed to observe an effect. It limits the detectability of moderately potent test compounds with limited solubility in screening campaigns. Furthermore, ATP, an inhibitor of eN (Ki = 8.9 lM) [20], needs to be present. The most frequently used colorimetric method is the malachite green assay, which detects phosphate produced by the hydrolysis of AMP [26– 28]. However, because phosphate is ubiquitously present, a high background is typically observed, and it is difficult to achieve a limit of detection (LOD) below 1 lM phosphate. Only after careful optimization of the assay conditions have we been able to achieve an LOD of 0.2 lM phosphate. Especially, the effects of antibodies as eN inhibitors are difficult to measure because their solutions typically contain phosphate [25]. Moreover, many colored compounds may interfere with the assay. Further assay procedures are based on high-performance liquid chromatography (HPLC) or capillary electrophoresis (CE) separation coupled with ultraviolet (UV) detection [20,29,30]. These methods are also not ideally suited for HTS because they are time-consuming. Two different coupled enzyme reaction methods have been described [31–35]. These assays are complex and vulnerable to false-positive results because it is not clear whether the enzyme of interest or the auxiliary enzyme is affected by an inhibitory test compound [36]. Assays using radiolabeled substrate have also been described applying different methods for the separation of substrate from product, including HPLC, thin layer chromatography (TLC), anion exchange or column chromatography, precipitation methods followed by centrifugation, or determination of residual substrate concentrations after filtration [37–45]. However, all of the described methods have drawbacks limiting their suitability for the screening of large libraries. Therefore, we developed and optimized a new procedure based on radioactive [3H]AMP as a substrate, quantifying formed [3H]adenosine, which is simple, highly sensitive, reproducible, and suitable for HTS.
from BIOLOG Life Science Institute (Bremen, Germany). Tris(hydroxymethyl)aminomethane (Tris), dimethyl sulfoxide (DMSO), 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid (Hepes), lanthanum chloride, and sodium chloride were obtained from Carl Roth (Karlsruhe, Germany). [2,8-3H]AMP (solution in ethanol/water, 1:1, 22.9 Ci/mmol, 1.0 mCi/ml, 849 GBq/mmol) was obtained from Hartmann Analytic (Braunschweig, Germany). [2-3H]adenosine (aqueous solution containing 0.1% ethanol, 23.0 Ci/mmol, 1.0 mCi/ml, 851 GBq/mmol) was obtained from Amersham Biosciences (Freiburg, Germany). Recombinant expression of rat ecto-50 -nucleotidase Catalytically active recombinant soluble glutathione S-transferase/eN fusion protein was expressed in insect cells using the baculovirus system and purified by affinity chromatography using agarose-coupled glutathione as described previously [46]. The active sites of rat and human enzyme are nearly identical, with only one minor difference (a Phe [human]/Tyr [rat] exchange) [47]. Preparation of standard solutions Nucleotides and nucleosides were dissolved in deionized water. For the radiometric assay, aqueous solutions with final concentrations of 1 mg/ml and 0.1 mg/ml were obtained. The standard solutions were further diluted in assay buffer as required for the calibration curves and enzyme assays. Radiometric ecto-50 -nucleotidase assay Reactions were carried out in assay buffer consisting of 25 mM Tris, 140 mM sodium chloride, and 50 mM sodium phosphate with a pH value adjusted to 7.4. [3H]AMP diluted to a specific activity of 3.7 109 Bq/mmol (100 mCi/mmol) was used as a substrate. Substrate solution (10 ll) was added to 80 ll of assay buffer (final concentration in assay 5 lM), and the mixture was preincubated for approximately 5 min at 37 °C. The reaction was then initiated by adding 10 ll of enzyme solution (in enzyme buffer containing 40 mM Hepes and 4 mM iodoacetamide, pH 7.4) to each test tube (final concentration in assay 0.3 ng/ll). The mixture was incubated for 25 min at 37 °C. To stop the reaction, 500 ll of ice-cooled precipitation buffer containing 100 mM aqueous lanthanum chloride and 100 mM aqueous sodium acetate solution (pH 4.0) was added. After approximately 30 min of cooling on ice, the precipitation was completed and the reaction mixture was filtered through GF/B glass fiber filters using a Brandel cell harvester (M-24, Brandel, Gaitherburg, MD, USA) equipped with an individual collection box for collecting the filtrates containing the separated product. After washing three times with 1 ml each of the precipitation buffer, the filtrates were poured into scintillation vials containing 6 ml of the scintillation cocktail ULTIMA Gold XR and were quantified by scintillation counting (TRICARB 2900 TR, Packard/PerkinElmer). Calibration curve
Materials and methods Reagents and chemicals Adenosine, calcium chloride dihydrate, sodium acetate, magnesium chloride, ammonium molybdate tetrahydrate (Bio Ultra), and malachite green oxalate were obtained from Sigma–Aldrich (Steinheim, Germany). AMP was obtained from Merck (Darmstadt, Germany). Disodium adenosine 50 -diphosphate monohydrate was obtained from Acros Organics (Nidderau, Germany). Adenosine 50 -O-(a,b-methylene)diphosphate (AMPCP, AOPCP) was obtained
To determine the sensitivity of the test system, a calibration curve for [3H]adenosine using a fixed concentration of [3H]AMP was determined. The detected counts per minute (cpm) was plotted versus the adenosine concentrations. Six different concentrations of adenosine ranging from 0.01 to 3 lM were used in combination with a fixed concentration of 5 lM AMP. The experiments were performed as described above for the assay, but without the addition of enzyme (three separate experiments, each in triplicate). The results were fitted to a linear regression curve using the GraphPad Prism 5 program (GraphPad Software, La Jolla, CA,
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
USA). As an indicator for the sensitivity of this assay, the LOD was calculated using the following formula:
LOD ¼
3r ; slope
where r is the standard deviation of the blank samples and ‘‘slope’’ is the slope of the corresponding calibration curve. Optimization of assay parameters For determination of the optimal enzyme concentration, an enzyme titration was conducted. Seven different enzyme concentrations between 0.032 and 2.0 lg/ml were incubated with a substrate concentration of 5 lM [3H]AMP for 15 min at 37 °C. To stop the enzymatic reaction, 500 ll of the precipitation buffer was added. Separation of the product was carried out by filtration as described above for the enzyme assay. To determine the optimal period of incubation resulting in a level of substrate conversion between 10 and 20%, the enzymatic reaction was carried out as described above but was terminated after different periods of time between 0 and 60 min of incubation. The substrate concentration was 5 lM [3H]AMP, and the enzyme concentration was 0.3 lg/ ml, based on the results of the enzyme titration conducted before. Three separate experiments were performed, each in triplicate. Determination of apparent Km value For the determination of the Michaelis–Menten constant (Km), seven different substrate concentrations of [3H]AMP between 1 and 1000 lM (final concentrations) were used. The enzymatic reaction was carried out using a final concentration of 0.3 lg/ml eN following the method described above. Three separate assays were performed, each in triplicate. Km and Vmax values were calculated by nonlinear fitting of the Michaelis–Menten equation (see below) using GraphPad Prism 5 software:
Km ¼
V max v ½S
v
:
Assay validation The validation of the newly developed assay was performed by measuring a series of positive and negative controls. In the negative controls, 10 ll of substrate solution was added to 90 ll of assay buffer, resulting in a final concentration of 5 lM. In the positive controls, 10 ll of substrate solution (final concentration 5 lM) and 10 ll of eN solution (final concentration 0.3 lg/ml protein) were added to 80 ll of assay buffer. The assay was performed as described above. Three separate assays were performed. For each of the positive and negative controls, 24 wells were analyzed. Assay performance was quantified using the Z0 factor calculated according to the formula of Zhang and coworkers [11] as shown below:
3 rþ þ 3 r ; Z 0 ¼ 1 lþ þ l where r+ and r– are the standard deviations of the positive and the negative control and l+ and l– are the mean values of the positive and negative controls, respectively. Inhibition assay For further validation of the radiometric assay, two standard inhibitors of eN, ADP and AOPCP, were tested under the established assay conditions as described above. Briefly, 10 ll of the aqueous inhibitor solution was added to 70 ll of assay buffer and 10 ll of
55
substrate solution (final substrate concentration 5 lM). To start the enzymatic reaction, 10 ll of the enzyme solution was added (final concentration 0.3 lg/ml). To obtain full concentration–response curves, 11 inhibitor concentrations ranging from 0.001 to 100 lM were tested. Three separate inhibition assays were performed, each in triplicate. Concentration–response curves were fitted using the sigmoidal dose–response curve of GraphPad Prism 5 software from which IC50 values were deduced. Ki values were calculated using the formula below:
Ki ¼
IC50 : ð1 þ ½S=K m Þ
Screening For screening of a small chemical library, compound solutions in DMSO (1 mM, 1 ll) were added to 79 ll of buffer, 10 ll of substrate solution, and 10 ll of enzyme solution, resulting in a final concentration in the assay of 10 lM of the test compound, 1% DMSO, 5 lM substrate, and 0.3 ng/ll enzyme. On each filter, the following controls were included: one containing only DMSO (no inhibitor) and one in the absence of enzyme. We showed in previous experiments that in the presence of the standard inhibitor AOPCP, the same cpm values were obtained as without enzyme. The percentage inhibition was calculated using the formula below:
% inhibition ¼
cpmnegative control cpmcompound
cpmnegative control cpmpositive control
:
Results and discussion Development and optimization of radiometric ecto-50 -nucleotidase assay AMP is the main substrate of eN. For the development of a new assay procedure, we used [3H]AMP that contained a tritium atom either in the 2-position only or in both the 2- and 8-positions. The radioactive substrate is converted to the radiolabeled product [3H]adenosine by eN (see Fig. 1). A method needed to be developed for the separation of the product [3H]adenosine (present in low amounts) from the substrate [3H]AMP (present in excess). The idea was to separate the labeled product from the radioactive substrate by precipitation of unreacted [3H]AMP using lanthanum chloride, which leads to the formation of a poorly soluble salt. An analogous procedure had previously been used for thymidine kinase assays [48]. However, in this reaction, the precipitated nucleotide was the product, which could be collected by filtration on glass fiber filters. In the case of eN, the product is [3H]adenosine, which would remain in solution. Because the lanthanum–AMP salt still shows some solubility in water, coprecipitation with sodium phosphate was investigated, yielding a lanthanum phosphate–AMP complex that possesses negligible solubility. For lanthanum chloride to achieve complete precipitation of [3H]AMP, the ratio of substrate to sodium phosphate and the pH value of the solution are important parameters that needed to be carefully optimized. We found that rather high amounts of both lanthanum chloride and sodium phosphate are required and that the precipitation needs to be performed in an acidic milieu of pH 4.0 rather than at neutral pH or an alkaline pH value of 9.0. The enzymatic reactions were performed in 25 mM Tris–HCl buffer containing 140 mM sodium chloride and 50 mM sodium phosphate with a pH value adjusted to 7.4 at 37 °C. We did not add magnesium or calcium cations because these tend to form precipitates with phosphate. The addition of these divalent cations
56
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
Fig.1. Hydrolysis of AMP by ecto-50 -nucleotidase (eN).
to the buffer was not required because the cations already bound to the added enzyme appeared to be sufficient for full enzymatic activity. Sodium phosphate was required as a buffer component to facilitate precipitation after the addition of lanthanum chloride. Although phosphate had been described to have an inhibitory effect on 50 -nucleotidases, we could observe high enzymatic activity in the applied buffer. Simulating a typical substrate–product ratio of the enzymatic reaction by analyzing a mixture of 0.5 lM [3H]adenosine and 4.5 lM [3H]AMP, a suitable measurement window was obtained in preliminary experiments (3-fold over background, which was measured with [3H]AMP only). Furthermore, we determined the DMSO sensitivity of the assay because many test compounds are only moderately or even poorly soluble in water. DMSO was tolerated up to a concentration of 2% without any effect.
Fig.2. Enzyme titration to determine the optimal enzyme concentration. The reaction was carried out with a substrate concentration of 5 lM AMP. Data points represent means ± standard errors from three different experiments, each run in triplicate. The LOD was calculated to be 0.013 lg/ml ecto-50 -nucleotidase.
Calibration of test system To determine the LOD as a parameter for the sensitivity of the test system, a calibration curve for adenosine in the presence of AMP at a final concentration of 5.00 lM was obtained by plotting the radioactivity versus the adenosine concentrations. The linear regression was calculated, and the following regression equation was obtained: y = 2512x + 996. A correlation coefficient of 0.9997 (n = 3) was obtained. The LOD for adenosine was found to be 0.028 ± 0.002 lM; thus, it was 30-fold lower than that for the recently developed CE method [20] and also more than 30-fold lower than the LOD for phosphate in colorimetric malachite green assays, which is described in the literature to be approximately 1 lM [26,28]. Optimization of assay parameters To optimize the assay parameters, we initially determined a suitable substrate concentration. To achieve appropriate assay conditions for testing competitive as well as uncompetitive inhibitors, the substrate concentrations should range between 0.1 Km and 10 Km. Because a high sensitivity is desired, a concentration of 5 lM AMP was chosen, which is slightly lower than the Km value of 25 lM determined for rat eN [46]. Another very important parameter to improve the assay sensitivity and robustness is the enzyme concentration, which is directly proportional to the initial velocity. If the enzyme concentration is very high and the substrate is quickly depleted, product formation might not be linear. To define the optimal enzyme concentration, an enzyme titration was performed at the selected substrate concentration of 5 lM. A linear increase of the detected signal in relation to an increase in enzyme concentration was observed (Fig. 2). For the determination of the optimal enzyme concentration, the LOD was calculated following the formula listed in Materials and Methods. The LOD corresponds to the minimal concentration of enzyme that can be reliably differentiated from the background. The optimal enzyme concentration is selected to be at least 20-fold higher than the LOD. The result of the enzyme titration (Fig. 2) yielded the regression equation y = 9.75x + 0.0464, with a correlation coefficient of 0.9978 (n = 3). The LOD was found to be 0.013 lg/ml; therefore, an enzyme concentration of 0.3 lg/ml was chosen as the final assay concentration. The third parameter
determined for assay optimization was the incubation time. To achieve an optimal substrate conversion rate between 10 and 20%, a time course analysis was performed. The range of 10 to 20% of product formation was obtained between 15 and 40 min of incubation; therefore, we chose an incubation time of 25 min for further experiments. Thus, the optimized parameters were as follows: 5 lM AMP as a substrate, 0.3 lg/ml enzyme, and 25 min of incubation time. Determination of kinetic parameters of ecto-50 -nucleotidase Because the kinetic parameters can vary between test systems, the assay-specific Km value was determined using the optimized parameters. The kinetic parameters of the purified rat eN for AMP were calculated using the Michaelis–Menten equation (Fig. 3). The determined Km value was 59 lM and, thus, was roughly in the same range as those determined with other assay methods (1–50 lM) [2,20,46] and an enzyme turnover number of 1.1 s1 was calculated.
Fig.3. Michaelis–Menten plot for the determination of the apparent Km value of AMP using the newly developed radiometric method. Data points represent means ± standard errors from three separate experiments, each run in triplicate. The Km value was calculated to be 59 ± 10 lM.
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
57
Method validation The Z0 factor is a widely used statistical parameter to evaluate robustness and quality of HTS assays. To investigate whether the new radiometric assay was suitable for HTS, a pilot Z0 factor study was conducted prior to applying the assay for extensive compound screening (Fig. 4). We determined a high Z0 factor of 0.73. This result demonstrates a very high assay robustness of the applied radiometric method and its excellent suitability for HTS campaigns. Enzyme inhibition assays
Fig.4. Z0 factor study to determine the quality and robustness of the assay.
For further validation of the radiometric assay, the inhibitory effects of two known competitive inhibitors, AOPCP and ADP, were determined using the optimized conditions. Concentration–inhibition curves are shown in Fig. 5. An IC50 value of 0.236 lM for AOPCP and an IC50 value of 4.66 lM for ADP were determined. These values are well in agreement with published data [3,20,42]. Screening of a compound library As a next step, we screened a small compound library consisting of 321 synthetic, drug-like, mostly non-ionic heterocyclic compounds. The compounds were screened at a concentration of 10 lM. Hit compounds were defined as those inhibiting at least 70% of the enzymatic reaction. Applying this high threshold, we identified 12 hit compounds (4%) as inhibitors. One compound had shown an increase in the signal in initial screening; however, it could not be confirmed in subsequent assays. The results are shown in Fig. 6. These results show that the newly developed and optimized assay procedure is suitable for library screening.
Fig.5. Concentration–inhibition curves for the eN inhibitors AOPCP and ADP. Data points represent means ± standard errors from three different experiments, each run in triplicate. The IC50 values were determined to be 0.236 ± 0.006 lM for AOPCP and 4.66 ± 0.13 lM for ADP. The corresponding Ki values were calculated as 0.197 ± 0.005 lM for AOPCP and 3.88 ± 0.11 lM for ADP.
Conclusions eN has recently gained considerable attention as a novel potential drug target. Therefore, a radiometric eN assay using recombinant purified rat eN was developed to provide a sensitive test system suitable for HTS. The rat eN is believed to be predictive
Fig.6. Graphical presentation of the screening results of a small library of 321 drug-like compounds tested at a concentration of 10 lM in the developed eN assay. Each dot represents one compound. Compounds above the red line showing more than 70% inhibition were defined as hit compounds (12 compounds, 4% of library). (For interpretation of the reference to color in this figure legend, the reader is referred to the web version of this article.)
58
Ecto-50 -nucleotidase assay / M. Freundlieb et al. / Anal. Biochem. 446 (2014) 53–58
for the human eN because the active sites of both enzymes are nearly identical, differing only in a single amino acid: a tyrosine in the rat, which is homologously exchanged for a phenylalanine in the human enzyme. With the newly developed radiometric method presented here, we were able to develop a new sensitive and fast assay for the analysis of eN activity and to identify and characterize inhibitors of eN. Advantages of the new assay include a very low LOD of 0.03 lM for adenosine. This is much lower than those of all previously developed eN assays and allows screening around or below the Km value. This is particularly useful for identifying hit compounds that possess low water solubility and for fragment-based screening approaches. Moreover, colored compounds and inorganic phosphate do not interfere with the assay, in contrast to the standard malachite green assay. The optimized assay is accurate, robust, convenient, inexpensive, and fast. The reaction product is directly detected and quantified. The developed radiometric eN assay will be highly useful for identifying and characterizing novel eN inhibitors that have the potential to become novel anti-cancer and anti-inflammatory drugs. Acknowledgments We thank Nicole Florin, Peter Ripphausen, Inês Ferreira, and Karin Pelka for preliminary experiments. H.Z. acknowledges support from grants from the Cluster of Excellence (EXC 115, Frankfurt, Germany) and the Gutenberg Research College (GRC) University Mainz. References [1] H. Zimmermann, M. Zebisch, N. Sträter, Cellular function and molecular structure of ecto-nucleotidases, Purinergic Signal. 8 (2012) 437–502. [2] H. Zimmermann, 50 -Nucleotidase: molecular structure and functional aspects, Biochem. J. 285 (1992) 345–365. [3] S.A. Hunsucker, B.S. Mitchell, J. Spychala, The 50 -nucleotidases as regulators of nucleotide and drug metabolism, Pharmacol. Ther. 107 (2005) 1–30. [4] K. Knapp, M. Zebisch, J. Pippel, A. El-Tayeb, C.E. Müller, N. Sträter, Crystal structure of the human ecto-50 -nucleotidase (CD73): insights into the regulation of purinergic signaling, Structure 20 (2012) 2161–2173. [5] G.G. Yegutkin, Nucleotide- and nucleoside-converting ectoenzymes: important modulators of purinergic signalling cascade, Biochim. Biophys. Acta 1783 (2008) 673–694. [6] U. Flögel, S. Burghoff, P.L.E.M. van Lent, S. Temme, L. Galbarz, Z. Ding, A. ElTayeb, S. Huels, F. Bönner, N. Borg, C. Jacoby, C.E. Müller, W.W. van den Berg, J. Schrader, Selective activation of adenosine A2A receptors on immune cells by a CD73-dependent prodrug suppresses joint inflammation in experimental rheumatoid arthritis, Sci. Transl. Med. 4 (2012) 146ra108. [7] A. El-Tayeb, J. Iqbal, A. Behrenswerth, M. Romio, M. Schneider, H. Zimmermann, J. Schrader, C.E. Müller, Nucleoside-50 -monophosphates as prodrugs of adenosine A2A receptor agonists activated by ecto-50 nucleotidase, J. Med. Chem. 52 (2009) 7669–7677. [8] S.P. Colgan, H.K. Eltzschig, T. Eckle, L.F. Thompson, Physiological roles for ecto50 -nucleotidase (CD73), Purinergic Signal. 2 (2006) 351–360. [9] B. Zhang, CD73: A novel target for cancer immunotherapy, Cancer Res. 70 (2010) 6407–6411. [10] J. Stagg, U. Dicisekera, N. McLaughlin, J. Sharkey, S. Pommey, D. Denoyer, K.M. Dwyer, M.J. Smyth, Anti-CD73 antibody therapy inhibits breast tumor growth and metastasis, Proc. Natl. Acad. Sci. USA 107 (2010) 1547–1552. [11] J.H. Zhang, T.D.Y. Chung, K.R. Oldenburg, A simple statistical parameter for use in evaluation and validation of high throughput screening assays, J. Biomol. Screen. 4 (1999) 67–73. [12] X. Zhi, Y. Wang, X. Zhou, J. Yu, R. Jian, S. Tang, L. Yin, P. Zhou, RNAi-mediated CD73 suppression induces apoptosis and cell-cycle arrest in human breast cancer cells, Cancer Sci. 101 (2010) 2561–2569. [13] S. Gessi, S. Merighi, V. Sacchetto, C. Simioni, P.A. Borea, Adenosine receptors and cancer, Biochim. Biophys. Acta 2011 (1808) 1400–1412. [14] J. Spychala, Tumor-promoting functions of adenosine, Pharmacol. Ther. 87 (2000) 161–173. [15] J. Stagg, M.J. Smyth, Extracellular adenosine triphosphate and adenosine in cancer, Oncogene 29 (2010) 5345–5358. [16] V. Kumar, A. Sharma, Adenosine: an endogenous modulator of innate immune system with therapeutic potential, Eur. J. Pharmacol. 616 (2009) 7–15. [17] R. Resta, L.F. Thompson, T cell signalling through CD73, Cell. Signal. 9 (1997) 131–139. [18] G.G. Yegutkin, F. Marttila-Ichihara, M. Karikoski, J. Niemelä, J.P. Laurila, K. Elima, S. Jalkanen, M. Salmi, Altered purinergic signaling in CD73-deficient mice inhibits tumor progression, Eur. J. Immunol. 41 (2011) 1231–1241.
[19] J. Stagg, U. Divisekera, H. Duret, T. Sparwasser, M.W.L. Teng, P.K. Darcy, M.J. Smyth, CD73-deficient mice have increased antitumor immunity and are resistant to experimental metastasis, Cancer Res. 71 (2011) 2892–2900. [20] J. Iqbal, D. Jirovsky, S. Lee, H. Zimmermann, C.E. Müller, Capillary electrophoresis-based nanoscale assays for monitoring ecto-50 -nucleotidase activity and inhibition in preparations of recombinant enzyme and melanoma cell membranes, Anal. Biochem. 373 (2008) 129–140. [21] Y. Baqi, S.Y. Lee, J. Iqbal, A. Lehr, A.B. Scheiff, H. Zimmermann, J. Bajorath, C.E. Müller, Development of potent and selective inhibitors of ecto-50 -nucleotidase based on athraquinone scaffold, J. Med. Chem. 53 (2010) 2076–2086. [22] E. Braganhol, A.S.K. Tamajusuku, A. Bernardi, M.R. Wink, A.M.O. Battastini, Ecto-50 -nucleotidase/CD73 inhibition by quercetin in the human U138MG glioma cell line, Biochim. Biophys. Acta 1770 (2007) 1352–1359. [23] M. Kavutcu, M.F. Melzig, In vitro effects of selected flavonoids on the 50 nucleotidase activity, Pharmazie 54 (1999) 457–459. [24] P. Ripphausen, M. Freundlieb, A. Brunschweiger, H. Zimmermann, C.E. Müller, J. Bajorath, Virtual screening identifies novel sulfonamide inhibitors of ecto-50 nucleotidase, J. Med. Chem. 55 (2012) 6576–6581. [25] K.F. Sachenmeier, C. Hay, E. Brand, L. Clarke, K. Rosenthal, S. Guillard, S. Rust, R. Minter, R. Hollingsworth, Development of a novel ecto-nucleotidase assay suitable for high-throughput screening, J. Biomol. Screen. 17 (2012) 993–998. [26] A.A. Baykov, O.A. Evtushenko, S.M. Avaeva, A malachite green procedure for orthophosphate determination and its use in alkaline phosphatase-based enzyme immunoassay, Anal. Biochem. 171 (1988) 266–270. [27] P.A. Lanzetta, L.J. Alvarez, P.S. Reinach, O.A. Candia, An improved assay for nanomolar amounts of inorganic phosphate, Anal. Biochem. 100 (1979) 95–97. [28] E.B. Cogan, G.B. Birrell, O.H. Griffith, A robotics-based automated assay for inorganic and organic phosphates, Anal. Biochem. 271 (1999) 29–35. [29] A. Amici, M. Emanuelli, N. Raffaelli, S. Ruggieri, G. Magni, One-minute highperformance liquid chromatography assay for 50 -nucleotidase using a 20-mm reverse-phase column, Anal. Biochem. 216 (1994) 171–175. [30] J.N. Stolk, R.A. De Abreu, A.M.T. Boerbooms, D.G. de Koning, R. de Graaf, P.J.S.M. Kerstens, L.B.A. van de Putte, Purine enzyme activities in peripheral blood mononuclear cells: comparison of a new non-radiochemical high performance liquid chromatography procedure and a radiochemical thin-layer chromatography procedure, J. Chromatogr. 666 (1995) 33–43. [31] S. Perez, N. Courtis, D. Kokkinopoulos, M. Papamichail, C.M. Tsiapalis, T. Trangas, A colorimetric assay for the determination of 50 -nucleotidase activity, J. Immunol. Methods 101 (1987) 73–78. [32] C.P. Price, P.G. Hill, H.G. Sammons, A comparison of two colorimetric procedures for the assay of 50 -nucleotidase, Clin. Chim. Acta 33 (1971) 260– 263. [33] V.G. Bethune, M. Fleisher, M.K. Schwarz, Automated method for determination of serum 50 -nucleotidase activity, Clin. Chem. 18 (1972) 1524–1528. [34] F. Heinz, R. Pilz, S. Reckel, J.R. Kalden, R. Haeckel, A new spectrophotometric method for the determination of 50 -nucleotidase, J. Clin. Chem. Clin. Biochem. 18 (1980) 781–788. [35] A.H. Chalmers, C. Hare, A semi-automated method for 50 -ecto-nucleotidase measurement in lymphocytes, Immunol. Cell Biol. 68 (1990) 75–79. [36] N.J. Kruger, Errors and artifacts in coupled spectrophotometric assays of enzyme activity, Phytochemistry 38 (1995) 1065–1071. [37] P.H. Ellims, L. Bailey, M.B. Van der Weyden, An improved method for the determination of human erythrocyte pyrimidin 50 -nucleotidase activity, Clin. Chim. Acta 88 (1978) 99–103. [38] E.P. Garvey, G.T. Lowen, M.R. Almond, Nucleotide and nucleoside analogues as inhibitors of cytosolic 50 -nucleotidase I from heart, Biochemistry 37 (1998) 9043–9051. [39] G.G. Yegutkin, T. Henttinen, S. Jalkanen, Extracellular ATP formation on vascular endothelial cells is mediated by ecto-nucleotide kinase activities via phosphotransfer reactions, FASEB J. 15 (2001) 251–260. [40] A.A. Suran, A simple microradioisotopic assay for 50 -nucleotidase activity, Anal. Biochem. 55 (1973) 593–600. [41] S.K. Chatterjee, M. Bhattacharya, J.J. Barlow, A simple, specific radiometric assay for 50 -nucleotidase, Anal. Biochem. 95 (1979) 497–506. [42] M.K. Gentry, R.A. Olsson, A simple, specific, radioisotopic assay for 50 nucleotidase, Anal. Biochem. 64 (1975) 624–627. [43] C. Tucker-Pian, B. Bakay, W.L. Nyhan, 50 -Nucleotidase: solubilization, radiochemical analysis, and electrophoresis, Biochem. Genet. 17 (1979) 995–1005. [44] T.S. Shenoy, A.J. Cliffod, Adenine nucleotide metabolism in relation to purine enzymes in liver, erythrocytes, and cultured fibroblasts, Biochim. Biophys. Acta 411 (1975) 133–143. [45] D.R. Janero, C. Yarwood, J.K. Thakkar, Application of solid-phase extraction on anion-exchange cartridges to quantify 50 -nucleotidase activity, J. Chromatogr. 573 (1992) 207–218. [46] J. Servos, H. Reilander, H. Zimmermann, Catalytically active soluble ecto-50 nucleotidase purified after heterologous expression as a tool for drug screening, Drug Dev. Res. 45 (1998) 269–276. [47] N. Furtmann, J. Bajorath, Evaluation of molecular model-based discovery of ecto-50 -nucleotidase inhibitors on the basis of X-ray structures, Bioorg. Med. Chem. 21 (2013) 6616–6622. [48] S.T. Gammon, M. Bernstein, D.P. Schuster, D. Piwnica-Worms, A method for quantification of nucleotides and nucleotide analogues in thymidine kinase assays using lanthanum phosphate coprecipitation, Anal. Biochem. 369 (2007) 80–86.
