Analytical Biochemistry 447 (2014) 146–155
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
Development of a highly sensitive, high-throughput assay for glycosyltransferases using enzyme-coupled fluorescence detection Kazuo Kumagai a,b,⇑, Hirotatsu Kojima a, Takayoshi Okabe a, Tetsuo Nagano a a b
Open Innovation Center for Drug Discovery, The University of Tokyo, Tokyo 113-0033, Japan Genomic Science Laboratories, Dainippon Sumitomo Pharma Co. Ltd., Osaka 554-0022, Japan
a r t i c l e
i n f o
Article history: Received 10 August 2013 Received in revised form 20 November 2013 Accepted 22 November 2013 Available online 1 December 2013 Keywords: Glycosyltransferase Nucleotide determination Enzyme-coupled assay Fluorescence assay High-throughput screening
a b s t r a c t Glycosyltransferases catalyze transfer of sugar moieties from activated donor molecules to specific acceptor molecules, forming glycosidic bonds. Identification of selective modulators of glycosyltransferases is important both to provide new tools for investigating pathophysiological roles of glycosylation reactions in cells and tissues, and as new leads in drug discovery. Here we describe a universal enzyme-coupled fluorescence assay for glycosyltransferases, based on quantification of nucleotides produced in the glycosyl transfer reaction. GDP, UDP, and CMP are phosphorylated with nucleotide kinase in the presence of excess ATP, generating ADP. Via coupled enzyme reactions involving ADP-hexokinase, glucose-6-phosphate dehydrogenase, and diaphorase, the ADP is utilized for conversion of resazurin to resorufin, which is determined by fluorescence measurement. The method was validated by comparison with an HPLC method, and employed to screen the LOPAC1280 library for inhibitors in a 384-well plate format. The assay performed well, with a Z0 -factor of 0.80. We identified 12 hits for human galactosyltransferase B4GALT1 after elimination of false positives that inhibited the enzyme-coupled assay system. The assay components are all commercially available and the reagent cost is only 2 to 10 US cents per well. This method is suitable for low-cost, high-throughput assay of various glycosyltransferases and screening of glycosyltransferase modulators. Ó 2013 Elsevier Inc. All rights reserved.
Glycosyltransferases are ubiquitous enzymes that are involved in the biosynthesis of disaccharides, oligosaccharides, and polysaccharides. They catalyze the transfer of a sugar moiety from donor molecules to glycosyl acceptor molecules, which range from proteins, lipids, sugars, and nucleic acids to small molecules [1]. Defects in glycosyltransferase activities are associated with various human diseases, including cancer, inflammation, neurological disorders, and congenital disorders of glycosylation [2,3]. Glycosyltransferases also influence the distribution and clearance of various biological molecules by regulating their glycosylation [4,5]. Thus, smallmolecular modulators of glycosyltransferases are expected to be useful research tools and may also serve as leads for drug discovery. However, few potent and selective inhibitors of glycosyltransferases are known [6], largely due to the lack of suitable high-throughput screening (HTS)1 assay methods to screen large chemical libraries. ⇑ Corresponding author. Address: Open Innovation Center for Drug Discovery, The University of Tokyo, Pharmaceutical Science Main Building, Tokyo 113-0033, Japan. Fax: +81 3 5841 1959. E-mail address: [email protected] (K. Kumagai). 1 Abbreviations used: HTS, high-throughput screening; B4GALT1, b-1,4-galactosyltransferase 1; FUT7, fucosyltransferase 7; ST6GAL1, b-galactoside a-2,6-sialyltransferase 1; NDP, nucleoside 50 -diphosphate; NDPK, nucleoside diphosphate kinase; CMPK1, CMP kinase 1; CMP-NeuAc, CMP-sialic acid; G6PDH, glucose-6-phosphate dehydrogenase; DTT, dithiothreitol; NEM, N-ethylmaleimide; BSA, bovine serum albumin; DMSO, dimethyl sulfoxide. 0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ab.2013.11.025
Currently, the most widely used assay methods for glycosyltransferases are based on the detection of glycosylated products by using chromatographic, radiochemical, spectrophotometric, or immunological techniques [7]. However, most of these methods are low throughput, expensive, and labor-intensive. Compared with these methods, an alternative approach based on determination of nucleoside di(mono)phosphate generated from the sugar donor substrate would be both simpler and generally applicable to all glycosyltransferases. Most glycosyltransferases in mammals utilize uridine diphosphate (UDP)-sugars (e.g., galactosyltransferases), guanosine diphosphate (GDP)-sugars (e.g., fucosyltransferases), or cytidine monophosphate (CMP)-sugars (e.g., sialyltransferases) as sugar donor substrates [8]. Hence, the nucleotides UDP, GDP, and CMP liberated from these glycosyl donors are common products of glycosyltransferase reactions. Several methods other than chromatographic methods have been developed to detect these nucleotides for glycosyltransferase activity assays. Fluorescence polarization-based assays have been developed for the quantification of UDP, GDP, and CMP, and are available as commercial assay kits, but they are expensive for use in HTS [7]. Wu et al. determined inorganic phosphate released from these nucleotides by nucleotide phosphatase reactions by means of colorimetry with a malachite-based reagent [9]. This method is less costly and applicable to a wide range of glycosyltransferases, but
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
is not sensitive. Zhang et al. reported an enzyme-coupled fluorescence assay for N-acetylglucosaminyltransferase activity in which UDP was determined fluorometrically after coupling with pyruvate kinase reaction: detection was done by a commercial pyruvate assay kit [10]. UDP, GDP, and CMP can also be determined spectrophotometrically or fluorometrically by enzyme-coupled NADH detection [11–13], but the fluorescence intensity of NADH is relatively low. Here we report a highly sensitive, easy-to-operate, universal assay based on enzyme-coupled fluorescence detection of UDP, GDP, and CMP. This method is applicable in principle to any glycosyltransferase that utilizes UDP-, GDP-, or CMP-sugars as substrates. It employs nucleotide kinase reaction to generate ADP from the three nucleotides, followed by a cascade of coupled enzyme reactions that utilize the ADP for conversion of resazurin to highly fluorescent resorufin, which is determined by fluorescence measurement. The optimized method requires a volume of only a few microliters per well for glycosyltransferase assay in a 384-well plate format, and provides a high signal-to-background ratio. The method was validated by comparison with HPLC assay, and was applied to screen the LOPAC1280 library for inhibitors of human galactosyltransferase B4GALT1. We believe that this assay system provides a simple, cost-effective tool that will be applicable to many glycosyltransferases.
Materials and methods Materials Recombinant human b-1,4-galactosyltransferase 1 (B4GALT1), fucosyltransferase 7 (FUT7), and b-galactoside a-2,6-sialyltransferase 1 (ST6GAL1) were purchased from R&D Systems (Minneapolis, MN, USA). Nucleoside diphosphate kinase (NDPK, 1.9 kU/mg protein) from baker’s yeast, fetuin from fetal bovine serum, N-acetyl-D-lactosamine, and the LOPAC1280 compound library were obtained from Sigma–Aldrich (St. Louis, MO, USA). Recombinant human cytidine monophosphate kinase 1 (CMPK1) was from Prospec (Ness Ziona, Israel). ADP-hexokinase (52.6 U/mg solid) from Thermococcus litoralis was from Asahi Kasei Pharma (Tokyo, Japan). Diaphorase-I (1.8 kU/mg protein) from Bacillus stearothermophilus was from Unitika (Osaka, Japan). Recombinant glucose-6-phosphate dehydrogenase (G6PDH, 754 U/mg protein) from Leuconostoc sp. and NADP+ were from Oriental Yeast (Tokyo, Japan). GDP-fucose (GDP-Fuc), UDP-galactose (UDP-Gal), and CMP-sialic acid (CMP-NeuAc) were from Yamasa (Choshi, Japan), MP Biomedicals (Solon, OH, USA), and Calbiochem (San Diego, CA, USA), respectively. ATP, GDP, UDP, dithiothreitol (DTT), N-acetylglucosamine (GlcNAc), N-ethylmaleimide (NEM), bovine serum albumin (BSA), resazurin, gallic acid, and other chemicals were from Wako Pure Chemical Industries (Osaka, Japan). Nonbinding, small-volume 384-well black microtiter plates were from Greiner Bio-One (Frickenhausen, Germany).
Determination of ADP by enzyme-coupled fluorescence assay ADP in 7.5 ll of 50 mM Tris-HCl (pH 7.5) was mixed with 7.5 ll aliquots of 2X ADP detection reagent (2 mM glucose, 200 lM NADP+, 100 lM resazurin, 2 U/ml ADP-hexokinase, 2 U/ml G6PDH, and 2 U/ml diaphorase I in buffer A (100 mM Tris-HCl, 10 mM MgCl2, pH 7.5)) in 384-well plates and incubated at room temperature for 60 min. Fluorescence intensity due to the formation of resorufin was measured with a microplate reader, PHERAstar (BMG Labtech, Offenburg, Germany), with excitation at 540 nm and emission at 590 nm.
147
Enzyme-coupled fluorescence assay of GDP-fucosyltransferase and UDP-galactosyltransferase The reactions of fucosyltransferase FUT7 and galactosyltransferase B4GALT1 were performed in 5 ll of buffer B (50 mM Tris-HCl, 5 mM MgCl2, 5 mM MnCl2, pH 7.5) in 384-well plates at 37 °C for 60 min. FUT7 was reacted with 3 mg/ml fetuin and 50 lM GDP-Fuc. B4GALT1 was reacted with 5 mM GlcNAc and 50 lM UDP-Gal. To detect GDP and UDP produced by these reactions, 5 ll aliquots of 200 lM ATP and 2 U/ml NDPK in buffer A were added and the mixtures were incubated at room temperature for 60 min. Then 10 ll aliquots of 2X ADP detection reagent were added and incubation was continued at room temperature for 60 min. The fluorescence intensity (ex 540/em 590) was measured as described above.
Enzyme-coupled fluorescence assay of CMP-sialyltransferase Sialyltransferase ST6GAL1 reaction was carried out by adding the enzyme to 5 ll of buffer C (25 mM Tris-HCl, 5 mM MgCl2, 5 mM MnCl2, 150 mM NaCl, 0.1% (w/v) Triton X-100, pH 7.5) supplemented with 250 lM N-acetyl-D-lactosamine and 50 lM CMP-NeuAc in 384-well plates at 37 °C for 90 min. To detect CMP produced by the reaction, 5 ll aliquots of 200 lM ATP, 3 lg/ ml CMPK1, and 4 mM DTT in buffer D (175 mM Tris-HCl, 20 mM MgCl2, 150 mM KCl, 0.01% (w/v) BSA, 0.1% (w/v) Triton X-100, pH 9.0) were added. Incubation was continued at 37 °C for 60 min, and then 10 ll aliquots of 2X ADP detection reagent supplemented with 20 mM NEM were added. After further incubation at room temperature for 60 min, the fluorescence intensity (ex 540/em 590) was measured.
HPLC analysis HPLC analysis for nucleotides and nucleotide-sugars was performed using a YMC-Triart C18 column (150 4.6 mm, S-5 lm, YMC, Kyoto, Japan) with 50 mM KH2PO4–K2HPO4 (100:0 (v/v) for CDP, CMP, and CMP-NeuAc; 67:33 (v/v) for ADP, ATP, GDP, GTP, GDP-Fuc, UDP, UTP, and UDP-Gal) as an eluent at a flow rate of 1 ml/min with UV detection at 260 nm.
Assay validation Assay of inhibitory activity toward FUT7 was performed by both the enzyme-coupled fluorescence method and the HPLC method for validation of the new method. FUT7 (1 lg/ml) was incubated with 3 mg/ml fetuin and 50 lM GDP-Fuc in the presence or absence of the test compound gallic acid in 20 ll of buffer B at 37 °C for 60 min. The reaction was stopped by chilling the tubes on ice, and samples (10 ll) were analyzed by HPLC to determine liberated GDP. Remaining samples (5 ll) were transferred to wells of a 384-well plate. Aliquots of 5 ll of 200 lM ATP and 2 U/ml NDPK in buffer A supplemented with 0.01% (w/v) BSA were added and the plate was incubated at room temperature for 60 min. Then 10 ll of 2X ADP detection reagent supplemented with 0.01% (w/v) BSA was added to each well and incubation was continued at room temperature for 60 min, followed by fluorescence measurement. The percentage inhibition was calculated relative to controls treated with DMSO alone. In the control assay to determine the effect of the test compound on the enzyme-coupled assay reactions, 5 ll aliquots of 20 lM GDP in buffer B in the presence or absence of test compound were pipetted into wells of a 384-well plate instead of the FUT7 reaction mixtures, and enzyme-coupled fluorescence detection of GDP was performed as described above.
148
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
Library screening The LOPAC1280 library of 1280 pharmacologically active compounds (1 mM in DMSO) was dispensed into wells of columns 3 through 22 of 384-well plates (320 compounds/plate) at 50 nl/well and the same amount of DMSO was dispensed into wells of columns 2 and 23 using an Labcyte Echo 555 acoustic liquid handler (Sunnyvale, CA, USA). Aliquots of 2.5 ll of 10 mM GlcNAc and 100 lM UDP-Gal in buffer B supplemented with an additional 5 mM MnCl2 and 0.02% (w/v) Triton X-100 were added into columns 2–23 by a Multidrop Combi dispenser (Thermo Fisher Scientific, Vantaa, Finland), followed by addition of 2.5 ll of 0.05 lg/ml B4GALT1 in the same buffer (columns 3–22 for test compounds and column 23 for high signal control) or buffer only (column 2 for low signal control). After incubation at 37 °C for 60 min, 5 ll aliquots of 200 lM ATP and 2 U/ml NDPK in buffer A supplemented with 0.02% (w/v) BSA and 0.02% (w/v) Triton X-100 were added to all wells and the plates were incubated at room temperature for 60 min. Next, 10 ll of 2X ADP detection reagent supplemented with 0.02% (w/v) BSA and 0.02% (w/v) Triton X-100 was added, and incubation was continued at room temperature for 60 min, followed by fluorescence measurement. Two runs of the assay were conducted. The activity of compounds was calculated as the average of the two runs. The Z0 -factor was calculated by using the equation Z0 = 1 (3rh + 3rl)/|lh ll| [14], where rh and rl are the standard deviations of the high and low signal controls, respectively, and lh and ll are the mean signal intensities of the high and low signal controls, respectively. In the control assay, 5 ll of 20 lM UDP in buffer B was added to wells containing the same amount of LOPAC1280 library compounds or DMSO only, and UDP detection by enzyme-coupled fluorescence assay was performed in the same manner as described above. The activity
of hit compounds was validated by HPLC assay by determining UDP produced by B4GALT1 reaction.
Results Optimization of enzyme-coupled fluorescence assay for determination of ADP The present assay for glycosyltransferase activity depends on the measurement of GDP, UDP, and CMP liberated from glycosyl donors. These nucleotides are phosphorylated by nucleotide kinase in the presence of excess ATP with stoichiometric formation of ADP, and the amount of ADP is measured fluorometrically after a series of coupled enzyme reactions that serve to generate a fluorescent product (Fig. 1). We first optimized the enzyme-coupled fluorescence measurement of ADP. The ADP measurement assay consists of the following coupled reactions: ADP was first coupled to the conversion of glucose to glucose-6-phosphate by ADP-hexokinase; then glucose-6-phosphate was coupled to the conversion of NADP+ to NADPH by glucose-6-phosphate dehydrogenase (G6PDH); finally NADPH was coupled to the conversion of resazurin to a highly fluorescent end product, resorufin, by diaphorase I. Because all three enzymes work at neutral pH in the presence of Mg2+, we tested the detection of 15 lM ADP in buffer A (100 mM Tris-HCl, 10 mM MgCl2, pH 7.5) using the enzymes at various concentrations in the presence of 1 mM glucose, 100 lM NADP+, and 50 lM resazurin. Addition of the enzymes at final concentrations higher than 1 U/ml each did not result in any further increase in the fluorescence intensity (data not shown), suggesting that 1 U/ ml of each enzyme is sufficient for detection of 15 lM ADP. Therefore, we set the composition of ADP detection reagent as 1 U/ml
Fig.1. Schematic representation of enzyme-coupled fluorescence assay for determination of glycosyltransferase activity. Glycosyltransferase reactions liberate nucleotides UDP, GDP, and CMP from the respective glycosyl-donor substrates. These nucleotides are then phosphorylated by NDP kinase or CMP kinase in the presence of an excess amount of ATP to produce phosphorylated products and ADP. Via a series of coupled enzyme reactions involving ADP-hexokinase, glucose-6-phosphate dehydrogenase, and diaphorase in the presence of excess amounts of glucose, NADP+, and resazurin, the ADP is utilized for formation of highly fluorescent resorufin, which is determined by fluorescence measurement. Therefore, the fluorescence intensity reflects the amount of UDP, GDP, or CMP formed by glycosyltransferase reaction.
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
each of ADP-hexokinase, G6PDH, and diaphorase I, 1 mM glucose, 100 lM NADP+, and 50 lM resazurin (final concentrations). A typical standard curve and time course of ADP detection reaction are shown in Fig. 2. The standard curve was linear up to 32 lM ADP (Fig. 2A). The signal/background (S/B) ratio was 55.0 for 32 lM ADP. The reactions were completed in 10 min at room temperature and the fluorescence intensity remained unchanged for at least 2 h thereafter (Fig. 2B). This indicates that there is no need to stop the reaction for signal measurement. When the assay was carried out in the presence of N-ethylmaleimide (NEM) at a final concentration of 10 mM, almost the same results were obtained (Fig. 2A and B), indicating that the enzyme-coupled reactions are resistant to this thiol reagent. Additives such as NaCl, MgCl2, CaCl2, MnCl2, DMSO, Na3VO4, Triton X-100, EDTA, NaF (640 mM), and ATP did not affect the fluorescence intensity at the concentrations tested (Fig. 3). NaF decreased the signal by 31% at 50 mM.
Optimization of enzymatic phosphorylation of GDP and UDP to produce ADP GDP and UDP were phosphorylated in the presence of a 5-fold molar excess of ATP using nucleoside diphosphate kinase (NDPK) to produce ADP. In the presence of 0.5 U/ml of NDPK, 87 and 86%
A
149
of 40 lM GDP and UDP were converted to nucleoside triphosphates GTP and UTP, respectively, during reaction at room temperature for 60 min and an equivalent amount of ADP was produced as determined by HPLC (Table 1). There was no increase in conversion rate at higher concentrations of NDPK for either GDP or UDP, indicating that 0.5 U/ml of NDPK is sufficient for quantitative conversion of the nucleotides. In a control assay without GDP or UDP, no increase in ADP concentration was observed even at 50 U/ml NDPK. This demonstrates that the NDPK used has no ATPase activity.
Optimization of enzymatic phosphorylation of CMP to produce ADP CMP was phosphorylated in the presence of a 5-fold molar excess of ATP by cytidine monophosphate kinase CMPK1 to produce ADP. Because CMPK1 is a DTT-dependent enzyme (Fig. 4), the phosphorylation reaction was performed in the presence of 2 mM DTT. About 58% of 40 lM CMP was converted to CDP during reaction at 37 °C for 60 min in the presence of 1 lg/ml CMPK1, as determined by HPLC (Table 2). At 2 lg/ml CMPK1, the increase of conversion rate was only about 5% over that at 1 lg/ml CMPK1, suggesting that a conversion rate of around 60% is the maximum under this assay condition. Thus, a CMPK1 concentration of
B
Fig.2. Typical standard curves (A) and time courses (B) of enzyme-coupled fluorescence assay determination of ADP in the presence (N) or absence (d) of N-ethylmaleimide (NEM). (A) To a solution of the indicated concentrations of ADP in 7.5 ll of 50 mM Tris-HCl, pH 7.5, was added 7.5 ll of 2X ADP detection reagent (2 mM glucose, 200 lM NADP+, 100 lM resazurin, 2 U/ml ADP-hexokinase, 2 U/ml G6PDH, and 2 U/ml diaphorase I in buffer A with or without 20 mM NEM) in a 384-well plate. The plate was incubated at room temperature for 60 min and the fluorescence intensity of each well was measured. (B) To a solution of 25 lM ADP in 7.5 ll of 50 mM Tris-HCl, pH 7.5, was added 7.5 ll of 2X ADP detection reagent (with or without 20 mM NEM) in a 384-well plate. The plate was incubated at room temperature and the fluorescence intensity of each well was measured with a PHERAstar (excitation at 540 nm and emission at 590 nm).
Fig.3. Effect of additives on enzyme-coupled fluorescence assay determination of ADP. To a solution of ADP (10 lM) in 5 ll of 50 mM Tris-HCl, 5 mM MgCl2, pH 7.5, supplemented with the indicated additives at the indicated concentrations was added 5 ll of 2X ADP detection reagent in a 384-well plate. The plate was incubated at room temperature for 60 min and the fluorescence intensity of each well was measured.
150
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
Table 1 HPLC analysis of phosphorylation of GDP and UDP by NDPK in the presence of ATP to produce GTP, UTP, and ADP. Substrates
NDPK (U/ml)
Conversion rate (%)
GDP + ATP
0 0.5 5 50 0 0.5 5 50 50
Contents lists available at ScienceDirect
Analytical Biochemistry journal homepage: www.elsevier.com/locate/yabio
Development of a highly sensitive, high-throughput assay for glycosyltransferases using enzyme-coupled fluorescence detection Kazuo Kumagai a,b,⇑, Hirotatsu Kojima a, Takayoshi Okabe a, Tetsuo Nagano a a b
Open Innovation Center for Drug Discovery, The University of Tokyo, Tokyo 113-0033, Japan Genomic Science Laboratories, Dainippon Sumitomo Pharma Co. Ltd., Osaka 554-0022, Japan
a r t i c l e
i n f o
Article history: Received 10 August 2013 Received in revised form 20 November 2013 Accepted 22 November 2013 Available online 1 December 2013 Keywords: Glycosyltransferase Nucleotide determination Enzyme-coupled assay Fluorescence assay High-throughput screening
a b s t r a c t Glycosyltransferases catalyze transfer of sugar moieties from activated donor molecules to specific acceptor molecules, forming glycosidic bonds. Identification of selective modulators of glycosyltransferases is important both to provide new tools for investigating pathophysiological roles of glycosylation reactions in cells and tissues, and as new leads in drug discovery. Here we describe a universal enzyme-coupled fluorescence assay for glycosyltransferases, based on quantification of nucleotides produced in the glycosyl transfer reaction. GDP, UDP, and CMP are phosphorylated with nucleotide kinase in the presence of excess ATP, generating ADP. Via coupled enzyme reactions involving ADP-hexokinase, glucose-6-phosphate dehydrogenase, and diaphorase, the ADP is utilized for conversion of resazurin to resorufin, which is determined by fluorescence measurement. The method was validated by comparison with an HPLC method, and employed to screen the LOPAC1280 library for inhibitors in a 384-well plate format. The assay performed well, with a Z0 -factor of 0.80. We identified 12 hits for human galactosyltransferase B4GALT1 after elimination of false positives that inhibited the enzyme-coupled assay system. The assay components are all commercially available and the reagent cost is only 2 to 10 US cents per well. This method is suitable for low-cost, high-throughput assay of various glycosyltransferases and screening of glycosyltransferase modulators. Ó 2013 Elsevier Inc. All rights reserved.
Glycosyltransferases are ubiquitous enzymes that are involved in the biosynthesis of disaccharides, oligosaccharides, and polysaccharides. They catalyze the transfer of a sugar moiety from donor molecules to glycosyl acceptor molecules, which range from proteins, lipids, sugars, and nucleic acids to small molecules [1]. Defects in glycosyltransferase activities are associated with various human diseases, including cancer, inflammation, neurological disorders, and congenital disorders of glycosylation [2,3]. Glycosyltransferases also influence the distribution and clearance of various biological molecules by regulating their glycosylation [4,5]. Thus, smallmolecular modulators of glycosyltransferases are expected to be useful research tools and may also serve as leads for drug discovery. However, few potent and selective inhibitors of glycosyltransferases are known [6], largely due to the lack of suitable high-throughput screening (HTS)1 assay methods to screen large chemical libraries. ⇑ Corresponding author. Address: Open Innovation Center for Drug Discovery, The University of Tokyo, Pharmaceutical Science Main Building, Tokyo 113-0033, Japan. Fax: +81 3 5841 1959. E-mail address: [email protected] (K. Kumagai). 1 Abbreviations used: HTS, high-throughput screening; B4GALT1, b-1,4-galactosyltransferase 1; FUT7, fucosyltransferase 7; ST6GAL1, b-galactoside a-2,6-sialyltransferase 1; NDP, nucleoside 50 -diphosphate; NDPK, nucleoside diphosphate kinase; CMPK1, CMP kinase 1; CMP-NeuAc, CMP-sialic acid; G6PDH, glucose-6-phosphate dehydrogenase; DTT, dithiothreitol; NEM, N-ethylmaleimide; BSA, bovine serum albumin; DMSO, dimethyl sulfoxide. 0003-2697/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ab.2013.11.025
Currently, the most widely used assay methods for glycosyltransferases are based on the detection of glycosylated products by using chromatographic, radiochemical, spectrophotometric, or immunological techniques [7]. However, most of these methods are low throughput, expensive, and labor-intensive. Compared with these methods, an alternative approach based on determination of nucleoside di(mono)phosphate generated from the sugar donor substrate would be both simpler and generally applicable to all glycosyltransferases. Most glycosyltransferases in mammals utilize uridine diphosphate (UDP)-sugars (e.g., galactosyltransferases), guanosine diphosphate (GDP)-sugars (e.g., fucosyltransferases), or cytidine monophosphate (CMP)-sugars (e.g., sialyltransferases) as sugar donor substrates [8]. Hence, the nucleotides UDP, GDP, and CMP liberated from these glycosyl donors are common products of glycosyltransferase reactions. Several methods other than chromatographic methods have been developed to detect these nucleotides for glycosyltransferase activity assays. Fluorescence polarization-based assays have been developed for the quantification of UDP, GDP, and CMP, and are available as commercial assay kits, but they are expensive for use in HTS [7]. Wu et al. determined inorganic phosphate released from these nucleotides by nucleotide phosphatase reactions by means of colorimetry with a malachite-based reagent [9]. This method is less costly and applicable to a wide range of glycosyltransferases, but
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
is not sensitive. Zhang et al. reported an enzyme-coupled fluorescence assay for N-acetylglucosaminyltransferase activity in which UDP was determined fluorometrically after coupling with pyruvate kinase reaction: detection was done by a commercial pyruvate assay kit [10]. UDP, GDP, and CMP can also be determined spectrophotometrically or fluorometrically by enzyme-coupled NADH detection [11–13], but the fluorescence intensity of NADH is relatively low. Here we report a highly sensitive, easy-to-operate, universal assay based on enzyme-coupled fluorescence detection of UDP, GDP, and CMP. This method is applicable in principle to any glycosyltransferase that utilizes UDP-, GDP-, or CMP-sugars as substrates. It employs nucleotide kinase reaction to generate ADP from the three nucleotides, followed by a cascade of coupled enzyme reactions that utilize the ADP for conversion of resazurin to highly fluorescent resorufin, which is determined by fluorescence measurement. The optimized method requires a volume of only a few microliters per well for glycosyltransferase assay in a 384-well plate format, and provides a high signal-to-background ratio. The method was validated by comparison with HPLC assay, and was applied to screen the LOPAC1280 library for inhibitors of human galactosyltransferase B4GALT1. We believe that this assay system provides a simple, cost-effective tool that will be applicable to many glycosyltransferases.
Materials and methods Materials Recombinant human b-1,4-galactosyltransferase 1 (B4GALT1), fucosyltransferase 7 (FUT7), and b-galactoside a-2,6-sialyltransferase 1 (ST6GAL1) were purchased from R&D Systems (Minneapolis, MN, USA). Nucleoside diphosphate kinase (NDPK, 1.9 kU/mg protein) from baker’s yeast, fetuin from fetal bovine serum, N-acetyl-D-lactosamine, and the LOPAC1280 compound library were obtained from Sigma–Aldrich (St. Louis, MO, USA). Recombinant human cytidine monophosphate kinase 1 (CMPK1) was from Prospec (Ness Ziona, Israel). ADP-hexokinase (52.6 U/mg solid) from Thermococcus litoralis was from Asahi Kasei Pharma (Tokyo, Japan). Diaphorase-I (1.8 kU/mg protein) from Bacillus stearothermophilus was from Unitika (Osaka, Japan). Recombinant glucose-6-phosphate dehydrogenase (G6PDH, 754 U/mg protein) from Leuconostoc sp. and NADP+ were from Oriental Yeast (Tokyo, Japan). GDP-fucose (GDP-Fuc), UDP-galactose (UDP-Gal), and CMP-sialic acid (CMP-NeuAc) were from Yamasa (Choshi, Japan), MP Biomedicals (Solon, OH, USA), and Calbiochem (San Diego, CA, USA), respectively. ATP, GDP, UDP, dithiothreitol (DTT), N-acetylglucosamine (GlcNAc), N-ethylmaleimide (NEM), bovine serum albumin (BSA), resazurin, gallic acid, and other chemicals were from Wako Pure Chemical Industries (Osaka, Japan). Nonbinding, small-volume 384-well black microtiter plates were from Greiner Bio-One (Frickenhausen, Germany).
Determination of ADP by enzyme-coupled fluorescence assay ADP in 7.5 ll of 50 mM Tris-HCl (pH 7.5) was mixed with 7.5 ll aliquots of 2X ADP detection reagent (2 mM glucose, 200 lM NADP+, 100 lM resazurin, 2 U/ml ADP-hexokinase, 2 U/ml G6PDH, and 2 U/ml diaphorase I in buffer A (100 mM Tris-HCl, 10 mM MgCl2, pH 7.5)) in 384-well plates and incubated at room temperature for 60 min. Fluorescence intensity due to the formation of resorufin was measured with a microplate reader, PHERAstar (BMG Labtech, Offenburg, Germany), with excitation at 540 nm and emission at 590 nm.
147
Enzyme-coupled fluorescence assay of GDP-fucosyltransferase and UDP-galactosyltransferase The reactions of fucosyltransferase FUT7 and galactosyltransferase B4GALT1 were performed in 5 ll of buffer B (50 mM Tris-HCl, 5 mM MgCl2, 5 mM MnCl2, pH 7.5) in 384-well plates at 37 °C for 60 min. FUT7 was reacted with 3 mg/ml fetuin and 50 lM GDP-Fuc. B4GALT1 was reacted with 5 mM GlcNAc and 50 lM UDP-Gal. To detect GDP and UDP produced by these reactions, 5 ll aliquots of 200 lM ATP and 2 U/ml NDPK in buffer A were added and the mixtures were incubated at room temperature for 60 min. Then 10 ll aliquots of 2X ADP detection reagent were added and incubation was continued at room temperature for 60 min. The fluorescence intensity (ex 540/em 590) was measured as described above.
Enzyme-coupled fluorescence assay of CMP-sialyltransferase Sialyltransferase ST6GAL1 reaction was carried out by adding the enzyme to 5 ll of buffer C (25 mM Tris-HCl, 5 mM MgCl2, 5 mM MnCl2, 150 mM NaCl, 0.1% (w/v) Triton X-100, pH 7.5) supplemented with 250 lM N-acetyl-D-lactosamine and 50 lM CMP-NeuAc in 384-well plates at 37 °C for 90 min. To detect CMP produced by the reaction, 5 ll aliquots of 200 lM ATP, 3 lg/ ml CMPK1, and 4 mM DTT in buffer D (175 mM Tris-HCl, 20 mM MgCl2, 150 mM KCl, 0.01% (w/v) BSA, 0.1% (w/v) Triton X-100, pH 9.0) were added. Incubation was continued at 37 °C for 60 min, and then 10 ll aliquots of 2X ADP detection reagent supplemented with 20 mM NEM were added. After further incubation at room temperature for 60 min, the fluorescence intensity (ex 540/em 590) was measured.
HPLC analysis HPLC analysis for nucleotides and nucleotide-sugars was performed using a YMC-Triart C18 column (150 4.6 mm, S-5 lm, YMC, Kyoto, Japan) with 50 mM KH2PO4–K2HPO4 (100:0 (v/v) for CDP, CMP, and CMP-NeuAc; 67:33 (v/v) for ADP, ATP, GDP, GTP, GDP-Fuc, UDP, UTP, and UDP-Gal) as an eluent at a flow rate of 1 ml/min with UV detection at 260 nm.
Assay validation Assay of inhibitory activity toward FUT7 was performed by both the enzyme-coupled fluorescence method and the HPLC method for validation of the new method. FUT7 (1 lg/ml) was incubated with 3 mg/ml fetuin and 50 lM GDP-Fuc in the presence or absence of the test compound gallic acid in 20 ll of buffer B at 37 °C for 60 min. The reaction was stopped by chilling the tubes on ice, and samples (10 ll) were analyzed by HPLC to determine liberated GDP. Remaining samples (5 ll) were transferred to wells of a 384-well plate. Aliquots of 5 ll of 200 lM ATP and 2 U/ml NDPK in buffer A supplemented with 0.01% (w/v) BSA were added and the plate was incubated at room temperature for 60 min. Then 10 ll of 2X ADP detection reagent supplemented with 0.01% (w/v) BSA was added to each well and incubation was continued at room temperature for 60 min, followed by fluorescence measurement. The percentage inhibition was calculated relative to controls treated with DMSO alone. In the control assay to determine the effect of the test compound on the enzyme-coupled assay reactions, 5 ll aliquots of 20 lM GDP in buffer B in the presence or absence of test compound were pipetted into wells of a 384-well plate instead of the FUT7 reaction mixtures, and enzyme-coupled fluorescence detection of GDP was performed as described above.
148
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
Library screening The LOPAC1280 library of 1280 pharmacologically active compounds (1 mM in DMSO) was dispensed into wells of columns 3 through 22 of 384-well plates (320 compounds/plate) at 50 nl/well and the same amount of DMSO was dispensed into wells of columns 2 and 23 using an Labcyte Echo 555 acoustic liquid handler (Sunnyvale, CA, USA). Aliquots of 2.5 ll of 10 mM GlcNAc and 100 lM UDP-Gal in buffer B supplemented with an additional 5 mM MnCl2 and 0.02% (w/v) Triton X-100 were added into columns 2–23 by a Multidrop Combi dispenser (Thermo Fisher Scientific, Vantaa, Finland), followed by addition of 2.5 ll of 0.05 lg/ml B4GALT1 in the same buffer (columns 3–22 for test compounds and column 23 for high signal control) or buffer only (column 2 for low signal control). After incubation at 37 °C for 60 min, 5 ll aliquots of 200 lM ATP and 2 U/ml NDPK in buffer A supplemented with 0.02% (w/v) BSA and 0.02% (w/v) Triton X-100 were added to all wells and the plates were incubated at room temperature for 60 min. Next, 10 ll of 2X ADP detection reagent supplemented with 0.02% (w/v) BSA and 0.02% (w/v) Triton X-100 was added, and incubation was continued at room temperature for 60 min, followed by fluorescence measurement. Two runs of the assay were conducted. The activity of compounds was calculated as the average of the two runs. The Z0 -factor was calculated by using the equation Z0 = 1 (3rh + 3rl)/|lh ll| [14], where rh and rl are the standard deviations of the high and low signal controls, respectively, and lh and ll are the mean signal intensities of the high and low signal controls, respectively. In the control assay, 5 ll of 20 lM UDP in buffer B was added to wells containing the same amount of LOPAC1280 library compounds or DMSO only, and UDP detection by enzyme-coupled fluorescence assay was performed in the same manner as described above. The activity
of hit compounds was validated by HPLC assay by determining UDP produced by B4GALT1 reaction.
Results Optimization of enzyme-coupled fluorescence assay for determination of ADP The present assay for glycosyltransferase activity depends on the measurement of GDP, UDP, and CMP liberated from glycosyl donors. These nucleotides are phosphorylated by nucleotide kinase in the presence of excess ATP with stoichiometric formation of ADP, and the amount of ADP is measured fluorometrically after a series of coupled enzyme reactions that serve to generate a fluorescent product (Fig. 1). We first optimized the enzyme-coupled fluorescence measurement of ADP. The ADP measurement assay consists of the following coupled reactions: ADP was first coupled to the conversion of glucose to glucose-6-phosphate by ADP-hexokinase; then glucose-6-phosphate was coupled to the conversion of NADP+ to NADPH by glucose-6-phosphate dehydrogenase (G6PDH); finally NADPH was coupled to the conversion of resazurin to a highly fluorescent end product, resorufin, by diaphorase I. Because all three enzymes work at neutral pH in the presence of Mg2+, we tested the detection of 15 lM ADP in buffer A (100 mM Tris-HCl, 10 mM MgCl2, pH 7.5) using the enzymes at various concentrations in the presence of 1 mM glucose, 100 lM NADP+, and 50 lM resazurin. Addition of the enzymes at final concentrations higher than 1 U/ml each did not result in any further increase in the fluorescence intensity (data not shown), suggesting that 1 U/ ml of each enzyme is sufficient for detection of 15 lM ADP. Therefore, we set the composition of ADP detection reagent as 1 U/ml
Fig.1. Schematic representation of enzyme-coupled fluorescence assay for determination of glycosyltransferase activity. Glycosyltransferase reactions liberate nucleotides UDP, GDP, and CMP from the respective glycosyl-donor substrates. These nucleotides are then phosphorylated by NDP kinase or CMP kinase in the presence of an excess amount of ATP to produce phosphorylated products and ADP. Via a series of coupled enzyme reactions involving ADP-hexokinase, glucose-6-phosphate dehydrogenase, and diaphorase in the presence of excess amounts of glucose, NADP+, and resazurin, the ADP is utilized for formation of highly fluorescent resorufin, which is determined by fluorescence measurement. Therefore, the fluorescence intensity reflects the amount of UDP, GDP, or CMP formed by glycosyltransferase reaction.
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
each of ADP-hexokinase, G6PDH, and diaphorase I, 1 mM glucose, 100 lM NADP+, and 50 lM resazurin (final concentrations). A typical standard curve and time course of ADP detection reaction are shown in Fig. 2. The standard curve was linear up to 32 lM ADP (Fig. 2A). The signal/background (S/B) ratio was 55.0 for 32 lM ADP. The reactions were completed in 10 min at room temperature and the fluorescence intensity remained unchanged for at least 2 h thereafter (Fig. 2B). This indicates that there is no need to stop the reaction for signal measurement. When the assay was carried out in the presence of N-ethylmaleimide (NEM) at a final concentration of 10 mM, almost the same results were obtained (Fig. 2A and B), indicating that the enzyme-coupled reactions are resistant to this thiol reagent. Additives such as NaCl, MgCl2, CaCl2, MnCl2, DMSO, Na3VO4, Triton X-100, EDTA, NaF (640 mM), and ATP did not affect the fluorescence intensity at the concentrations tested (Fig. 3). NaF decreased the signal by 31% at 50 mM.
Optimization of enzymatic phosphorylation of GDP and UDP to produce ADP GDP and UDP were phosphorylated in the presence of a 5-fold molar excess of ATP using nucleoside diphosphate kinase (NDPK) to produce ADP. In the presence of 0.5 U/ml of NDPK, 87 and 86%
A
149
of 40 lM GDP and UDP were converted to nucleoside triphosphates GTP and UTP, respectively, during reaction at room temperature for 60 min and an equivalent amount of ADP was produced as determined by HPLC (Table 1). There was no increase in conversion rate at higher concentrations of NDPK for either GDP or UDP, indicating that 0.5 U/ml of NDPK is sufficient for quantitative conversion of the nucleotides. In a control assay without GDP or UDP, no increase in ADP concentration was observed even at 50 U/ml NDPK. This demonstrates that the NDPK used has no ATPase activity.
Optimization of enzymatic phosphorylation of CMP to produce ADP CMP was phosphorylated in the presence of a 5-fold molar excess of ATP by cytidine monophosphate kinase CMPK1 to produce ADP. Because CMPK1 is a DTT-dependent enzyme (Fig. 4), the phosphorylation reaction was performed in the presence of 2 mM DTT. About 58% of 40 lM CMP was converted to CDP during reaction at 37 °C for 60 min in the presence of 1 lg/ml CMPK1, as determined by HPLC (Table 2). At 2 lg/ml CMPK1, the increase of conversion rate was only about 5% over that at 1 lg/ml CMPK1, suggesting that a conversion rate of around 60% is the maximum under this assay condition. Thus, a CMPK1 concentration of
B
Fig.2. Typical standard curves (A) and time courses (B) of enzyme-coupled fluorescence assay determination of ADP in the presence (N) or absence (d) of N-ethylmaleimide (NEM). (A) To a solution of the indicated concentrations of ADP in 7.5 ll of 50 mM Tris-HCl, pH 7.5, was added 7.5 ll of 2X ADP detection reagent (2 mM glucose, 200 lM NADP+, 100 lM resazurin, 2 U/ml ADP-hexokinase, 2 U/ml G6PDH, and 2 U/ml diaphorase I in buffer A with or without 20 mM NEM) in a 384-well plate. The plate was incubated at room temperature for 60 min and the fluorescence intensity of each well was measured. (B) To a solution of 25 lM ADP in 7.5 ll of 50 mM Tris-HCl, pH 7.5, was added 7.5 ll of 2X ADP detection reagent (with or without 20 mM NEM) in a 384-well plate. The plate was incubated at room temperature and the fluorescence intensity of each well was measured with a PHERAstar (excitation at 540 nm and emission at 590 nm).
Fig.3. Effect of additives on enzyme-coupled fluorescence assay determination of ADP. To a solution of ADP (10 lM) in 5 ll of 50 mM Tris-HCl, 5 mM MgCl2, pH 7.5, supplemented with the indicated additives at the indicated concentrations was added 5 ll of 2X ADP detection reagent in a 384-well plate. The plate was incubated at room temperature for 60 min and the fluorescence intensity of each well was measured.
150
Fluorescence assay for glycosyltransferases / K. Kumagai et al. / Anal. Biochem. 447 (2014) 146–155
Table 1 HPLC analysis of phosphorylation of GDP and UDP by NDPK in the presence of ATP to produce GTP, UTP, and ADP. Substrates
NDPK (U/ml)
Conversion rate (%)
GDP + ATP
0 0.5 5 50 0 0.5 5 50 50
