Development and characterization of the assay

Considering that the sensitivity in phospho-ELISA methods is mainly limited by the number of substrate binding sites available for immobilization, we reasoned that this problem could be overcome by using the phosphospecific antibody for capturing the small amounts of phosphopeptide from the complex reaction mixture. We decided to use a substrate peptide mimicking the sequence around Thr212 in the tau protein (also called microtubule-associated protein tau, MAPT). This is a well characterized phosphorylation site of DYRK1A [23, 24], and phosphospecific antibodies directed against this site are commercially available. We used a biotinylated substrate peptide to allow for colorimetric detection of the bound phosphopeptide with the help a streptavidin-HRP conjugate (Figure 1A). Concentrations of the capture antibody and the streptavidin-HRP conjugate were optimized to develop the standard protocol used in all experiments shown here (see Methods section).

We performed a titration experiment to determine the key parameters of the assay, i.e. the power to discriminate between the phosphorylated and unphosphorylated biotinylated tau peptide (hereafter referred to as tau_207-219), the absolute lower detection limit, and the linear range of the assay. As shown in Figure 1B, the detection limit was about 5-10 fmol phosphopeptide per well. A comparable signal was obtained with 320 pmol of the unphosphorylated peptide, indicating that this combination of phosphospecific antibody and substrate peptide offers excellent selectivity for the detection of the phosphorylated substrate (~ 10⁵-fold discrimination). Saturation of the assay was reached at about 1 pmol of phosphopeptide, but the linear plot (Figure 1C) illustrates that the useful measuring range was between 0.01 pmol and 0.1 pmol. Thus, the sensitivity of the ELISA compares well with radiometric assays, where about 0.1 pmol of phosphopeptide can be routinely detected. For comparison, we analysed the same concentrations of phosphorylated and unphosphorylated tau_207-219 in the inverse configuration, in which the phosphospecific antibody was used to detect the biotinylated peptide after binding to streptavidin-coated plates (Figure 1D). This assay was much less sensitive (detection limit of 1 pmol phosphorylated peptide) and suffered from crossreaction of the antibody with the unphosphorylated peptide at concentrations greater than ~ 40 pmol per well.

Next we tested whether small amounts of the phosphopeptide can be detected in mixtures with a large excess of the unphosphorylated peptide. Although the experiment shown in Figure 1B suggested that 50-100 pmol of unphosphorylated peptide should only marginally contribute to the total signal, these concentrations reduced signal intensities obtained with 10-1000 fmol phosphopeptide (Figure 1E). We decided to apply a maximum of 50 pmol total peptide per well in subsequent experiments, and to keep this amount constant in all samples of an assay (including the standards).

Measurement of DYRK1A kinase activity

We performed in vitro-kinase reactions with varying concentrations of recombinant GST-DYRK1A-ΔC to determine the minimal detectable amount of kinase activity (Figure 2). The titration revealed a useable linear measuring range between 10 and 100 μU kinase activity, corresponding to 20-200 pg of the recombinant kinase. This results is in the range that could be predicted from the detection limit of the ELISA (> 10 fmol phosphopeptide), because 10 μU of kinase should phosphorylate 300 fmol substrate within 30 min, given that a sample of 1/10 of the reaction mix was loaded per well and that the excess of unphosphorylated peptide does not severely affect detection of the phosphopeptide (Figure 1E). A similar sensitivity was achieved for GST-DYRK2, consistent with the previous finding that both DYRK1A and DYRK2 can phosphorylate Thr212 in the tau protein [23].

Activity measurements of a DYRK1-related kinase in Xenopus laevis oocytes

The mRNA originally detected in Xenopus oocytes by Qu et al. [14] corresponds to xDYRK1B (see discussion). To immunoprecipitate this kinase, we took advantage of a polyclonal antiserum we had previously raised against a peptide with the sequence of the 15 C-terminal amino acids of mouse DYRK1B [26], of which 9 are identical with xDYRK1B (Figure 3A). To test whether this antiserum also recognized xDYRK1B, we overexpressed xDYRK1B in mammalian cells. As shown in Figure 3B, the antiserum was indeed suitable for immunoprecipitation and subsequent immunodetection of recombinant xDYRK1B by Western blot analysis. No band was detected in a control immunoprecipitation with serum taken from the same rabbit before immunization. The immunopurified protein contained immunodetectable phosphotyrosine, a common feature of the members of the DYRK family. Taking advantage of the fact that mammalian and Xenopus DYRK1A have identical sequences in the region corresponding to the immunogenic peptide (Figure 3A), we overexpressed GFP-DYRK1A in Hela cells to assess whether the DYRK1B antiserum might crossreact with xDYRK1A. Importantly, anti DYRK1B did not immunoprecipate GFP-DYRK1A (Figure 3C), implying that it specifically immunoprecipitates xDYRK1B and not xDYRK1A from the oocytes.

We used the DYRK1B antibody for immunocomplex kinase assays to determine whether unstimulated Xenopus oocytes contained active xDYRK1B. First we determined the phosphorylation of tau_207-219 by the immunoprecipitates over time (Figure 4). Kinase activity was well detectable in the anti DYRK1B immunoprecipitates, with little background in the preimmune serum control. We decided to use an incubation time of 30 min for the following assays because the turnover was flattening after this time point. The detection of xDYRK1B activity was reproduced in three independent assays with different batches of oocytes and was also confirmed in a radiometric assay with a different substrate peptide (DYRKtide) (Table 3). In spite of this unequivocal detection of active xDYRK1B in Xenopus oocytes, we failed to detect the immunopreciptated xDYRK1B on Western blots (data not shown), apparently because the immunodetection of xDYRK1B on Western blots was less sensitive than the assay of its catalytic activity.

Peptides and antibodies

A biotinylated substrate peptide mimicking the sequence around Thr212 of the tau protein (Biotin-Ttds-GSRSRT PSLPTPP-NH₂, Thr212 is underlined, where Ttds is a proprietary linker) was selected from the Single Kinase Substrate Peptides commercially available from JPT Peptide Technologies (Berlin, Germany). For use in standard curves, a phosphorylated version was custom synthesized by Eurogentec (Seraing, Belgium). The optimized substrate peptide DYRKtide and the rabbit antibody directed against the C-terminus of human DYRK1B have been described previously [26, 34]. The following antibodies were obtained from commercial sources: rabbit anti TAU(pT212) antibody (Cat. No. 44-740G; Invitrogen, Carlsbad, CA), mouse monoclonal antibodies for phosphotyrosine (PY99, Santa Cruz Biotechnology, Santa Cruz, CA), FLAG epitope (FLAG-M2, Sigma, Munich, Germany) and DYRK1A (clone 7D10; Abnova, Taipei City, Taiwan), and goat GFP-specific antibody (Rockland, Gilbertsville, PA).

Preparation of mouse heart and Xenopus oocyte lysates

Hearts from TW-104 mice were weighed and 2 ml of lysis buffer (50 mM Tris-HCl pH7.5, 150 mM NaCl, 0.5% NP-40, 15% glycerol, 1 mM EDTA, 1 mM NaF, 1 mM Na₃VO₄, 10 μg/ml aprotinin, 10 μg/mL pepstatin, 10 μg/mL leupeptin, 1 μL PMSF) were added per 400 mg of tissue. The hearts were cut into small pieces, subjected to 10 strokes of medium rpm (stage 6 out of 10) in a Potter Elvehjem homogenizer, and the homogenates were then sonified for 20 s. After centrifugation (21,000 g, 4°C, 10 minutes), the supernatant was taken and stored at -80°C. The lysates were cleared by pre-incubation for 30 minutes with 30 μL protein G agarose (EZview™ Red Protein G Affinity Gel, Sigma, Munich, Germany) and subsequent centrifugation before the supernatant was taken for immunoprecipitations. Follicle cell-free oocytes of oogenesis stages V or VI from Xenopus laevis [35] were kindly provided by the members of the Schmalzing group in the Institute of Pharmacology. The oocytes were lysed in 4 volumes of lysis buffer by up-and-down pipetting, and the lysates were cleared by repeated centrifugation to remove lipids and insoluble debris.

Immunoprecipitation

For the immunoprecipitation from heart lysates, 1 μg monoclonal anti DYRK1A or 1 μg anti-FLAG antibody were added to a volume of lysate containing 1 mg of total protein. For the oocyte experiments, the amount of lysate derived from a volume of 100 μl settled oocytes was used for immunoprecipitation with either 1 μg anti DYRK1A antibody, 1 μl anti DYRK1B serum [26] or 1 μl non-reacting serum (preimmune serum of the same rabbit used to raise the DYRK1B-specific serum).

Immuncomplexes were captured by over-night incubation with protein G agarose at 4°C in an end-over-end rotator. After washing the agarose beads twice with TBS (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM EDTA) supplemented with 0.1% Igepal CA-630, twice with TBS without Igepal and once with kinase buffer, bound immunocomplexes were assayed for in vitro-kinase activity (see below). In some experiments, the affinity matrix was divided equally between two test tubes during the last washing step to allow for the direct comparison of radioactive and non-radioactive assay.

Cell culture and transfection

HeLa cells were cultured in Quantum 101 Medium for HeLa cells (PAA Laboratories, Pasching, Austria) and COS7 cells in DMEM medium supplemented with 10% fetal calf serum (PAA) at 37°C in a humidified 5% CO₂ atmosphere. Transient transfections were performed using the FuGENE HD transfection reagent (Roche Molecular Biochemicals, Mannheim, Germany). Expression clones for GFP-DYRK1A and GFP-DYRK1B (p69 splicing variant) have been described previously [26, 36]. An expression clone for untagged full length xDYRK1B (IMAGE clone 4724858) was obtained from Imagenes (Berlin, Germany).

Western blotting

After washing the immunoprecipitates with TBS, bound proteins were eluted with Laemmli sample buffer, separated by SDS-PAGE (7%) and blotted onto nitrocellulose membranes. Immunodetection with commercial antibodies was performed according to the supplier's instructions. The DYRK1B antiserum was used at a dilution of 1:5000 for 2 h at room temperature, and the TBS for the washes following incubation with the primary antibody was supplemented with 0.1% SDS to reduce non-specific binding [26]. Western blots were developed using horseradish peroxidase-coupled secondary antibodies and chemiluminescence detection. The Clean-Blot IP Detection Reagent (Thermo Scientific, Rockford, IL) was used as a secondary antibody when proteins were immunodetected with the same antibody used for immunoprecipitation.

Kinase assays

Recombinant glutathione S-transferase (GST) fusion proteins of rat DYRK1A-ΔC and human DYRK2 were prepared as described previously [34, 36]. One unit of kinase activity was defined as the amount of enzyme that catalyzed the phosphorylation of 1 nmol DYRKtide per min at 30°C in kinase buffer (25 mM Hepes, pH 7.0, 5 mM MgCl₂, 0.5 mM dithiothreitol).

To determine the assay sensitivity (Figure 2), reactions were performed in a total volume of 10 μl with variable amounts of the recombinant kinase, 50 μM tau_207-219 and 100 μM ATP at 30°C for 30 min. Reactions were stopped by addition of 10 μl 10 mM EDTA and stored at -20°C until analysis by ELISA.

To prepare phosphorylated tau_207-219 for the use in the ELISA (Figure 1), we aimed to achieve maximal phosphorylation by incubating 2.5 nmol tau_207-219 with an excess of active kinase (4 mU/μL GST-DYRK1A-ΔC) in the presence of 0.82 mM ATP. The reaction mixture was incubated for 2 hours, and the same amount of kinase was supplemented after one hour. Total incorporation of phosphate was determined in a parallel reaction that was incubated in the presence of [γ-³³P]ATP under identical conditions. All other kinase assays were incubated for 30 min at 30°C.

Immunocomplex kinase assays were carried out in a total reaction volume of 50 μl in kinase buffer, 100 μM ATP and 50 μM tau_207-219 at 30°C for 30 min and stopped by addition of 50 μL 10 mM EDTA. Samples were stored at -20°C until evaluation by ELISA. Radiometric assays were run in the presence of [γ-³³P]ATP (~ 2000 cpm/pmol). Incorporation of ³³P into the biotinylated tau peptide was determined by spotting aliquots of the reaction mix onto a streptavidin membrane (SAM² Biotin Capture Membrane, Promega, Mannheim, Germany), which was then washed as described by the manufacturer and evaluated by scintillation counting. Radiometric immunocomplex kinase assays with the substrate peptide DYRKtide (at 100 μM) were run with 10 μM ATP as described before [37].

Direct Sandwich ELISA

MaxiSorp F96 MicroWell™ Plates (Nunc, Langenselbold, Germany) were coated with 100 ng/well TAU(pT212) antibody at 4°C overnight. To keep the antibody consumption low, we used 100 ng/well of capturing antibody for coating in the standard protocol, although 200 ng/well gave a slightly higher readout (1.2-fold). After blocking for 1 hour with 200 μl/well 2% (w/v) bovine serum albumin (fraction V, modified Cohn, Calbiochem) in phosphate buffered saline (PBS) (140 mM NaCl, 3 mM KCl, 8 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 7.4), the antigen was added in a volume of 100 μL/well and the plates were incubated for 2 h at room temperature. Bound antigen was detected with streptavidin-HRP conjugate (Pierce, Rockford, IL; 100 μL/well, 2 h at room temperature; 25 ng/ml) and subsequent incubation with 100 μL/well TMB (3,3',5,5'-tetramethylbenzidine) substrate solution until colour development was observed. Other detection reagents (HRP-linked anti-biotin antibody from CST; strepavidin-HRP from R&D systems) were less sensitive, and higher concentration of the streptavidin-HRP conjugate resulted in higher background in empty wells. The reaction was then stopped by adding 50 μl/well 1 M HCl, and plates were read at 450 nm (reference wavelength 620 nm). After each incubation step, plates were washed 4 times with PBS, pH 7.4, 0.1% Tween-20. Antibody and antigen were diluted in PBS.

ELISA format based on streptavidin-coated plates

Reacti-Bind™ Streptavidin High Binding Capacity Coated 96-Well Plates (Pierce; Rockford, IL, USA) were loaded with the indicated concentrations of the biotinylated peptides (100 μL/well) and incubated at RT for 2 h. After washing, bound phosphopeptide was detected with rabbit anti TAU(pT212) antibody (1:16.000) and HRP-linked goat anti rabbit IgG (Fc) (Pierce) (1:40.000) and TMB substrate as described above. PBS with 2% BSA and 0.1% Tween-20 was used for plate washing and for dilution of antigen and antibodies.

Data evaluation and curve fitting

Standard curves of phosphorylated tau peptide and the graphical representation of the time course shown in Figure 4 were generated by non-linear curve fitting using GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA).