AlphaScreen ® -Based Characterization of the Bifunctional Kinase/RNase IRE1α

Abstract

Assay technologies that were originally developed for high-throughput screening (HTS) have recently proven useful in drug discovery for activities located upstream (target identification and validation) and downstream (ADMET) of HTS. Here the authors investigated and characterized the biological properties of a novel target, IRE1α, a bifunctional kinase/RNase stress sensor of the endoplasmic reticulum (ER). They have developed a novel assay platform using the HTS technology AlphaScreen^® to monitor the dimerization/oligomerization and phosphorylation properties of the cytosolic domain of IRE1α. They show in vitro that dimerization/oligomerization of the cytosolic domain of IRE1 correlated with the autophosphorylation ability of this domain and its endoribonuclease activity toward XBP1 mRNA. Using orthogonal in vitro and cell-based approaches, the authors show that the results obtained using AlphaScreen^® were biologically relevant. Preliminary characterization of assay robustness indicates that both AlphaScreen^® assays should be useful in HTS for the identification of IRE1 activity modulators.

Introduction

Large-scale technologies used within the life science industry for drug discovery have recently proven useful in academic research laboratories, facilitating chemical genomics and the investigation of specific biological functions. These technologies, including homogeneous luminescence (e.g., AlphaScreen^® and Aequorin), fluorescence intensity (FI) or polarization (FP), time-resolved fluorescence (TRF), laser scanning cytometry, or microscopy-based approaches, have become enabling tools to characterize complex functions of biological targets.

In the present study, we used AlphaScreen^® to investigate cell signaling pathways ¹ and in particular those emanating from the endoplasmic reticulum (ER) under stress conditions. The ER ensures proper protein folding and export to later compartments of the secretory pathway. This is achieved through complex machineries, including protein synthesis, translocation/folding, quality control, ER-associated degradation (ERAD), and export. ² In addition to these functional attributes, the ER has evolved a highly conserved adaptive signaling pathway, referred to as the unfolded protein response (UPR), whose activation occurs upon accumulation of improperly folded proteins in the ER. ³

UPR signaling is mediated by 3 ER resident transmembrane proteins—the PKR-like ER kinase (PERK), the activating transcription factor 6 (ATF6), and the inositol requiring enzyme 1 alpha (IRE1α). ^3,4 Our work has focused on IRE1α, which is a transmembrane sensor of ER stress. The luminal domain of IRE1α contains binding sites for the chaperone BiP, whereas the cytosolic region has 2 main catalytic elements: a serine/threonine kinase and an endoribonuclease domain. ^4-6 Under basal conditions, IRE1α is thought to exist as a monomer, and in response to the accumulation of misfolded proteins in the ER, IRE1 proteins oligomerize, resulting in its trans-autophosphorylation and triggering of its endoribonuclease activity. ⁴ However, structural studies in Saccharomyces cerevisiae showed that the luminal domain of IRE1 exists as a dimer/oligomer and suggested that it could potentially bind directly to unfolded peptides similar to MHC class I molecules. ⁷ This was confirmed by structural studies on the cytosolic domain of IRE1 in S. cerevisiae, which revealed a dimer conformation that is subjected to conformational changes upon ER stress, leading to activation of its kinase and RNAse activities. ⁸ Recently, the cytosolic domain of IRE1 was crystallized in an oligomeric form, which was promoted by the presence of the kinase inhibitor Sunitinib. ^9,10 This highly ordered molecular structure was found to be physiologically relevant in S. cerevisiae. ¹¹ Therefore, a revised model for IRE1 activation process has been proposed in which different IRE1 dimers promote the trans-autophosphorylation that is observed in vivo in oligomeric structures. This would rearrange the RNAse domain and create a binding surface for the mRNA substrates. ⁹ However, this model does not take into account the capacity of IRE1 to induce JNK ¹² and ERK signaling cascades ¹³ and does not provide any insights into the mechanisms by which these kinases are activated downstream of IRE1α. ⁶ Regardless, IRE1α endoribonuclease activity then initiates an unconventional splicing of XBP1 mRNA that introduces a translational frame shift leading to the synthesis of a potent transcription factor. ¹⁴ Recently, IRE1 endoribonuclease activity has been shown to be partly independent of IRE1 phosphorylation, ¹⁵ suggesting the existence of alternative activation modes. ⁶

Thus far, the only in vitro or in vivo assays that have been used for identifying IRE1 modulators and reporting on IRE1 biological functions have mainly focused on the endoribonuclease activity of IRE1 toward XBP1 or its functional yeast homolog HAC1. ^16-19 In the present study, we have used AlphaScreen^® to investigate and characterize IRE1α signaling properties. We have developed 2 AlphaScreen^®-based assays for IRE1α oligomerization and phosphorylation and validated the results using orthogonal in vitro and cell-based experimental systems.

Materials and Methods

Cloning

IRE1^cyto cDNA (AA 470 to 977) was cloned from human liver cDNAs using either the Gateway^® technology (Invitrogen, Carlsbad, CA) in pGEX-2TK or pDEST17. IRE1^cyto cDNA devoid of ATG was amplified by PCR using the Platinium^® Taq DNA Polymerase High Fidelity (Invitrogen) and the following amplification scheme: denaturation 94 °C for at 40 s, annealing at 60 °C for 40 s, and elongation at 68 °C for 2 min, 35 cycles. The PCR products were precipitated using PEG8000 and recombined into pDONR201 using the Gateway^® BP clonase (Invitrogen). The plasmids were then transformed into competent DH5α cells and positive clones selected and sequenced. The clones were then recombined into destination vectors using LR clonase (Invitrogen).

Recombinant protein expression, purification, and quality control

The above-described plasmids were transformed into competent DH5α cells. Five individual colonies were selected and pooled, and plasmid DNA was amplified and subsequently transformed into competent BL21 bacterial cells. Recombinant protein expression in BL21 cells was induced using 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 3 h. Bacteria were then collected by centrifugation and lysed and recombinant proteins purified as described previously. ²⁰ The resulting purified proteins were then concentrated and dialyzed using Amicon ultra centrifugal filters (cut-off = 20,000 Da; Millipore, Billerica, MA). The recombinant purified protein was resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and either directly visualized following Coomassie Blue staining ( Fig. 1A ), or subjected to immunoblot using anti-IRE1 antibodies ( Fig. 1B ).

Thrombin cleavage

Glutathione-S-transferase (GST)–IRE1^cyto coupled to glutathione (GSH)–sepharose beads was incubated in the presence of thrombin (50 UI) for 16 h at room temperature. The supernatant was collected following centrifugation at room temperature for 10 min in a bench-top microfuge at maximal speed. Thrombin was removed following incubation with benzamidine-sepharose beads for 45 min at 4 °C. The resulting cleaved IRE1^cyto was then concentrated and dialyzed using Amicon ultra centrifugal filters (cut-off = 10,000 Da; Millipore). The purified products were then suspended in 6× Laemmli sample buffer, ²¹ resolved by SDS-PAGE, and either directly stained using Coomassie Brilliant Blue R250 (CBB) or immunoblotted using anti-GST (Hyperomics Farma, Pierrefonds, Quebec, Canada), anti-IRE1α (Santa Cruz Biotechnology, Santa Cruz, CA), or anti-phospho-IRE1 (Abcam, Inc., Cambridge, MA) antibodies. S. cerevisiae IRE1^cyto was cloned from genomic DNA and subcloned in pET15b by restriction cloning. The corresponding recombinant protein was expressed in BL21 cells and purified on nickel-coated beads NiNTA (QIAGEN, Inc., Valencia, CA) according to the manufacturer’s recommendations. Where designated, the 6xHis tag was removed using TEV protease cleavage and dialyzed/concentrated using Centricon concentrators (Millipore).

Fig. 1.

IRE1^cyto AlphaScreen^®-based dimerization/oligomerization assay. (A) Expression of GST-IRE1^cyto in BL21 bacterial cells was induced by 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 3 h. Bacterial recombinant proteins were then purified using appropriate chromatography matrices, eluted, dialyzed, concentrated, and then resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and stained with Coomassie Blue. (B) Identical experiment as (A) with the exception that following SDS-PAGE, proteins were transferred onto nitrocellulose and immunoblotted using anti-IRE1 antibodies (Santa Cruz Biotechnology). (C) AlphaScreen^®-based dimerization/oligomerization assay. An increased signal is observed in the presence of adenosine triphosphate (ATP). The assay was performed with glutathione (GSH)–coated donor beads and TAR GSH acceptor beads (GSH::GSH) (n > 4). (D) Similar experiments as those for (C) were performed using identical amounts of GST-IRE1^cyto previously subjected to dephosphorylation using alkaline phosphatase (n = 3). (E) Dimerization/oligomerization properties of GST-IRE1K599A^cyto analyzed using AlphaScreen^®. Comparison between wild-type GST-IRE1^cyto (red) and the kinase inactive GST-IRE1K599A^cyto (blue) is shown (n = 3). The experiment was carried out using GSH donor and Europium GSH acceptor beads.

AlphaScreen^®-Based Assays

AlphaScreen^® assays were performed in Costar 384-well microplates in a 25-µL final reaction volume. GSH donor beads or anti-GST acceptor beads (PerkinElmer, Inc., Waltham, MA) were used at a final concentration of 0.02 mg/mL per well. The assays were performed in 20 mM Tris-HCl (pH 7.4), 50 mM NaCl, 1 mM MgCl₂, 1 mM MnCl₂, 2 mM dithiothreitol (DTT), and 0.1% Tween-20. All incubations were performed at 23 °C. Laser excitations were carried out at 680 nm, and readings were performed at 520 to 620 nm using the EnVision^® (PerkinElmer) plate reader.

Dimer/oligomer formation assay

Increasing concentrations of GST-IRE1^cyto, dephosphorylated GST-IRE1^cyto, or GST were incubated for 45 min with 1 mM adenosine triphosphate (ATP; Sigma-Aldrich, St. Louis, MO) or 1 mM ATPγS (Sigma-Aldrich) or buffer and donor beads. Acceptor beads were then added to the reaction and incubated for an additional hour before reading.

Nucleotide specificity assay

Optimal concentrations of GST-IRE1^cyto or dephosphorylated GST-IRE1^cyto (determined in preliminary experiments) were incubated for 45 min with varying concentrations of ATP, adenosine diphosphate (ADP), adenosine monophosphate (AMP), guanosine triphosphate (GTP), guanosine diphosphate (GDP; 0-10 mM; Sigma-Aldrich), and donor beads. Acceptor beads were then added to the reaction and incubated for an additional hour before reading.

Oligomer formation assay

Optimal concentrations of GST-IRE1^cyto or dephosphorylated GST-IRE1^cyto were incubated for 45 min with increasing concentrations of 6xHis-hIRE1^cyto, bovine serum albumin (BSA) or 6xHis-ScIRE1^cyto in the presence of 1 mM ATP and donor beads. Acceptor beads were then added to the reaction and incubated for an additional hour before reading.

Phosphorylation assay

Increasing concentrations of GST-IRE1^cyto were incubated for 45 min in the presence or absence of ATP. GSH donor beads and Lewis Metal–Fe acceptor beads (PerkinElmer) were added and incubated for an additional hour prior to reading. The assays were performed using thioxene, anthracene, and rubrene containing acceptor beads in all the experiments except for the experiment carried out in Figure 1E , where AlphaLISA acceptor beads were used.

RNA cleavage assay

Total RNA (10 µg) from HepG2 cells was incubated with GST-IRE1^cyto (10 µg) at 37 °C for 0 to 10 h in 50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 1 mM MgCl₂, 1 mM MnCl₂, and 5 mM β-mercaptoethanol, supplemented with 1 mM ATP or 1 mM ATPγS. Cleaved or uncleaved RNAs were used as a template for reverse transcription using XBP1 primers and GAPDH as an internal control.

In vitro phosphorylation of GST-IRE1^cyto

Two micrograms of GST-IRE1^cyto purified as described above were incubated for 30 min at 37 °C with 20 mM Tris-HCl (pH 7.5) in the presence of 5 to 20 mM MgCl₂, 0 to 5 mM MnCl₂, 2 µCi of [³²Pγ]-ATP, and 100 µM ATP. The reaction was then stopped by the addition of 5 mM ATP and the reaction products analyzed by SDS-PAGE followed by Coomassie Brilliant Blue R-250 staining and radioautography. Phosphorylation was also detected using anti-phospho-IRE1 (Abcam) antibodies.

Cell culture and transfection

HeLa or Chinese hamster ovary (CHO) cells were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum and antibiotic. HeLa and CHO cells were transiently transfected with the pED-IRE1 wild-type (WT) plasmid or pED-IRE1 K599A plasmid. ²² Transfections were performed using LipofectAMINE (Invitrogen) according to the manufacturer’s recommendations.

In-cell cross-linking

HeLa cells transfected with pCDNA3 containing wild-type human IRE1 cDNA were washed twice with phosphate-buffered saline (PBS). The cross-linker DSS solution (final concentration 1 mM) was added and cells incubated for 2 h on ice. The reaction was stopped by the addition of 20 mM Tris-HCl (pH 7.4). Cells were lyzed with 1% Triton X-100 in 20 mM Tris-HCl (pH 7.4) buffer in the presence of protease inhibitors. Proteins were resolved by SDS-PAGE followed by immunoblot analysis using anti-IRE1α antibodies (Santa Cruz Biotechnology).

IRE1^cyto in vitro cross-linking

First, 4 µM of purified IRE1^cyto or dephosphorylated IRE1^cyto was incubated in a final volume of 50 µL for 30 min at 4 °C. Then, the cross-linker DSS (Thermo Fisher Scientific, Inc., Waltham, MA), 1 mM final concentration, or DMSO as a control was added and the samples incubated for 2 h at 4 °C. The reaction was stopped with 20 mM Tris-HCl (pH 7.4). The samples were then suspended in 6× Laemmli buffer and the proteins resolved by SDS-PAGE. The gel was then either silver stained or subjected to immunoblot analysis using anti-IRE1α (Santa Cruz Biotechnology) or anti-phospho-IRE1 (Abcam) antibodies.

Results

Because IRE1 could be a relevant therapeutic target in many diseases, it was important to design biological activity assays that could be adaptable to high-throughput screening (HTS). To this end, we designed an in vitro assay for IRE1 using the AlphaScreen^® technology to monitor its dimerization/ oligomerization and phosphorylation properties. AlphaScreen^® is a technology in which donor (photosensitizer) and acceptor (chemiluminescer) microbeads can be coated with target-specific antibodies, secondary antibodies, proteins, or any molecular entity of interest. A signal is produced when the acceptor and donor beads are brought into proximity (<200 nm) by a molecular interaction occurring between the binding partners captured on the beads. Laser excitation of the Donor beads at 680 nm causes ambient oxygen to be converted to the singlet state by photosensitizers (phtalocyanine). These react with chemiluminescent agents (thioxene, anthracene, rubrene) on the acceptor bead only when the latter is in close proximity. Upon energy transfer between those compounds, activated rubrene emits light at 520 to 620 nm, which is in turn detected by the photo-detector in a microplate reader. An excitation wavelength higher than the emission wavelength ensures a low fluorescent background by avoiding any autofluorescence from biological media or compounds.

Purification of recombinant IRE1α cytosolic domain and AlphaScreen^® dimerization/oligomerization assay development

We used the AlphaScreen^® technology to evaluate the dimerization/oligomerization steps of human IRE1α cytosolic domains, a prerequisite for IRE1 activation. ^7,8 IRE1^cyto was then assayed for its ability to dimerize/oligomerize using AlphaScreen^® as follows: donor beads coated with glutathione (GSH) were incubated in the presence of increasing amounts of GST-IRE1^cyto, and acceptor beads coated with GSH were then added in a second incubation step in the presence or absence of 1 mM ATP. A maximal interaction signal was observed when the assay was carried out in the presence of ATP at GST-IRE1^cyto concentrations of 4.1 nM (optimal concentration of GST-IRE1^cyto, arrowheads, Fig. 1C , brown curve) and 0.3 µM when using GST-IRE1^cyto subjected to dephosphorylation ( Fig. 1D ). Interestingly, in the absence of ATP, signal intensity was approximately 2-fold lower than in the presence of ATP ( Fig. 1C , orange curve). However, this signal peaked at a concentration of GST-IRE1^cyto identical to that observed in the presence of ATP. This observation suggested that IRE1 phosphorylation may affect its dimerization/oligomerization properties. To test this, GST-IRE1^cyto was subjected to alkaline phosphatase treatment prior to the AlphaScreen^® assay. A result comparable with that obtained in Figure 1C was observed ( Fig. 1D ). However, the signal obtained in the absence of ATP was approximately 3-fold lower than that obtained in the presence of ATP ( Fig. 1D ), thus supporting the idea that GST-IRE1^cyto was phosphorylated and that this phosphorylation may affect the formation of a dimer/oligomer. Identical data were obtained using acceptor beads coated with anti-GST antibodies (data not shown). To further test this hypothesis, we used a catalytically inactive GST-IRE1^cyto kinase mutant, GST-IRE1K599A^cyto, ²² in a similar assay ( Fig. 1E ). The autophosphorylation-impaired mutant IRE1^cyto (K599A) had a dimerization/oligomerization signal (blue curve) 4- to 5-fold lower than that obtained with the wild-type protein (red curve). Collectively, these data demonstrate that an AlphaScreen^®-based GST-IRE1^cyto dimerization/ oligomerization assay can be developed and that the signal observed is dependent on GST-IRE1^cyto phosphorylation status.

AlphaScreen^® dimerization/oligomerization assay characterization

To further characterize the assay, dimerization/oligomerization of the dephosphorylated GST-IRE1^cyto was measured 17 h after the addition of excess ATP. This treatment yielded identical signal intensities when either dephosphorylated or phosphorylated GST-IRE1^cyto proteins were used ( Fig. 2A ). Moreover, when the reactions were carried out at 15 °C, 23 °C, or 37 °C for 17 h, the signal intensities and GST-IRE1^cyto optimal concentrations were identical (data not shown). To test the impact of GST on GST-IRE1^cyto dimerization/oligomerization, GST alone was tested in our assay ( Fig. 2B ). No AlphaScreen^® signal increase was observed for GST alone when used at the same concentrations and conditions as that defined in the GST-IRE1^cyto dimerization/oligomerization assay ( Fig. 2B ), thus suggesting that the signal increase observed was independent of GST function or attributes. Finally, we tested the characteristics of this assay and its adaptability to HTS of compound libraries. To this end, we evaluated its tolerance toward DMSO ( Fig. 2C ). When the assay was performed in the presence of increasing concentrations of DMSO (0.1%-10%) a Z′ value of 0.85 was obtained with a signal-over-background (S/B) ratio of 9.26, numbers compatible with compound screening. Taken together, these results indicate that our assay is relevant to IRE1^cyto dimerization/oligomerization and is compatible with HTS.

Fig. 2.

Validation and characterization of the IRE1^cyto AlphaScreen^®-based dimerization/oligomerization assay. (A) Effect of incubation time on AlphaScreen^®-based dimerization/oligomerization assay. An increased signal is observed in the presence of adenosine triphosphate (ATP). The assay was performed with glutathione (GSH)–coated donor beads and TAR anti–glutathione-S-transferase (GST) acceptor beads (GSH::anti-GST) using either GST-IRE1^cyto (4 nM) or GST-IRE1^cyto previously subjected to dephosphorylation using alkaline phosphatase (4 nM). Plate reading was carried out after 17 h and shows no difference between both proteins (n = 3). (B) Effect of GST on AlphaScreen^® signals. Increasing amounts of either GST or GST-IRE1^cyto were incubated in the presence or absence of 1 mM ATP as described in Materials and Methods. GSH donor beads and GSH acceptor beads were then added to the reaction and incubated for an additional hour. An enhanced AlphaScreen^® signal was observed only when GST-IRE1^cyto was incubated in the presence of ATP (n = 3). (C) Effect of DMSO on AlphaScreen^® assay characteristics. Increasing concentrations of DMSO (0.1%-10%) were added to the IRE1 dimerization assay using 4 nM GST-IRE1^cyto. A calculated Z′ value of 0.85 and a signal-to-noise (S/N) ratio of 9.26 were obtained (n = 3).

These experiments defined a GST-IRE1^cyto dimerization/oligomerization assay. Because we observed that IRE1 dimerization/oligomerization was in part dependent on its phosphorylation state, we next sought to develop an AlphaScreen^®-based assay to monitor GST-IRE1^cyto phosphorylation.

IRE1^cyto phosphorylation properties in vitro

As our objective was to design an assay (or a combination of assays) to identify functional modulators of IRE1 activity, we first verified the biological activity of GST-IRE1^cyto and tested its autophosphorylation ability in vitro as a readout using [³²Pγ]-ATP ( Fig. 3A ). GST-IRE1^cyto was capable of autophosphorylation, and this was in part dependent on the presence of magnesium and manganese ( Fig. 3A ). To further characterize this recombinant protein’s phosphorylation status following purification, bacteria were transformed with the pGEX-IRE1^cyto plasmid and were subjected to metabolic labeling with ³²P-orthophosphate. This revealed that GST-IRE1^cyto was phosphorylated in vivo ( Fig. 3B , top panel). Moreover, treatment of the recombinant proteins with alkaline phosphatase led to the decreased phosphorylation of Serine 724 as assessed by immunoblotting using a specific phospho-IRE1 antibody ²³ ( Fig. 3C ). To test whether this characteristic of IRE1 was detectable using the AlphaScreen^® technology, we designed an additional assay in which protein phosphorylation was detected using acceptor beads coated with Lewis metal ions ( Fig. 3D ; PerkinElmer). When the experiment was carried out in the absence of ATP, an increased signal was observed for GST-IRE1^cyto at concentrations comparable with those required for maximal signal in the dimerization/oligomerization assay ( Fig. 3E ). When the assay was carried out in the presence of ATP, the same concentration of GST-IRE1^cyto resulted in a 2-fold higher signal, suggesting an enhanced phosphorylation of IRE1^cyto ( Fig. 3E ). Using this assay, we then tested the ability of GST-IRE1^cyto, dephosphorylated GST-IRE1^cyto, and GST-IRE1K599A^cyto to autophosphorylate in the presence or absence of ATP ( Fig. 3F ). When these experiments were carried out using the identical concentrations of recombinant protein (as defined in Fig. 3E ), a 2-fold increase in GST-IRE1^cyto phosphorylation was observed in the presence of ATP (as seen in Fig. 3E ). When dephosphorylated GST-IRE1^cyto was used, the presence of ATP led to a ~6-fold increase in phosphorylation. No increase was detected when GST-IRE1K599A^cyto was used in the assay ( Fig. 3F ). These results show that GST-IRE1^cyto is phosphorylated in vitro and is able to autophosphorylate. In addition, the phosphorylation data are consistent with the data obtained using the dimerization/oligomerization assay as maximal AlphaScreen^® signals observed in the phosphorylation assay and the dimerization/oligomerization assay are obtained with identical concentrations of GST-IRE1^cyto.

Fig. 3.

In vitro phosphorylation of IRE1^cyto. (A) In vitro autophosphorylation activity of purified GST-IRE1^cyto (2 µg) using [³²Pγ]–adenosine triphosphate (ATP) in the presence or absence of MnCl₂ and MgCl₂. GST-IRE1^cyto autophosphorylation was visualized by radioautography (R; top panel) and the corresponding proteins by Coomassie Brilliant Blue staining of the gels (CBB; bottom panel). (B) ³²P-orthophosphate metabolic labeling of BL21 Escherichia coli cells transformed with a plasmid expressing GST-IRE1^cyto under the control of IPTG. Induction was carried out in the presence (right lane) or not (left lane) of isopropyl β-d-1-thiogalactopyranoside (IPTG) and GST-IRE1^cyto purified using glutathione sepharose beads. Purified products were then resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and either directly stained with Coomassie Blue, dried, and exposed to X-ray films (R) or transferred to nitrocellulose and immunoblotted using anti-IRE1 antibodies. (C) Dephosphorylation of GST-IRE1^cyto was carried out using alkaline phosphatase, as described in Materials and Methods, and the resulting protein product was then immunoblotted with either anti-phospho-IRE1 Ser 724 (top blot) or anti-IRE1 (bottom blot) antibodies. Data are representative of at least 3 independent experiments (mean ± SD). Student’s t-test was performed to show statistical significance, p < 0.005. (D) Schematic representation of IRE1^cyto phosphorylation assay based on the AlphaScreen^® technology. The donor beads are coated with glutathione (GSH), whereas the acceptor beads are coated with Lewis metal ions (Fe3⁺). (E) Typical phosphorylation profile obtained using the AlphaScreen^®-based dimerization assay. An ATP-dependent increase in GST-IRE1^cyto phosphorylation is observed and is comparable with data generated in the dimerization/oligomerization assay (n = 3). (F) Comparison of the autophosphorylation ability of GST-IRE1^cyto, dephosphorylated GST-IRE1^cyto, and GST-IREK599A^cyto in the presence or not of ATP using the AlphaScreen^® phosphorylation assay shown in (D) (n = 3).

IRE1 cytosolic domain characteristics in vitro

IRE1^cyto dimer/oligomer formation was enhanced in the presence of ATP ( Fig. 4A ), whereas the ATP nonhydrolyzable analog ATPγS had no such effect ( Fig. 4A ), which correlated with reduced endoribonuclease activity ( Fig. 4C , lane 4 and Fig. 4D ). Moreover, when the GST-IRE1^cyto dimerization/ oligomerization assay was carried out in the presence of the nucleotides AMP, GTP, or GDP, a lower signal was observed ( Fig. 4B ), indicating that ATP was specifically required for dimer formation. ADP was also capable of promoting IRE1^cyto dimerization, albeit to a lesser extent than ATP ( Fig. 4B ). Interestingly, similar observations have previously been reported using other experimental systems, demonstrating the role of ADP as an inducer of IRE1 activation. ^5,24 This observation suggests that diphosphate or triphosphate adenine nucleotides may act as enhancers of IRE1 oligomerization and activation. In addition, our data suggest that ATP hydrolysis and IRE1 phosphorylation may act as stabilizing factors in IRE1 oligomerization by changing the kinetics of oligomer dissociation.

Fig. 4.

Nucleotide dependence of IRE1^cyto dimerization/oligomerization and RNAse activity. (A) The impact of adenosine triphosphate (ATP) hydrolysis on the formation of GST-IRE1^cyto dimers (using 4 nM GST-IRE1^cyto) was tested using the nonhydrolyzable ATP analogue, ATPγS. An increased AlphaScreen^® signal was observed only in the presence of ATP, whereas the signals observed in the absence of ATP or in the presence of ATPγS were identical (n = 3). (B) The impact of mono-, di-, or triphosphate-containing nucleotides was tested on GST-IRE1^cyto dimerization in the experimental setup used in (A). Increasing concentrations of adenosine monophosphate (AMP), adenosine diphosphate (ADP), ATP, guanosine diphosphate (GDP), or guanosine triphosphate (GTP) were added to the reaction and AlphaScreen^® signal monitored as described in Materials and Methods (n = 5). (C) IRE1^cyto endoribonuclease activity was tested in an in vitro XBP1 mRNA cleavage assay as described in Materials and Methods. Total RNA was incubated in the presence of 10 µg glutathione (GST) or GST-IRE1^cyto and in the presence of 1 mM ATP or ATPγS, as described in Materials and Methods. After a 2-h incubation at 37 °C, RNA reaction products were purified and the amount of XBP1 mRNA monitored by RT-PCR. GAPDH was used as an internal standard (n = 3). (D) Time course (0-10 h) analysis of XBP1 mRNA cleavage by GST-IRE1^cyto in the presence of ATP or ATPγS as described in (C) was compared with an identical experiment carried out in the presence of GST or XBP1 mRNA alone (n = 4).

Because IRE1 possesses a dual enzymatic activity, we next tested whether the phosphorylation and dimerization/ oligomerization status of IRE1 affected its endoribonuclease activity for XBP1 mRNA. To this end, an endoribonuclease assay was developed (see Materials and Methods), and the amount of XBP1 mRNA remaining after a 2-h incubation was monitored by RT-PCR ( Fig. 4C ). Incubation in the presence of GST-IRE1^cyto and ATP led to a loss of 25% of XBP1 mRNA, which was not observed when GST-IRE1^cyto was incubated in the presence of ATPγS. To further confirm these results, a time course analysis of XBP1 mRNA degradation was performed for 0 to 10 h ( Fig. 4D ). Greater than 80% of XBP1 mRNA was degraded over the 10-h time period only when incubated in the presence of GST-IRE1^cyto and ATP.

Together, these experiments revealed that IRE1 dimerization/oligomerization in vitro was under the control of (1) IRE1 kinase activity, (2) nucleotide concentration, and (3) IRE1 phosphorylation status. However, these results did not completely explain the AlphaScreen^® signals observed in the dimerization/oligomerization assay.

Oligomerization of IRE1 in vitro and in vivo

The above results led us to hypothesize that in our assay, the decrease in signal observed at higher GST-IRE1^cyto concentrations may reflect either the loss of detectable dimers due to an excess of free GST-IRE1^cyto (i.e., not bound to beads) or increased oligomer size leading to enhanced distance between donor and acceptor beads. Therefore, we further examined the ability of IRE1^cyto to dimerize/oligomerize using our assay. Here, we used 6xHis-tagged IRE1^cyto ( Fig. 5A ) as a modulator, which would be unable to bind to GSH-coated AlphaScreen^® beads but would be able to affect IRE1^cyto dimer/oligomer formation. Increasing concentrations of 6xHis-tagged IRE1^cyto were incubated in the presence of 0.3 µM GST-IRE1^cyto and 1 mM ATP. Dimer/oligomer formation was then evaluated using the AlphaScreen^® assay described in Figure 1 . Interestingly, a decrease in signal was observed when increasing concentrations of 6xHis-IRE1^cyto were added to the reaction ( Fig. 5B , black curve) but not when BSA was used as a control ( Fig. 5B , gray curve). These results show that 6xHis-IRE1^cyto interfered with IRE1^cyto dimer/oligomer formation with an apparent EC₅₀ of approximately 5 µM. When the same experiment was carried out using S. cerevisiae IRE1^cyto ( Fig. 5B , dashed line), an apparent EC₅₀ of approximately 200 µM was observed, suggesting that both proteins may behave differently in terms of affinity. This could in part explain the discrepancy between the affinity values obtained in our study with 6xHis-IRE1^cyto and the published data on ScIRE1. ^9,10 These results suggested that 6xHis-IRE1^cyto competed with GST-IRE1^cyto dimer formation or increased the distance between donor and acceptor beads by promoting oligomer formation.

Fig. 5.

Characteristics of IRE1^cyto dimer/oligomer using AlphaScreen^®. (A) The expression of 6xHis-IRE1^cyto in BL21 cells was induced by 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 3 h. Bacterial recombinant proteins were then purified using appropriate chromatography matrices, eluted, dialyzed, concentrated, and then resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunoblot analysis using anti-IRE1 antibodies. (B) Using the experimental setup described in Figure 1 for the optimal concentration of GST-IRE1^cyto (4 nM), increasing amounts of 6xHis-human IRE1^cyto (continuous black line) or 6xHis– Saccharomyces cerevisiae IRE1^cyto (dashed line) were added to the reaction and the loss of AlphaScreen^® signal monitored. In this assay, the IC₅₀ observed for 6xHis-human IRE1^cyto and 6xHis–S. cerevisiae IRE1^cyto was compared with the same amounts of bovine serum albumin (BSA). The experiment was carried out using glutathione (GSH) donor and Europium GSH acceptor beads (n = 3).

To further analyze the mode of interaction for the IRE1 cytosolic domains, we used IRE1^cyto devoid of GST ( Fig. 6A ) in cross-linking experiments with disuccinimidyl suberate (DSS; see Materials and Methods). Interestingly, in the presence of DSS, IRE1^cyto formed oligomeric structures ranging from dimers to at least tetramers ( Fig. 6B , lane 2). This reaction appears to be dependent on phosphorylation because the oligomer intensity was diminished when the experiment was carried out using dephosphorylated IRE1^cyto ( Fig. 6B , lane 4). This demonstrates that our assay can be used as a reporter for IRE1 oligomerization and reveals that IRE1 oligomerization is dependent on intact IRE1 kinase activity, phosphorylation, and the presence of ATP or ADP nucleotides. The signal decrease observed in Figure 5B would therefore not correspond to the competition for dimer formation but rather for oligomer build-up. Hence, the apparent EC₅₀ values calculated would then reflect a distance increase between the 2 beads rather than the disruption of IRE1 dimers.

Fig. 6.

In vitro and cell-based oligomerization of IRE1. (A) Coomassie Brilliant Blue staining of thrombin cleaved GST-IRE1^cyto following removal of thrombin, concentration/dialysis of the cleaved product, and migration on 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The arrow indicates the purified IRE1^cyto protein. (B) Oligomeric status of IRE1^cyto using disuccinimidyl suberate (DSS)–based cross-linking experiments. IRE1^cyto (4 µM) was purified as a glutathione-S-transferase (GST) fusion protein, and GST was removed using thrombin following (dP) or not following dephosphorylation with alkaline phosphatase. Oligomers (indicated by asterisks) were visualized by immunoblotting using anti-IRE1 antibodies. (C) Oligomerization of full-length IRE1 in HeLa cells. HeLa cells were transfected with a plasmid bearing human full-length wild-type (wt) or K599A mutant IRE1 cDNA. Twenty-four hours posttransfection, cells were subjected or not to cross-linking experiments using DSS as the cross-linker. Proteins from cell lysate were then resolved by SDS-PAGE and analyzed by immunoblotting using anti-IRE1 antibodies (oligomers are indicated by asterisks). (D) Chinese hamster ovary (CHO) cells were transfected with the indicated amounts of plasmid bearing either human full-length wild-type IRE1 cDNA or human K599A mutant IRE1 cDNA. Twenty-four hours posttransfection, the cells were lyzed and proteins resolved by SDS-PAGE, followed by immunoblotting using anti-phospho-IRE1 and anti-IRE1.

To test the relevance of IRE1 oligomer formation in a more physiological context, HeLa cells were transfected with a plasmid encoding wild-type human IRE1 or K599A IRE1. Following transfection, cells were subjected to chemical cross-linking using DSS. Cells were then lyzed in RIPA buffer and the resulting material resolved by SDS-PAGE followed by immunoblotting using anti-IRE1 antibodies. In the absence of DSS, only 1 immunoreactive band migrating at 110 kDa was observed ( Fig. 6C , right panel), whereas when cells were treated with DSS, at least 4 immunoreactive bands were detected in cells transfected with wild-type IRE1 ( Fig. 6C , left panel). Interestingly, the signals were much weaker in cells transfected with K599A IRE1, thus confirming the results obtained in vitro using the AlphaScreen^® technology. To further correlate these data with the phosphorylation results, CHO cells were transfected with the indicated amounts of plasmid bearing either wild-type IRE1 or K599A IRE1 ( Fig. 6D ). IRE1 expression and phosphorylation were monitored by immunoblotting with anti-IRE1 ( Fig. 6D , middle panel) and anti-phospho-IRE1 antibodies ( Fig. 6D , top panel). Phosphorylation of IRE1 was observed in cells transfected with wild-type IRE1 cDNA but not in cells transfected with K599A IRE1 cDNA, thus validating the results obtained in vitro and reinforcing the correlation between IRE1 phosphorylation and oligomerization.

In this work, we establish and make use of a novel AlphaScreen^®-based assay to monitor IRE1 dimerization/ oligomerization properties. This was confirmed using orthogonal assays both in vitro and cell based. In addition, we developed an AlphaScreen^®-based IRE1 phosphorylation assay, which could serve as a secondary screening tool to identify modulators of IRE1 activity.

Discussion

In this work, we used AlphaScreen^® to investigate the biochemical properties of the dual kinase/RNAse protein IRE1. We were able to monitor and correlate 2 biochemical properties of the cytosolic domain of IRE1 (oligomerization and autophosphorylation). Moreover, we show that enhanced phosphorylation and oligomerization observed using AlphaScreen^® are consistent with increased endoribonuclease activity toward XBP1 mRNA in vitro and are also relevant to events occurring in cultured cells. This provides evidence for the compatibility of using HTS technologies to address biological questions and opens new avenues to identify modulators of IRE1 activity.

Assay platform development

Thus far, the AlphaScreen^® technology has been mostly used in HTS contexts in industrial settings. ¹ In the past, we have used AlphaScreen^® to monitor in vitro the activity of Ras GTPase, ²⁵ but no functional significance was provided for these assays. In the present study, we focused on IRE1-mediated ER stress signaling, which is implicated in many diseases. ⁶ In addition, no IRE1 activity modulator has yet been reported in the literature following activity-based compound screening.

Here, our objectives were (1) to develop a series of in vitro IRE1 activity assays using an HTS technology such as AlphaScreen^®, (2) to provide evidence that these assays provide biochemically relevant information on the IRE1 activation process, and (3) to demonstrate that the assays were physiologically relevant by using cell culture models. We have shown that the cytosolic domain of human IRE1α can be expressed in E. coli as 6xHis or GST fusion proteins and purified in quantities of approximately 3 mg per liter of cell culture medium. Two AlphaScreen^®-based activity assays suitable for screening of IRE1 modulators were optimized. These assays were designed to monitor either the oligomerization or phosphorylation status of the cytosolic domain of IRE1. Both assays provided relative affinity values for the monomers and the autophosphorylation ability of the protein, respectively. In addition, using either the nonhydrolyzable forms of ATP, ATPγS or the catalytically inactive mutant of IRE1, IRE1 K599A, we provide evidence that both oligomerization and phosphorylation of the cytosolic domain of IRE1 are linked. We propose a model in which IRE1^cyto phosphorylation may stabilize IRE1^cyto oligomers.

Biological relevance of the assay platform

To confirm the information gathered from the AlphaScreen^®-based experiments, we first used orthogonal in vitro assays monitoring IRE1^cyto endoribonuclease activity toward XBP1 mRNA and IRE1^cyto oligomerization. These assays demonstrated that the ability of IRE1^cyto to form higher order oligomers correlated with its enhanced endoribonuclease activity and appeared to be dependent on IRE1^cyto autophosphorylation ( Figs. 5-6 ). These data were consistent with previous reports demonstrating that S. cerevisiae IRE1^cyto could form oligomeric structures in vitro ^8,9 and in vivo in yeast. ¹¹ In addition, these data were also consistent with results obtained in mammalian cells on IRE1 function in regulating cell fate. ¹⁰ We next investigated the existence of IRE1 oligomers in cultured cells ( Fig. 6 ). These experiments demonstrated that when transfected into cultured cells, IRE1 was able to form oligomeric structures that was dependent on its kinase activity. Furthermore, using these assays, we characterized the formation of IRE1^cyto oligomers as a function of the nucleotides present in the reaction and correlated these characteristics with IRE1 endoribonuclease activity. Our findings link IRE1 phosphorylation to its oligomerization. Moreover, we show that these events could tightly regulate IRE1 endoribonuclease activity toward XBP1 mRNA. As a consequence, altering either IRE1 phosphorylation ability or oligomerization capacity may represent an efficient manner to alter IRE1 downstream signaling pathways.

In addition, these data indicate that the results obtained using AlphaScreen^® were consistent with data obtained in vitro and using cell-based approaches. Moreover, because the AlphaScreen^® assays provide information on IRE1^cyto oligomeric and phosphorylation status and because these assays can be easily adapted for HTS, we believe that screening for IRE1 activity modulators can now be achieved. The identification of IRE1 activity modulators may certainly play significant roles in the regulation of disease progression. Indeed, it has been shown that IRE1 is involved in specific diseases involving protein misfolding and that prolonged IRE1 activation could modulate cell fate. ^5,10 In addition, as IRE1 activity has already been directly implicated in tumor growth and angiogenesis, ²⁶ inhibition of its activity may represent an interesting antitumor avenue. Finally, as illustrated by Hetz and Glimcher, ⁶ beyond modulation of IRE1 activity, specific targeting of IRE1-dependent signaling pathways could also constitute approaches of significant interest.

Applications to HTS

The AlphaScreen^® assays are performed in a 384 format, so they are fully amenable for HTS. In addition, we have demonstrated that the oligomerization assay is not sensitive to high concentrations of DMSO ( Fig. 3 ). Both assays developed here are homogeneous assays and could be miniaturized even further (1536-well plates) to decrease enzyme and reagent consumption. Several other AlphaScreen^® assays monitoring cell signaling pathways have been used successfully in screening for inhibitors. ¹ We anticipate that modulators of IRE1 activity will soon be identified.

In conclusion, we have demonstrated that the HTS technology AlphaScreen^® can be used to address biological questions in an academic research setting. This may consequently represent tools not only for screening purposes but also for more fundamental investigations, which could then be easily translatable to chemical genomics-based assays.