Mapping the Brain Interaction Network of the Dual-Specificity, Tyrosine Phosphorylation-Regulated Kinase 1A (DYRK1A) Targeted by Leucettinib-21 Using Affinity Chromatography

Abstract

Leucettinibs are substituted 2-aminoimidazolin-4-ones inspired by the marine sponge natural product Leucettamine B and developed as pharmacological inhibitors of DYRK1A (dual-specificity, tyrosine phosphorylation-regulated kinase 1A), a therapeutic target for indications such as Down syndrome, Alzheimer’s disease, Parkinson’s disease, diabetes, myocardial infarction, etc. Leucettinib-21 is currently being tested in a phase 1 clinical trial. In this study, four different affinity chromatography-based approaches were developed to identify the rat brain targets of Leucettinib-21: (1) Leucettinib-21 (and its kinase-inactive isomer as control) immobilized on agarose beads, (2) immobilized metal affinity chromatography, (3) KinAffinity bead competition assays, and (4) immunoprecipitation with DYRK1A-specific antibodies. Altogether, these complementary methods (1) confirm known targets of Leucettinib-21, and identify (2) new protein kinases and nonkinases interacting with Leucettinib-21, (3) potential new partners of DYRK1A, and (4) pathways and cellular mechanisms potentially modulated by Leucettinib-21. These methods can be expanded to various cells and tissues from models of pathologies where Leucettinib-21 demonstrates efficacy.

There are numerous reasons to develop and characterize pharmacological inhibitors of the dual-specificity, tyrosine phosphorylation-regulated kinase 1A (DYRK1A) (reviews in refs − ). First, DYRK1A is involved in a multitude of fundamental biological processes including intracellular signaling, DREAM complex regulation, , mRNA splicing, synaptic functions, axonal transport, proliferation of neural stem cells, neuronal apoptosis, mitochondrial functions, , DNA damage and aging, pancreatic β-cell proliferation, , multiciliogenesis, viral replication, B cell immune responses, etc. Well-characterized DYRK1A inhibitors are thus important as pharmacological tools to study these essential biological pathways. Second, dysregulated DYRK1A is involved in various pathological states (reviews in refs , ); various cancers , including pancreatic cancer, , myeloproliferative neoplasm (MPN), acute megakaryoblastic leukemia (AMKL) and acute lymphocytic leukemia (ALL), neurodegeneration, inflammation, diabetes, , congenital heart defect, myocardial infarction, osteoarthritis, Down syndrome (DS), − Alzheimer’s disease (AD), Parkinson’s disease (PD), , and others. Abnormally active DYRK1A is responsible for cognitive disorders (memory, learning, spatial localization) observed in DS (the DYRK1A gene is located on chromosome 21) and in AD (review in ref ). The links between DYRK1A and disease has led to increasing interest for this kinase as a potential therapeutic target and for the development of pharmacological inhibitors, some of which reaching clinical evaluation − (reviews in refs , ). The search for therapeutic inhibitors of DYRK1A is also encouraged by the global clinical success of kinase inhibitors: , over 100 have been approved (94 by the FDA) as of January 2026 , (https://www.icoa.fr/pkidb/).

Identifying the range of the cellular targets of kinase inhibitors is of key importance, but not an easy matter. − The selectivity of kinase inhibitors is usually evaluated in vitro on a battery of recombinant kinases (direct catalytic activity, indirect interaction, thermal shift assays, competition affinity chromatography) or, less frequently, in cellular kinase assays (various reporter systems for individual kinases). These approaches have three major weaknesses: (1) they cover only a (usually small) fraction of the human kinome (538 kinases), (2) in most instances the tested kinases are not in a native configuration (they are massively overexpressed, they do not have physiological post-translational modifications, and they are tested under nonphysiological conditions in terms of ATP concentration, substrates, associated proteins and molecular environment), (3) these assays bypass potential nonkinase targets. ,

Two affinity chromatography approaches provide alternative and complementary ways to investigate the native targets of a given kinase inhibitor. The first method is based on the use of Sepharose beads on which a set of broad-spectrum kinase inhibitors have been immobilized (multiplexed inhibitor beads (MIBs) or Kinobeads). ,, Such beads loaded with tissue/cell lysates will bind a large number of native, endogenous kinases and their associated proteins. Increasing concentrations of the studied inhibitor compete out its targets which can be identified by quantitative mass spectrometry and apparent interaction constants can be calculated for each target. This method has been applied, for example, to the cyclin-dependent kinase (CDK) inhibitors roscovitine and CR8, to the DYRK1A inhibitor Leucettine L41, to the Ca²⁺/calmodulin-dependent protein kinase kinase (CaMKK) inhibitor TIM-063 (see other examples in refs ,, ). The second method is based on the immobilization of the studied inhibitor on Sepharose beads through a linker (usually polyethylene glycol). Tissue/cell extracts are loaded on the beads and the bound proteins are identified by Western blotting and mass spectrometry or sequencing. We have used this approach with various kinase inhibitors including purvalanol B, paullones, indirubin, hymenialdisine, roscovitine, and Leucettine L41.

Leucettinib-21 is a clinical drug candidate developed for the treatment of cognitive disorders associated with Down syndrome (DS) and Alzheimer’s disease (AD) (review in ref ). It was selected among >670 Leucettinibs, a family of DYRK1A inhibitors inspired by Leucettamine B, a natural product identified from the marine sponge Leucetta microraphis. ,− The selectivity of Leucettinib-21 has been extensively studied with various panels of recombinant kinases.

This study delineates the endogenous target landscape of Leucettinib-21 and its kinase-inactive isomer, iso-Leucettinib-21, through a multimethod affinity chromatography and proteomics approach. Three distinct techniquesSepharose-immobilized Leucettinib-21/iso-Leucettinib-21, DYRK1A enrichment via immobilized metal affinity chromatography (IMAC), and KinAffinity bead competition assayswere combined with silver staining, Western blot, and mass spectrometry analyses.

To discriminate Leucettinib-21 targets from DYRK1A-associated proteins, pulled-down protein lists were cross-referenced with DYRK1A immunoprecipitates identified by mass spectrometry. The results offer insights into Leucettinib-21’s cellular mechanisms of action, with potential implications for its clinical effects.

Results

Chemistry: Preparation of Immobilizable Leucettinib-21 and iso-Leucettinib-21

According to cocrystal and modeling structures, the (R)-leucinol moiety of Leucettinib-21 is facing the outside of the ATP-binding pocket. This position was selected to tether poly­(ethylene glycol) (PEG)-based linkers to build various affinity chromatography reagents.

This strategy required the synthesis of various noncommercial PEG-functionalized (R)-leucinols. To achieve this goal, two options were considered: tethering the PEG on the “upper” hydroxylated region or the “lower” iso-butyl arm of (R)-leucinol. In both cases, we believed that an ether function or a triazole between the PEG and (R)-leucinol would provide the desired PEG-functionalized amines, without significantly impacting the inhibitory profiles of the affinity chromatography reagents.

The upper functionalization was carried out as follows: double-benzylation of (R)-leucinol (1.1) followed by a Garegg reaction provided the sensitive iodinated substrate (1.3). The latter was engaged in an SN₂ with commercial hydroxylated PEGs (3 or 5 units) bearing a terminal NHBoc (a function to be later used for coupling with the beads). A final debenzylation by transfer hydrogenation successfully provided the PEG-tethered (R)-leucinols (1.6) and (1.7). Various observations worthy of note were made during this first sequence: (1) the combination of Bn as a protective group and iodine as the leaving group were the only “couple” able to tolerate the addition of deprotonated BocNH-PEG–OH; (2) yield of the SN₂ rapidly decreased when the PEG length increased; (3) reversing the reactivity and using bis-benzylated leucinol as the nucleophile and tosylated or iodinated PEG’s as the electrophile only provided vinylated PEG’s (elimination product); (4) the final debenzylation step occurred much faster by palladium-catalyzed transfer hydrogenation compared to classical hydrogenolysis in an autoclave. As an alternative to its etherification, bis-benzylated (R)-leucinol (1.2) was also successfully engaged in a propargylation reaction yielding the sensitive propargyl ether (1.9). The latter required to be rapidly engaged in a Cu­(I)-catalyzed cycloaddition with a pegylated azide bearing a terminal NHBoc. A final debenzylation provided the chiral aminotriazole (1.10) in a short and high-yielding sequence (Scheme ). With respect to the lower functionalization of (R)-leucinol, we were unable to directly attach a PEG moiety by etherification. However, the diastereopure aminotriazole (2.14) was isolated after a 13-steps sequence starting from commercial formylated oxazolidine (2.1). Attempts to control the oxazolidine ring opening (BiBr₃ in ACN or cat. PTSA in MeOH) and avoid Boc deprotection proved unsuccessful. Unlike PTSA, deprotection of Boc with HCl lowered the yields and induced a partial epimerization of the stereogenic center α to the methyl ester (up to 10%). Diastereoisomers (2.11a) and (2.11b) were separated by flash chromatography (see S.I. for full synthesis and analytical data) (Scheme ).

1. Synthesis of Pegylated Amines: “Upper” and “Lower” Functionalization of (R)-Leucinol (Detailed Syntheses Descriptions in S.I.).

2. Synthesis of Pegylated (iso)-Leucettinibs Bis-Hydrochlorides (Detailed Syntheses Description, Yields, UPLCs (Intermediates) and HPLCs (Final Products) in S.I.).

PEG linkers from (R)-leucinol were tethered to Leucettinib-21 (benzothiazol-6-ylmethylene ring) and iso-Leucettinib-21 (benzothiazol-5-ylmethylene ring), its kinase-inactive isomer, leading to compounds 11, 13, 15, 17 and 12, 14, 16, 18, respectively. PEG linkers were also bound to the closely related compounds 1, 3 and 2, 4, leading to compounds 5, 7, 9 and 6, 8, 10, respectively (Scheme and Table ).

1. Structure of Leucettinibs and iso-Leucettinibs Ligands and Their PEGylated Forms Synthesized in This Study .

^a(Z)/(E) ratios were determined by HPLC (see S.I.). Compounds (5) to (18) were isolated as bis-hydrochlorides.

The syntheses of the pegylated compounds were carried out as follows (Scheme ). Final products (1)–(4) and intermediates (4.1)–(4.14) were synthesized according to our previously described S_NAr strategy ,, between isothioureas (3.1) and (3.2) and the desired amines. When needed, TEA.HCl (activator of the isothiourea) significantly enhanced the rate and quality of the aromatic substitution. Intermediates (4.1)–(4.14) bearing a BocNH-PEG spacer were finally engaged in a deprotection with HCl in dioxane. Removal of the Boc protective group and subsequent quaternarization of both the terminal amine and the aminoimidazolone smoothly resulted in bis-hydrochlorinated (iso)-Leucettinibs (5)–(18) in high yields and limited (Z)/(E) isomerization (Table ). Noteworthy, we tried to isolate the PEG-functionalized (iso)-Leucettinibs as free bases in our initial attempts. However, these free bases were found to be highly viscous solids which were difficult to handle and analyze. Moreover, HPLC data indicated a higher (Z)/(E) isomerization rate (up to 25%) compared to their bis-hydrochlorinated counterparts. Consequently, bis-hydrochlorides were chosen as the favored structures for both evaluation on kinases and coupling with agarose beads.

Kinase Inhibitory Activity of Leucettinib and iso-Leucettinib Ligands

All compounds from Table were tested for IC₅₀ determination on a panel of 12 kinases (CDK5/p25, CK1ε, CLK1–4, DYRK1A, 1B, 2–4, and GSK-3β) (Table ). As expected, like iso-Leucettinib-21, all iso-Leucettinibs (2, 4, 6, 8, 10, 12, 14, 16, 18) were completely inactive on all 12 kinases (highest dose tested: 10 μM). Compounds 1, 3 (partial analogy with the leucinol structure) showed reduced kinase inhibitory activity (submicromolar IC₅₀ values) compared to Leucettinib-21, indicating the crucial roles of both the methyl ethers and the iso-butyl side arms in the biological activity. Their pegylated versions (5, 7, 9) displayed even lower inhibitory activities (micromolar IC₅₀ values) (Table ). Compounds 11, 13, 15, 17 (various PEGylated Leucettinib-21) showed submicromolar inhibitory activities, down to 27/32 and 49/72 nM (toward DYRK1A/DYRK1B) for compounds 11 and 13, respectively (Table ). Selectivity among the 12 kinases tested was essentially similar to that of Leucettinib-21.

2. Kinase Selectivity (Radiometric Assay) of Leucettinibs and iso-Leucettinibs Synthesized in This Study .

^aAll products were tested on a panel of 12 purified, recombinant human kinases using various substrates and ATP concentrations (S.I., Table S13). Leucettinib-21 and iso-Leucettinib-21 were used as reference compounds. IC₅₀ values were determined from the dose-response curves and are shown in μM. IC₅₀ values <0.1 μM are shown in bold.

Affinity Chromatography on Immobilized Leucettinibs and iso-Leucettinibs

Coupling and Evaluation of the Various Immobilized Ligands

Compounds 5 to 18 were coupled to cyanogen bromide-activated agarose beads as described in Scheme and in the Experimental Section. After binding of the ligand in an alkaline buffer, ethanolamine was used to block remaining active sites. Ethanolamine beads (no ligand) were prepared to serve as negative controls.

3. Coupling of (iso)-Leucettinibs (5) to (18) to Cyanogen Bromide–Activated Agarose Beads, General Principle.

Rat brain extracts were next loaded on the various Leucettinib/iso-Leucettinib-agarose beads (see Experimental Section for extract preparation and protein loading on beads). After extensive washing and addition of electrophoresis sample buffer, the bound proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Western blotting with anti-DYRK1A antibodies (7D10, directed against the C-terminal domain (C-ter)). Results show that, as expected, none of the iso-Leucettinib ligands bound any DYRK1A (Figure A). While ligand 15 was only the third most potent inhibitor overall, it demonstrated the strongest binding affinity for DYRK1A, followed by ligands 13 and 11 (Figure A). To avoid using LCTB compounds containing a potentially reactive triazole motif (such as ligands 15–18), ligand 13 and its kinase-inactive control counterpart, ligand 14, were selected for the remainder of the study. The corresponding affinity chromatography reagents were named respectively Leucettinib-21 beads (LCTB beads) and iso-Leucettinib-21 beads (iso-LCTB beads) in the rest of the article.

1.

DYRK1A binding affinity of the various Leucettinib-21/iso-Leucettinib-21-agarose beads. 20 μL of beads were incubated for 30 min with rat brain extracts (500 μg proteins), followed by extensive washing and the bound (A) and unbound (first supernatant) (B) proteins were resolved by SDS-PAGE and Western blotting with anti-DYRK1A antibody 7D10 (top) or by silver staining (bottom). Ethanolamine beads were used as control. CE, crude extract.

Five protein bands between 70 and 90 kDa, corresponding to full-length DYRK1A (FL-DYRK1A) and detected using the DYRK1A C-terminal antibody (clone 7D10), were observed in the crude brain extracts of Sprague–Dawley rats (Figure A). The uppermost was not retained on Leucettinib-21 beads while the three intermediate protein bands did bind to Leucettinib-21 beads (Figure B). Silver staining showed that only a few proteins were enriched/purified on the various Leucettinib-21 beads (Figure A, bottom). A ∼150 kDa protein was found to bind to all Leucettinib and iso-Leucettinib beads. A ∼42 kDa specifically bound to Leucettinib-21 beads but not to iso-Leucettinib-21 beads. Many, but not all affinity beads bound a ∼13 kDa protein (Figure ).

Leucettinib-21/iso-Leucettinib-21-Binding Proteins from Rat Organs/Tissues

We next analyzed DYRK1A expression in extracts from 15 rat organs (brain, colon, heart, intestine, kidneys, liver, lung, gastrocnemius muscle, ovary, pancreas, skin, spleen, stomach, testis, and thymus) using two distinct anti-DYRK1A antibodies: D1694, targeting the N-terminal domain (N-ter), and 7D10, targeting the C-terminal domain (C-ter), while DYRK1A catalytic activity was assessed in parallel.

The 7D10 antibody produced more consistent signals near the expected DYRK1A molecular weight (three splice variants of 86.3, 85.5, and 84.5 kDa are described in the rat) compared to the D1694 antibody (Figure A). It revealed a characteristic four-band pattern across most organs, except for the intestine, kidneys, and spleen, with the most intense bands being detected in the brain. Although differing in pattern, low molecular weight (LMW) protein bands cross-reacting with both antibodies were also detected. Note that the uppermost band detected by the 7D10 antibody in brain extracts, which is not bound by Leucettinib beads, is absent in the other organs (Figure A).

2.

DYRK1A level and kinase activity in crude extracts of various rat organs. Tissue extracts were prepared and proteins (15 μg) were resolved by SDS-PAGE and Western blotting with anti-DYRK1A antibodies (D1694 and 7D10) (A) and DYRK1A catalytic activity was measured (B). Both mean kinase activity values (±S.D.) and individual values are represented (n = 6; 3 males: black dots; 3 females: white dots). Catalytic activity is expressed as μM of phosphorylated Dan-peptide (see Experimental Section) produced/min/mg of each tissue extract.

The global catalytic kinase activity of DYRK1A was measured in lysates from rat organs (6 animals per organ, 3 mg/mL protein concentration), both in the absence and presence of 10 μM Leucettinib-21. Unexpectedly, the brain was only the second organ exhibiting the highest DYRK1A activity, the highest activity was observed in the thymus and did not correlate with the DYRK1A levels detected by Western blot (Figure B). No activity was detectable in the heart, intestine, liver, lung, and spleen, although protein was detected in some of them by Western blots. Leucettinib-21 completely abolished the activity detected in all extracts except in the testis, where a minor kinase activity remained (3.5 ± 1.8% of initial activity) (data not shown). No differences were observed between male and female tissues in terms of DYRK1A activity (Figure B). The discrepancy between enzymatic activity and Western blot signals may reflect tissue-specific regulation: proteins detected by Western blot may display different catalytic activities depending on tissue-specific protein interactions, whereas active forms might not be detected if the antibody epitope is lost due to protein cleavage or post-translational processing.

Extracts were next run on either immobilized Leucettinib-21 (Figure A) or immobilized nickel beads (Figure B) (taking advantage of the poly histidine stretch present in the C-terminal domain of DYRK1A (13 contiguous histidines), a natural “histidine tag” which allows its purification/enrichment by immobilized metal affinity chromatography (IMAC) , ). Bound proteins were analyzed by SDS-PAGE/silver staining and Western blotting with the two DYRK1A antibodies 7D10 and D1694 (Figure ). With both affinity chromatography methods, purified DYRK1A levels were higher in the brain than in other tissues. Low molecular weight bands cross-reacting with the anti-N-ter-DYRK1A antibody (D1694) were observed, especially in the thymus. The DYRK1A profile seen with the C-ter-directed antibody (7D10) on Leucettinib-21 and nickel beads were similar among different tissues. Silver staining revealed distinct protein expression profiles across different tissues.

3.

Affinity purification of DYRK1A from crude extracts of various rat tissues. Tissue extracts were prepared in homogenization buffer and centrifuged. A total of 2 mg or 0.1 mg of total proteins from the supernatant was loaded onto LCTB beads (A) or immobilized nickel beads (B), respectively. Beads were then washed extensively and the bound proteins were resolved by SDS-PAGE followed by Western blotting with anti-DYRK1A antibodies D1694 or 7D10 (upper figures) or by silver staining (lower figures). Vertical bars in the Western blots separating the brain from the other organs indicate that a longer exposure time was required for the other organs, which express less DYRK1A than the brain.

A ∼42 kDa protein band (Figure A, arrow) strongly cross-reacting with the D1694 antibody was observed in brain, kidney, liver, spleen, stomach and testis on LCTB beads, but not on immobilized nickel. It was identified as glutamine synthetase (also called glutamate-ammonia ligase, GLUL) by in-gel tryptic digestion coupled to mass spectrometry and by Western blotting (Figures S1, , and ). Purified recombinant human GLUL (expressed in Escherichia coli) selectively bound to LCTB beads, but not to iso-LCTB or nickel beads, and was detected by the DYRK1A D1694 antibody (Figure S1B). However, Leucettinib-21 did not affect the catalytic activity of GLUL, as determined by a colorimetric assay (Figure S1C).

4.

Affinity chromatography purification of rat brain proteins on immobilized ethanolamine, Leucettinib-21, iso-Leucettinib-21 and nickel. Rat brain extracts were prepared and 0.5 mg total proteins were loaded on ethanolamine (control), Leucettinib-21 (LCTB), iso-Leucettinib-21 (iso-LCTB) and immobilized metal (IMAC) beads. After extensive washing of the beads, the bound proteins were resolved by SDS-PAGE and analyzed by Western blotting with various antibodies (directed against DYRK1A (N-ter (D1694) and C-ter (7D10) domains), DYRK1B, DCAF7, FAM53C, GAK, GSK-3, PRKG2, GLUL, CDKL4, CK2α, CK2β). A crude extract (CE) sample (10 μg proteins) was run as reference.

5.

Affinity chromatography purification of rat brain proteins on immobilized Leucettinib-21: competition study. Rat brain extracts were prepared and 0.5 mg total proteins samples were mixed with increasing concentrations (0.01 to 100 μM) of free Leucettinib-21 (LCTB-21), free iso-Leucettinib-21 (iso-LCTB-21), or with a corresponding concentration of DMSO (control), prior to loading on LCTB beads. After extensive washing of the beads, the bound proteins were resolved by SDS-PAGE and analyzed by Western blotting with various antibodies (directed against DYRK1A (N-ter (D1694) and C-ter (7D10) domains), DYRK1B, GSK-3, DCAF7, GLUL). A crude brain extract sample (CE) (10 μg proteins) was run as reference.

Leucettinib-21/iso-Leucettinib-21-Binding Proteins in Rat Brain

We next focused on the analysis of rat brain proteins. Brain extracts were prepared and incubated with ethanolamine, LCTB, iso-LCTB and nickel beads and bound proteins were identified by Western blotting with various antibodies following SDS-PAGE (Figure ). No anti-DYRK1A cross-reacting proteins were detected on the control ethanolamine or iso-LCTB beads. Four major FL-DYRK1A bands were detected in the crude brain extract with the N-ter-directed antibody (D1694), but only two on LCTB beads and on immobilized nickel beads (Figure ). In contrast, five major FL-DYRK1A bands were detected in the crude brain extract with the C-ter-directed antibody (7D10), but only the central three bands were observed on LCTB beads and on immobilized nickel (Figure ), the uppermost band corresponding to a cross-reacting protein or being undetectable DYRK1A due to epitope masking.

As expected, DYRK1B was detected in the crude extract and on LCTB beads, but not on iso-LCTB beads (Figure ). DYRK1B was also purified by immobilized nickel although it does not contain the poly histidine present in DYRK1A. Indeed, Ni-NTA matrices bind not only proteins with exposed patches of histidine residues, but also those containing cysteine and tryptophan residues. It is therefore possible that DYRK1B has an affinity for nickel via other amino acid motifs than histidines, or that its binding is mediated by a complex of proteins, one of which has an affinity for nickel.

DCAF7 (DDB1 and CUL4 associated factor 7, also known as tryptophan-aspartic (WD) repeat-containing protein 68 (WDR68)) is known to bind both DYRK1A and DYRK1B. − DCAF7 was detected in crude brain extract, on LCTB beads (probably through DYRK1A and/or DYRK1B) and on immobilized nickel (most likely through DYRK1A) (Figure ). The recently identified DYRK1A-interacting protein, FAM53C, − was detected in crude brain extract on LCTB beads and on immobilized nickel. GSK-3α/β (glycogen synthase kinase 3α/β) was detected in crude brain extract and on LCTB beads (as observed with immobilized Leucettine L41), but not on immobilized iso-Leucettinib-21 or nickel (Figure ). Like Leucettine L41, Leucettinib-21 has rather limited inhibitory effect on GSK-3 (IC₅₀: 4.7 μM). The presence of GSK-3 on both LCTB and Leucettine L41 beads suggests a direct or indirect interaction of GSK-3 with DYRK1A. DYRK1A phosphorylates and inhibits GSK-3β and DYRK1A acts as a priming kinase for GSK-3. , Since a direct binding of Leucettinib-21 to GSK-3β has already been seen by thermal shift assays, GSK-3 most likely interacts directly with the LCTB beads. cGMP-dependent protein kinase 2 (PRKG2) was also identified as another interacting target on LCTB beads but not on immobilized iso-LCTB or nickel (Figure ). PRKG2 had been identified previously as a modest target of Leucettinib-21 which displayed an IC₅₀ value of 118 nM (Cerep assay). We retested Leucettinib-21 on PRKG1, PRKG2, CSNK2A1, CSNK2A2 and found IC₅₀ values of 1.20, 0.36, 1.13, and 1.25 μM, respectively (Reaction Biology assay). Iso-Leucettinib-21 was inactive on these kinases (IC₅₀ > 10 μM). Cyclin G-associated kinase (GAK) was also found to bind to LCTB beads but not to iso-LCTB beads (Figure ). This interaction was further confirmed in the double affinity chromatography, Kinobeads and immunoprecipitation approaches (see further down). GAK is only poorly inhibited by Leucettinib-21 (IC₅₀: 4 μM). Another previously identified target of Leucettinib-21, CDKL4 (cyclin-dependent kinase like 4), was enriched on LCTB beads, but not on iso-LCTB or nickel beads (Figure ). As observed with immobilized Leucettine L41, casein kinase 2α (CSNK2A1/CK2α) was found to bind to immobilized LCTB beads, but not to iso-LCTB or nickel beads. Although CSNK2A1 was identified as a potential target in the thermal shift selectivity screen, Leucettinib-21, at 1 μM, showed only 20% and 15% inhibition of CSNK2A1 and CSNK2A2, respectively. The purification of GLUL on LCTB beads was confirmed by Western blotting. Much less or no GLUL was detected on iso-LCTB or nickel beads, respectively (Figure ).

To validate these results, we ran the same affinity chromatography experiments on LCTB beads with brain extracts supplemented with free Leucettinib-21, free iso-Leucettinib-21 or dimethyl sulfoxide (DMSO) as control (Figure ). Increasing concentrations of Leucettinib-21, but not of iso-Leucettinib-21, led to a dose-dependent, competitive inhibition of DYRK1A, DYRK1B, and DCAF7 binding to LCTB beads (Figure ). The competition effect was much more limited with GSK-3 suggesting a DYRK1A-independent binding mode, yet ligand-dependent (competition with Leucettinib-21, but not with iso-Leucettinib-21). GLUL, detected with both the antibody against the N-ter part of DYRK1A (D1694; Figure , arrow) and the antibody against GLUL, could not be competed out with either Leucettinib-21 or iso-Leucettinib-21, suggesting nonspecific binding to the matrix.

Mass Spectrometry (MS) Analysis of LCTB-21 Protein Interactors

Analysis of Proteins Bound to Immobilized Leucettinib-21 and iso-Leucettinib-21 by Affinity Purification by Mass Spectrometry (AP-MS)

To gain a broader view of the brain proteins that may interact directly or indirectly with LCTB-21, we performed mass spectrometry on proteins bound to LCTB beads, iso-LCTB beads, and control ethanolamine beads. To further refine the identification of DYRK1A partners, we also conducted an initial enrichment step using nickel IMAC to enrich DYRK1A and its associated proteins before loading them to LCTB or iso-LCTB beads. The experimental protocol was first validated by assessing the binding of DYRK1A and its partner DCAF7 to the various affinity chromatography methods using Western blot analysis (Figure S2). High DYRK1A and DCAF7 levels were observed in LCTB, IMAC, and IMAC-LCTB bead fractions, and reduced levels in iso-LCTB beads (with or without prior enrichment by IMAC, Figure S2), validating their use for mass spectrometry analysis.

An initial comparative analysis was conducted on the proteins identified by MS in LCTB-21, iso-LCTB-21, and control bead samples. A total of 1,374 proteins were detected across all samples. Following data filteringretaining only proteins with valid values in at least five replicates of any single condition1,085 proteins were identified and quantified for further analysis.

A series of t-tests was first performed to compare proteins significantly enriched in LCTB and iso-LCTB beads versus control (Ctrl) beads (Figure ). Proteins showing a significant difference with a log₂ fold change (log₂FC) greater than 2 were considered for further analysis (LCTB beads, n = 449, Figure A and Table S1; iso-LCTB-21, n = 462, Figure B and Table S2). This analysis revealed distinct groups of proteins specifically associated with each ligand: 92 proteins were uniquely bound to LCTB-21, including 26 kinases, while 105 proteins were exclusive to iso-LCTB beads, none of which being a kinase (Figure C and Table S3). The majority of detected proteins (n = 357, including DYRK1A) were common to both bead types, with 22 kinases, resulting in a total of 48 kinases bound across all conditions (Table S3). In general, as is the case for DYRK1A, proteins enriched in both conditions were slightly less enriched in iso-LCTB beads. Notably, among the ten most enriched proteins in the LCTB pulldown, three were exclusively identified in this conditionfructosamine-3-kinase-related protein (FN3KRP), cAMP-dependent protein kinase catalytic subunit α (PRKACA), and casein kinase II subunit α (CSNK2A2), highlighting their specific interaction with LCTB beads.

6.

Mass spectrometry analysis of rat brain proteins enriched with Leucettinib-21 and iso-Leucettinib-21 beads (A, B). Volcano plots comparing protein enrichment on LCTB (A) and iso-LCTB (B) beads relative to control beads. Proteins identified as significantly different are highlighted in red (t-test, FDR = 0.01). Venn diagram comparing proteins enriched (Log₂FC > 2 compared to ctrl beads) between immobilized LCTB (n = 449) and iso-LCTB (n = 462) conditions (C). Top 20 nonredundant enrichment clusters of all 449 proteins associated with LCTB beads (D). Biological processes are ranked by p-value, with darker colors indicating higher statistical significance (lower p-values). Protein–protein interaction (PPI) networks were constructed using PPI databases, and Gene Ontology (GO) enrichment analysis (D) was performed to identify biologically meaningful processes using Metascape. The MCODE algorithm applied to the PPI network of proteins linked to LCTB beads (log₂FC > 2) enables the detection of densely connected protein neighborhoods (E).

Gene ontology (GO) pathway analysis of the 92 proteins uniquely associated with LCTB revealed enrichment in synaptic transmission, axon guidance, insulin signaling, and long-term potentiation pathways, among others (Figure S3A). These processes are central to neuronal communication, plasticity, metabolic regulation, and kinase-driven pathways. The strongest term in the iso-LCTB specific pathway analysis is “pathways of neurodegeneration (multiple diseases)”, indicating a broader association with neuronal health rather than specific synaptic signaling pathways (Figure S3B).

GO analysis of all 449 proteins purified on LCTB beads (92 proteins uniquely associated with LCTB beads and 357 in common to both LCTB and iso-LCTB beads) revealed enrichment in pathways related to Parkinson’s disease, modulation of chemical synaptic transmission, and mitochondrial organization, and among others (Figure D).

Overall, these results suggest that LCTB-21 engages proteins involved in neuronal health, energy metabolism, and synaptic signaling, indicating potential therapeutic benefits but also serving as a warning for possible side effects. Protein–protein interaction enrichment analysis was performed using Metascape, and the Molecular Complex Detection (MCODE) algorithm was applied to identify densely connected network components (Figure E). It revealed several clusters of highly interconnected proteins bound to LCTB beads, primarily associated with neurodegenerative disease pathways, transmembrane proton transport, oxidative phosphorylation, and related biological processes. These findings suggest that proteins predominantly associate with the beads as part of multiprotein complexes, with LCTB-21 showing a selective affinity for protein networks linked to neuronal function and mitochondrial processes.

Double Affinity Chromatography: Enrichment by IMAC Followed by Affinity Purification on Immobilized Leucettinib-21

We next enriched DYRK1A and associated proteins by IMAC. The proteins bound to Ni-NTA beads were eluted with imidazole and loaded on LCTB or iso-LCTB beads. The prepurification step using Ni-NTA beads substantially reduced the complexity of protein mixtures bound to LCTB or iso-LCTB beads. t tests identified 21 and 80 proteins significantly enriched under the IMAC → LCTB (Figure A and Table S4) and IMAC→iso-LCTB (Figure B and Table S5) beads conditions, respectively, compared to IMAC alone. Among the 69 IMAC-prepurified proteins that bound to LCTB and iso-LCTB beads, 16 were shared by both groups, including DYRK1A (Table S6), and 17 were potential DYRK1A interactors (BioGRID database). When the prepurified proteins bound to LCTB beads were compared with the prepurified proteins bound to iso-LCTB beads, DYRK1A, DCAF7 and GAK stood out among the major enriched targets (Figure C and Table S7), thereby validating the method.

7.

Volcano plot analysis of protein enrichment following sequential affinity purification. (A) Protein enrichment in IMAC–LCTB-21 relative to IMAC alone. (B) Protein enrichment in IMAC–iso-LCTB-21 relative to IMAC alone. (C) Comparative protein enrichment between IMAC–LCTB-21 and IMAC–iso-LCTB-21. Proteins identified as significantly different are highlighted in red in panels (A, B) (t-test, FDR = 0.01). In panel (C), proteins with a–log­(p) > 2 are indicated in red.

KinAffinity/LC-MS Kinome Profiling of Leucettinib-21

We next wanted to compare the list of kinases bound to LCTB beads to kinases displaced by LCTB-21 in the KinAffinity (Evotec) assay, using the same tissue samples.

Rat brain extracts were loaded on KinAffinity resin to which five broad-spectrum kinase inhibitors are immobilized. LC-MS analysis of the bound proteins provided an overview of the kinases expressed and detectable in rat brains.

To maximize kinase coverage, protein annotation was performed using the highly curated mouse FASTA database, which shares approximately 98% sequence identity with the rat proteome and contains nearly twice as many validated proteins. K _d values derived from the rat database are also provided (Figure and Table S8).

8.

Competition assay identification of Leucettinib-21 targets from rat brain. Left, rat brain extracts were loaded on an affinity matrix (KinAffinity, Evotec) comprising a set of 5 broad-spectrum kinase inhibitors designed to affinity purify endogenously expressed kinases of cells or tissues. Bound proteins (49 kinases–36 shown here-, 14 nonkinase proteins or kinase regulatory proteins) were identified following exposure to increasing concentrations of Leucettinib-21 (S.I. Table S8). The K _d values were calculated and are plotted in a Log scale. Protein kinases (blue), and nonkinase proteins (orange) are ranked from low (top) to high (bottom) K _d values, determined using the Swissprot database. K _d values obtained with rat database searches are also indicated. For some proteins with good enrichment curve fitting failed. In these cases, the K _d for resin binding was set to the lowest resin concentration (10 μM) to estimate the K _d value. Such proteins are marked with a “*”. For an unknown reason, PRKCQ (“**”) was among the targets in the rat data set, but not when the mouse fasta file was used. For proteins marked with “#”, some data points are missing in the competition assays, most likely due to complete displacement from the matrix at higher compound concentrations. Right, Leucettinib-21 was tested in dose–response curves on recombinant kinases corresponding to native kinases identified on the affinity matrix using the mouse swissprot database. Dotted lines in the table separate proteins with K _d values 0.001–0.01, 0.01–0.1, 0.1–1, and 1–10 μM. n.a., nonapplicable. n.t., not tested.

In total, 275 protein kinases and 9 lipid kinases were detected in rat brain lysate, of which 49 were identified as targets of Leucettinib-21 (Figure and Table S8). The remaining 235 kinases were not displaced by Leucettinib-21 at concentrations up to 30 μM. While K _d values were comparable between rat and mouse Swiss-Prot database searches, the mouse database yielded a higher number of identified proteins, with K _d values ranging from 0.002 to 29.961 μM (Figure and Table S8). Among the identified targets, eight protein kinases (DYRK1B, DYRK1A, RPS6KA4, CLK2, CDC42BPB, CAMK2G, PRKG1, and CLK3) exhibited high affinity (K _d < 70 nM) for Leucettinib-21, with DYRK1B (K _d: 0.002 μM) and DYRK1A (K _d: 0.003 μM) showing the strongest binding. The remaining kinases demonstrated lower affinity. Of the 49 proteins identified, 32 had a K _d > 1 μM and were considered poor or modest binders, including CSNK2A1, GSK-3α, GSK-3β, and TAOK1.

Interestingly, 14 nonkinase proteins were also identified as targets of Leucettinib-21, some with high affinity, such as DCAF7 and CDC37. The K _d value for DCAF7 (0.003 μM), comparable to those of DYRK1A/1B, suggests that DCAF7 was displaced in complex with DYRK1A and/or DYRK1B. Similarly, the high affinity measured for CDC37 (Cell Division Cycle 37; K _d: 0.008 μM) likely reflects its association with its partners RPS6KA4 (Ribosomal Protein S6 Kinase A4; K _d: 0.011 μM) and/or DYRK1B. − Other nonkinase targets include the postsynaptic protein PSD (PH and SEC7 domain-containing protein 1), IQSEC1 (IQ motif and SEC7 domain-containing protein 1) and GIT1 (ARF GTPase-activating protein GIT1), VCPIP1 (Deubiquitinating protein VCIP135), ZWINT (ZW10 interactor), and MAP2 (Microtubule-associated protein 2). These proteins either directly interact with LCTB-21 or are partners/substrates of its kinase interactors.

AP-MS Analysis of DYRK1A Immunoprecipitates

To discriminate between direct Leucettinib-21 targets and proteins that associate with DYRK1A, we performed immunoprecipitations (IP) of endogenous DYRK1A from rat brain extracts with the two same DYRK1A antibodies used throughout the study (D1694: rat polyclonal, targeting the N-ter region; 7D10: mouse monoclonal, targeting the C-ter region). Mass spectrometry analysis, combined with statistical filtering, identified 1,219 proteins significantly enriched (Log₂FC ≥2) compared to the no-antibody control (Ctrl) (Figure S4 and Tables S9 and S10). Of these, 311 were listed among the 1,024 DYRK1A interactors documented in BioGRID (Table S11). Among the enriched proteins, 648 were common to both antibodies, while 514 were uniquely detected with D1694 and 57 with 7D10 (Table S11). As expected, several proteins with strong evidence of interaction with DYRK1A, including DCAF7, FAM117B, FAM53C, DYNLL2, PRKACB, and DYNLL1, were coimmunoprecipitated with both antibodies (Table S11). In contrast, other interactors were detected only with one of the two antibodies, such as YWHAG (14-3-3γ) with D1694 and GLCCI1 (Glucocorticoid-induced transcript 1) with 7D10, likely due to epitope masking or differential antibody binding.

To refine the interactome of DYRK1A and elucidate the target profile of LCTB-21, we cross-referenced proteins coimmunoprecipitated with DYRK1A antibodies against those enriched in LCTB-21 pull-downs and previously documented interactors (Figure and Table S12). This analysis identified 43 proteins common to all three data sets, confirming their role as established DYRK1A interactors (Figure and Table S12). Generally, proteins pulled down with DYRK1Awhether using DYRK1A-specific antibodies or LCTB-21 beadsinclude proteins involved in endocytosis and vesicular trafficking (AP2A1, AP2B1, AP2M1, AP2S1, CLTA, CLTB, DNAJA1, DNAJB11, GAK), signal transduction (CAMK2D, PRKACB, PRKAR1B, DYRK1A), cytoskeletal dynamics (MAP6, SPTAN1, TUBB4A, TUBB4B), and mitochondrial function (OPA1, PHB2, VDAC2). Among these, GLUL was included, despite previous evidence demonstrating its ability to bind LCTB-21 independently of DYRK1A (Figure S1). Additionally, CAMK2D, DCAF7, USP5, and PRKACB were detected, consistent with their identification in the KinAffinity assay (Figure and Table S8). The distinct dissociation constants observed for CAMK2D, USP5, and PRKACB, compared to DYRK1A suggest that they are likely both partners of DYRK1A and interactors of LCTB-21. In line with this hypothesis, the interaction between DYRK1A and CAMK2 has been previously established.

9.

Discrimination of DYRK1A-associated proteins from LCTB-21 targets by multiple approaches. Venn diagram comparing proteins bound to LCTB beads, coimmunoprecipitated with N-ter (D1694) and C-ter (7D10) DYRK1A antibodies, and BioGRID-referenced partners. Specific proteins intersecting with LCTB-21 bead binding are listed in boxes. The protein targets of LCTB-21 identified by the KinAffinity assay are underlined.

Additionally, 34 proteins were identified on LCTB beads and described as DYRK1A interactors but were not immunoprecipitated in our study, suggesting they may represent either direct targets of LCTB-21 or indirect partners of DYRK1A (Figure and Table S12). A further 112 proteins were common to both LCTB bead pull-downs and DYRK1A immunoprecipitates, indicating potential novel interactors of DYRK1A (Figure and Table S12), although they include 6 protein kinases displaced by LCTB-21 in the KinAffinity assay. Notably, 260 proteins were exclusively bound to LCTB beads, pointing to possible new targets of LCTB-21 or partners of undiscovered pathways. Within this group, 31 proteins with kinase activity were identified, 6 of which were validated as LCTB-21 targets using KinAffinity assays (Figure and Table S12). While CDK14, CSNK2A2, GSK-3α, and GSK-3β displayed weak binding affinities, CDC42BPB and RPS6KA4 exhibited strong binding (K _d values of 0.018 μM and 0.011 μM, respectively, Figure ), albeit with moderate inhibition of catalytic activity (IC₅₀ = 0.140 μM for CDC42BPB, Figure ). Sevenof them correspond to metabolic kinases (ADK, AGK, BCKDK, FN3K, FN3KRP, GK, and HKDC1). These enzymes were not previously characterized as targets of LCTB-21 in in vitro assays and did not interact with the KinAffinity matrix, which uses a set of immobilized protein kinase inhibitors. Nevertheless, several of these metabolic kinases may remain of interest in specific disease contexts (ADK in epilepsy, AGK, HKDC1, FN3K in cancers − ).

More generally, the proteins identified constitute a diverse repertoire essential to fundamental cellular processes, ranging from metabolic regulation to intracellular signaling, vesicular transport and cytoskeletal organization. Among these, a prominent subset of metabolic enzymes, including ACACA, ACADL, ACADM, and GCDH, plays pivotal roles in fatty acid β-oxidation, lipid biosynthesis, and mitochondrial energy production. The data set also features a robust representation of signaling molecules, such as the kinases and regulatory proteins BRAF, MAPK1, PRKAG2, and STAT3, which orchestrate critical cellular responses to growth factors, stress stimuli, and inflammatory signals, thereby influencing cell fate decisions. , Additionally, the analysis highlights proteins essential for intracellular trafficking and membrane dynamics, such as RAB14, RAB18, VAMP1, and VTI1B, which collectively ensure the precise delivery of cargo within cells and maintain the integrity of membrane-bound compartments. , Several identified proteinsincluding GABRA1, the HSP40 cochaperone DNAJB6, and ALDH5A1are implicated in neurological and metabolic disorders, such as epilepsy and neurodegenerative diseases.

These results provide valuable insights into the molecular targets of LCTB-21 and expand the known interactome of DYRK1A.

Discussion

Identifying the targets of any given drug and the binding partners of any given target is no trivial work! The range of targets and associated proteins indeed depends on their expression which varies according to species, organs, tissue type, cell type and intracellular compartment. In each of these settings, expression varies according to development stage, physiological state, hormonal status, health and disease condition, etc.

Although often initially selected for potent and selective interaction with a unique target, most drugs interact with a wider set of targets in real life. Many drugs also show nonselective binding due to nonselective interactions (e.g., hydrophobic). These nonexpected and/or nondesired targets are collectively named “off targets”, although the physiological outcome of any drug treatment is merely the result of these combined, complex interactions.

Understanding the selectivity of drug candidates is crucial for clarifying their mechanism of action, preventing dilution of therapeutic effects, and reducing toxicity and adverse side effects while maximizing potential benefits.

In this article, we have investigated the rat brain native targets of a drug candidate, Leucettinib-21, and their associated partners, using different affinity chromatography approaches: (1) Sepharose-immobilized Leucettinib-21 (and its negative control, iso-Leucettinib-21), directly or after DYRK1A enrichment using immobilized metal (nickel) affinity chromatography, (2) immunoprecipitation with various anti-DYRK1A antibodies and (3) competition affinity on KinAffinity beads (set of immobilized wide-range selectivity kinase inhibitors). Altogether these methods provide a global view of two overlapping protein populations: (1) Leucettinib-21 targets and their associated proteins, and (2) DYRK1A-associated proteins.

Immobilized Leucettinib-21 Approach

The synthesis of an immobilized Leucettinib-21 matrix was driven by molecular modeling and cocrystal structures of Leucettinib-21 with DYRK1A and CLK1, respectively. The linker had to be attached to the leucinol part of the molecule, which faces the solvent. We explored various lengths and structures for the linker: a 3 ethylene glycol units linker appeared to be optimal for DYRK1A binding from tissue extracts (Table and Figure ). The kinase-inactive isomer, iso-Leucettinib-21 was similarly tethered to Sepharose beads to be used as a negative control. The first limitation of this approach is that while ligand selection was based on DYRK1A binding and inhibition, the PEG linkerirrespective of its positionmust restrict interactions between Leucettinib-21 and unidentified but biologically relevant proteins. This distinction renders immobilized Leucettinib-21 inevitably functionally different from its free counterpart. Competition with free Leucettinib-21 partially compensates this limitation.

A first group of direct targets was identified:

DYRK1A was detected on LCTB beads, but various antibodies revealed various bands, some of which did not bind to the beads. The highest MW DYRK1A form appears to be unable to bind to this affinity reagent nor to nickel beads. The most likely explanation is that the antibodies cross-react with a protein distinct from DYRK1A. Alternatively, a post-translational modification or a binding partner prevents interaction with the inhibitor or the metal chelate. Alternatively, the antibodies may cross-react with a protein distinct from DYRK1A. The existence of multiple forms of DYRK1A may reflect alternative splicing variants or various post-translational modifications. Not surprisingly, two well established DYRK1A-binding partners, DCAF7 and FAM53C, copurified on LCTB and nickel beads but not or poorly on iso-LCTB beads (Figure ).

The binding of direct targets like DYRK1B, CDKL4, or PRKG2 on LCTB beads was not a surprise, as they had been identified in previous kinase panel screens. CDKL4 and PRKG2 were not retained by immobilized nickel, suggesting that they probably bind directly to Leucettinib-21.

More difficult to understand targets are GSK-3 and CSNK2 (CK2), two kinases which are very poorly inhibited by Leucettinib-21. Both had been previously been detected on Sepharose -immobilized Leucettine L41. These two kinases are not found on nickel beads, suggesting that they are not purified on LCTB beads through DYRK1A or DYRK1A-binding partners. In contrast to DYRK1A and DYRK1B, GSK-3 can be competed out from binding to LCTB beads by only high concentrations of Leucettinib-21 (Figure ).

A more intriguing target is GLUL. This protein not only binds to LCTB beads, and to a lower extent to iso-LCTB beads, but it appears to cross react with anti-DYRK1A antibodies (Figure S1). GLUL cannot be competed out from binding to LCTB beads by high concentrations of Leucettinib-21 (Figure ), suggesting an alternative binding mode such as through the linker associated with a fragment (leucinol?) common to Leucettinib-21 and iso-Leucettinib-21. Furthermore, Leucettinib-21 only weakly inhibited GLUL in vitro, though its effects could differ in cells where multiple interacting partners are involved (Figure S1).

GAK is another interesting target binding to LCTB beads, but not to iso-LCTB. This interaction was further confirmed with the competition affinity on KinAffinity beads and anti-DYRK1A immunoprecipitation approaches. GAK also stands out as a target in the comparison between the IMAC → LCTB versus IMAC → iso-LCTB double-affinity approaches (Figure C and Table S7), suggesting a valid interaction. GAK is listed in the Biogrid DYRK1A interactome. Yet, a nonspecific interaction cannot be completely excluded. GAK is only poorly inhibited by Leucettinib-21 (IC₅₀: 4 μM). Given the identification of GAK, and its gene DNAJC26, in genome-wide association studies as a risk factor for Parkinson’s disease , and its role in regulating synaptic trafficking and clathrin dynamics, the link between Leucettinib-21 and GAK certainly deserves further investigation.

Competition with free Leucettinib-21 in the LCTB beads affinity approach provides a visual representation of the high affinity targets of Leucettinib-21 (DYRK1A, DYRK1B) and their associated proteins (DCAF7), as well as of low affinity targets (GSK-3) and ‘pseudotargets’ (GLUL) which bind to the affinity reagent independently of Leucettinib-21 itself (Figure ).

Anti-DYRK1A Immunoprecipitation Approach

Immunoprecipitation with various anti-DYRK1A (N-terminal and/or C-terminal domains) or anti-tagged DYRK1A antibodies followed by mass spectrometry has been the classical method used to identify DYRK1A interactors (review in ref ). It has been used with various biological materials, essentially dividing cells: HeLa cells, HeLa cell nuclei, , HEK-293-T-Rex overexpressing DYRK1A tagged with Strep-Tag and hemagglutinin (HA), human glioblastoma T98G cells overexpressing HA-tagged DYRK1A, HEK-923T and HCT116 cells, human neural stem cells, and human neuroblastoma SH-SY5Y cells differentiated as neuron-like cells. These detailed analyses have unravelled a large diversity of interacting partners with limited overlap between studies, despite the fact that most studies were carried out with dividing cells. This diversity is likely a consequence of the diverse cell lines used, the use of native or overexpressed (and tagged) DYRK1A, and numerous false positives which are not easily identified. Nevertheless, many of the identified interactors relate to cell cycle control, cell survival, DNA damage response, cell signaling, cytoskeleton, and gene expression. Among the constantly identified interactors are DCAF7, FAM53C, FAM117C/GLCCl1 (glucocorticoid-induced transcript 1) (and the related FAM117A, FAM117B), E3 ubiquitin ligase RNF169, TROAP (trophinin associated protein), LZTS2 (leucine zipper tumor suppressor 2).

In contrast to previous studies, we studied the interactome of native DYRK1A from Sprague–Dawley rat brain extracts. Two antibodies, one directed against the N-terminal domain of DYRK1A (D1694), the other directed against the C-terminal domain (7D10) were used (Figure S4 and Tables S9 and S10). Most DYRK1A interactors (648) were detected with both antibodies, with much less proteins uniquely detected with 7D10 (57) compared to proteins uniquely detected with D1694 (514), suggesting a differential distribution of interacting proteins between the N-terminal and C-terminal domains of DYRK1A. Among the 648 immunoprecipitated partners (both antibodies), 189 are found in the BioGRID DYRK1A partners, 155 are purified on LCTB-21 beads, and 43 are shared by both (Figure and Table S12). Compared to previous immunoprecipitation studies carried out with dividing cells, DCAF7, FAM53C, FAM117B, but not RNF169, TROAP and LZTS2, were immunoprecipitated from rat brain. Gene ontology pathway analysis reveal enrichment in neurotransmitter receptors, synaptic transmission, axon guidance, Parkinson’s disease, etc.

Competition Affinity on KinAffinity Beads Approach

The KinAffinity competition affinity approach provides a semiquantitative view of proteins that interact directly or indirectly with Leucettinib-21 (Figure ). This assay confirmed DYRK1A and DYRK1B as the highest-affinity brain targets of Leucettinib-21. This competition approach showed secondary targets identified from the Leucettinib-21 beads such as CLK2, CSNK2A, GSK-3β or PRKG2. Leucettinib-21 was tested on most of the 27 kinases identified by competition on KinoAffinity beads using rat database for protein identification. Besides DYRK1A, DYRK1B, and CLK2, most kinase targets were only modestly inhibited by Leucettinib-21. Among the indirect interactors, DCAF7 and CDC37 showed the highest affinity. CDC37 is a cochaperone that binds to numerous kinases and promotes their interaction with the Hsp90 complex, resulting in stabilization and promotion of their activity. , CDC37 has been previously demonstrated to interact with DYRK1A and DYRK1B. ,, CDC37 was not found on LCTB or IMAC beads nor in immunoprecipitates. Another potentially interesting protein is IQSEC1 (IQ motif and SEC7 domain-containing protein 1) also known as BRAG2 (Brefeldin A-resistant ARF-GEF 2). This protein was enriched in the KinAffinity beads approach (Figure and Table S8), but also with LCTB beads (especially in the comparison between the IMAC → LCTB versus IMAC→iso-LCTB double-affinity approachessee Figure ). IQSEC1 is an ADP ribosylation factor-guanine nucleotide exchange factor (ARF-GEF) which interacts with PSD-95, and plays a role in synaptic functions, axon guidance, dendritic projection, membrane trafficking, cytoskeletal organization, signal transduction, and cell adhesion. , The link between IQSEC1/BRAG2 and DYRK1A/B certainly deserves further investigation.

In summary, Leucettinib-21 binds to rat brain proteins involved in networks linked to multiple neuronal and mitochondrial functions, often implicated in neurodegenerative processes.

One major limitation of our study is that proteins identified in tissue lysates using LCTB-21 beads may not fully reflect their true in vivo interactions. In intact biological systems, cellular compartmentalization and protein accessibility differ substantially, which can alter binding profiles. In addition, certain protein classessuch as cytoskeletal or membrane-associated proteins - may be underrepresented due to the extraction and isolation methods used. Validation with purified proteins would be essential to confirm these interactions under physiological conditions. Additionally, the expression levels of DYRK1A and its interacting partners vary across brain areas, potentially influencing binding patterns and functional outcomes, thereby limiting the generalizability of our findings. While mass spectrometry enabled the identification of hundreds of proteins, adjusting the stringency of analysis - based on enrichment and p-value thresholdsremains challenging, risking false positives or the exclusion of relevant low-abundance targets. We also observed discrepancies between mass spectrometry and Western blot data, likely due to differences in sensitivity, specificity, or detection limits between these methods. Furthermore, the selectivity and specificity of antibodies may introduce variability, affecting the reliability of protein identification. Finally, protein annotation and functional interpretation can vary depending on whether rat, mouse, or human databases are used, potentially leading to inconsistencies in our results.

Next, we will apply these techniques to (1) different brain regions, (2) different cell compartments following cellular fractionation, (3) genetically modified cell lines (including DYRK1A knockout, overexpression, or kinase-dead mutants) and (4) disease model tissues (such as those from Down syndrome, myocardial infarction, Parkinson’s disease, Alzheimer’s disease and diabetes) to further investigate the targets of Leucettinib-21 and the interacting partners of DYRK1A.

Conclusion

This study provides a comprehensive characterization of the protein targets associated with Leucettinib-21 in rat brain using four complementary affinity-based proteomic approaches. The combined strategies not only validated previously described targets of Leucettinib-21, including DYRK1A, but also expanded the target landscape by identifying additional kinases and nonkinase interactors. Moreover, the data reveal potential new protein partners of DYRK1A and highlight signaling pathways and cellular processes that may be modulated by Leucettinib-21. Together, these findings contribute to a better understanding of the molecular mechanisms underlying the pharmacological effects of Leucettinib-21 and provide a broader view of its interactome. The methodological framework established here can be readily applied to other tissues and disease models in which Leucettinib-21 shows therapeutic potential, thereby facilitating the exploration of its mechanism of action and supporting its continued development as a candidate therapeutic agent.

Experimental Section

Chemistry

Syntheses of Immobilizable Leucettinibs

The synthesis and chemical characterization of intermediate and final compounds presented here are described in the S.I. section. The two global synthetic routes are described above (Schemes and ). All compounds are >95% pure by HPLC.

Preparation of Immobilized Leucettinib Analogs

Bead buffer: 50 mM Tris pH 7.4, 5 mM NaF, 250 mM NaCl, 5 mM EDTA, 5 mM ethylene glycol-bis (β-aminoethyl ether)-N,N,N,N-tetraacetic acid, tetrasodium salt (EGTA), 0.1% Nonidet-P40, 0.05% NaN₃ and protease inhibitor cocktail (cOmplete Sigma 11697498001, 1×).

Blocking buffer: 1 M ethanolamine in coupling buffer, pH 10.

Coupling buffer: 0.1 M H₃BO₃, 0.2 M NaCl, pH 10.

Washing buffer: 0.1 M CH₃COONa, pH 4.

CNBr Activated Sepharose 4B beads (Cytiva 17–0430–01) were first washed for 30 min with 1 mM HCl (200 mL/g beads) to remove the lactose used in the beads lyophilization process. Beads swell immediately, 1 g of lyophilized powder yielding about 3.5 mL of packed beads. Beads were then washed extensively with water and twice with coupling buffer. The ligand, solubilized in DMSO:coupling buffer (1:1), was added to the beads (1, 3.3, or 10 μmol/mL of packed beads), and the mixture was incubated overnight at 4 °C in the dark, under constant rotation. Beads were washed twice with coupling buffer and incubated with blocking buffer for 2 h at room temperature (RT), in the dark, to saturate the potentially remaining active sites. Beads were then washed 4 times (vortex and 1 min centrifugation at 3,000 rpm, 4 °C) with, alternately, 10 volumes coupling buffer and 10 volumes washing buffer, and finally twice with bead buffer. Beads were stored as a suspension in bead buffer at 20% (packed bead volume/total volume) at 4 °C, in the dark.

Biology–Biochemistry

Preparation of Organs and Lysates

Homogenization buffer: 60 mM β-glycerophosphate, 15 mM p-nitrophenylphosphate, 25 mM 3-(N-morpholino)-propanesulfonic acid (MOPS) pH 7.2, 15 mM EGTA, 15 mM MgCl₂, 2 mM dithiothreitol (DTT), 1 mM Na₃VO₄, 1 mM NaF, 1 mM phenylphosphate disodium, 0.5% Nonidet-P40 and protease inhibitor cocktail (cOmplete, Sigma 11697498001, 1×).

Sprague–Dawley rats (3 males (380–390 g), 3 females (230–240 g), 2 months old), were euthanized using carbon dioxide and immediately dissected. Sixteen organs were collected (brain hemispheres, colon (rinsed with physiologic media), heart (rinsed with physiologic media), intestine (rinsed with physiologic media), kidneys, liver (left lobe), lung, muscle (gastrocnemius), ovary, pancreas, skin, spleen, stomach, testis, thymus) and snap-frozen in liquid nitrogen and stored at −80 °C until use. Organs were reduced to powder while kept frozen with liquid nitrogen and dry ice and stored at −80 °C.

Organ powders were weighed (150 mg (kidneys and livers) or 300 mg (other organs)), and homogenized in homogenization buffer (1.5 mL/300 mg powder) either for 20 min under constant agitation at 4 °C, and sonicated (Branson SLPe 4C15; 1 min cycle: 10 s ON, 10 s OFF, 20% amplitude); or powders were homogenized by vortexing 10 s and then sonicated three times. Homogenates were either centrifuged for 10 min at 13,000 rpm at 4 °C or ultracentrifuged for 10 min at 100.000 g at 4 °C (Sorvall MTX 150). The supernatant containing proteins was kept and stored at −80 °C until use or used immediately. The protein concentration was determined with BioRad DC Protein Assay.

Affinity Chromatography on Immobilized Leucettinib Analogs

100 μL of the beads + ligand 20% suspension in bead buffer were transferred into a 1.5 mL tube thus providing a final volume of 20 μL packed beads. Beads were washed once with 1 mL of bead buffer. Tissue extract supernatants (0.5, 1, or 2 mg total proteins) were added to beads and volumes were adjusted to 1 mL with bead buffer. The tubes were rotated at 4 °C for 30 min in the dark. After a quick spin at 3,000 rpm and removal of the supernatant, beads were washed 5 times with bead buffer before addition of 50 μL of LDS (lithium dodecyl sulfate) 1× (Invitrogen) containing 100 mM DTT to elute the proteins from the beads. To study unbound proteins, 50 μL of supernatant were kept and mixed with 50 μL of LDS 2× containing 200 mM DTT. Proteins were denaturated at 70 °C for 10 min. Proteins were analyzed by NuPAGE Bis-Tris 4 to 12% and Western blotting or silver staining.

Competition Experiments

An excess of free Leucettinib-21 or iso-Leucettinib-21 was used to compete with the ligand immobilized on the beads. Solutions ranging from 0.02 to 200 μM of Leucettinib-21 or iso-Leucettinib-21 were prepared in bead buffer. Rat brain proteins were diluted into bead buffer to reach 0.5 mg of total proteins per tube. Protein solutions and free Leucettinib-21 or iso-Leucettinib-21 were mixed (1:1) to obtain solutions at concentrations ranging from 0.01 to 100 μM free competing ligand and incubated for 10 min in darkness at 4 °C. DMSO was used as a control. The proteins were then subjected to affinity chromatography on agarose-immobilized Leucettinib-21 as described above.

Immobilized Metal Affinity Chromatography (IMAC)

IMAC lysis buffer: 60 mM β-glycerophosphate, 15 mM p-nitrophenylphosphate, 25 mM MOPS pH 7.2, 1 mM EGTA, 15 mM MgCl₂, 1 mM Tris­(2-carboxyethyl)­phosphine (TCEP), 1 mM Na₃VO₄, 1 mM NaF, 1 mM phenylphosphate disodium, 0.1% Nonidet-P40, 10 mM imidazole and protease inhibitor cocktail (cOmplete EDTA-free, Sigma 11873580001, 1×).

IMAC washing buffer: IMAC lysis buffer with 40 mM imidazole.

100 μL of HisPurJust Ni-NTA magnetic beads (Thermo Scientific) were used per tube. Beads were washed twice with 500 μL of IMAC washing buffer and once with 500 μL of IMAC lysis buffer. Either 0.1 mg or 0.5 mg of proteins were added to the beads in a final volume of 1 mL.

The tubes were rotated at 4 °C for 30 min. After a quick spin at 3,000 rpm and removal of the supernatant (or collection to study unbound proteins), IMAC beads were washed 3 times with 500 μL of IMAC washing buffer before addition of 50 μL of LDS 1× (Invitrogen), 100 mM DTT. Following heat denaturation (10 min at 70 °C), the bound proteins were analyzed by NuPAGE Bis-Tris 4 to 12% and Western blotting or silver staining.

Double Affinity Chromatography: IMAC Followed by Leucettinib-21/iso-Leucettinib-21 Beads

BB IMAC: Bead buffer (50 mM Tris pH 7.4, 5 mM NaF, 250 mM NaCl, 1 mM EDTA, 1 mM ethylene glycol-bis (β-aminoethyl ether)-N,N,N,N-tetraacetic acid, tetrasodium salt (EGTA)), 0.5% Nonidet-P40, 0.05% NaN₃ and protease inhibitor cocktail (cOmplete EDTA-free, Sigma 11873580001, 1×).

BB IMAC lysis buffer: BB IMAC, 10 mM imidazole.

BB IMAC washing buffer: BB IMAC without protease inhibitor cocktail, 40 mM imidazole.

BB IMAC elution buffer: BB IMAC, 500 mM imidazole.

2 mL of HisPur Ni-NTA Superflow (Thermo Scientific) were used in spin columns of 10 mL. Beads were washed twice with 3 mL of BB IMAC washing buffer and once with 3 mL of BB IMAC lysis buffer.

Beads were resuspended in BB IMAC lysis buffer (1:1). Nine mg of rat brain extract proteins were added to reach a final concentration of 2 mg/mL within 4.5 mL. The tubes were closed with parafilm and incubated horizontally at 4 °C for 30 min. After a quick spin at 3,000 rpm and removal of the supernatant, beads were washed 3 times with 3 mL of BB IMAC washing buffer. Bound proteins were eluted with 1.44 mL of BB IMAC elution buffer (10 min at 4 °C, horizontal platform). IMAC elution is then either transferred onto Leucettinib-21 or iso-Leucettinib-21 beads, or mixed directly with homemade Laemmli buffer. In parallel, affinity chromatography on LCTB, iso-LCTB or ethanolamine beads was performed as previously explained in the “Affinity chromatography on immobilized Leucettinib analogs” section, except that the Bead buffer washing did not contain Nonidet P40 and the proteins were eluted with Laemmli buffer (4% SDS, 20% glycerol, 10% DTT, 125 mM Tris HCl pH 6.8, 0.01% bromophenol blue) (95 °C, 5 min) to be compatible with mass spectrometry analysis.

Immunoprecipitation of DYRK1A

IP washing buffer: 50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 5 mM EDTA, 5 mM EGTA, 0.1% Nonidet-P-40.

IP binding buffer: 50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 5 mM EDTA, 5 mM EGTA, 5 mM NaF, 0.1% Nonidet-P-40 and protease inhibitor cocktail (cOmplete, Sigma 11697498001, 1×).

IP elution buffer: 0.1 M glycine, pH 3.0, 0.1% Nonidet-P40.

Brain proteins were extracted on the day of the experiment from two male Sprague–Dawley rats in Homogenization buffer as described in “Preparation of Organs and Lysates section”. A volume of 240 μL of Dynabeads Protein G (Thermo Fisher, 100004D) was washed twice with 1.2 mL IP wash buffer and once with 1.2 mL IP binding buffer. For each wash, the tube stayed 1 min onto the magnet to separate the beads from the solution. First, the crude extract (2.55 mg) was incubated for 60 min at 4 °C with 120 μL Dynabeads protein G in a final volume of 770 μL IP binding buffer. All incubation steps were performed under rotation. The precleared extract was then mixed with 15 μg DYRK1A antibodies targeting either the N-terminal (D1694, Sigma) or the C-terminal (7D10, Abnova, H00001859-M01) regions of the protein. All samples were in a final volume of 900 μL loaded into1.5 mL tubes. After a 60 min incubation at 4 °C, the antibody-protein samples were added to the previously washed beads (3.6 mg) and incubated for another 30 min. After removing of the unbound proteins, beads were washed once with 600 μL IP wash buffer, then four times with PBS buffer. During the last wash, the solution was moved into a clean 1.5 mL tube and separated in two tubes, 25% for mass spectrometry analysis and the remaining beads for Western blot analysis. For mass spectrometry, the last wash supernatant was removed and beads were dried out for 15 min before freezing until analysis. For Western blotting, immunoprecipitated proteins were eluted with 60 μL IP elution buffer, 375 mM Tris-HCl, pH 6.8, 0.03% bromophenol blue; 1:3, v:v. Proteins were denaturated by heating at 95 °C for 5 min. As a negative control, free beads without antibodies were used. Experiments under the three conditions (IP Control, IP D1694, IP 7D10) were run in five independent replicates.

On-bead digestion

The proteins retained on the beads were reduced 30 min at 56 °C with 5 mM TCEP in Urea 2M, 0.1 M Tris pH8, then alkylated 30 min at RT with 10 mM IAA (Iodoacetamide). Trypsin (Promega, V511) digestion was carried out at 37 °C, OVN with 500 ng enzymes. After acidification with TFA, the peptide supernatant was desalted on C18 spin column and dried on Speed-Vacuum before LC-MS/MS analysis.

KinAffinity/LC-MS Kinome Profiling

Powdered rat brain was dissolved in lysis buffer (20 mM HEPES, pH 7.5, 150 mM NaCl, 0.25% Triton X-100, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 10 mM NaF, 1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/mL aprotinin, 10 μg/mL leupeptin), and incubated on a rotator for 30 min at 4 °C. Lysate was centrifuged at 14,000 rpm at 4 °C. Supernatant was kept and proteins from the pellet were extracted using lysis buffer with 400 mM NaCl, and 1.5 mM MgCl₂. Debris were removed by centrifugation. Both supernatants were combined to one lysate prior to protein determination using the bicinchoninic acid (BCA) assay (ThermoFisher).

2 mg of protein lysate were used per data point for binding to the KinAffinity resin (Evotec), which consists of five broad selectivity ATP-competitive kinase inhibitors. Leucettinib-21 was used for competition at several concentrations between 3 nM and 30 μM. Incubation was performed for 2.5 h at 4 °C. Beads were washed twice with lysis buffer containing 150 mM NaCl and phosphate-buffered saline (PBS) followed by an on-bead digestion of bound proteins with trypsin/endoproteinase LysC in sodium deoxycholate (SDC) buffer. Samples were acidified and desalted using styrenedivinylbenzenereversed phase sulfonate (SDB-RPS) in Stagetip format, and recovered peptides were analyzed by LC-MS/MS analysis on an Orbitrap mass spectrometer (ThermoFisher).

Raw data generated by LC-MS/MS were searched against a combined forward and reverse (decoy) peptide database. MaxQuant was employed to identify proteins and obtain quantitative abundance data for all in vitro associations with immobilized and soluble compounds. For the matter of covering as many kinases as possible, we used the mouse fasta database for protein annotation (the mouse proteome has close to 98% sequence overlap with the rat proteome but contains twice the number of reviewed proteins compared to rat). Swissprot served for database searches (version 2024_01_mouse_swissprot_and_varsplic). The quantitative competition experiments allow to determine IC₅₀ values, at which as many target protein molecules are bound to the free compound as to the KinAffinity resin. The final K _d values for the free compound are calculated for each target protein using the Cheng-Prusoff equation.

Electrophoresis and Western Blotting

Cell pellets were lysed on ice in homogenization buffer and centrifuged at 17,000 g for 1 min at 4 °C. Protein extracts were mixed (1:1 v/v) with sample buffer (2× NuPAGE LDS sample buffer, 200 mM DTT). Following heat denaturation, 30 μg of proteins were loaded on NuPAGE precast 4–12% Bis-Tris protein gels (ThermoFisher). Electrophoresis was performed in MOPS buffer. Rapid blot transfers were carried out at 2.5 A/25 V for 7 min on Trans-Blot Turbo 0.2 μm nitrocellulose membrane (BioRad). Membranes were blocked in skimmed milk (5% Regilait in Tris Buffered Saline with 0.1% Tween (TBS-T)) for 1 h. Membranes were then washed 3 times in TBS-T and incubated with primary antibodies overnight at 4 °C. Two antibodies directed against DYRK1A were used: 7D10 (Abnova, H00001859-M01; 1:500); DYRK1A D1694 (Sigma, D1694, 1:1,000). Other antibodies were directed against DYRK1B (Cell signaling, D40D1(#5672), 1:1,000), GSK-3α/β (Enzo Life Sciences, ADI-KAM-ST002, clone 1H8, 1:2,000), DCAF7/WDR68 (Abcam, Ab138490, 1:2,000), PRKG2 (Proteintech, 55138–1-AP, 1/1,500), Glutamine synthetase GLUL (Proteintech, 66323–1-Ig, 1/2,000), CDKL4 (Sigma, SAB4501511, 1/5,000), CSNK2A (Proteintech, 10992–1-AP, 1/10,000), CSNK2B (Proteintech, 22418–1-AP, 1/1,500), FAM53C (Miyata), GAK (Cell Signaling, E7W7Z, 1/1,000). Finally, membranes were incubated for 1 h at RT with HRP (horseradish peroxidase)-conjugated secondary antibodies (1:2,500) (goat antirabbit, or goat antimouse antibodies; Bio-Rad, 1706515 and 1706516, respectively) in TBS-T, and chemiluminescent detection was achieved with homemade ECL-Tris buffer (100 mM Tris pH 8.5, 0.01% H₂O₂, 0.225 mM p-coumaric acid, 1.25 mM luminol) with Fusion Fx7 camera software.

Silver Staining

Electrophoresis gels were processed as described in the Pierce Silver Stain Kit instructions (ThermoFisher). Briefly: fixation with 30% ethanol:10% acetic acid (2 × 15 min), washes with 10% ethanol, then with water (2 × 5 min) followed by sensitization (Sensitizer Working Solution, 1/500 in water, 1 min), washes with water (2 × 1 min), staining (Silver Stain Working Solution, 1/50 of Enhancer into Stain solution, 30 min), quick washes with water (2 × 20 s) and Development (Developer Working Solution, 1/50 of Enhancer into Developer solution, 1 min). Staining was stopped by addition of 5% acetic acid (2 washes of 20 s then 10 min).

Mass Spectrometry

Mass Spectrometry Analysis

Samples were analyzed using an Ultimate 3000 nano-RSLC (Thermo Scientific, San Jose California) coupled in line with a quadrupoleorbitrap Exploris 480 via a nanoelectrospray ionization source (Thermo Scientific, San Jose California) and the FAIMS pro interface. The tryptic peptides were loaded on a C18 Acclaim PepMap100 precolumn (75 μm × 2 cm, 3 μm, 100 Å, Thermo Fisher Scientific) for 2 min at 15 μL/min with 2% ACN, 0.1% FA in H₂O and then separated on a C18 PepMap nanocolumn (75 μm ID × 25 cm, 2.6 μm, 150 Å, Thermo Fisher Scientific) with a 25 min linear gradient from 7% to 45% buffer B (A: 0.1% FA in H₂O; B: 0.1% FA in 80% ACN) at 450 nL/min followed by a regeneration step at 95% B and a equilibration at 7% B. The total chromatography was 35 min. The mass spectrometer was operated in positive ionization mode in Data-Dependent Acquisition (DDA) for one FAIMS compensation voltages (CV = −45 V). The DDA cycle consisted of one survey scan or MS1 (330–1400 m/z, 90,000 fwhm) followed by MS2 spectra (HCD; 28% normalized energy; 8 m/z window; Scan Range 140–1400 m/z; 22,500 FWMH). The cycle time was limited at 1.6 s. The Automatic Gain Control (AGC target) were set to 3E6 and 1E5 respectively for MS1 and MS2 scans, and the maximum injection time (IT) was set to 50 ms for both scan modes. Unassigned and single charged states were rejected. The exclusion duration was set for 30s with mass width was ±10 ppm.

Mass Spectrometry Data Processing

Proteins were identified with Proteome Discoverer 2.5 software (Thermo Fisher Scientific) and Rattus norvegicus proteome database (UniProt, reviewed, release 2025_01_17 with 22367 entries). Precursor and fragment mass tolerances were set at 10 ppm and 0.02 Da respectively, and up to 2 missed cleavages were allowed. Oxidation (M) and N-terminal Acetylation were set as variable modification, and Carbamidomethylation (C) as fixed modification. Peptides were filtered with a false discovery rate (FDR) at 1%, rank 1. Proteins were quantified with a minimum of 1 unique peptide based on the XIC (sum of the Extracted Ion Chromatogram). The quantification values were exported in Perseus for statistical analysis involving a log[2] transform, normalization on the median, imputation of missing values and followed by t test calculation (FDR = 0.01).

In Vitro Radiometric Kinase Profiling Assays (Reaction Biology)

Kinase assays were first performed by Reaction Biology GmbH (Freiburg, Germany. www.reactionbiology.com). The IC₅₀ profile of all compounds was determined using 12 protein kinases (CDK5/p25, CK1ε, CLK1, 2, 3, 4, DYRK1A, 1B, 2, 3, 4, GSK-3β). IC₅₀ values were measured by testing 10 concentrations (from 30 nM to 10 μM) of each compound in singlicate. In addition, Leucettinib-21 was tested on PRKG1, PRKG2, and CSNK2 as well as all the kinases identified with the KinoAffinity beads (Figure ).

Compounds were provided as 1 mM stock solutions in 100% DMSO. Prior to testing, the stock solutions were subjected to a serial, semilogarithmic dilution using 100% DMSO as a solvent. This resulted in 10 distinct concentrations, with a dilution end point of 3 × 10 nM/100% DMSO, with 100% DMSO as controls. In the process, 90 μL H₂O were added to each well of each compound dilution plate. To minimize potential precipitation, H₂O was added to each plate only a few min before the transfer of the compound solutions into the assay plates. The plate was shaken thoroughly, resulting in a compound dilution plate/10% DMSO.

For the assays (see below), 5 μL solution from each well of the compound dilution plates/10% DMSO were transferred into the assay plates. The final volume of the assay was 50 μL. All compounds were tested at 10 final assay concentrations in the range from 30 nM to 10 μM. The final DMSO concentration in the reaction cocktails was 1% in all cases.

All recombinant protein kinases provided by Reaction Biology were expressed in Sf9 insect cells or in E. coli as recombinant GST-fusion proteins or His-tagged proteins, either as full-length or enzymatically active fragments. All kinases were produced from human cDNAs and purified by either GSH-affinity chromatography or immobilized metal. Affinity tags were removed from a number of kinases during purification. The purity of the protein kinases was examined by SDS-PAGE/Coomassie-Blue staining, the identity was checked by mass spectroscopy.

A radiometric protein kinase assay (33 PanQinase Activity Assay) was used for measuring the kinase activity of the 12 protein kinases. All kinase assays were performed in 96-well FlashPlates from PerkinElmer (Boston, MA, USA) in a 50 μL reaction volume. The reaction cocktail was pipetted in four steps in the following order: 25 μL of assay buffer (standard buffer/[γ-³³P]-ATP), 10 μL of ATP solution (in H₂O), 5 μL of test compound (in 10% DMSO), 10 μL of enzyme/substrate mixture. The assay for all protein kinases contained 70 mM HEPES-NaOH pH 7.5, 3 mM MgCl₂, 3 mM MnCl₂, 3 μM Na-orthovanadate, 1.2 mM DTT, 50 μg/mL PEG20000, ATP (variable concentrations, corresponding to the apparent ATP-Km of the respective kinase, Table S13), [γ-³³P]-ATP (approximately 6.5 × 10⁵ cpm per well), protein kinase (variable amounts), and substrate (variable amounts). The reaction cocktails were incubated at 30 °C for 60 min. The reaction was stopped with 50 μL of 2% (v/v) H₃PO₄, plates were aspirated and washed two times with 200 μL 0.9% (w/v) NaCl. Incorporation of ³³Pi was determined with a microplate scintillation counter (Microbeta, Wallac). The IC₅₀ values for all compounds were calculated from the dose–response curves

As a parameter for assay quality, the Z′-factor for the low and high controls of each assay plate (n = 8) was used. Reaction Biology′s criterion for repetition of an assay plate is a Z′-factor below 0.4.

DYRK1A Activity Measurement

Dosage of DYRK1A activity was performed on organ lysates (3 mg/mL; supplemented with 10 μM Leucettinib-21 or 4.5 μL DMSO/mL of lysate) by the Laboratory of Biological Chemistry, team Metabolism, Pharmacochemistry and Neurochemistry (Paris Descartes University, CNRS UMR8601) as described previously.

To assess the activity of DYRK1A by HPLC, a Dansyl-conjugated peptide was designed, Dansyl–KKISGRLSPIMTEQ-COOH (Dan–peptide), the sequence of which is derived from the human forkhead transcription factor (FKHR). This transcription factor is known to be phosphorylated by DYRK1A on the Ser329 residue of the GRLSPIM motif. This peptide and its Ser329 phosphorylated product can be easily separated by reverse-phase HPLC and specifically quantified by detection of the fluorescence of the Dansyl moiety.

The purity and identity of the fluorescein-labeled peptide substrate (Dan–peptide) was initially assessed by reverse-phase HPLC (Prominence Shimadzu UFLC (ultrafast liquid chromatography) system interfaced with LabSolutions software). Samples were injected into a Nucleodur C18 column (length: 150 mm; internal diameter: 4.6 mm; particle size: 5 μm) at 45 °C. The mobile phase used for the separation consisted of two eluents; solvent A was water with 0.1% trifluoroacetic acid (TFA), and solvent B was acetonitrile with 0.1% TFA. Compounds were separated by an isocratic flow (85% A/15% B) rate of 1.0 mL/min. The products were monitored by fluorescence emission (λ: 537 nm) after excitation at λ 375 nm and quantified by integration of the peak absorbance area, employing a calibration curve established with various known concentrations of peptides.

Incubation of DYRK1A with the Dan–peptide in the presence of ATP leads to two peaks corresponding to the Dan-peptide (retention time of 5.5 min) and its phosphorylated form (retention time of 4.8 min). As expected, the amount of phosphorylated Dan-peptide calculated by integration of the peak (area under the curve, AUC) increases linearly with the time of incubation. Assays were performed in a 96-well ELISA plate in a total volume of 50 μL consisting of kinase buffer (50 mM Tris–HCl, 10 mM DTT, and 5 mM MgCl₂), 1 mM ATP, 100 μM Dan–peptide substrate, and tissue extracts (25 μg total protein extract). Briefly, samples containing the enzyme were preincubated with peptide substrate at 37 °C for 1 min, and the reaction was started by the addition of ATP. At different time points (up to 30 min), 50 μL of HClO₄ (15% in water) was added to stop the reaction, and 20 μL was automatically injected into the HPLC column. IC₅₀ values were calculated from the dose–response curves using Graphpad Prism 5.0.3.

Glutamine Synthetase Activity Assay

The inhibition of glutamine synthetase activity by LCTB-21 was assessed according to the manufacturer’s instructions (Abcam, ab284572), using purified recombinant GLUL (Abcam, ab222354) and the Glutamine Synthetase Activity Assay Kit (Colorimetric) (ab284572) (K2056). In this assay, GLUL hydrolyzes glutamate to glutamine and ADP. ADP in the subsequent enzymatic reaction, in the presence of ADP Converter, ADP Developer Mix and ADP Probe, forms a colorimetric product that is measured at OD 570 nm.

Glossary

Abbreviations

ACACA

acetyl-CoA carboxylase α

ACADL

acyl-CoA dehydrogenase long chain

ACADM

acyl-Coa dehydrogenase medium chain

ACN

acetonitrile

AD

Alzheimer’s disease

ADK

adenosine kinase

AGK

acylglycerol kinase

ALL

acute lymphoblastic leukemia

ALDH5A1

aldehyde dehydrogenase 5 family member A1

AP-MS

affinity purification mass spectrometry

AMKL

acute megakaryoblastic leukemia

AP2

adaptor related protein complex 2

ARF-GEF

ADP-ribosylation factor guanine nucleotide exchange factor

ATP

adenosine triphosphate

BCA

bicinchoninic acid

BCKDK

branched chain keto acid dehydrogenase kinase

BRAG2

brefeldin A-resistant ARF-GEF2

CaMKK, CaMK2

Ca²⁺/calmodulin-dependent protein kinase kinase

CDK

cyclin-dependent kinase

CDC37

cell division cycle 37

CDC42BPB

Cdc42 binding protein kinase β

CDKL

cyclin-dependent kinase like

CE

crude extract

cGMP

cyclic monophosphate

CK1ε

casein kinase 1ε

CK2α/ CSNK2A1

casein kinase 2α 1

CSNK2A2

casein kinase 2α 2

CLKs

cdc2-like kinases

C-ter

C-terminal domain

CLT

Clathrin Light Chain

DCAF7

DDB1 and CUL4 associated factor 7

DDA

data-dependent acquisition

DMSO

dimethyl sulfoxide

DNAJ

DnaJ Heat shock protein family (hsp40) member

DS

Down syndrome

DREAM

dimerization partner

RB-like

E2F and multivulval class B

DSCAM

down syndrome cell adhesion molecule

DTT

dithiothreitol

DYNLL2

dynein light chain LC8-type 2

DYNLL1

dynein light chain LC8-type 1

DYRK1A

dual-specificity, tyrosine phosphorylation-regulated kinase 1A

EGTA

ethylene glycol-bis (β-aminoethyl ether)-N,N,N,N-tetraacetic acid, tetrasodium salt

FAM53C

family with sequence similarity 53 member C

FAM117B

family with sequence similarity 117 member B

FDA

Food and Drug Administration

FDR

false discovery rate

FN3KRP

fructosamine-3-kinase-related protein

FKHR

forkhead transcription factor

FL-DYRK1A

full-length DYRK1A

GABRA1

γ-aminobutyric acid type A receptor subunit alpha1

GAK

cyclin G-associated kinase

GCDH

glutaryl-CoA dehydrogenase

GIT1

ARF GTPase-activating protein GIT1

GK

glycerol kinase

GLCCI1

glucocorticoid induced 1

GLUL

glutamine synthetase, glutamate-amonia ligase

GO

gene ontology

GSK-3

glycogen synthase kinase-3

HA

hemagglutinin

HKDC1

hexokinase domain containing 1

HPLC

high performance liquid chromatography

HRP

horseradish peroxidase

IC ₅₀

half-maximal inhibitory concentration

IMAC

immobilized metal affinity chromatography

IQSEQ1

IQ motif and SEC7 domain-containing protein 1

LCTB

leucettinib

LDS

lithium dodecyl sulfate

LMW-DYRK1A

low molecular weight form of DYRK1A

LZTZ2

leucine zipper, putative tumor suppressor 2

MAP

microtubule-associated protein

MAPPIT

mammalian protein–protein interaction trap

MCODE

molecular complex detection

MIBs

multiplexed inhibitor beads

MOPS

3-(N-morpholino)­propanesulfonic acid

MPN

myeloproliferative neoplasm

MS

mass spectrometry

N-ter

N-terminal domain

Ni-NTA

nickel-nitrilotriacetic acid

Opa1

OPA1 mitochondrial dynamin like GTPase

PBS

phosphate-buffered saline

PEG

poly­(ethylene glycol)

PMSF

phenylmethylsulfonyl fluoride

PPI

protein-protein interactions

PRKACA

cAMP-dependent protein kinase catalytic subunit α

PRKACB

cAMP-dependent protein kinase catalytic subunit β

PRKG2

cGMP-dependent protein kinase 2

PRKAR1B

protein kinase cAMP-dependent type I regulatory subunit β

RAB

member RAS oncogene family

RPS6KA4

ribosomal protein S6 kinase A4

PHB2

prohibitin 2

PSD

PH and SEC7 domain-containing protein 1

SDB-RPS

styrenedivinylbenzenereversed phase sulfonate

PTSA

p-toluenesulfonic acid

SDC

sodium deoxycholate

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

SPTAN1

spectrin α, non-erythrocytic 1

STAT3

signal transducer and activator of transcription 3

SNAr

nucleophilic aromatic substitution

S.I.

supporting information

TAOK1

thousand and one amino acid (TAO) kinase 1

TBST

tris buffered saline with 0.1% Tween

TCEP

tris­(2 carboxyethyl)­phosphine

TEA

triethylamine

TFA

trifluoroacetic acid

TP53

tumor protein P53

TROAP

trophinin associated protein

TUBB4

tubulin β 4

UFLC

ultrafast liquid chromatography

USP5

ubiquitin specific peptidase 5

VAMP1

vesicle associated membrane protein 1

VCPIP1

deubiquitinating protein VCIP135

VDAC2

voltage dependent anion channel 2

VTI1B

vesicle transport through interaction with T-SNAREs 1B

WDR68

tryptophan-aspartic (WD) repeat-containing protein 68

ZWINT

ZW10 interactor

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.6c00066.

• Description of the 18 synthesized compounds reported here together with their corresponding ¹H and ¹³C NMR analyses. It also details the identification of glutamine synthetase (GLUL) as an LCTB-21 interactor and the effect of LCTB-21 on GLUL activity, representative rat brain protein fractions purified by affinity chromatography and analyzed by AP-MS, the Metascape analysis of proteins uniquely associated with LCTB beads, and the immunoprecipitation of DYRK1A from rat brain extracts using anti-DYRK1A antibodies (PDF)

• Detailed lists of proteins identified using the various chromatography methods and their comparisons (immobilized LCTB, iso-LCTB, IMAC followed by immobilized LCTB or iso-LCTB beads, and the KinAffinity assay), lists of proteins immunoprecipitated with the 7D10 and D1694 DYRK1A antibodies and their comparisons, assay parameters for the tested protein kinases in the radiometric assay, and the molecular formula strings (SMILES) of the synthesized compounds (XLSX)

∇.

E.D. and C.S. contributed equally to this work. E.D. designed and synthesized the affinity chromatography reagents. C.S. coupled all the Leucettinibs to the beads, extracted the proteins from rat organs, ran all the affinity chromatography and competition experiments with Leucettinibs and nickel beads, ran the Western blots and the GLUL activity assay, prepared and validated the samples for MS. J.D. ran the DYRK1A activity assays. B.M. ran the MS experiments. Y.M. provided the FAM53C antibody. J.C. and Y.H. provided dissected rat organs. G.H. ran the IP experiments and prepared the samples for MS. L.M. searched and analyzed the literature, provided guidance in the design of immobilized inhibitors, supervised and analyzed the in-house and outsourced assays, the results, and the literature, and wrote the article. M.F.L. supervised the affinity chromatography experiments, analyzed the results, and wrote the article. All authors read and contributed to the manuscript.

This research was supported by grants from the “Fondation Jérôme Lejeune”, the “Agence Nationale pour la Recherche (ANR)” (DYRK-DOWN, TRANSBIOROYAL, KINHIB-DIAB), and France 2030–BpiFrance (i-Nov vague 9, Leucettinib-21 project). This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 848077 (GO–DS21), a EUROSTARS grant (T2DiaCURE) and the European Innovation Council Accelerator (EIC) Accelerator Program (DOWN-AUTONOMY project, 190138295), by the Interdisciplinary Thematic Institute IMCBio+, as part of the ITI 2021–2028 program of the University of Strasbourg, CNRS, and Inserm, by IdEx Unistra (ANR-10-IDEX-0002), SFRI-STRAT’US project (ANR-20-SFRI-0012), EUR IMCBio (ANR-17-EURE-0023), INBS PHENOMIN (ANR-10-IDEX-0002–02) under the framework of the France 2030 Program and by grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Views and opinions expressed here are those of the authors only and do not necessarily reflect those of the European Union, which cannot be held responsible for the information it contains.

The authors declare the following competing financial interest(s): L.M. is a founder of Perha Pharmaceuticals. E.D. and L.M. are co-inventors in the Leucettinibs patents. L.M, M.L., and E.D. are stockholders of Perha Pharmaceuticals.