Structure-Based Design of a Potent and Selective YTHDC1 Ligand

N6-Adenosine methylation (m6A) is a prevalent post-transcriptional modification of mRNA, with YTHDC1 being the reader protein responsible for recognizing this modification in the cell nucleus. Here, we present a protein structure-based medicinal chemistry campaign that resulted in the YTHDC1 inhibitor 40, which shows an equilibrium dissociation constant (Kd) of 49 nM. The crystal structure of the complex (1.6 Å resolution) validated the design. Compound 40 is selective against the cytoplasmic m6A-RNA readers YTHDF1–3 and YTHDC2 and shows antiproliferative activity against the acute myeloid leukemia (AML) cell lines THP-1, MOLM-13, and NOMO-1. For the series of compounds that culminated into ligand 40, the good correlation between the affinity in the biochemical assay and antiproliferative activity in the THP-1 cell line provides evidence of YTHDC1 target engagement in the cell. The binding to YTHDC1 in the cell is further supported by the cellular thermal shift assay. Thus, ligand 40 is a tool compound for studying the role of YTHDC1 in AML.


■ INTRODUCTION
Post-transcriptional changes of eukaryotic mRNA play a pivotal role in cellular processes. 1Before leaving the nucleus, mRNA undergoes a series of chemical alterations, such as methylation, acetylation, and splicing. 2,3These modifications directly affect mRNA stability, processing, and translation efficiency. 4,5Recent investigation of post-transcriptional events have given rise to a dynamic research field known as epitranscriptomics. 3Among the diverse mRNA modifications in the human transcriptome, the most prevalent and extensively studied is the N 6 methylation of adenosine (m 6 A). 6 This reversible process is facilitated by proteins referred to as "writers", such as METTL3/14, 7,8 while the demethylation is achieved by "erasers" like FTO or ALKBH5. 9,10Our research group has made recent contributions to this field through the development of small-molecule inhibitors targeting METTL3/14, demonstrating the modulation of interconnected cellular events. 11,12ecognition of the m 6 A modification is mediated by "readers" that subsequently impact downstream processes. 13he YTH family, comprising five proteins (YTHDC1, YTHDC2, and YTHDF1−3), is the best characterized family of m 6 A readers. 14 Due to their crucial role in diverse biological processes, these proteins hold great promise as therapeutical targets. 15,16Predominantly localized in the nucleus, YTHDC1 is responsible for the regulation of pre-mRNA splicing and mRNA export from the nucleus. 17A growing number of reports associate the activity of YTHDC1 with various biological functions, including embryonic development, neuronal development, and others. 18Furthermore, its critical role has been identified in various types of cancer. 19,20ere, we focus on the function of YTHDC1 in acute myeloid leukemia (AML).The gene was identified as essential for AML in a genome-wide CRISPR knockout screening. 20In another study, Sheng et al. indicate an oncogenic role of YTHDC1 in the regulation of leukemogenesis by MCM4, a component of the MCM complex responsible for DNA replication. 21Chen et al. showed that m 6 A-dependent export of mRNA from the nucleus, mediated by YTHDC1, in combination with lncRNA MALAT1 promotes the expression of fusion genes such as PML-RARA, MLL-ENL, or MLL-AP9 characteristic for particular types of leukemia. 22These examples describe the distinct role of YTHDC1 in blood cancers, indicating the need for therapeutics targeting this specific m 6 A reader protein.
Our group recently reported fragments binding to YTHDC1 or YTHDF2, respectively. 23,24Other studies in the literature have reported inhibition using rather promiscuous binders providing limited selectivity toward proteins of the YTH family, 25−27 except for a recent report which describes a selective YTHDC1 inhibitor identified by in vitro highthroughput screening followed by a structure-based optimization. 28n this study, we present a structure-based design campaign aimed at developing a potent and selective ligand of YTHDC1.Additionally, we provide biochemical evaluation based on the homogeneous time-resolved fluorescence assay (HTRF), isothermal titration calorimetry (ITC), thermal shift assay (TSA), and protein X-ray crystallography.We also report a direct connection between YTHDC1 inhibition and antiproliferative activity on acute myeloid leukemia cell lines (THP-1, MOLM-13, NOMO-1).

■ RESULTS AND DISCUSSION
The present medicinal chemistry campaign builds upon our initial in silico screening for YTHDC1. 23,29The structure-based medicinal chemistry optimization started here with the ligandefficient fragment 1 (Figure 1C) which had been characterized by biochemical assays and crystallography. 23Structurally, this fragment contains a pyrazolo [4,3-d]pyrimidine core that mimics the natural ligand m 6 adenine (we use m 6 adenine for nucleotide, in contrast to m 6 A for nucleoside).The recognition of fragment 1 is achieved by the aromatic cage consisting of two tryptophan residues (Trp428, Trp377) and five hydrogen bonds between the pyrazolopyrimidine ring and the binding pocket (Asn367, Asp476, backbone N-H of Asn363, backbone C=O of Ser378, and structural water-bridging side chains of Trp377 and Asp476).Notably, unlike the natural ligand, the pyrazolopyrimidine 1 forms an additional hydrogen bond with the side chain of Asp476 (Figure 1, PDB: 7P8F).This interaction significantly enhances the YTHDC1 affinity of fragment 1 by nearly 10-fold compared to m 6 -adenine (IC 50 = 39 vs 306 μM). 23Upon evaluating the binding pose of pyrazolopyrimidine 1 within the binding pocket, we identified two positions on the aromatic core that are conducive to ligand growth and optimization.The first position lies between N 6 and N 4 of the pyrimidine ring.The substituent in this position leads to a small pocket, suggesting that a suitable substituent could occupy the space and enhance binding.The second substituent leads from C 3 toward a shallow, positively charged pocket that binds the negatively charged part of RNA in the natural ligand.
During the initial stages of the campaign, we identified that a chloride substituent located between N 6 and N 4 fits into the small pocket forming a halogen-hydrogen bond donor interaction 30 with the side chain hydroxyl of Ser362 (C− Cl•••H angle = 76°; Cl•••O distance = 3.8 Å) (Figure 1A), which substantially improves the binding affinity.(Note the difference in the numbering of purine and pyrazolopyrimidine core as shown in Figure 1C).This improvement was demonstrated by comparing the potency of nonchlorinated 2a and chlorinated m 6 adenine 2b resulting in IC 50 values of 306 and 37 μM, respectively.However, the adenine aromatic core cannot form a conventional hydrogen bond with Asp476, unlike pyrazolopyrimidine 1 moiety.Therefore, we decided to merge the two structural features and synthesize 5-chloropyrazolopyrimidine fragment 3.This compound exhibited improved potency together with ligand efficiency (IC 50 = 2 μM, LE = 0.60) and submicromolar equilibrium dissociation constant (K D = 146 nM, Figure S6) measured by isothermal titration calorimetry (ITC).As mentioned above, an alternative position for ligand growth involved extending or replacing the methyl substituent at position 3 of 5chloropyrazolopyrimidine 3. Considering the synthetic challenges, we decided to pursue the optimization using 2chloropurine 2b instead of 5-chloropyrazolopyrimidine 3.This approach offered the advantage of convenient derivatization at N 9 of the purine moiety, as compared to C 3 in the equivalent position of the pyrazolopyrimidine core.The primary rationale behind this choice was to focus on extending the purine scaffold and synthesizing only the most promising examples in combination with the 5-chloropyrazolopyrimidine 3.This strategy was based on the observation of essentially identical binding poses among the fragments (see Figure 1B) and the hypothesis that interactions outside the aromatic cage would be preserved across different fragments.
Ligand Growing�First Round of Optimization.In the initial screening phase, we opted to replace/extend the methyl group at position 9 with various aliphatic and aromatic rings, including both −CH 2 − bridged and nonbridged structures (Table 1).Notably, the substituents with a methylene linker (5, 6, 8, 9, and 11) exhibited low micromolar potency (IC 50 = 3, 6, 8, 11, and 9 μM, respectively) in comparison with compounds devoid of the methylene: 4 (IC 50 > 100 μM) and 7 (IC 50 = 47 μM).These results suggest the presence of van der Waals interaction between the rings and lipophilic residues (Leu380, Pro431, Met434).The flexibility provided by the methylene allows the aromatic ring to achieve better interactions within the binding site residues, which is not the case for compounds 4 and 7 (Figure 2A,B).However, there was one exemption among the tested molecules.Compound Table 1.2-Chloropurine Derivatives with Different Substituents on N 9 a Homogeneous time-resolved fluorescence (HTRF).b Ligand efficiency (kcal•mol −1 heavy atom count −1 ). 31 c Lipophilic ligand efficiency (pIC 50 − c log P). 31 d Growth inhibition 50 (GI 50 ) values after 72 h treatment (THP-1).
10 does not feature a methylene bridge and has a carboxyl group in meta position which resulted in enhanced potency (IC 50 = 11 μM) with respect to compounds 4 and 7.This observation indicates that the absence of hydrophobic interactions is compensated with ion−ion or dipole−ion interactions.Such interactions are likely to be formed between negatively charged carboxyl group and positively charged residues within the binding pocket.The significant impact of the carboxylic group is evident when comparing the potency of compound 4, featuring a nonsubstituted phenyl ring to its carboxylic acid derivative 10.
Second Round of Optimization�Benzyl Rings SAR Study.Due to its promising potency, the benzyl substituent was retained for further optimization.The easily accessible and commercially available nature of benzyl halides facilitated an efficient structure−activity relationship (SAR) study, enabling identifications of potent functionalization.The benzylic scaffolds we synthesized and tested could be divided into subgroups based on the position and number of substituents (Table 2).In the following, ortho, meta, and para positions refer to the substitution relative to the methylene connecting the purine heterocycle.
Among the ortho-substituted derivatives, compound 12, containing free amine moiety, and its sulfonamide derivatives (13, 14) exhibited an IC 50 value of 9, 0.46, and 2 μM, respectively.The potency of compound 13 could be potentially attributed to an additional hydrogen bond formed between the sulfonamide oxygen and the backbone N−H of Asp476 (Figure 2C).On the other hand, the presence of trifluoroacetic amide 15 and methoxy derivative 18 provided the IC 50 value above 100 and 13 μM, respectively.This suggested that the carbonyl of 15 is unable to adopt the favorable geometry orientation for hydrogen bond formation observed between the sulfonamide oxygen of 13 and Asp476.Although perfluorinated alkyl substituents (16, 17) exhibited lower potency in comparison with the most promising compound from this set (IC 50 = 2 and 3 μM versus 0.46 μM for compound 14), compound 16 showed antiproliferative activity against THP-1 cell line (GI 50 = 14 μM).Additionally, compounds 12, 13, 15, 16, and 18 were soaked to YTHDC1 enabling X-ray crystallography validation of their binding poses with high resolution (<1.5 Å).
Among the meta-substituted derivatives, our focus primarily centered on carboxylic acid derivatives, as encouraged by the aforementioned significant potency increase between compounds 4 and 10.To expand structural and functional diversity, we also evaluated purine analogues containing halogen and methoxy groups.As expected, meta-carboxylic acid analogue 20 demonstrated the highest affinity with an IC 50 value of 0.51 μM.In contrast, the tetrazole heterocycle 21, serving as carboxylic acid bioisoster, displayed lower affinity (IC 50 = 1 μM) compared to the free carboxylic acid compound   (22,  23) and methyl ester moiety 19 exhibited affinity (IC 50 = 2, 1, 1 μM) compared to nitrile group 26 (IC 50 = 2 μM) and did not surpass the potency of compound 20.We hypothesized that the improved binding toward YTHDC1 might be a result of ion−ion or ion−dipole interaction involving the positively charged side chain of Arg475.However, X-ray crystallography could not confirm this, due to the lack of density for both the carboxylic group and the Arg475 side chain.Regarding the antiproliferative activity against THP-1, the methyl ester derivative 19 exhibited low micromolar GI 50 (6.2μM) while carboxylic acid derivative 20, which showed greater potency in the HTRF assay, had a minimal effect (GI 50 > 100 μM), likely due to poor cell permeation of the charged carboxylic residue at physiological pH.Among the compounds tested, we also discovered that a chloride substituent in the meta position exhibited enhanced binding with an IC 50 value of 0.96 μM (compound 25) and displayed antiproliferative activity against THP-1 (GI 50 = 11 μM).This compound also exhibited a very favorable LE value (0.41).The improved affinity could potentially arise from the formation of a halogen bond.X-ray crystallography did not provide evidence for an interaction between the chlorine substituent and the backbone carbonyl of Ala432 for compound 25.However, the presence of halogen bond was later confirmed by similar compounds (31�Figure 2C, and 40�Figure 3A).On the other hand, a bigger and more lipophilic bromide substituent present in compound 24 resulted in a 2-fold decrease in potency (IC 50 = 2 μM) in comparison to compound 25.Because of the positive effect observed with carboxylic acid functionality, we also prepared derivatives of carboxylic acid (28, 29) and benzyl alcohol derivative 30 in the para position.However, these compounds did not exhibit improved binding compared to the ones with functional groups in meta position, presumably due to an unfavorable orientation of the substituents that lead toward solvent exposed area out of the pocket.
The combination of the most potent aromatic ring substitutions was tested by compound 31, namely, methylsulfonamide in ortho position (compound 13) and chloride in meta (compound 25), but did not improve potency with respect to the monosubstituted analogue 13.Although compound 31 showed an IC 50 of 0.45 μM, the presence of the sulfonamide moiety once again resulted in mediocre activity against THP-1 growth (GI 50 = 22 μM).On the other

Journal of Medicinal Chemistry
hand, compound 32 containing meta-chloro and meta-carboxyl substitution showed greater enhancement in binding resulting in the most potent compound (IC 50 = 0.18 μM).Unfortunately, the biochemical potency increase was not transferred into the improved antiproliferative activity (GI 50 > 100 μM) which is also true for its methyl ester analogue 33 (GI 50 = 9.6 μM).Furthermore, while the combination of paramethoxy and meta-chloro substitution provided submicromolar potency (compound 34, 0.59 μM), the presence of two chloro substituents in meta position diminished the binding (compound 35, IC 50 = 2 μM).
Headgroup Optimization.After optimizing the benzylic side chain and identifying suitable substituents, we shifted our focus back to the heterocyclic core optimization.For this reason, we explored further modifications on the purine fragment while maintaining the meta-chlorobenzyl substituent at N 9 (Table 3).First, we confirmed that the presence of a hydrogen bond between the methylamino group and the backbone of Ser378 is essential for the binding.This was verified by testing compounds 38 and 39, both containing a chlorine substituent at position 6 of the purine ring.The presence of chlorine atom instead of methylamino moiety resulted in a notable decrease in potency.Additionally, compound 36, bearing fluoro substituent instead of chloro (position 2 of purine ring), pointing toward a small lipophilic pocket, also exhibited weakened binding (IC 50 = 2 μM).Moreover, the substitution of the methylamino group with cyclopropyl amino moiety 37 also led to weakened binding, indicating that the aromatic cage of YTHDC1 is intolerant to increased bulkiness.
As outlined in our strategy, our primary objective was to combine the most promising (LE, GI 50 ) substitution pattern optimized on purine heterocycle with the 5-chloropyrazolopyrimidine core.Thus, we decided to obtain compound 40, which combines 5-chloropyrazolopyrimidine and chlorobenzyl substituent in position 3.As anticipated, X-ray analysis confirmed the interactions between compound 40 and YTHDC1, which are the same as observed with the purine analogue 25, along with an additional hydrogen bond with Asp476 (Figure 3A).Furthermore, compound 40 exhibits an IC 50 of 0.35 (Figure 3B), good LE (0.44), and antiproliferative activity against THP-1 (GI 50 = 3.2 μM), MOLM-13 (5.6 μM) and NOMO-1 (8.2 μM) (Figures 4A and S2A).We further confirmed the binding of compound 40 to YTHDC1 by thermal shift assay (TSA), where the protein exhibited a thermal stabilization upon binding of pyrazolopyrimidine 40, increasing the melting temperature by 12 °C at a compound concentration of 100 μM (Figure 3C).Moreover, an equilibrium dissociation constant (K D ) of 49 nM was measured by ITC for compound 40 and YTHDC1 (Figure 3D).
Next, we tested the selectivity of 40 against YTHDF1−3.Compound 40 displayed an IC 50 value of 89, 60, and 83 μM against YTHDF1, YTHDF2, and YTHDF3, respectively (Figure S2) which is about 200-fold difference in binding preference toward YTHDC1.To provide further evidence of the selectivity of compound 40, its activity was evaluated against a panel of 58 protein kinases.At a concentration of 2 μM, compound 40 did not display significant inhibition of kinase activity across the entire panel (Figure S1).
With the evidence of target engagement and selectivity against off-targets in hand, we decided to further evaluate compound 40 by comparing its antiproliferative activity against cancer cell lines with its toxicity toward noncancerous cells.The GI 50 value of compound 40 for THP-1 cancer (AML) cells is 3.2 μM (Figure 4A), whereas for human peripheral blood mononuclear cells (PBMCs, Figure 4B) and human embryonic kidney cells (HEK293T, Figure S2B), it is 14 μM.The 4-fold difference in activity is modest but acceptable considering the high nanomolar on-target potency of compound 40.The comparatively low GI 50 for noncancerous cells may reflect the importance of the YTHDC1 protein under physiological conditions, which is not very well studied beyond embryonic development. 18,32ompound 39 (adenine scaffold) features the same substituents as ligand 40 (pyrazolopyrimidine scaffold) with swapped −Cl and −NHCH 3 groups (Table 3).Thus, we selected it as a negative control considering that it is inactive in the biochemical assay.As expected, the negative control 39 was also inactive in the antiproliferative assay with the THP-1 cells (Figure 4A) and the PBMC cells (Figure 4B).
The correlation between biochemical IC 50 values measured by the HTRF assay and the GI 50 values (THP-1) provides indirect evidence for the engagement of the YTHDC1 target in the cell (Figure 4C).To further validate target engagement, we performed the cellular thermal shift assay (CETSA) with compounds 25 and 40 (Figure 4E).The results demonstrated thermal stabilization of YTHDC1 upon ligand binding, confirming target engagement.Additionally, to gain insight into the mechanism of cell death underlying the observed cytotoxicity, we measured the cleavage of poly(ADP-ribose) polymerase (PARP), a recognized marker of apoptosis (Figure 4D).Notably, we observed concentration-dependent increases in the signal from cleaved PARP after 24 h of incubation with compound 40 in the THP-1 cell line.Apoptosis onset following the binding of a small-molecule antagonist (ligand 40) to the m 6 A pocket of YTHDC1 aligns with earlier observations of the YTHDC1 gene knockout in AML cells. 21

■ CHEMISTRY
During the initial stages of the campaign focused on fragment optimization, we successfully developed a synthetic route to prepare compound 3. Compounds 1 and 2 had been purchased during the screening phase prior to this campaign. 23o synthesize chloropyrazolopyrimidine heterocycle 3, first, we performed a cyclization reaction of 41 by heating it in the presence of urea forming a condensed ring 42.The following deoxygenation and chlorination of the pyrimidine ring was achieved by POCl 3 .Finally, we carried out regioselective S N Ar using an ethanolic solution of MeNH 2 (33%), yielding the desired fragment 3 with an overall yield of 9% (Scheme 1).
The main synthetic part comprises the preparation of N 9substituted-4-methylaminopurines as promising YTHDC1 inhibitors.The strategy was based on the convenient purine N 9 derivatization followed by a regioselective S N Ar.To achieve N 9 arylations, we used reported Chan-Lam coupling between commercially available 2,6-dichloropurine and boronic acid derivatives. 33For the N 9 alkylation, corresponding alkyl halides were used.During these reactions, we observed a formation of both N 9 and N 7 regioisomers.However, the desired N 9 isomer was formed as the main product and could be easily separated during the purification.The final synthetic step, S N Ar, was carried out using MeNH 2 , except for compound 37, where cyclopropylamine was used instead.If not stated otherwise, all of the final molecules were synthesized according to the Table 3.Additional Optimization of the Original Fragment a a Same as Table 1.For compound 40, the K D determined by ITC is reported below the IC 50 value.e Tested only in single dose at 5 μM (103%) f Compound 40 was tested in the form of HCl salt.The K d determined by ITC is reported below the IC 50 value.
General synthetic scheme (Scheme 2).However, compounds 8, 9, and 40 were acquired from a commercial supplier. 34For the preparation of compound 7, the General synthetic procedure could not be applied.Instead, the targeted compound was obtained by a two-step reaction process using pTSA-catalyzed reaction of 2,6-dichloropurine 44 and 3,4-dihydropyran followed by nucleophilic aromatic substitution.
Compounds 12−15 were prepared from a shared intermediate 46.The synthesis of 12 followed the reaction sequence outlined in the General synthetic scheme (Scheme 2), with subsequent removal of the Boc protecting group from 47.The preparation of compounds 13−15 involved the initial Boc removal and formation of 48, followed by free amino group modification using standard conditions for sulfonation and acylation, respectively.The S N Ar was carried out as the final step of the reaction sequence (Scheme 3).On the contrary, compound 31, which also contains a sulfonamide moiety, was prepared according to the General synthetic scheme (Scheme 2).The 2,6-dichloropurine 44 alkylation was carried out with an intermediate already containing the sulfonamide group.Even though, this approach provided the final compound 31 in limited yield (6%, after three steps), the reaction sequence was not further optimized.These findings suggest, that the modified reaction sequence, used for compounds 13−15, shall be preferentially used in combination with other sulfonamide derivatives.

Journal of Medicinal Chemistry
Compounds 20, 29, and 32 that contain carboxylic groups, as well as compounds 22, and 23, with amide moiety, were synthesized from corresponding alkyl esters 19, 28, and 33, respectively.The methyl ester hydrolysis was achieved by heating the reaction mixture in the presence of 37% HCl for compounds 19, 28, 33.In the case of compound 10, its tertbutyl ester analogue was subjected to hydrolysis with TFA.For the synthesis of compounds 22, and 23, a combination of COMU as the coupling agent along with the presence of NH 3 or MeNH 2 was used (Scheme 4).
Lastly, due to the nature of fluorine substituent, the nucleophilic aromatic substitution of 51 was not regioselective.As a result, this reaction led to the formation of two products 36 and 39 that were isolated and tested against YTHDC1 (Scheme 5).

■ CONCLUSIONS
In summary, we used structure-based design to develop a small-molecule inhibitor targeting YTHDC1.Through step-bystep ligand optimization and leveraging the most potent substituents identified on a purine heterocycle, we successfully combined the structural features together with the 5chloropyrazolopyrimidine scaffold.Supporting the campaign with high-resolution X-ray data, we evaluated the efficacy of the most potent compound 40 using TSA and ITC assays.Compound 40 exhibits a K D value of 49 nM, very good LE and LLE values, and antiproliferative activity against acute myeloid leukemia cell lines (THP-1, MOLM-13, NOMO-1).Compound 40 is selective against the cytoplasmic YTHDF1−3 readers according to two different assays (HTRF assay, TSA) and against YTHDC2 according to TSA.Furthermore, compound 40 is inactive in a panel of 58 human protein kinases.The correlations between the biochemical SAR (HTRF IC 50 values) and cellular SAR (THP-1 GI 50 values) for the series that led to ligand 40, together with CETSA validation, provide evidence of target engagement in the cell.Thus, we propose compound 40 as a small-molecule tool to study the role of the YTHDC1 m 6 A-reader in AML.

■ MATERIALS AND METHODS
HTRF Assay (IC 50 ).GST-YTHDC1, GST-YTHDF1, GST-YTHDF2, and GST-YTHDF3 were purified as previously reported. 35he HTRF assay was assembled as detailed in ref 23 with the only difference being that the starting concentration of the dose−response experiments used for the IC 50 determination was variated dependently from the tested compound.The same protocol applies to the four proteins.The competitive inhibition data of GST-YTHDC1 in the presence of the compounds were normalized using a blank assembled with all of the components of the assay, including DMSO, except for GST-YTHDC1.The competitive inhibition data of GST-YTHDF1, GST-YTHDF2, and GST-YTHDF3 in the presence of compound 40 were normalized using a blank assembled with all of the components of the assay, minus the protein, and 2-fold serial dilutions of compound 40.This measure was adopted to mitigate the interference arising from high concentrations of compound 40.In all cases, the signal was measured using a Spark plate reader (Tecan), with a 320 nm excitation filter and 620 nm (measurement 1) or 665 (measurement 2) emission filters, a dichroic 510 mirror, 75 flashes, and applying a lag time of 100 μs and an integration time of 400 μs.
Thermal Shift Assay.The YTH domain of YTHDF1 (residues 361−559), YTHDF2 (residues 383−579), YTHDF3 (residues 387− 585), YTHDC1 (residues 345−509), and YTHDC2 (residues 1285− 1424) was cloned into pET-based vector harboring N-terminal hexahistidine tag and a TEV cleavage site.All recombinant proteins were overexpressed at 20 °C in Escherichia coli BL21 (DE3) upon induction with 0.4 mM IPTG and purified by HiTrap nickel column.The His tag was cleaved by the addition of TEV protease (1:100) to the purified recombinant protein while dialysis to remove imidazole at 4 °C overnight.The samples were then passed through nickel column and further purified by size exclusion chromatography.YTHDC1, YTHDF1, YTHDF2, YTHDF3, and YTHDC2 were buffered in 50 mM HEPES pH 7.5, 150 mM NaCl, and tested in a white 96-well plate at a final concentration of 2 μM.SYPRO Orange dye (Sigma-Aldrich, S5692) was added to the mix with a final concentration of 3.75×.Compound 40 was also added to the mix and tested as a set of 2-fold dilutions.The fluorescence monitoring was performed using a

Journal of Medicinal Chemistry
LightCycler 480 System.The temperature was set up to increase with a ramp rate of 0.06 °C/s from 20 to 85 °C and 10 acquisitions per °C were taken in dynamic integration time mode and using red 610 (498−610) filter combination.The melting curves were calculated using the T m calling analysis of the LightCycler 480 software release 1.5.1.62SP3.
Isothermal Titration Calorimetry.The Isothermal titration calorimetry (ITC) experiment was carried out at 18 °C using MicroCal ITC200 (GE Healthcare).Protein and compound were dissolved in 20 mM Tris pH 7.4, 150 mM NaCl along with 0.2% DMSO.Protein at the concentration of 100 μM was titrated into the sample cell containing 10 μM compound.After an initial injection of 1 μL, 13 injections of 3.0 μL each were performed.The raw data were integrated and analyzed using a single-binding site model, provided in the MicroCal Origin software package. 36rystallography.The crystals of YTHDC1 YTH domain were obtained by mixing 1 μL protein solution at 10 mg/mL with mother liquor containing 0.1 M Bis-Tris at pH 6.5, 0.2 M ammonium sulfate and 25% PEG 3350 at 22 °C in a hanging drop vapor diffusion setup.To obtain crystals of protein complexed with fragments, the crystals were transferred to a 1 μL drop containing 50−200 mM (depending on the solubility) fragment directly dissolved in 0.1 M Bis-Tris at pH 6.5, 0.2 M ammonium sulfate, and 30% PEG 3350, soaked overnight at 22 °C, harvested, and frozen in liquid nitrogen without additional cryoprotection.Diffraction data were collected at the Swiss Light Source (Villigen, Switzerland) using the beamline X06DA (PXIII) and processed using XDS. 37The structures were solved by molecular replacement using Phaser program 38 from the Phenix package. 39The unliganded structure of YTHDC1 (PDB ID: 4R3H) was used as a search model.The model building and refinements were performed using COOT 40 and phenix.refine. 41ell Culture.THP-1, MOLM-13, and NOMO-1 cell lines were obtained from DSMZ-German Collection of Microorganisms and Cell Cultures GmbH.Cells were cultured in RPMI 1640 medium (11875093, Thermo Fisher Scientific) containing 10% FBS (16140071, Thermo Fisher Scientific) and 1% penicillin-streptomycin (15140122, Thermo Fisher Scientific) in 5% CO 2 at 37 °C in a humidified incubator.Cell lines were tested negative for mycoplasma contamination (PCR-based assay by Microsynth, Switzerland).Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll Paque Plus (17−1440−02, Cytiva) density centrifugation according to the manufacturer's instructions.Human blood for PBMCs isolation was obtained from the Blood Donation Center Zurich (Ethics Committee approval number: 2021−00024).
Cytotoxicity (GI 50 ).Cells were seeded in white clear-bottom 96well plates at a density of 6 × 10 3 cells/well in 50 μL of the complete RPMI medium and treated with 50 μL of increasing concentrations of the indicated compounds dissolved in DMSO (final concentration of compounds 0.6−160 μM) or DMSO only (0.5% (v/v)) as a negative control and incubated for 72 h at 37 °C with 5% CO2.Cell viability was determined using a CellTiter-Glo luminescent cell viability assay (Promega) based on the detection of ATP according to the manufacturer's instructions.100 μL of the reagent was added to each well and incubated for 10 min at room temperature on an orbital shaker.The luminescence was recorded using a Tecan Infinite 3046 M1000 microplate reader from the top.Background luminescence value was obtained from wells containing the CellTiter-Glo reagent and medium without cells.Cell viability curves were plotted in GraphPad Prism 9 and fitted with nonlinear regression, from which GI 50 values were determined.The assay was carried out with two technical replicates for each concentration and repeated three to four times on different days.PBMCs were seeded in 96-well plates in technical triplicates at a density of 1 × 10 5 cells/well in 100 μL of complete RPMI medium and repeated three times.Cell viability was determined using a CellTiter-Glo luminescent cell viability assay (Promega).
Cellular Thermal Shift Assay (CETSA).One million MOLM-13 cells were suspended in 100 μL of PBS (10010023, Thermo Fisher Scientific) supplemented with a 2× protease inhibitor cocktail (11697498001, Roche) for each experimental condition.The cells were incubated with compounds or DMSO control (1% (v/v)) for 1 h at 37 °C.Samples were then heated up to 48 °C for 3 min followed by cooling to room temperature.Next, samples were lysed by three freeze−thaw cycles in liquid nitrogen and centrifuged at 16,000g for 30 min, 4 °C.Equal volumes of control and tested samples (12 μL) were analyzed by Western blot.The protein signal was quantified by densitometry in Image Studio Lite software.The amount of YTHDC1 (ab259990, Abcam, 1:1000) protein was first normalized to β-actin (ab8226, Abcam, 1:2000) and then to DMSO control and analyzed in GraphPad Prism 9.
Apoptosis Induction by Western Blot.THP-1 cells were seeded into 6-well plates at a density of 1 × 10 6 cells/mL in 1 mL of complete RPMI media.The cells were treated with indicated concentrations of Compound 40 or DMSO control (0.1% (v/v)).After 24 h, cells were collected by centrifugation, washed twice with PBS, and resuspended in 50 μL of RIPA buffer (89900, Thermo Fisher Scientific) with added 2× protease inhibitor cocktail (11697498001, Roche).Cell lysates were centrifuged for 30 min at 16,000g at 4 °C, and the supernatant was collected.The protein concentration was quantified with Micro BCA Protein Assay Kit (23235, Thermo Fisher Scientific) and 20 μg of protein was loaded per well on a 10% Tris-Glycine polyacrylamide gel.Following electrophoresis, the proteins were transferred to a nitrocellulose membrane, blocked with 5% nonfat milk, 0.5% BSA in TBST buffer, and incubated with cleaved PARP (5625, Cell Signaling, 1:1000), βactin (ab8226, Abcam, 1:2000), and GAPDH (2118S, Cell Signaling, 1:4000) antibodies overnight at 4 °C.For the detection, IRDye 800CW goat anti-rabbit IgG and IRDye680RD donkey anti-mouse IgG secondary antibodies (1:10,000) were used.Fluorescence signal was detected on Odyssey CLx Imaging System (LI-COR).The band intensity in each lane was quantified using Image Studio LiteVersion 5.2.5 (LI-COR) and analyzed in GraphPad Prism 9.
Kinases Selectivity.For kinase inhibition testing, the Diversity Kinase [10 uM ATP] KinaseProfiler LeadHunter Panel�FR (50− 015 KP10) was performed by Eurofins Discovery.Compound 40 was tested at a single concentration of 2 μM, in two replicates, against a panel of 58 protein kinases at an ATP concentration of 10 μM.
Chemistry.All reagents were purchased from commercial suppliers and used as received.Reactions run at elevated temperatures were carried out in an oil bath.Our research group successfully synthesized all of the compounds as described, except for compounds 1, 2a, 2b, 8, 9, and 40 which were obtained from a commercial supplier. 34All compounds have >95% purity (HPLC).All reactions were monitored by thin-layer chromatography (Aluminum plates coated with silica gel 60 F 254 ).Flash column chromatography was carried out over silica gel (0.040−0.063 mm).1 H and 13 C { 1 H} NMR spectra were recorded on AV2 400 MHz and AV600 Bruker spectrometers (400, 101, and 600, 150 MHz, respectively) in DMSO or CDCl 3 Chemical shifts are given in ppm and their calibration was performed to the residual 1 H and 13 C signals of the deuterated solvents.Multiplicities are abbreviated as follows: singlet (s), doublet (d) multiplet (m), and broad signal (bs).The purity was acquired by Liquid chromatography high resolution electrospray ionization mass spectrometry (LC-HR-ESI-MS): Acquity UPLC (Waters, Milford) connected to an Acquity eλ diode array detector and a Synapt G2 HR-ESI-QTOF-MS (Waters, Milford); injection of 5-Chloro-N,3-dimethyl-1H-pyrazolo [4,3-d]pyrimidin-7-amine (3).To a powder of 4-amino-3-methyl-1H-pyrazole-5-carboxamide 41 (0.13 g, 0.97 mmol), which was prepared following the reported procedure, 42 was added urea (389 mg, 6.4 mmol).The neat reaction mixture was heated and stirred at 195 °C for 5 h.Upon the temperature increase, the solid reactants melted and after the product formation, the reaction mixture solidified.The reaction vessel was cooled to rt and the crude product 42 was used in the next step without further purification.
The pyrazolo [4,3-d]pyrimidine-5,7(6H)-dione 42 was suspended in POCl 3 (4.6 mL) followed by the addition of DIPEA (0.403 mL, 2.3 mmol).The reaction mixture was heated at 70 °C for 14 h.The volatiles were removed in vacuo and the residue was poured over ice.The mixture was extracted into EtOAc (3 × 6 mL) and the combined organic layers were dried over MgSO 4 and filtered.Activated charcoal was added to the filtrate, and the mixture was stirred for 10 min.After the charcoal removal (filtration paper), the solvent was removed under reduced pressure.The crude product 43 was dissolved in EtOH and 33% MeNH 2 in EtOH (0.2 mL) was added into the reaction vessel.The reaction mixture was stirred at rt for 1 h and after the reaction completion (TLC), the volatiles were removed in vacuo.The crude product 3 was purified using flash column chromatography (SiO 2 ; EtOAc/MeOH = 10:1) and the desired compound was obtained as a white solid (0.018 g, 9% after three steps). 1H NMR (400 MHz, MeOH -d 4 ) δ 3.12 (s, 3H), 2.49 (s, 3H). 13 General Procedure 1 (N 9 -Alkylation of Purines).2,6-Dichloropurine 44 (1 equiv) was dissolved in DMF (0.5 M) and K 2 CO 3 (2 equiv) was added.Corresponding alkyl halide was subsequently added to the reaction mixture (1 equiv).The resulting reaction mixture was stirred at rt until reaction completion (Monitored by TLC).The reaction mixture was quenched by the addition of water and extracted into EtOAc.Combined organic layers were washed by 10% aq.sol. of LiCl, dried over MgSO 4 , filtrated, and evaporated.

Figure 4 .
Figure 4. Biological evaluation of compound 40.(A) Dose−response curves for the antiproliferative effect of compounds 25 and 40 against the THP-1 cell line.Compound 39 is a negative control.(B) Dose−response curves for the antiproliferative effect of compounds 25, 39, and 40 against human peripheral blood mononuclear cells (PBMCs).(C) Scatter plot of GI 50 values for THP-1 vs biochemical IC 50 values for those compounds that were measured in both cellular and biochemical assays.The correlation provides evidence of target engagement in the cell for this series of compounds.Note that the negative control 39 was excluded from the fitting.Compounds 20 and 32 were also excluded from the fitting as they feature a carboxylic acid which hinders the passage through the cell membrane.(D) Induction of apoptosis by compound 40 in THP-1 cell line after 24 h treatment.Apoptosis was monitored by detecting cleaved PARP using a Western blot.The protein signal was normalized to DMSO control.(E) Thermal stabilization of YTHDC1 upon treatment was quantified by CETSA at 48 °C in MOLM-13 cells.Both compounds 25 and 40 stabilize YTHDC1 in a concentration-dependent manner.The dashed line represents the protein level of the DMSO control used for normalization.

Table 2 .
Structure−Activity Relationship: Benzyl Ring Substitutions a

Journal of Medicinal Chemistry 20
. Conversely, meta-substituted compounds with amide