Nanoscale Direct-to-Biology Optimization of Cdk2 Inhibitors

Abstract

Modern hit-to-lead optimization winnows down vast chemical spaces of virtual compounds into a selection of potent and selective compounds that can be further profiled with in vitro assays. Today, miniaturized chemical synthesis can be performed in high-throughput, shifting the bottleneck to compound purification. Direct-to-biology (D2B) approaches seek to overcome this hurdle by omitting the purification step and submitting reaction mixtures directly to bioassays. Here, we explore nanoscale hit-to-lead optimization through the multistep synthesis of a library of Cdk2/CycE inhibitors, utilizing ultrahigh-throughput experimentation (ultraHTE) in 1,536-well plates. Library performance is assessed by D2B in functional biochemical, bioaffinity, and X-ray crystallographic assays. A selection of potent lead Cdk2/CycE inhibitors identified by D2B was submitted to a phenotypic cell painting assay, which showed cell cycle arrest at G₀, consistent with Cdk2 inhibition. This miniaturized workflow allows the upper tiers of a typical optimization screening cascade to be performed in a single experiment.

Graphical Abstract

INTRODUCTION

Efficient drug discovery relies on rapid design-make-test (DMT) cycles.^1,2 Automation and machine learning are well poised to accelerate the DMT cycle, ideally leading to better drugs faster.^3–6 In recent years, drug hunters have sought to integrate the DMT cycle into a single experiment in a process termed “direct-to-biology” (D2B),^7,8 wherein freshly synthesized reaction mixtures are submitted to bioassay with minimal or no purification. This accelerated approach has been successfully deployed for several bioassay methods, including affinity selection,^9–13 biophysical,^14–17 structural biology,^15,18–21 biochemical,^20,22–28 in cellulo,^29–39 ADME,^40,41 and even in vivo⁴² assays. While the approach accelerates the DMT cycle, care must be taken in accounting for the presence of unreacted reagents, byproducts, and solvents present in reaction mixtures to ensure assay integrity and minimize false-positive and -negative results. As such, carefully selected control experiments are typically used to prevalidate the chemistry with the bioassay of choice. Many studies have now demonstrated such validations, suggesting that the D2B technology will feature prominently in AI-guided drug discovery campaigns, especially if multistep high-throughput experimentation (HTE) and diverse bioassays could be performed simultaneously.

Ultrahigh-throughput experimentation (ultraHTE)^43–45 is an emerging tool that goes hand in hand with D2B studies.^9,27,31,46 By using well plates and liquid-handling robotics to transfer microliter-sized droplets, thousands of reactions can be rapidly executed and submitted directly to bioassay without purification, facilitating the systems-level link between reaction conditions, chemical structure, physicochemical properties, and biological function.^47,48

In 2018, we reported a D2B technology wherein affinity-selection mass spectrometry⁴⁹ (ASMS) was used to rank compound affinity of miniaturized libraries through protein titration-induced competitive binding.⁹ While this method expedites hit synthesis and identification, interrogation of larger compound libraries would benefit from D2B assay methods that provide information beyond compound affinity, such as biochemical function and structural insights (Figure 1A). Here, we report a D2B workflow to rapidly and thoroughly explore the chemical space around a class of cyclin-dependent kinase (Cdk) inhibitors. The complementary assays illuminate how subtle changes to molecular structure can impact binding affinity, structural binding mode, functional biochemical activity on purified protein complexes, and functional activity in cells.

Figure 1.

(A) Previous work optimized reaction conditions, building blocks, and protein affinity using direct-to-biology ultraHTE. This work adds multistep synthesis, bioactivity, X-ray structure, and cell morphology readouts. (B) Previously reported amino-pyrazole synthesis uses volatile solvents and harsh acidic conditions at elevated temperatures, which are incompatible with ultraHTE. (C) Multistep miniaturized synthesis was used in this work. All reagents and solvents are compatible with ultraHTE and D2B.

Cdks regulate cell cycle phases in association with cyclin regulatory subunits.⁵⁰ The Cdk2/Cyclin E (CycE) complex controls cell cycle progression and DNA replication during the G1/S transition through phosphorylation of specific substrates.⁵¹ In many cancers, CycE is overexpressed, leading to replication stress, DNA damage, and carcinogenesis associated with poor patient outcomes.^52,53 Additionally, compensatory pathways involving Cdk2 have frequently been observed in cancers with resistance to Cdk4/6-based therapies.^54–56 Therefore, the Cdk2/CycE complex is a promising therapeutic target in cancer treatment^57–63 among other indications.^64–66 Here, we present the multistep miniaturized library synthesis of 768 potential Cdk2/CycE inhibitors with relevant D2B profiling. Potent hits were subsequently evaluated in a cell painting assay to examine the inhibitors’ effect on cancer cell morphology.⁶⁷

RESULTS AND DISCUSSION

We chose to explore chemical space around a recently reported amino-pyrazole class of Cdk2/CycE inhibitors 4 (Figure 1B,C).^68–71 Thus, amide coupling of 128 commercially available carboxylic acids (see Supporting Information for full list) with 6 amino-pyrazole cores (5–10) was employed to generate a 768-compound library (Figure 2A). Amino-pyrazoles 5–10 were chosen for diversity and the anticipated Cdk2/CycE potency of the amide-coupled products. Using Morgan fingerprints, visualization of the 768 targets alongside compounds enumerated from 5–10 and a library of commercial carboxylic acid building blocks (Enamine) showed broad sampling of the chemical space (Figure 2B). Compared to previously reported Cdk2/CycE inhibitor libraries of this class,⁶⁸ our library used a greater proportion of aliphatic and propionic acids and explored more deeply the impact of aryl ring substituent patterns. Importantly, the synthesis needed to be at once amenable to ultraHTE and the D2B assays. Generally, successful ultraHTE requires high-boiling solvents and room temperature reactions to prevent evaporation. Additionally, the solvent must be compatible with the plastic well plates employed by liquid handling robots, and it is preferable that all reagents and reaction mixtures be soluble to ensure accurate dosing.^46,47,72,73 Previously reported syntheses for 4 did not meet these requirements (Figure 1B).⁶⁸ For instance, the reported solvent, dichloromethane, presents volatility and plastic compatibility challenges in the amide coupling step, and heating at 75 °C was required for tBu-pyrazole deprotection, which is close enough to the boiling point of formic acid to be problematic to reaction miniaturization. Additionally, strongly acidic conditions may present issues with respect to hydrolysis of the substrate carbamate moiety, well plate seal integrity, reaction quenching, and damage to metallic liquid-handler components. Therefore, a synthesis strategy that removed the need for volatile solvents and pyrazole protection was designed (Figure 1C). A critical modification was the addition of two equivalents each of carboxylic acid and coupling reagent. In this way, the first equivalent of acid building block acts as an in situ protecting group for the pyrazole N–H as in 2, then the second equivalent forms the desired amide bond (3). Subsequent addition of aqueous lithium hydroxide selectively deprotects the pyrazole while leaving the target amide bond intact (4), and hydrolyses any residual activated-ester byproducts. Upon reaction completion, acetic acid was added to neutralize the excess base. This method negates the need for a protecting group and avoids volatile solvents and harsh acidic conditions. The small scale of ultraHTE means a negligible amount of the commercially available acid building blocks are needed when used in 2-fold excess. Additionally, N-methyl-2-pyrrolidone (NMP) and water are used as solvents, which are high boiling and compatible with direct injection onto UPLC-MS for reaction analysis, and compatible with bioassays (see Supporting Information for control experiments). Classical HTE in a 96-well format²⁷ was used to identify suitable reagents for the ultraHTE protocol (Figure S1). Heating to 80 °C was required for a productive reaction. Sealing studies of the 1536-well plate showed that NMP does not leach or evaporate at this temperature (see Supporting Information). Coupling various acids with ent–6 (the less bioactive enantiomer was used for validation studies) and a model amine system S2 identified HATU as a viable coupling reagent. Nonetheless, CDI was selected for further study owing to a distinct byproduct profile compared to HATU and its derivatives. With respect to bases, DIPEA was chosen due to its efficacy and common use in amide coupling,⁷⁴ and NMI, which gave high yields and has been shown to promote challenging amide bond formations by acting as a nucleophilic catalyst.⁷⁵ Thus, two coupling reagents and two bases gave a total of four possible reaction conditions, increasing the chances that each compound would be synthesized at least once. We then executed all 3,072 ultraHTE reactions (128 acids × 6 amines × 2 coupling reagents × 2 bases) utilizing mosquito liquid handling robotics. With this approach, 691 compounds (691/768 = 90%) were successfully produced in at least one reaction condition, as determined by UPLC-MS analysis with ultraviolet (UV), mass spectrometric (MS), and evaporative light scattering (ELS) detection (Figure S2). The reaction conversion for the three detection methods had a strong correlation, with R² values between 0.77 and 0.89 (Figures 2C and S30–S32). For the purpose of selecting compounds for further study, the average conversion from the three detection methods was used.

Figure 2.

Library design and synthesis; (A) Six amine cores 5–10 were chosen for amide coupling with 128 commercially available acids. The reported acids were selected from ref 68. (B) t-SNE plot of Morgan fingerprints (radius 4, 2,048 bits), comparing 768 targeted compounds to 233,736 virtual compounds generated from enumeration of amides formed using a commercial acid catalog (Enamine) and 5–10. The targeted compounds cover a large portion of chemical space. (C) Relative conversion to product for the ultraHTE amide coupling of 5–10 with 128 different acids, as determined by UPLC-MS. Three detection methods, UV absorbance, mass spectrometric (MS) total ion count in positive mode, and evaporative light scattering detection (ELSD), were utilized. Each point represents one reaction, colored by percent inhibition observed when subjected to ADP-Glo biochemical assay.

With the miniaturized library of compounds in hand, we sought to assess bioactivity for Cdk2/CycE inhibition in a D2B biochemical assay. We chose ADP-Glo—a bioluminescence assay that monitors the conversion of ATP to ADP.⁷⁶ To test ADP-Glo’s suitability as a D2B assay, control plates were designed to interrogate the effects of the crude reaction mixtures on assay readout (see Supporting Information). These controls comprised titration of a known potent inhibitor of Cdk2/CycE S24 in both pure DMSO and “dummy” reaction mixtures designed to mimic the final conditions of our amide coupling protocol, and subsequent treatment with scavenger resins, buffers, and various workup procedures. In all cases, little impact on assay readout was observed. The synthesis of S24 using the four chosen amide coupling conditions was evaluated in a D2B format, along with reactions in which reagents were deliberately left out. Standard reactions and those lacking LiOH·H₂O gave high inhibition values as expected. Surprisingly, inhibition was observed for reactions in the absence of base. LCMS confirmed the presence of S24 in these reactions, showing that amide coupling can proceed without base, but at a lower conversion. Reactions without amine exhibited no inhibition. Reactions lacking coupling reagents or carboxylic acid building block gave approximately 10–40% background inhibition in the ADP-Glo assay. These data suggest that the amine starting material, reagents, or intermediates could have a small inhibitory effect on Cdk2/CycE. Examination of all starting materials, reagents, and solvents revealed that only amines 5–10 had a small, consistent inhibition across repeated measurements (<15%). These values were used to correct measured inhibition values in subsequent experiments. With biochemical control data in hand and background inhibition accounted for, all 3,072 crude ultraHTE reactions were submitted to single-point ADP-Glo assay. The different reaction conditions used to synthesize each compound provided a range of reaction conversions. Typically, for a given active compound, assay inhibition should increase with reaction conversion, which we observed with the ADP-Glo assay (Figures 2C and S52, R² = 0.56). During this work, we simultaneously evaluated the library using a Caliper electrophoretic separation biochemical assay (including appropriate D2B controls; see Supporting Information).⁷⁷ However, the correlation between conversion and inhibition was poor (Figures S53 and S54, R² = 0.10), and compounds synthesized from the 20 acids reported to be potent showed low inhibition. Therefore, the Caliper assay was not explored further. The distribution of ADP-Glo inhibition values showed the library to be of high potency, in line with previous findings for Cdk2/CycE inhibitors with this substructure.^68–70 By considering the highest-yielding reaction for each target compound, we observed that ~450 of the 768-compound library gave ≥80% inhibition in the bioassay (Figure S3). To resolve the library further, we introduced an ASMS affinity assay into the D2B workflow⁹ to identify compounds with high affinity and activity, and to confirm that ADP-Glo inhibition was coming from protein–ligand binding and not a false positive readout based on a byproduct or reaction component.

For ASMS binding experiments, crude ultraHTE reaction mixtures were pooled and titrated against decreasing concentrations of Cdk2/CycE to identify protein-complex-bound compounds using exact mass-to-charge ratio (m/z). Decreasing protein concentration induces binding competition, allowing only compounds with the highest affinity to be observed at lower protein concentrations. Given a larger library size compared to the previous study,⁹ initial attempts involved high pooling compression into wells containing 768 compounds each, with the aim of minimizing the number of ASMS injections required (see Supporting Information). However, this led to inconsistent data with many false positives and negatives, illuminated when compared to confirmatory data obtained on pure compounds. We hypothesize this is due to the relatively low concentration of each compound—particularly when compared to remaining starting materials 5–10, which were at high relative concentration due to their commonality in hundreds of reactions. To overcome this problem, a lower compression strategy was employed in which, typically, the reaction with the highest conversion for each compound was selected and then grouped into mass-encoded pools of 18–25 compounds using an acoustic ejection liquid handler. To limit false negatives for low-conversion reactions⁷⁸ compounds were first binned by conversion, and then distributed into pools using ADP-Glo inhibition values (see Supporting Information). This ensured that reactions with low conversion were pooled over a Gaussian distribution of inhibitory values. Given the overall potency of the library, higher conversion pools were weighted toward compounds that showed high inhibition values. ASMS protein titration of the crude reaction pools was performed to achieve affinity ranking⁹ at Cdk2/CycE concentrations of 0.0, 0.05, 0.10, 0.50, 1.0, 2.0, 5.0, 10, and 20 μM, and the lowest concentration at which protein binding was observed for each compound was labeled. The distribution of the lowest protein concentration at which inhibitor was observed versus ADP-Glo inhibition shows a good correlation between biochemical activity and affinity (Figure 3). A broad inhibition distribution is observed for high-affinity compounds that bind at 0.05 μM. This may indicate that at low protein concentrations, the ASMS assay is unable to resolve pools of high-potency compounds, or that “breakthrough” (vide infra) is operative. 54 compounds had no observable binding at 20 μM Cdk2/CycE (NB category, Figure 3), and 31 of these had <30% reaction conversion. Some modest inhibition was still observed in these reactions, which, in line with our control experiments, could be a weak false-positive readout. Several compounds were classified as “breakthrough” false positives as they bound to the invertase protein included at a constant concentration as a detector trigger in the ASMS assay and were disregarded from further analysis.

Figure 3.

Kernel density estimates show the distribution of ADP-Glo inhibition values at each of the lowest Cdk2/CycE concentrations at which ASMS compound binding was observed. NB = no binding observed at any Cdk2/CycE concentration tested, BT = breakthrough compound.

The entire data set was collectively visualized (Figure 4) to analyze UPLC-MS reaction conversion, ADP-Glo bioactivity, and ASMS bioaffinity results. Each segment of the circular plot corresponds to one amine starting material (5–10) with outer heatmaps displaying reaction conversion for the 128 acids using the four distinct reaction conditions (labeled A–D). The reaction chosen for ASMS analysis is highlighted with a black outline, and the lowest protein complex concentration at which the compound bound is indicated in the concentric inner circles, where each compound is shown as a dot colored by its ADP-Glo inhibition. In this way, higher-order trends such as the overall synthetic efficiency and potency of the library can readily be visualized, as can the correlation between potency and affinity, as the frequency of warmer colors increases toward the center of the plot. This is in contrast to the previous study on CHK1 inhibitors, in which the majority of compounds did not bind to the target protein.⁹ Comparing the color of a dot to its position in the circle also provides some information on binding. For example, if a dot is purple but is seen to bind at 0.05 μM, it may indicate a nonactive or allosteric binder. The majority of compounds in this study bound at a lowest protein concentration of 0.50 μM. Compounds that bound at lower concentration than this threshold are likely highly potent and warrant further study. In particular, products generated from amines 9 and 10 show relatively high populations of the 0.10 and 0.05 μM protein concentrations. Compounds derived from the 6 and 7 amine class are generally the poorest inhibitors, with no products detected in the 0.05 μM protein concentration bucket, a greater population in the 10 and 20 μM protein concentrations, and overall lower ADP-Glo inhibition values. Coloring the dots by physicochemical properties instead of ADP-Glo inhibition shows that less potent compounds have slightly more polar surface area (PSA) (Figure S60).

Figure 4.

Circos mapping of reaction conditions, bioactivity, and bioaffinity data. 3,072 reactions were run using only 84 mg total of amines 5–10. The outer heatmap shows the reaction conversion, as an average of values based on UV, MS, and ELSD peak area for a survey of 4 amide coupling conditions (A-D), with each amine coupling to 128 carboxylic acids. 691 compounds (90%) were successfully synthesized. The reaction chosen for direct submission to ADP-Glo Cdk2/CycE inhibition assay is highlighted with a black box in the heatmap. The inner data points are colored by ADP-Glo inhibition values, with those closer to the center of the circle representing compounds whose ASMS binding signals persisted at lower Cdk2/CycE concentrations, corresponding to higher Cdk2/CycE affinity. Values outside of 0–100% were colored as 0 or 100%. NT = not tested owing to 0% yield, NB = no binding observed at 20 μM Cdk2/CycE, BT = disregarded breakthrough compound that bound to control protein (see Supporting Information for details). Exemplary compounds (11–18 and Figure S5) were resynthesized on a 50 μmol scale and purified to obtain IC₅₀ values. Figure S4 displays filtered data for just these compounds. Isolated yields are in parentheses.

To validate the D2B strategy of using multiple assays, select compounds with a range of bioactivities, bioaffinities, and physicochemical properties (Figure S4) were resynthesized on a 50 μmol scale, and pure samples were isolated—a selection of which (11–18) are shown in Figure 5 (see Figure S5 for all isolated compounds). IC₅₀ values were obtained for each pure compound using the ADP-Glo inhibition assay and plotted against the lowest ASMS protein concentration and the ADP-Glo inhibition obtained directly from reaction mixtures (Figure S6). A good correlation was observed between the three data sets, as indicated by the positive gradient of the plot, with data points graduating from yellow to purple. Of the 120 compounds we produced that utilized an acid building block previously reported to be potent, 103 were successfully synthesized using ultraHTE, 46 of which bound at an ASMS Cdk2/CycE concentration of 0.5 μM or lower. 12 of the 103 successfully synthesized compounds have affinity data reported for Cdk2/CycE⁶⁸ (21, S18–S28). A plot of these reported data against our D2B data (lowest ASMS protein concentration and ADP-Glo inhibition) showed a good affinity correlation for compounds synthesized in more than 12% reaction conversion (Figure S7). From these compounds, D2B successfully identified S18 as having the highest affinity. Compounds synthesized in less than 12% conversion (S24–S28) were generally not observed to bind by ASMS, likely due to being outcompeted by higher concentration ligands in their pool. The exception was S24, which was successfully detected at 5 μM Cdk2/CycE despite being produced in only 3% conversion. This is confirmed by the reported high affinity for this compound and lends support for our chosen pooling strategy.

Figure 5.

Overlayed protein-ligand X-ray crystal structures were obtained directly from crude miniaturized reactions. Three conformations of the Cdk2 P-loop (pink) and TYR-15 (magenta) were observed: EE, in which the P-loop and TYR-15 adopt an extended conformation; CC, in which the P-loop is collapsed and TYR-15 adopts a tucked position forming an amide-π interaction with the bound ligand; and EC, in which the P-loop is in an extended position and TYR-15 adopts a tucked position forming a cation-π interaction with LYS-33, which forms a bridging hydrogen-bonding interaction to the carbamate carbonyl group of the ligand. Compound color: 11 = violet, 16 = yellow, 19 = gold, 20 = orange, 21 = purple. PDB Codes: 9OB2 (11), 9OB3 (16), 9OB4 (19), 9OB5 (20), 9OB6 (21).

Data on pure compounds confirmed many of the D2B trends observed. For instance, lipophilic benzoic or phenylacetic acid moieties, in particular those with para-substituents, gave the most potent compounds. Lower potency was observed for compounds with ortho-substituents (e.g., 18 and S15) presumably due to unfavorable steric interactions within the binding site. Similarly, compounds with a heteroatom at the ortho-position of the acid, for example compounds 16 and S16, were of lower potency. D2B successfully predicted that compounds 13 and S6 derived from meta and para tolyl benzoic acids would be more potent than the ortho derivative S11. Similarly, D2B could successfully rank the potency of isomers 14, 15, and S8. Methyl groups can have a significant effect on the potency of drug molecules.⁷⁹ The D2B protocol successfully distinguished between two highly potent compounds, 21 and 22, which vary only by a methyl group and have IC₅₀ values of 1.2 and 2.8 nM, respectively. Compound 19 is an outlier, with an ADP-Glo IC₅₀ value of 41.9 nM and a lowest protein binding concentration of 0.10 μM, indicative of a compound with high affinity but relatively low biochemical potency. Therefore, 19 was included in the next stage of investigations to further understand its binding.

X-ray crystallography in a D2B format was next used to examine which protein conformations were prevalent when compounds from the miniaturized library were bound.^18–20 This is an important demonstration for D2B, as it would augment bioaffinity and biochemical activity data with structural biology data. Reaction mixtures were selected to sample each acid family depicted in Figure 2A: acetic (11), benzoic (16), propionic (20) and aliphatic (19), as well as 21, which is reported to produce potent Cdk2/CycE inhibition.⁶⁸ By directly soaking protein crystals with crude reaction mixtures, ligand-bound crystal structures were obtained for all five compounds (Figure 5). In this study, and drawing from a substantial number of internal data sets, we were able to differentiate between four distinct P-loop conformations. These conformations were modeled in either a single state or in a combination of states, denoted by the abbreviations EE, EC, CE, and CC, where E and C each refer to the extended or collapsed conformations of the P-loop and the positioning of TYR15 at the tip of the loop, respectively (Figure S8). The EE, CE, and CC conformations have been previously observed in published structures of Cdk2/CycE.^80–82 For this study, all compounds bound in the frequently observed EE conformation, and only the CE conformation was not clearly observed. Compounds produced from acids in the families 11, 20, and 21, all of which derive from a longer chain acetic or propanoic acid, showed multistate conformations. The CC conformation was modeled in a 50:50 proportion with the EE conformation, in which TYR15 adopts a tucked position and the P-loop is collapsed into the orthosteric site. Cyclohexane carboxylic acid-derived 19 was the only compound tested that bound in the EC conformation, in which a bridging interaction can be seen between TYR15, the side chain of LYS-33, and the carbamate carbonyl oxygen atom of 19. This binding mode may explain the high affinity observed in the protein titration experiment despite relatively low biochemical potency in the ADP-Glo assay. During scale-up synthesis, 19 was unstable to standard workup conditions and was not successfully isolated on the first attempt. Only by implementing a modified workup procedure could a 42% yield be obtained. This exemplifies how D2B assay protocols can highlight compounds that could be disregarded in traditional library synthesis due to difficult purification. Even a low-conversion nanoscale reaction, 26% in the case of 19, can be used to measure activity and affinity and provide crystallographic information while consuming only submilligram amounts of material.

Finally, we sought to explore the activity of selected potent lead compounds in the HCT-116 colon cancer cell line, which has a high expression of Cdk1/2.⁸³ Cell painting⁶⁷ is a powerful phenotypic image-based assay that can be used to profile the morphology of cells when treated with various fluorescent dyes and is an area of expertise on our team.^84–86 Our Cdk2 inhibitor library covers a range of physicochemical properties that could impact distribution to different organelles within the cell and cause differing cell morphologies. We were particularly interested in observing the effects on the nucleus and histones, given the role of Cdk2/CycE in DNA replication. Compounds from the library with the highest affinity for Cdk2/CycE—those found bound in the 0.05 and 0.10 μM lowest protein concentration pools—were plotted as t-SNE and KDE plots using physicochemical properties (Figure S9) and six were selected (11, 12, 19, 22, S4, and S5) to cover a diverse range of property space for cell painting studies. The compounds were tested for cell-based activity via multiplexed high-content screening to assess induction of cell death, cell cycle arrest, and morphologic perturbations in HCT-116 cells. HCT-116 was chosen as a robust cell line that has been widely employed in the study of Cdk2 inhibition because of its dependence on CycE.^87–89 Compounds were dispensed from 100 μM to 20 nM and were incubated for 24 h, at which time 5-ethynyl-2’-deoxyuridine (EdU) was added to label cells in S-phase. At 48 h, the cells were fixed and stained with Hoechst-33342 to label nuclei, EdU-Alexa-488 to label cells in S-phase, phospho-Histone H3 (pHH3) to label cells in M-phase, and HCS CellMask Orange to label the entire cell. The assay workflow is shown in Figure S10A and representative images are shown in Figure 6A,B. A dose-responsive reduction in cell viability, EdU positivity, and pHH3 positivity was observed for all compounds (Figures S10B, C and S11). Nonlinear dose-response curve fitting was performed for reduction in viability, and the fit half-maximal cytotoxic concentrations (CC₅₀) are shown in Figure S11A with a range of values from 800 nM to 210 nM. Additionally, we observed a dose-dependent reduction in EdU positivity (EC₅₀), indicating cell cycle arrest at G₀, consistent with direct Cdk1/2 inhibition preventing cells from entering S-phase. A linear correlation (Figure S11B) between CC₅₀ and EC₅₀ values was observed for all compounds—except for 19 that had a 6-fold reduction in potency for EdU positivity relative to the CC₅₀—consistent with the high affinity and low biochemical potency observed in the D2B assays, and alternative binding of 19 in the EC protein conformation (Figure 5).

Figure 6.

High-content screening to assess the cell-based efficacy of Cdk1/2 inhibitors in HCT116 colon carcinoma cells. Cellular imaging and analysis were performed to quantify cytotoxicity, cell cycle inhibition, and morphologic perturbations. (A) Cells were imaged, and nuclei were identified using the Hoechst-stained channel, and each nucleus was classified for EdU positivity and pHH3 positivity (scale bar = 50 μm). (B) Representative images of cells treated with Cdk2 inhibitor 22 and DMSO vehicle. EdU = 5-ethynyl-2’-deoxyuridine.

Cell-level morphologic profiling was performed to identify characteristics of inhibitor-treated cells versus DMSO vehicle-treated cells. UMAP dimensionality reduction was performed on 500 cellular features to observe the differential phenotypic presentation of cells upon treatment with our Cdk2 inhibitors or taxol relative to the unperturbed DMSO vehicle control (Figure S11C). The UMAP analysis showed strong clustering of taxol-treated cells and no distinguishable phenotypic perturbation of compound-treated cells relative to the DMSO control, which indicates there is no pronounced phenotypic perturbation other than cell cycle arrest via Cdk1/2 inhibition. Figure S11D shows representative images of cells stained with HCS CellMask Orange under DMSO treatment, compound 22, and taxol-treated cells. This morphologic analysis suggests high target specificity and low off-target or general xenobiotic cytotoxicity, indicating that 22 is a highly selective Cdk2 inhibitor with a desirable phenotypic profile.

CONCLUSIONS

We present here the development of a multibioassay miniaturized D2B platform, which enables the identification of high value leads through expedited synthesis and evaluation of compound libraries, making the drug discovery process faster and more efficient. Through judicious selection of reaction conditions, we successfully executed an automated multistep synthesis of the target compound library, with 90% of compounds synthesized. This included heating reactions in a 1,536-well plate to 80 °C with no obvious leaching of reaction components or evaporation of solvent. Evaluation of the library in both ADP-Glo and Caliper kinase inhibitory assays revealed the former to be the most suitable for D2B. Carefully selected controls were used to validate the D2B assay results. ASMS affinity ranking of the crude reaction mixtures was able to resolve potent hits. The crude bioactivity and bioaffinity data sets correlated well with each other, with IC₅₀ values generated here, and with affinity data from the literature for purified compounds. Protein–ligand X-ray crystal structures of five compounds obtained from crude reaction mixtures revealed three protein-binding conformations, with binding of 19 in the EC conformation suggestive of a distinct interaction mechanism. Finally, a cell painting assay of high-affinity compounds showed no pronounced phenotypic perturbation of cell morphology other than a dose-responsive reduction in cell viability and EdU positivity, consistent with cell cycle arrest at G₀ by Cdk1/2 inhibition, except for 19. Collectively, these results highlight the D2B strategy as an opportunity to streamline many of the frontline investigations of hit-to-lead optimization. Future incorporation of ADME profiling and high-throughput cellular assays for target engagement or cell-cycle monitoring into this D2B workflow could further expand utility down the pipeline to late-stage lead optimization campaigns. Using this technology, it will be possible to rapidly identify cell-active inhibitors with diverse structures and properties or cast a wide net over accessible chemical space to protect intellectual property.

EXPERIMENTAL SECTION

General Information on Chemical Synthesis

Unless otherwise stated, reactions were conducted in oven- or flame-dried glassware in air. Tetrahydrofuran, dichloromethane, toluene, acetonitrile, and N,N-dimethylformamide were purified using an MBraun SPS solvent purification system by purging with nitrogen, and then passing the solvent through a column of activated alumina. Other solvents were purchased as anhydrous and used as received. Reagents were purchased from Sigma-Aldrich, Thermo Fischer Scientific, Alfa Aesar, Oakwood Chemical, or TCI Chemical. All chemicals were used as received. Glass 1-dram (Fisherbrand part No. 03-339-21B) or 2-dram vials (Fisherbrand part No. 03-339-21D) were used as reaction vessels, fitted with standard screwcaps (#03-452-225 or #03-452-300) or with Teflon-coated silicone septa (#CG-4910-02), and magnetic stir bars (Fisher Scientific #14-513-93 or #14-513-65; stirbars.com #SBM-0803-MIC or #SBM-1003-MIC). The synthesis of intermediates 5–10 and ent–6 was performed at WuXi AppTec Co., Ltd. as detailed below.

¹H NMR spectra were recorded on a Varian MR-500 MHz, Varian MR-400 MHz or Bruker Avance Neo 500 MHz spectrometer. Chemical shifts are reported in parts per million (ppm) and the spectra are calibrated to the resonance resulting from incomplete deuteration of the solvent (CDCl₃: 7.26 ppm; DMSO-d₆: 2.50 ppm; MeOD-d₄: 3.31 ppm; Acetone-d₆: 2.05 ppm; D₂O: 4.79 ppm). ¹³C NMR spectra were recorded on the same spectrometers with complete proton decoupling. Chemical shifts are reported in ppm with the solvent resonance as the internal standard (¹³CDCl₃: 77.16 ppm, t; DMSO-d₆: 39.52 ppm, sept; MeOD-d₄: 49.00 ppm, sept). For certain compounds, as specified below, signals for ¹³C pyrazole environments could not be observed due to peak broadening.⁹⁰ Data are reported using the abbreviations: s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, h = hextet, m = multiplet, br = broad, Rot. = rotamer. Coupling constant(s) are reported in Hz. ¹⁹F NMR spectra were recorded on the same spectrometers as above. ¹³C and ¹⁹F signals are singlets unless otherwise stated. ¹H–COSY, HSQC, and HMBC were used where appropriate to facilitate structural determination.

High-resolution mass spectrometry data (HRMS) were obtained either on an Agilent 6230 TOF LC/MS equipped with an ESI detector in positive mode, on a Micromass AutoSpec Ultima Magnetic Sector instrument with an EI detector in positive mode, or by reversed-phase liquid chromatography–high-resolution mass spectrometry (RPLC-HRMS) containing Waters Acquity I-class components and a Synapt XS mass spectrometer. The Acquity system consisted of an I-class binary solvent manager, an FTN-I sample manager, and a column manager connected to the Synapt XS. Compound samples were solubilized in DMSO, then diluted to 0.2 μM in 0.1% formic acid with 1% DMSO. Compounds were then placed into the sample manager, which was chilled to 4 °C. A total of 10 μL of the sample was injected onto a C18 column (Waters BEH C18, 1 mm × 50 mm, 1.7 μm) for compound separation and detection by mass spectrometry. Compound separation was achieved using a linear 5 min gradient of water (A phase) and acetonitrile (B phase) with 0.1% formic acid at 60 °C flowing at 200 μL/min. The mass spectrometer was equipped with a lockspray exact mass ESI source. The instrument was operated in ToF-only resolution mode with source temperature and desolvation gas temperature at 100 and 500 °C, respectively, cone gas flow and desolvation gas (N2) flow at 100 and 1000 L/h, respectively, and sampling cone voltage at 40 V, with a capillary voltage of 3.0 kV in positive ion mode. The ToF was set to monitor a mass range from 100 to 1000 m/z. Reaction analysis was typically performed by thin-layer chromatography on silica gel or using a Waters I-class ACQUITY UPLC-MS (Waters Corporation, Milford, MA, USA) equipped with an in-line photodiode array detector (PDA), an evaporative light scattering detector (ELSD), and a QDa mass detector (both ESI positive and negative ionization modes). Typically, 0.1 μL sample injections were taken from acetonitrile solutions of reaction mixtures or products (~1 mg/mL). A partial loop injection mode was used with the needle placement at 2.0 mm from the bottom of the wells and a 0.2 μL air gap at preaspiration and postaspiration. Column used: Waters Cortecs UPLC C18+ column, 2.1 mm × 50 mm (Waters #186007114) with Waters Cortecs UPLC C18+ VanGuard Precolumn, 2.1 mm × 5 mm (Waters #186007125). Mobile Phase A: 0.1% formic acid in Optima LC/MS-grade water, Mobile Phase B: 0.1% formic acid in Optima LC/MS-grade MeCN. Flow rate: 0.8 mL/min. Column temperature: 45 °C. The PDA sampling rate was 20 points/sec. The QDa detector monitored m/z 150–750 with a scan time of 0.06 s and a cone voltage of 30 V. The ELSD had a gain of 750, data rate of 10 pps, a time constant “normal” (0.2000 s), a gas pressure of 40.0 psi, with the nebulizer in cooling mode at 75% power level, and the drift tube temperature set to 50 °C. The PDA detector range was between 210 nm and 400 nm, with a resolution of 1.2 nm. 2 minute and 8 minute methods were used. The method gradients are below: 2 min method, 0 min: 0.8 mL/min, 95% 0.1% formic acid in water/5% 0.1% formic acid in acetonitrile; 1.5 min: 0.8 mL/min, 0.1% 0.1% formic acid in water/99.9% 0.1% formic acid in acetonitrile; 1.91 min: 0.8 mL/min, 95% 0.1% formic acid in water/5% 0.1% formic acid in acetonitrile. 8 min method, 0 min: 0.8 mL/min, 95% 0.1% formic acid in water/5% 0.1% formic acid in acetonitrile; 7.5 min: 0.8 mL/min, 0.1% 0.1% formic acid in water/99.9% 0.1% formic acid in acetonitrile; 7.91 min: 0.8 mL/min, 95% 0.1% formic acid in water/5% 0.1% formic acid in acetonitrile.

Flash chromatography was performed on silica gel (230 – 400 mesh, grade 60) under a positive pressure of nitrogen. Thin-layer chromatography was performed on 25 μm TLC silica gel 60 F254 glass plates purchased from Fisher Scientific (part number: S07876). Visualization was performed using ultraviolet light (254 and 365 nm) and/or potassium permanganate (KMnO₄) stain. Reversed-phase prep-HPLC was performed on a Teledyne ISCO CombiFlash EZ Prep (RediSep Prep C18, 100 Å, 5 μm, 150 mm × 20 mm; part no. 692203810) using water and acetonitrile eluent, or water +0.1% formic acid and acetonitrile +0.1% formic acid eluent.

Microscale HTE Reagent Screen

Stock solutions, or suspensions, were prepared as shown in each table in the Supporting Information. Reagents were weighed and dissolved or suspended in anhydrous solvent and stirred until either a clear solution or a uniform slurry was formed to achieve the listed concentrations. A 96-well aluminum microvial plate (Analytical Sales and Services, cat. no. 25243) was equipped with oven-dried shell vials (Analytical Sales and Services, cat. no. 884001). Stock solutions were dosed to the appropriate shell vials using single-channel micropipettes. A parylene-coated stir dowel (Analytical Sales and Services, cat. no. 13258) was then added to each vial. The microvial plate was sealed and stirred on a tumble stirrer with heating to the indicated temperature for the planned reaction time in a heating block.

Upon completion, the reactions were diluted with NMP, and an aliquot was taken and filtered through a cotton wool plug, washing with ethyl acetate into a 96-well polypropylene collection plate (Analytical Sales and Services cat. no. 17P687). The solvent was evaporated by blowing nitrogen down on the analytical plate. A solution of benzophenone in MeCN (100 μL) as an internal standard was added, followed by MeCN (200 μL) and water (100 μL). The plate was then mixed by pipet aspiration, and another amount of MeCN (300 μL) and water (100 μL) was added, followed by further pipet aspiration mixing. The reactions were then analyzed by UPLC-MS. The assay yields were produced by measuring the UV absorbance of the desired product relative to the internal standard.

UltraHTE Synthesis

General UltraHTE Information.

Reagents were prepared as stock solutions in oven-dried 1-dram glass vials (Table S9) with significant overage, using concentrations determined by phactor.²⁷ The stock solutions were dispensed using a micropipette into a 384-well source plate (Analytical Sales part #38120) as mapped by NanoChem (Figure S26). Two 384-well source plates were created, each containing stock solutions of the 6 amine cores, 2 coupling reagents, 2 bases, and 64 acids. Reactions were prepared at 1 μL scale (0.1 M, 100 nmol) in a polypropylene 1536-well microplate (Analytical Sales #15020) using an SPT Labtech mosquito liquid handing robot. Plate sealing was done with a peel heat seal (4titude part #4ti-0521) using a 4titude 4s³ Semi-Automatic Sheet Heat Sealer equipped with a 384/1536 support adapter (4titude parts #4ti-0615 and #4ti-0625) and sealing frame (4titude part #4ti-0612). The mosquito was used for addition of LiOH· H₂O, reaction quenching, dilution, and reformatting for UPLC-MS and bioassays. UPLC-MS analysis was performed using the protocols outlined in the General Information. Data were processed using Virscidian Analytical Studio and visualized using phactor.

UltraHTE Protocol.

The mosquito robot was loaded with a 384-source plate and a 1,536-reaction plate. Using the multiaspirate function, 245 nL of acid (0.9 M), coupling reagent (0.9 M), and base (1.8 M) stock solutions were aspirated into the pipet tips from the source plate, each separated by pure NMP (5 nL). The mixture was dispensed to the reaction plate and mixed with aspirate/dispense cycles (5 × 300 nL). The pipet tips were changed, and 250 nL amine (0.4 M) stock solution was transferred from the source plate to the reaction plate, followed by aspirate/dispense mixing (5 × 300 nL). The pipet tips were changed, and the above steps were repeated for all combinations of amines, acids, coupling reagents, and bases (Figure S27). The reaction plate was sealed, centrifuged (10 min), and heated overnight at 80 °C from above and 70 °C from below. The reaction plate was then cooled to room temperature, centrifuged (10 min), the seal removed, and loaded onto the mosquito along with a 384-source plate containing LiOH·H₂O in water (1.5 M). From the source plate, 400 nL LiOH·H₂O solution was dosed to each well of the reaction plate, followed by aspirate/dispense mixing (5 × 300 nL). The reaction plate was then sealed, centrifuged (10 min), and heated to 30 °C overnight. The reaction plate was then cooled to room temperature, centrifuged (10 min), the seal removed, and loaded onto the mosquito along with a 384-source plate containing pure DMSO. From the source plate, 6 μL of DMSO was dosed to each well of the reaction plate, followed by aspirate/dispense mixing (5 × 1 μL). The mosquito was then used to transfer 9 μL of solution from each well of the 1,536-reaction plate to four 384-well bioassay plates (Labcyte part #LP-0200). The mosquito was then used to transfer 1 μL from each of the bioassay plates to a 384-well analytical plate (Fisherbrand part #12566615) predosed by micropipette with 30 μL of 0.0128 mg/mL biphenyl internal standard in DMSO. The bioassay plates were then sealed, centrifuged, and stored at −80 °C for future use. The analytical plates were sealed and analyzed by UPLC-MS. The data was processed using Viscidian Analytical Studio and visualized using phactor

Assay Plate Creation

Crude reaction mixtures (CRMs) at 10 mM were diluted down to 1 mM using the Agilent Bravo, 384ST Bravo head, and 10 μL tips (10734–202). 1 μL was transferred to LDV and backfilled with 9 μL DMSO using a small tube dispensing cassette (Cat#: 24073290) on Thermoscientific MultiDrop Combi. CRMs were plated for bioassay using Acoustic Droplet Ejection on the Echo 555 in conjunction with the corresponding software suite (Echo Dose Response, Echo Cherry Pick). CRMs were plated in Echo-qualified low-dead-volume plates (LP-0200), and varying volumes were dispensed according to desired concentration and plate layout. ASMS pools and dose-response curves were plated using Echo Cherry Pick on the Echo 555, and the Tempo scheduling software was utilized to move plates to and from the Echo on the Beckman Access system. All source plates were covered by MicroClime lids saturated with DMSO during plating. Source, intermediate, and destination plates were heat-sealed on the Agilent PlateLoc with an aluminum peelable seal (24214–001) at 170 °C for 1.5 s. All samples were stored at −80 °C.

Protein Expression and Purification

Human Cdk2 and cyclin E1 constructs were expressed separately and then combined during the purification process to form the phosphorylated complex. Cdk2 (1–298) was prepared as an N-terminally tagged fusion (TwinStrep-MBP-3C-His10-Avi-TEV) and coexpressed with CAK1 (yeast, 1–368) using a pFastBac Dual vector. Cdk2 was expressed in Sf9 cells (BV) and was initially enriched using a Talon column prior to tag removal with TEV. The protein was then further treated with previously isolated CAK1 (1:20 - CAK1:protein, 5 mM MgCl₂, 10 mM ATP) at 4 °C overnight to drive the phosphorylation as close to 100% as possible. The cleaved protein was further cleaned up by flowing through tandem HisTrap HP columns and collecting the flow-through.

Cyclin E1 (96–378) was prepared as an N-terminally tagged fusion protein (His6-TEV) expressed in Sf9 cells (BV, pFastBac1). Cyclin E1-containing lysates were loaded onto tandem Talon columns, followed by a wash step, and then the purified pCdk2 was passed over the cyclin E immobilized resin to allow for complex formation. The complex was eluted off the column, followed by cleavage overnight with TEV. Desalting and further cleanup by another pass over tandem Talon columns were completed, collecting the complex in the flow-through. The complex was further purified using a Mono Q column before a final sizing step on a Superdex 75 column.

ADP-Glo Bioluminescence Kinase Assay

Materials.

50 mM HEPES, pH 7.5 (Boston Bioproducts, BBH-75), 20 mM NaCl (Sigma, S6546), 3 mM MgCl₂ (ThermoFisher Scientific, AM9530G), 3 mM MnCl₂ (Sigma, M1787), Tween-20 (TekNova, T0124), and DTT (Sigma, 43816). The peptide substrate for Cdk2 reactions is from Enzo Life Sciences (cat#: BML-P265–0001, aa seq: HHASPRK). Proxiplates (PerkinElmer, 6008260), benchtop plate centrifuge (Beckman, Allegra X-14R), ADP-Glo Kinase Assay (Promega, V9101) and Envision Plate Reader (PerkinElmer, 2104–0010).

Procedure.

Enzyme solution (2X) in assay buffer (1X) was added to plates via multichannel pipet (2.5 μL/well), followed by substrate solution (2X) in assay buffer (1X). The plates were incubated at room temperature for 5, 10, 15, 30, and 60 min. Then, ADP-Glo Reagent solution (5 μL/well) was added to stop reactions at different incubation times. The plates were then centrifuged at 1000g for 1 min, then incubated at room temperature for 40 min. Then, 10 μL ADP-Glo detection solution was added, and the plates were centrifuged at 1000g for 1 min and incubated at room temperature for 30 min. Finally, the plates were read on an Envision plate read with Luminescence protocol.

Results.

Assay linearity was determined up to 60 min at room temperature using varying Cdk enzyme concentrations (2.5, 5, 10, and 20 nM) at a fixed 5 nM cyclin concentration. In the case of the Cdk/cyclin complexes, total protein concentrations were used. 5 nM Cdk protein and 5 nM cyclin protein, or 5 nM Cdk/cyclin complexes, were linear up to at least 60 min and are the conditions used in the current experiment. ADP-Glo assay kit (Promega) was used to measure the amount of ATP converted to ADP in the reaction. A standard ADP curve was used to determine amounts of ADP produced (see manufacturer’s protocol⁹¹).

ASMS Instrumentation and Methods

ALIS System.

The affinity selection mass spectrometry system, ALIS, is based on the original system described by Annis et al.⁹² The 2D LC-MS system is an Acquity H-class UPLC instrument platform (Waters Corp., Milford, MA). The ALIS system contained a bioSample Manager (FTN-H, plate holder, and sample injection system), bioQuaternary solvent manager (SEC elution), column manager (valving system with sample catch loop), PDA detector (protein/complex detection), TUV detector (ensuring complete protein transition to sample catch loop), binary solvent manager (reversed-phase chromatography, RPC), and a Xevo G2-XS QTof mass spectrometer (compound detection). Sample and sample plates were placed onto the sample manager, and the system was chilled to 4 °C. A total of 10 μL of sample was injected onto the SEC column (Waters BEH 2.5 μm, 125Å, 4.6 mm ID × 30 mm guard column), and protein and small molecule SEC separation was achieved at 10 °C with 0.1 M ammonium acetate (NH4OAc) running buffer, pH 7.5, flowing at 400 μL/min. Once protein was detected by the PDA, the valving system diverted the flow onto a sample loop, collecting ~0.1 mL of protein/complex, which was subsequently diverted onto the RPC column (Waters BEH C18 1 mm × 50 mm, 1.7 μm) for compound separation and detection by mass spectrometry. Compound separation was achieved by RPC using a linear 5-min gradient of water (A phase) and acetonitrile (B phase) with 0.1% formic acid at 60 °C flowing at 400 μL/min. The mass spectrometer was equipped with a lockspray exact mass ESI source. The instrument was operated in ToF-only resolution mode with source temperature and desolvation gas temperature at 100 and 500 °C, respectively, cone gas flow and desolvation gas (N₂) flow at 100 and 1000 L/h, respectively, and sampling cone voltage at 40 V, with a capillary voltage of 3.0 kV in positive ion mode. The ToF was set to monitor a mass range from 100 to 1000 m/z.

Sample Preparation.

31 different compound pools, each containing 18–25 different crude reactions, were plated (0.4 μL in 100% DMSO) into a 384-well plate. Before running the ALIS experiment, each of the 31 pools was analyzed for quality control using ToF spectrometry, and all expected masses were detected. The 31 different pools were plated repeatedly 9 times across the plate (Figure S59, Table S16), with 3 total plates created. To each well, 40 μL of protein solution (50 mM HEPES, 150 mM NaCl, 0.5 mM DTT, pH 7.5) was added. The protein Invertase (Millipore Sigma, I4504) was used as a control and marker protein for each injection and was kept constant at 5 μM across all wells. Cdk2/CycE was titrated across each pool at 9 different concentrations (0, 0.05, 0.1, 0.5, 1, 2, 5, 10, 20 μM).

Data Analysis.

ALIS data was acquired using MassLynx software (Waters Corp.) and analyzed using Analytical Studio Express and Analytical Studio Pro (ASpro) software (Virscidian, Inc., Cary, NC). The search was performed by looking for each mass contained in each unique pool. The search criteria were set up in ASpro as described in Table S17. The exact masses in each pool were identified, and from that, the lowest enzyme concentration at which each compound bound identified.

Direct-to-Biology X-ray Crystallography

Crystallization, Soaking, and Data Collection.

The phosphorylated complex was crystallized using an optimized screen consisting of 14–19% PEG 3350 and 110–145 mM sodium citrate. 150 nL of pCdk2-cyclin E1 at 6.9 mg/mL was combined with 150 nL of a reservoir solution in SWISSCI 3 lens plates using a mosquito LCP robot, and incubated against 45 μL reservoir solution at 20 °C. Initial crystal growth was observed in 24–48 h, and crystals reached their maximum size in about 7–10 days.

The crude reaction mixtures were supplied at a concentration of 10 mM (assuming 100% conversion) in DMSO. The DMSO was removed using a GeneVac EZ-2 ELITE evaporator, and the mixtures were resuspended in 100% DMSO at a final concentration of 40 mM. The mixtures were then diluted to 1 mM in a soaking solution consisting of 26% PEG 3350, 85 mM sodium citrate, and 16% ethylene glycol, with little to no issue with compound solubility. 1.5 μL of each soaking solution was transferred to an empty well of the crystallization plate, and crystals were then transferred directly into the solution. The plate was resealed, and the crystals were left to incubate for 6–8 h prior to harvesting. The soaking solution was sufficient for cryoprotection, and crystals were mounted directly from the soaking drop, flash frozen in liquid nitrogen.

Data sets were collected on the 8.3.1 beamline at the ALS (Lawrence Berkeley National Laboratory) at 100 K with a wavelength of 1.11582 Å. Molecular replacement was completed using DIMPLE with an internal, unbound search model of the phosphorylated complex. Model building was completed in Coot and multiple rounds of refinement using Phenix. Ramachandran statistics: 9OB2 (97.17% Favored, 2.3% Allowed, 0.53% Outliers), 9OB3 (97.35% Favored, 2.3% Allowed, 0.35% Outliers), 9OB4 (97.52% Favored, 1.95% Allowed, 0.53% Outliers), 9OB5 (96.99% Favored, 2.65% Allowed, 0.35% Outliers), 9OB6 (97.17% Favored, 2.3% Allowed, 0.53% Outliers).

Substrate Synthesis

rac-Benzyl (1-(tert-Butyl)-3-((1S,3R)-3-hydroxycyclopentyl)-1H-pyrazol-5-yl)carbamate (S30).

Two batches were carried out in parallel. To a solution of benzyl (1-(tert-butyl)-3-(3-oxocyclopentyl)-1H-pyrazol-5-yl)carbamate S29 (100 g, 281 mmol, 1.00 equiv) in THF (1.00 L) was added LiBHEt₃ (1.00 M, 422 mL, 1.50 equiv) dropwise at −65 °C. The mixture was stirred at −65 °C for 1 h. Two batches of the mixture were poured into sat. NH₄Cl (2.50 L) and extracted with ethyl acetate (1.50 L × 2). The combined organic layers were washed with HCl (1.00 M, 1.50 L), brine (1.50 L × 2), and concentrated to obtain a residue. The residue was triturated with petroleum ether/ethyl acetate = 1/1 (1.00 L × 2) at 20 °C for 30 min, filtered, and the filtrate was concentrated to yield S30 (120 g, 320 mmol, 57% yield, 95.3% purity) as a white solid. LCMS: Rt = 0.857 min, m/z = 358.3 (M+H)⁺; HPLC: Rt = 2.729 min, 95.3% purity under 220 nm; SFC: (Racemate), product: Rt = 2.023 min, product: Rt = 2.263 min; ¹H NMR (400 MHz, DMSO-d₆) δ 9.07 (s, 1H), 7.38–7.32 (m, 5H), 5.93 (s, 1H), 5.12 (s, 2H), 4.56 (s, 1H), 4.18–4.12 (m, 1H), 2.94–2.88 (m, 1H), 2.21–2.18 (m, 1H), 1.87–1.86 (m, 1H), 1.74–1.71 (m, 2H), 1.58–1.57 (m, 2H), 1.48 (s, 9H).

Benzyl (1-(tert-Butyl)-3-((1S,3R)-3-hydroxycyclopentyl)-1H-pyrazol-5-yl)carbamate (S31) and Benzyl (1-(tert-Butyl)-3-((1R,3S)-3-hydroxycyclopentyl)-1H-pyrazol-5-yl)carbamate (S32).

S30 (120 g, 336 mmol) was purified by SFC (column: DAICEL CHIRALPAK IG; 250 mm × 50 mm, 10 μm); mobile phase: [ACN/MeOH (0.1%NH₃H₂O)]; B%: 25%–25%, 3.6 min) to give S31 (51.0 g, 142 mmol, 42% yield, 99.7% purity) as a white solid and S32 (50.0 g, 139 mmol, 42% yield, 99.6% purity) as a white solid. MS: product: m/z = 358.1 (M+H)⁺; HPLC: product: Rt = 2.684 min, 99.7% purity under 220 nm; SFC: (Homochiral), product: Rt = 2.010 min, 100% ee under 220 nm; (¹H NMR: 400 MHz, DMSO-d₆) δ 9.06 (s, 1H), 7.38–7.33 (m, 5H), 5.92 (s, 1H), 5.11 (s, 2H), 4.56 (d, J = 4.4 H_Z, 1H), 4.16–4.12 (m, 1H), 2.91–2.87 (m, 1H), 2.20–2.17 (m, 1H), 1.86–1.81 (m, 1H), 1.73–1.70 (m, 2H), 1.57–1.50 (m, 2H), 1.47 (s, 9H); MS: product: m/z = 358.1 (M+H)⁺; HPLC: product: Rt = 2.672 min, 99.6% purity under 220 nm; SFC: (Homochiral), product: Rt = 2.225 min, 98.7% ee under 220 nm; ¹H NMR (400 MHz, DMSO-d₆) δ 9.06 (s, 1H), 7.38–7.33 (m, 5H), 5.92 (s, 1H), 5.12 (s, 2H), 4.56 (d, J = 4.4 HZ, 1H), 4.16–4.12 (m, 1H), 2.91–2.87 (m, 1H), 2.20–2.17 (m, 1H), 1.86–1.81 (m, 1H), 1.73–1.70 (m, 2H), 1.57–1.50 (m, 2H), 1.47 (s, 9H).

Benzyl (1-(tert-Butyl)-3-((1S,3R)-3-(((4-nitrophenoxy)-carbonyl)oxy)cyclopentyl)-1H-pyrazol-5-yl)carbamate (S33).

To a solution of S31 (8.00 g, 22.4 mmol, 1.00 equiv) and 4-nitrophenyl chloroformate (9.02 g, 44.8 mmol, 2.00 equiv) in DCM (80.0 mL) was added DMAP (273 mg, 2.24 mmol, 0.10 equiv) and Py (5.31 g, 67.2 mmol, 5.42 mL, 3.00 equiv) at 0 °C. The reaction mixture was then stirred at 25 °C for 12 h. The reaction mixture was concentrated under reduced pressure to give a residue. The residue was purified by column chromatography (SiO₂, petroleum ether: ethyl acetate = 1:0 to 0:1) to give S33 (13.4 g, crude) as yellow oil. ¹H NMR (400 MHz, CDCl₃) δ 8.30–8.24 (m, 2H), 7.45–7.30 (m, 7H), 6.40–6.05 (m, 2H), 5.20–5.15 (m, 3H), 3.08–3.25 (m, 1H), 2.70–2.55 (m, 1H), 2.20–2.10 (m, 1H), 2.08–1.90 (m, 4H), 1.60–1.50 (s, 9H).

Benzyl (3-((1S,3R)-3-(((4-Nitrophenoxy)carbonyl)oxy)-cyclopentyl)-1H-pyrazol-5-yl)carbamate (S34).

A solution of S33 (13.4 g, 25.6 mmol, 1.00 equiv) in formic acid (100 mL) was stirred at 80 °C for 12 h. The reaction mixture was concentrated under reduced pressure to remove most of the formic acid and then quenched by saturated sodium bicarbonate solution. It was extracted with dichloromethane (40.0 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure to give a residue. The residue was purified by column chromatography (SiO₂, petroleum ether: ethyl acetate = 1:0 to 0:1) to give S34 (4.14 g, 7.62 mmol, 30% yield, 86% purity) as a colorless oil. LCMS: product: RT = 0.945 min, m/z = 467.2 (M+H)⁺.

Benzyl (3-((1S,3R)-3-(((1-Methylcyclopropyl)carbamoyl)-oxy)cyclopentyl)-1H-pyrazol-5-yl)carbamate (S35).

To a solution of S34 (1.50 g, 3.22 mmol, 1.00 equiv) in THF (15.0 mL) was added 1-methylcyclopropan-1-amine hydrochloride (692 mg, 6.43 mmol, 2.00 equiv) and DIPEA (2.08 g, 16.1 mmol, 2.80 mL, 5.00 equiv) dropwise at 25 °C. The mixture was heated to 60 °C and stirred at 60 °C for 2 h. The reaction mixture was poured into water (20.0 mL) and extracted with DCM (15.0 mL × 2). The combined organic layers were washed with brine (20.0 mL), dried over Na₂SO₄, filtered, and concentrated under reduced pressure to give a residue. The residue was purified by column chromatography (SiO₂, petroleum ether/ethyl acetate = 5/1 to 1/1, petroleum ether/ethyl acetate = 0/1) to give S35 (848 mg, 2.04 mmol, 64% yield, 96.0% purity) as a white solid. LCMS: product: RT = 0.866 min, 96.0% purity under 220 nm.

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl (1-methylcyclopropyl)carbamate (5).

To a solution of S35 (848 mg, 2.13 mmol, 1.00 equiv) in MeOH (20.0 mL) was added Pd/C (170 mg, 2.13 mmol, 10% purity) at 25 °C under N₂. The mixture was then degassed and purged with H₂ 3 times at 25 °C. Subsequently, the mixture was heated to 30 °C and stirred for 12 h under H₂ (50 psi). After completion, the reaction mixture was cooled to 25 °C and filtered. The filtrate was concentrated to yield a crude product. The residue was purified by preparative HPLC (basic condition column: Waters XBridge 150 × 25 mm × 5 μm; mobile phase: [water (ammonium hydroxide v/v)—ACN]; B%: 10%–38%, 8 min) and lyophilized to give 5 (434 mg, 1.61 mmol, 76% yield, 97.8% purity) as a white solid. LCMS: product: RT = 0.615 min, m/z = 265.2 (M+H)⁺; MS: m/z = 265.0 (M+H)⁺; HPLC: product: RT = 0.779 min, 97.8% purity under 220 nm; SFC: product: RT = 1.329 min, 100% ee under 220 nm; the compound appears as a 4.8:1 mixture of rotamers on the NMR time scale, ¹H NMR (500 MHz, DMSO-d₆) δ 11.05 (s, br, 1H), 7.33 (s, 1H, Rot. A), 7.08 (s, 1H, Rot. B), 5.18 (s, 1H), 5.00 (s, 1H, Rot. B), 4.94 (s, 1H, Rot. A), 4.51 (s, 2H, br), 3.00–2.80 (m, 1H), 2.44–2.30 (m, 1H), 2.00–1.87 (m, 1H), 1.87–1.78 (m, 1H), 1.78–1.57 (m, 2H), 1.57–1.45 (m, 1H), 1.23 (s, 3H), 0.64–0.55 (m, 2H), 0.50–0.43 (m, 2H); ¹³C NMR (126 MHz, DMSO-d₆) δ 155.8, 87.9, 75.1 (Rot. B), 74.6 (Rot. A), 39.1 (under DMSO solvent peak), 35.2, 32.9 (Rot. B), 32.3 (Rot. A), 31.2 (Rot. B), 30.8 (Rot. A), 28.6, 23.8 (Rot. B), 22.9 (Rot. A), 14.6 (Rot. B), 13.8 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₁₃H₂₁N₄O₂]⁺ expected 265.1659; found 265.1658. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

Benzyl (1-(tert-Butyl)-3-((1S,3R)-3-((isopropylcarbamoyl)-oxy)cyclopentyl)-1H-pyrazol-5-yl)carbamate (S36).

To a solution of S33 (2.00 g, 3.83 mmol, 1.00 equiv) in THF (20.0 mL) was added DIPEA (2.47 g, 19.2 mmol, 3.33 mL, 5.00 equiv) and isopropylamine (679 mg, 11.5 mmol, 987 μL, 3.00 equiv) at 25 °C. The solution was heated to 60 °C and stirred at 60 °C for 12 h. The mixture was cooled to 25 °C and poured into sat. NH₄Cl (20.0 mL), then extracted with ethyl acetate (20 mL × 2). The combined organic layer was washed with brine (20 mL), dried over Na₂SO₄, filtered, and concentrated to give a residue. The residue was purified by column chromatography (SiO₂, petroleum ether/ethyl acetate = 100/1 to 3/1, petroleum ether/ethyl acetate = 2/1) twice to give S36 (1.10 g, crude) as a colorless oil.

(1R,3S)-3-(5-Amino-1-(tert-butyl)-1H-pyrazol-3-yl)-cyclopentyl isopropylcarbamate (S37).

To a solution of S36 (1.10 g, crude) in MeOH (11.0 mL) was added Pd/C (0.22 g, 497 μmol, 10% purity, 0.20 equiv) at 25 °C. The mixture was degassed and purged with H₂ 3 times at 25 °C. Then, the mixture was stirred at 25 °C for 12 h under H₂ (50 psi). The mixture was filtered, and the filtrate was concentrated to give S37 (800 mg, crude) as a colorless oil. LCMS: product: RT = 0.700 min, m/z = 309.3 (M+H)⁺.

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl isopropylcarbamate (6).

A solution of S37 (800 mg, crude) in HCl (20.0 mL, 36% aqueous solution) was stirred at 25 °C for 84 h. The solution was adjusted to pH 9 with sat. Na₂CO₃ at 0 °C and extracted with ethyl acetate (20 mL × 3). The combined organic layer was dried over Na₂SO₄, filtered and concentrated to give a residue. The residue was purified by prep-HPLC (column: Waters XBridge C18 150 × 50 mm × 10 μm; mobile phase: [water (NH₄HCO₃)—ACN]; B%: 10%–40%, 10 min) and lyophilized to give 6 (392 mg, 1.54 mmol, 40% over 3 steps, 99.4% purity) as a white solid. LCMS: product: RT = 0.727 min, m/z = 253.1 (M+H)⁺; MS: product: m/z = 253.0 (M+H)⁺; HPLC: product: RT = 1.700 min, 99.4% purity under 220 nm; the compound appears as a 4.8:1 mixture of rotamers on the NMR time scale, ¹H NMR (500 MHz, DMSO-d₆) δ 11.10 (s, 1H, Rot. A), 10.95 (s, 1H, Rot. B), 6.94 (d, J = 6.6 Hz, 1H, Rot. A), 6.59 (s, 1H, Rot. B), 5.23 (s, 1H, Rot. A), 5.11 (s, 1H, Rot. B), 4.96 (s, 1H), 4.79 (s, 2H, Rot. B), 4.33 (s, 2H, Rot. A), 3.65–3.48 (m, 1H), 2.90 (s, 1H), 2.37 (s, 1H), 1.92 (s, 1H), 1.90–1.78 (m, 1H), 1.75–1.60 (m, 2H), 1.55 (s, 1H), 1.04 (d, J= 6.5 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 155.2 (Rot. A), 155.1 (Rot. B), 146.8, 89.0, 74.7, 42.1, 39.1, 34.5, 32.3, 30.8, 22.6; HRMS: m/z: [M + H]⁺ calc’d for [C₁₂H₂₁N₄O₂]⁺ expect 253.1659; found 253.1658. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(5-(((Benzyloxy)carbonyl)amino)-1H-pyrazol-3-yl)-cyclopentyl isopropyl(methyl)carbamate (S38).

To a solution of S34 (2.50 g, 5.36 mmol, 1.00 equiv) and DIPEA (3.46 g, 26.8 mmol, 4.67 mL, 5.00 equiv) in THF (25.0 mL) was added N-methylpropan-2-amine (588 mg, 8.04 mmol, 837 μL, 1.50 equiv) at 25 °C. The mixture was heated to 60 °C and stirred at 60 °C for 12 h. The reaction mixture was poured into water (20.0 mL) at 25 °C and then extracted with ethyl acetate (30.0 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure to give a residue. The mixture was purified by column chromatography (SiO₂, petroleum ether/ethyl acetate = 1/0 to 0/1) to afford S38 (1.76 g, 4.21 mmol, 78.5% yield, 95.8% purity) as light-yellow oil. LCMS: product: RT = 0.908 min, m/z= 401.2 (M+H)⁺; ¹H NMR (400 MHz, CDCl₃) δ 10.0 (s, 1H), 8.37 (s, 1H), 7.45–7.30 (m, 5H), 6.39 (s, 1H), 5.19 (s, 3H), 4.30 (br, d, J = 92 Hz, 1H), 3.20–3.05 (m, 1H), 2.68 (s, 3H), 2.55–2.40 (m, 1H), 2.15–2.05 (m, 1H), 2.00–1.80 (m, 4H), 1.06 (d, J = 6.4 Hz, 6H).

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl isopropyl-(methyl)carbamate (7).

To a solution of S38 (1.76 g, 4.39 mmol, 1.00 equiv) in MeOH (20.0 mL) was added Pd/C (200 mg, 10% purity) under N₂. The suspension was degassed under vacuum and purged with H₂ several times. The mixture was stirred at 30 °C under H₂ (50 psi) for 12 h. The mixture was filtered, and the filtrate was concentrated to give a residue. The crude product was purified by prep-HPLC (column: Waters XBridge 150 × 25 mm × 5 μm; mobile phase: [water (ammonium hydroxide v/v)-ACN]; B%: 14%–44%, 8 min) and lyophilized to afford 7 (797 mg, 2.95 mmol, 67% yield, 98.7% purity) as an orange solid. LCMS: product: RT = 0.695 min, m/z = 267.3 (M+H)⁺; MS: product: m/z = 267.1 (M+H)⁺; HPLC: product: RT = 1.210 min, 98.7% purity under 220 nm; SFC: product: RT = 0.758 min, 100% ee under 220 nm; The compound appears as a 1.7:1 mixture of rotamers on the NMR time scale,¹H NMR (500 MHz, DMSO-d₆) δ 11.08 (s, 1H), 5.18 (s, 1H), 4.99 (s, 1H), 4.44 (s, br, 2H), 4.24 (s, 1H, Rot. A), 4.13 (s, 1H, Rot. B) 2.92 (p, J = 8.4 Hz, 1H), 2.64 (s, 3H), 2.40–2.30 (m, 1H), 1.98–1.89 (m, 1H), 1.89–1.78 (m, 1H), 1.76–1.64 (m, 2H), 1.64–1.56 (m, 1H), 1.04 (d,J= 6.8 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 155.0, 87.8, 75.9, 45.8, 39.1 (under DMSO solvent peak), 35.2, 32.5, 30.9, 27.1 (Rot. B), 26.8 (Rot. A), 19.7 (Rot. B), 19.4 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₁₃H₂₃N₄O₂]⁺ expect 267.1816; found 267.1813. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

Benzyl (3-((1S,3R)-3-((((R)-1-Cyclopropylethyl)carbamoyl)-oxy)cyclopentyl)-1H-pyrazol-5-yl)carbamate (S39).

To a solution of S34 (2.00 g, 4.29 mmol, 1.00 equiv) and DIPEA (2.77 g, 21.4 mmol, 3.73 mL, 5.00 equiv) in THF (20.0 mL) was added (R)-1-cyclopropylethan-1-amine (782 mg, 6.43 mmol, 1.50 equiv., HCl) at 25 °C. The mixture was heated to 60 °C and stirred for 12 h. The reaction mixture was poured into water (30.0 mL) at 25 °C and then extracted with ethyl acetate (30.0 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure to give a residue. The residue was purified by column chromatography (SiO₂, petroleum ether/ethyl acetate = 1/0 to 0/1, petroleum ether/ethyl acetate = 0/1) to afford compound S39 (1.85 g, 3.98 mmol, 93% yield, 88.7% purity) as a light-yellow solid. LCMS: product: RT = 0.909 min, m/z = 413.2 (M+H)⁺; ¹H NMR (400 MHz, CDCl₃) δ 9.78 (s, 1H), 8.13 (d, J= 9.2 Hz, 1H), 7.45–7.30 (m, 5H), 6.39 (s, 1H), 5.27–5.10 (m, 3H), 3.20–3.05 (m, 2H), 2.53–2.37 (m, 1H), 2.15–2.05 (m, 1H), 2.00–1.80 (m, 4H), 1.17 (d,J= 6.0 Hz, 3H), 0.85–0.70 (m, 1H), 0.55–0.40 (m, 2H), 0.34 (s, 1H), 0.25–0.15 (m, 1H).

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl ((R)-1-cyclopropylethyl)carbamate (8).

To a solution of S39 (1.85 g, 4.49 mmol, 1.00 equiv) in MeOH (20.0 mL) was added Pd/C (200 mg, 795 μmol, 10% purity) at 25 °C under N₂. The suspension was degassed under vacuum and purged with H₂ several times. The mixture was stirred at 30 °C under H₂ (50 psi) for 12 h. The mixture was filtered, and the filtrate was concentrated to give a residue. The crude product was purified by prep-HPLC (column: Waters XBridge 150 × 25 mm × 5 μm; mobile phase: [water (ammonium hydroxide v/v)-ACN]; B%: 15%–45%, 8 min) and lyophilized to afford 8 (733 mg, 2.61 mmol, 58% yield, 99.0% purity) as a white solid. LCMS: product: RT = 0.692 min, m/z = 279.3 (M+H)⁺; MS: product: m/z = 279.1 (M+H)⁺; HPLC: product: RT = 1.207 min, 99.0% purity under 220 nm; SFC: product: RT = 1.294 min, 100% ee under 220 nm; the compound appears as a 9:1 mixture of rotamers on the NMR time scale, ¹H NMR (500 MHz, DMSO-d₆) δ 11.07 (s, 1H), 6.99 (d, J = 7.3 Hz, 1H, Rot. A), 6.66 (s, 1H, Rot. B), 5.19 (s, 1H), 4.94 (s, 1H), 4.36 (s, br, 2H), 3.05–2.93 (m, 1H), 2.92–2.80 (m, 1H), 2.45–2.28 (m, 1H), 2.00–1.88 (m, 1H), 1.88–1.78 (m, 1H), 1.74–1.60 (m, 2H), 1.59–1.46 (m, 1H), 1.08 (d, J = 6.7 Hz, 3H), 0.86–0.74 (m, 1H), 0.42–0.34 (m, 1H), 0.34–0.27 (m, 1H), 0.27–0.20 (m, 1H), 0.14–0.05 (m, 1H); ¹³C NMR (126 MHz, DMSO-d₆) δ 155.5, 74.7, 49.9, 39.1 (under DMSO solvent peak), 32.3, 30.8, 20.6, 17.3, 2.7, 2.6; HRMS: m/z: [M + H]⁺ calc’d for [C₁₄H₂₃N₄O₂]⁺ expect 279.1816; found 279.1807. Three pyrazole ¹³C peaks could not be observed due to peak broadening.

(1R,3S)-3-(5-(((Benzyloxy)carbonyl)amino)-1H-pyrazol-3-yl)-cyclopentyl 2,2-dimethylazetidine-1-carboxylate (S40).

To a solution of S34 (2.88 g, 3.70 mmol, 60% purity, 1.00 equiv) and 2,2-dimethylazetidine hydrochloride (496 mg, 4.11 mmol, 1.10 equiv) in THF (28.9 mL) was added DIPEA (2.40 g, 18.6 mmol, 3.24 mL, 5.00 equiv) at 25 °C. The reaction mixture was heated to 45 °C and stirred for 4 h. The mixture was poured into H₂O (30 mL) and extracted with DCM (30 mL × 2). The combined organic layers were dried over Na₂SO₄, filtered and concentrated to give a crude product. The crude product was purified by prep-HPLC (column: Welch Ultimate XB—SiOH 250 × 50 × 10 μm; mobile phase: [hexane—EtOH (0.1% FA)]; B%: 1%–30%, 15 min), then concentrated to give S40 (1.29 g, 1.65 mmol) as yellow oil. S40 (1.29 g, 1.65 mmol) was purified by prep—HPLC (column: Kromasil Eternity XT 250 × 80 mm × 10 μm; mobile phase: [water (NH₄HCO₃)—ACN]; B%: 32%–62%, 20 min). The mixture was concentrated and extracted with dichloromethane (6.0 mL × 3). The combined organic layers were washed with brine (6.0 mL), dried over Na₂SO₄, filtered and concentrated to give S40 (651 mg, 42% yield, 1.55 mmol, 98.0% purity) as yellow oil. LCMS: product: RT = 0.731 min, m/z = 413.1 (M+H)⁺.

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl 2,2-dimethylazetidine-1-carboxylate (9).

A mixture of S40 (650 mg, 1.58 mmol, 98.0% purity, 1.00 equiv) in MeOH (6.00 mL) was added to Pd/C (70.0 mg, 10% purity) under N₂. The suspension was degassed under vacuum and purged with H₂ several times. The mixture was stirred under H₂ (50 psi) at 30 °C for 12 h. The reaction mixture was filtered and concentrated under reduced pressure to give a residue. The crude product was purified by prep—HPLC (column: Waters XBridge 150 × 25 mm × 5 μm; mobile phase: [water (ammonium hydroxide v/v)—ACN]; B%: 13%–43%, 9 min) and concentrated to remove most of ACN and then lyophilized. The product was lyophilized again to give 9 (327 mg, 1.16 mmol, 74% yield, 98.7% purity) as a light-yellow solid. LCMS: product: RT = 0.570 min, m/z = 279.1 (M+H)⁺; MS: m/z = 279.2 (M+H)⁺; HPLC: product: RT = 2.693 min, 98.7% purity under 220 nm; SFC: product: RT = 1.265 min, 100% ee under 220 nm; the compound appears as a 1.7:1 mixture of rotamers on the NMR time scale, ¹H NMR (500 MHz, DMSO-d₆) δ 11.08 (s, 1H), 5.17 (s, 1H), 5.03–4.96 (m, 1H, Rot. A), 4.96–4.89 (m, 1H, Rot. B), 4.46 (s, br, 2H), 3.73 (t, J = 7.6 Hz, 1H, Rot. B), 3.65 (t, J = 7.4 Hz, 1H, Rot. A), 2.90 (p, J = 8.2 Hz, 1H), 2.43–2.27 (m, 1H), 2.00–1.87 (m, 1H), 1.95 (t, J= 7.6 Hz, 2H), 1.87–1.76 (m, 1H), 1.75–1.52 (m, 3H), 1.37 (s, 6H, Rot. B), 1.34 (d, J = 3.6 Hz, 6H, Rot. A); ¹³C NMR (126 MHz, DMSO-d₆) δ 154.9 (Rot. A), 153.8 (Rot. B), 87.9, 75.2 (Rot. A), 74.9 (Rot. B), 65.2 (Rot. A), 64.7 (Rot. B), 43.8 (Rot. B), 42.4 (Rot. A), 39.1 (Rot. B, under DMSO solvent peak), 39.1 (Rot. A, under DMSO solvent peak), 35.4, 32.8 (Rot. A), 32.4 (Rot. B), 31.1 (Rot. A) 30.9 (Rot. B), 30.4 (Rot. B), 30.1 (Rot. A), 27.0 (Rot. A), 26.7 (Rot. A), 26.2 (Rot. B), 26.1 (Rot. B); HRMS: m/z: [M + H]⁺ calc’d for [C₁₄H₂₃N₄O₂]⁺ expect 279.1816; found 279.1812. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

(1R,3S)-3-(5-(((Benzyloxy)carbonyl)amino)-1-(tert-butyl)-1H-pyrazol-3-yl)cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (S41).

To a solution of S33 (1.92 g, 3.67 mmol, 1.00 equiv) and DIPEA (2.37 g, 18.3 mmol, 3.20 mL, 5.00 equiv) in THF (20.0 mL) was added (2S,5S)-2,5-dimethylpyrrolidine (546 mg, 5.51 mmol, 1.50 equiv) at 25 °C. The mixture was heated to 60 °C ion mixture was poured into water (40 mL) at 25 °C and extracted with ethyl acetate (50 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄, filtered, and concentrated under reduced pressure to give a residue. The crude product was purified by column chromatography (SiO₂, petroleum ether/ethyl acetate = 1/0 to 1/1) to give S41 (1.50 g, 3.11 mmol, 85% yield, 100% purity), which was obtained as colorless oil. LCMS: product: RT = 1.024 min, m/z = 483.4 (M+H)⁺; LCMS: product: RT = 0.952 min, m/z = 483.4 (M+H)⁺; ¹H NMR (400 MHz, CDCl₃) δ 7.35 (s, 5H), 6.10 (s, 1H), 5.20–5.14 (m, 3H), 4.03–3.90 (m, 2H), 3.07–3.03 (m, 1H), 2.54–2.48 (m, 1H), 2.10–1.84 (m, 9H), 1.58 (s, 9H), 1.16 (d, J = 6.4 Hz, 3H), 1.08 (d, J = 6.4 Hz, 3H).

(1R,3S)-3-(5-Amino-1-(tert-butyl)-1H-pyrazol-3-yl)-cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (S42).

To a solution of S41 (1.50 g, 3.11 mmol, 1.00 equiv) in MeOH (15.0 mL) was added Pd/C (150 mg, 10.0% purity) under N₂. The suspension was degassed under vacuum and purged with H₂ several times. The mixture was stirred at 30 °C under H₂ (50 psi) for 12 h. The mixture was filtered, and the filtrate was concentrated to give S42 (1.10 g, crude) as brown oil. LCMS: product: RT = 0.776 min, m/z = 349.3 (M+H)⁺; ¹H NMR (400 MHz, CDCl₃) δ 5.42 (s, 1H), 5.16 (s, 1H), 4.00 (t, J = 6.4 Hz, 1H), 3.93 (t, J = 6.4 Hz, 1H), 3.48 (s, 2H), 3.02–2.75 (m, 1H), 2.15–2.48 (m, 1H), 2.20–2.10 (m, 2H), 2.06–1.73 (m, 7H), 1.61 (s, 9H), 1.17 (d, J = 6.0 Hz, 3H), 1.11 (d, J = 6.0 Hz, 3H).

(1R,3S)-3-(5-Amino-1H-pyrazol-3-yl)cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (10).

A solution of S42 (1.10 g, 3.16 mmol, 1.00 equiv) in H₂SO₄ (18.4 g, 187 mmol, 10.0 mL, 59.4 equiv) was stirred at 25 °C for 2 h. The mixture was adjusted to pH = 7–8 with 1 N NaOH and extracted with dichloromethane (50 mL × 3). The combined organic layers were dried over Na₂SO₄, filtered, and concentrated to give a residue. The crude product was purified by prep-HPLC (column: Phenomenex Luna C18 150 × 40 mm × 15 μm; mobile phase: [water (FA) —ACN]; B%: 10%–40%, 10 min). The product-containing fractions were combined, and the mixture was adjusted to pH = 7–8 with saturated NaHCO₃. The mixture was then concentrated to remove most of the ACN and extracted with dichloromethane (50 mL × 3). The combined organic layers were dried over Na₂SO₄, filtered, and concentrated to afford compound 10 (484 mg, 1.65 mmol, 52% yield, 99.5% purity) as a brown solid. LCMS: product: RT = 0.700 min, m/z = 293.2 (M+H)⁺; MS: product: m/z = 293.2 (M+H)⁺; HPLC: product: RT = 1.740 min, 99.5% purity under 220 nm; SFC: product: RT = 1.330 min, 100% ee under 220 nm; ¹H NMR (500 MHz, DMSO-d₆) δ 5.17 (s, 1H), 5.03–4.96 (m, 1H), 3.83 (h, J = 6.8 Hz, 2H), 2.96–2.85 (m, 1H), 2.43–2.33 (m, 1H), 2.14–1.98 (m, 2H), 1.98–1.89 (m, 1H), 1.87–1.78 (m, 1H), 1.78–1.62 (m, 2H), 1.62–1.53 (m, 1H), 1.53–1.37 (2H), 1.06 (d, J = 6.7 Hz, 3H), 1.05 (d,J = 6.8 Hz, 3H); ¹³C NMR (126 MHz, DMSO-d₆) δ 153.3, 87.7, 75.3, 52.8, 52.5, 39.2 (under DMSO solvent peak), 35.4, 32.6, 31.0, 29.8, 28.8, 20.4, 19.2; HRMS: m/z: [M + H]⁺ calc’d for [C₁₅H₂₅N₄O₂]⁺ expect 293.1972; found 293.1965. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

Synthesis of Compounds Selected for IC₅₀ Measurement

General Procedure.

A flame-dried 2-dram vial equipped with a stir bar was charged with carboxylic acid (2.2 equiv), coupling reagent (2.2 equiv), and NMP (0.2 M with respect to amine). To the resulting mixture, base (4.4 equiv) was added dropwise with stirring. Amine substrate (1.0 equiv) was dissolved in NMP (0.2 M with respect to amine) in a second flame-dried vial. The amine solution was then added dropwise to the carboxylic acid solution with stirring. The resulting reaction mixture was capped and heated to 80 °C overnight. The reaction was then allowed to cool, and an aliquot was taken for UPLC-MS analysis. If multiple isomers or diacetylated products were observed, 1.5 M LiOH·H₂O solution (6.0 equiv) was added, and the reaction was stirred at 30 °C and monitored by UPLC-MS. When multiple isomers and diacetylated products were no longer detected (typically after 1–2 h), the reaction was neutralized with 0.2 M AcOH in H₂O (6.0 equiv). EtOAc was then added, and the organic phase was washed with water and brine, then dried over Na₂SO₄. The resulting crude residue was purified either via silica column chromatography or RP-HPLC, as stated for each specific compound below. The compound was then dissolved in a small amount of acetonitrile and water, frozen at −78 °C, and lyophilized. Any variations to this protocol are specified for each specific compound below.

(1R,3S)-3-(3-(2-(4-(Methylsulfonyl)phenyl)acetamido)-1H-pyrazol-5-yl)cyclopentyl (1-methylcyclopropyl)carbamate (11).

Following the General Procedure using 5 as the amine (13.2 mg, 0.050 mmol), 4-(methylsulfonyl)phenylacetic acid (23.6 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (26% formate salt, 4.70 mg, 0.010 mmol, 20%) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.08 (s, 1H), 10.62 (s, 1H), 7.87 (d, J = 8.3 Hz, 2H), 7.57 (d, J = 8.3 Hz, 2H), 7.33 (s, 1H), 6.25 (s, 1H), 5.01 (s, 1H, Rot. B), 4.95 (s, 1H, Rot. A), 3.72 (s, 2H), 3.18 (s, 3H), 3.05–2.94 (m, 1H), 2.47–2.38 (m, 1H), 2.05–1.91 (m, 1H), 1.91–1.80 (m, 1H), 1.71–1.57 (m, 2H), 1.57–1.45 (m, 1H), 1.22 (s, 3H), 0.62–0.51 (m, 2H), 0.45 (dd, J = 6.0, 4.3 Hz, 2H); ¹³C NMR (126 MHz, DMSO-d₆) δ 167.2, 155.7, 147.0, 146.6, 142.2, 139.1, 130.0, 127.0, 93.3, 74.5, 43.6, 41.9, 39.5 (under DMSO solvent peak), 34.5, 32.3, 30.7, 28.7, 23.0, 13.8; HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₂₉N₄O₅S]⁺ expect 461.1853; found 461.1818; ASMS protein concentration first observed: 0.05 μM; measured IC₅₀: 1.7 nM.

(1R,3S)-3-(3-(Nicotinamido)-1H-pyrazol-5-yl)cyclopentyl 2,2-dimethylazetidine-1-carboxylate (12).

Following the general procedure using 9 as the amine (13.9 mg, 0.050 mmol), nicotinic acid (13.5 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (5–10% MeOH in DCM) yielded the target compound (16.4 mg, 0.043 mmol, 86%) as a pale orange solid. The compound appears as a 1.4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d6) δ 12.25 (s, 1H), 10.97 (s, 1H), 9.11 (d,J = 1.7 Hz, 1H), 8.73 (dd, J = 4.8, 1.4 Hz, 8.31 (dt, J = 8.0, 1.8 Hz, 1H), 7.53 (dd, J = 8.0, 4.8 Hz, 1H), 6.48 (s, 1H), 5.10–5.02 (m, 1H, Rot. A), 5.02 (m, 1H, Rot. B), 3.77 (h,J = 8.0 Hz, 1H), 3.66 (t,J = 7.6 Hz, 1H), 3.12 (p,J = 8.7 Hz, 1H), 2.50–2.38 (m, 1H), 2.10–2.01 (m, 1H), 1.96 (t,J = 8.0 Hz, 2H), 1.92–1.84 (m, 1H), 1.82–1.72 (m, 2H), 1.72–1.63 (m, 1H), 1.39 (s, 6H, Rot. B), 1.35 (s, 3H, Rot. A), 1.34 (s, 3H, Rot. B); ¹³C NMR (126 MHz, DMSO-d6) δ 163.0, 155.0 (Rot. A), 153.8 (Rot. B), 152.1, 148.8, 147.1, 146.9, 135.5, 130.0, 123.5, 94.0, 75.2 (Rot. A), 75.0 (Rot. B), 65.4 (Rot. A), 64.8 (Rot. B), 43.9 (Rot. B), 42.4 (Rot. A), 39.1 (under DMSO solvent peak), 34.7 (Rot. A), 34.4 (Rot. B), 32.8 (Rot. A), 32.5 (Rot. B), 31.11 (Rot. A), 31.0 (Rot. B), 30.5 (Rot. B), 30.1 (Rot. A), 27.0 (Rot. A), 26.8 (Rot. A), 26.3 (Rot. B), 26.1 (Rot. B); HRMS: m/z: [M + H]⁺ calc’d for [C₂₀H₂₆N₅O₃]⁺ expect 384.2030; found 384.2019; ASMS protein concentration first observed: 0.10 μM; measured IC₅₀: 5.2 nM.

(1R,3S)-3-(3-(3-Methylbenzamido)-1H-pyrazol-5-yl)cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (13).

Following the general procedure using 10 as the amine (14.6 mg, 0.050 mmol), 3-methylbenzoic acid (15.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (Col 1:100% EtOAc, Col 2:3–5% MeOH in DCM) and RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (13.4 mg, 0.033 mmol, 65%) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.23 (s, 1H), 10.66 (s, 1H), 7.81 (s, 1H), 7.77 (s, 1H), 7.43–7.27 (m, 2H), 6.46 (s, 1H), 5.05 (s, 1H), 3.84 (p, J = 6.3 Hz, 2H), 3.18–3.05 (m, 1H), 2.52–2.44 (m, 1H), 2.37 (s, 3H), 2.15–1.98 (m, 3H), 1.96–1.85 (m, 1H), 1.83–1.73 (m, 2H), 1.72–1.62 (m, 1H), 1.51–1.38 (m, 2H), 1.07 (d, J = 6.2 Hz, 3H), 1.04 (d, J = 6.2 Hz, 3H);¹³C NMR (126 MHz, DMSO-d₆) δ 164.3, 153.2, 147.1, 137.5, 134.2, 132.0, 128.3, 128.2, 124.8, 94.0, 75.2, 52.8, 52.5, 39.6 (under DMSO solvent peak), 34.8, 32.6, 30.9, 29.8, 28.8, 20.9, 20.4, 19.2; HRMS: m/z: [M + H]⁺ calc’d for [C₂₃H₃₁N₄O₃]⁺ expect 411.2391; found 411.2389; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 11.3 nM. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(3-(Cyclopropanecarboxamido)-1H-pyrazol-5-yl)-cyclopentyl isopropyl(methyl)carbamate (14).

Following the general procedure using 7 as the amine (13.3 mg, 0.050 mmol), cyclo-propanecarboxylic acid (9.5 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, and DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (3–7% MeOH in DCM) yielded the target compound (2.8 mg, 0.008 mmol, 17%) as a white solid. The compound appears as a 1.3:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.01 (s, 1H), 10.51 (s, 1H), 6.28 (s, 1H), 5.02 (s, 1H), 4.24 (s, 1H, Rot. A), 4.10 (s, 1H, Rot. B), 3.13–3.00 (m, 1H), 2.63 (s, 3H), 2.45–2.35 (m, 1H), 2.07–1.95 (m, 1H), 1.94–1.84 (m, 1H), 1.84–1.79 (m, 1H), 1.78–1.71 (m, 2H), 1.71–1.62 (m, 1H), 1.02 (d, J = 6.7 Hz, 6H), 0.78–0.67 (m, 4H); ¹³C NMR (126 MHz, DMSO-d6) δ 170.6, 154.9, 147.5, 146.7, 93.2, 75.9, 45.8, 39.5 (under DMSO solvent peak), 34.3, 32.5, 30.8, 27.0 (Rot. B), 26.7 (Rot. A), 19.70 (Rot. B), 19.4 (Rot. A), 13.6, 6.9; HRMS: m/z: [M + H]⁺ calc’d for [C₁₇H₂₇N₄O₃]⁺ expect 335.2078; found 335.2064; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 4.2 nM.

(1R,3S)-3-(3-(3-Methylbut-2-enamido)-1H-pyrazol-5-yl)-cyclopentyl isopropylcarbamate (15).

Following the general procedure using 6 as the amine (12.6 mg, 0.050 mmol), 3-methylbut-2-enoic acid (11.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (3–7% MeOH in DCM) yielded the target compound (4.3 mg, 0.013 mmol, 26%) as a white solid. ¹H NMR (500 MHz, DMSO-d6) δ 12.01 (s, 1H), 10.15 (s, 1H), 6.96 (d,J = 6.7 Hz, 1H), 6.37 (s, 1H), 5.88 (s, 1H), 4.99 (s, 1H), 3.58 (q,J = 6.7 Hz, 1H), 3.13–2.94 (m, 1H), 2.52–2.40 (m, 1H), 2.14 (s, 3H), 2.06–1.95 (m, 1H), 1.94–1.84 (m, 1H), 1.82 (s, 3H), 1.77–1.64 (m, 2H), 1.64–1.53 (m, 1H), 1.03 (d,J = 6.3 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 164.0, 155.6, 151.2, 147.9, 146.7, 119.2, 93.8, 75.1, 42.6, 39.5 (under DMSO solvent peak), 35.0, 32.9, 21.3, 27.6, 23.0, 19.9; HRMS: m/z: [M + H]⁺ calc’d for [C₁₇H₂₇N₄O₃]⁺ expect 335.2078; found 335.2081; ASMS protein concentration first observed: 2.00 μM; measured IC₅₀: 61.5 nM.

(1R,3S)-3-(3-(Pyrazine-2-carboxamido)-1H-pyrazol-5-yl)-cyclopentyl ((R)-1-cyclopropylethyl)carbamate (16).

Following the general procedure using 8 as the amine (13.9 mg, 0.050 mmol), pyrazinecarboxylic acid (13.7 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (10.3 mg, 0.027 mmol, 54%) as a pale-yellow solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.33 (s, 1H), 10.46 (s, 1H), 9.26 (s, 1H), 8.92 (d,J = 2.1 Hz, 1H), 8.79 (s, 1H), 7.02 (d,J = 8.2 Hz, 1H), 5.10–4.92 (m, 1H), 3.17–3.05 (m, 1H), 3.04–2.92 (m, 1H), 2.54–2.42 (m, 1H), 2.11–1.99 (m, 1H), 1.99–1.84 (m, 1H), 1.84–1.68 (m, 2H), 1.68–1.55 (m, 1H), 1.08 (d, J = 6.7 Hz, 3H), 0.88–0.72 (m, 1H), 0.42–0.34 (m, 1H), 0.34–0.28 (m, 1H), 0.28–0.16 (m, 1H), 0.14–0.05 (m, 1H); ¹³C NMR (126 MHz, DMSO-d₆) δ 160.5, 155.5, 147.8, 144.5, 143.7, 143.5, 93.7, 74.7, 50.0, 39.1 (under DMSO solvent peak), 34.7, 32.3, 30.8, 20.6, 17.3, 2.7, 2.6; HRMS: m/z: [M + H]⁺ calc’d for [C₁₉H₂₅N₆O₃]⁺ expect 385.1983; found 385.1988; ASMS protein concentration first observed: 5.00 μM; measured IC₅₀: 45.3 nM. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

(1R,3S)-3-(3-(2-Benzamidoacetamido)-1H-pyrazol-5-yl)-cyclopentyl isopropyl(methyl)carbamate (17).

Following a modified general procedure using 7 as the amine (26.6 mg, 0.100 mmol), hippuric acid (39.4 mg, 0.220 mmol), CDI (35.7 mg, 0.220 mmol) as the coupling reagent, and NMI (35.1 μL, 36.1 mg, 0.440 mmol) as the base, the compound was directly injected into RP-HPLC (water/MeCN) with no addition of LiOH·H₂O and no subsequent quench with AcOH and workup to yield the target compound (19.4 mg, 0.045 mmol, 45%) as a white solid. The compound appears as a 1.4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.06 (s, 1H), 10.3 (s, 1H), 8.76 (t,J = 5.8 Hz, 1H), 7.88 (d, J = 7.4 Hz, 2H), 7.55 (tt, J = 7.4, 1.0 Hz, 1H), 7.48 (t, J = 7.7 Hz, 2H), 6.30 (s, 1H), 5.02 (s, 1H), 4.22 (s, 1H, Rot. A), 4.09 (s, 1H, Rot. B), 4.01 (d,J = 5.8 Hz, 2H), 3.09 (p, J = 8.4 Hz, 1H), 2.62 (s, 3H), 2.47–2.37 (m, 1H), 2.08–1.95 (m, 1H), 1.94–1.82 (m, 1H), 1.79–1.71 (m, 2H), 1.71–1.62 (m, 1H), 1.00 (d,J = 6.8 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 166.8, 166.5, 154.9, 146.9, 134.0, 131.3, 128.3, 127.3, 93.1, 75.9, 46.0 (Rot. B), 45.8 (Rot. A), 42.8, 39.1, 34.4, 32.5, 30.9, 27.0 (Rot. B), 26.8 (Rot. A), 19.7 (Rot. B), 19.4 (Rot. B); HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₃₀N₅O₄]⁺ expect 428.2292; found 428.2265; ASMS protein concentration first observed: 10.00 μM; measured IC₅₀: 58.4 nM. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(3-(2-Methoxybenzamido)-1H-pyrazol-5-yl)cyclopentyl isopropyl(methyl)carbamate (18).

Following a modified general procedure using 7 as the amine (26.6 mg, 0.100 mmol), 2-methoxybenzoic acid (33.5 mg, 0.220 mmol), HATU (83.7 mg, 0.220 mmol) as the coupling reagent, and DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, twelve equivalents 1.5 M LiOH·H₂O and 60 °C were used for the hydrolysis step. Purification via RP-HPLC (water/MeCN) yielded the target compound (35.0 mg, 0.087 mmol, 87%) as a white solid. The compound appears as a 1.2:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d6) δ 12.17 (s, 1H), 10.19 (s, 1H), 7.80 (dd, J = 7.6, 1.5 Hz, 1H), 7.55–7.48 (m, 1H), 7.19 (d,J = 8.4 Hz, 1H), 7.08 (t,J = 7.5 Hz, 1H), 6.46 (s, 1H), 5.04 (s, 1H), 4.24 (s, 1H, Rot. A), 4.13 (s, 1H, Rot. B), 3.94 (s, 3H), 3.13 (p,J = 7.7 Hz, 1H), 2.65 (s, 3H), 2.48–2.38 (m, 1H), 2.10–2.00 (m, 1H), 1.97–1.83 (m, 1H), 1.83–1.66 (m, 3H), 1.03 (d,J = 6.8 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 162.4, 156.9, 155.0, 147.5, 146.8, 132.8, 130.5, 122.7, 120.8, 112.3, 93.6, 76.0, 56.2, 46.1 (Rot. A), 45.9 (Rot. B), 39.1 (under DMSO solvent peak), 34.4, 32.6, 30.9, 27.0 (Rot. B), 26.8 (Rot. A), 19.7 (Rot. B), 19.4 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₂₁H₂₉N₄O₄]⁺ expect 401.2183; found 401.2151; ASMS protein concentration first observed: 20.00 μM; measured IC₅₀: 772.2 nM.

(1R,3S)-3-(3-(Cyclohexanecarboxamido)-1H-pyrazol-5-yl)-cyclopentyl ((R)-1-cyclopropylethyl)carbamate (19).

Following a modified general procedure using 8 as the amine (27.8 mg, 0.100 mmol), cyclohexanecarboxylic acid (52.0 mg, 0.405 mmol), HATU (83.7 mg, 0.440 mmol) as the coupling reagent, and DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, the compound was directly purified by RP-HPLC with no workup (water/MeCN). The target compound was isolated as a white solid (16.2 mg, 0.042 mmol, 42%). The compound appears as an 8.6:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 11.98, 10.11, 7.00 (d,J = 8.1 Hz, 1H, Rot. A), 6.66 (s, 1H, Rot. B), 6.28 (s, 1H), 4.97 (s, 1H), 3.10–2.91 (m, 2H), 2.47–2.37 (m, 1H), 2.37–2.26 (m, 1H), 2.03–1.94 (m, 1H), 1.93–1.81 (m, 1H), 1.78–1.51 (m, 8H), 1.36 (q, 2H,J = 10.5 Hz, 2H), 1.28–1.11 (m, 3H), 1.08 (d,J = 6.6 Hz, 3H), 0.88–0.72 (m, 1H), 0.42–0.34 (m, 1H), 0.34–0.27 (m, 1H), 0.27–0.15 (m, 1H), 0.13–0.04 (m, 1H); ¹³C NMR (126 MHz, DMSO-d6) δ 173.3, 155.4, 147.0, 146.6, 93.2, 74.7, 49.9, 43.8, 39.1, 34.6, 32.3, 30.8, 29.1, 25.4, 25.2, 20.5, 17.2, 2.7, 2.6; HRMS: m/z: [M + H]⁺ calc’d for [C₂₁H₃₃N₄O₃]⁺ expect 389.2547; found 389.2533; ASMS protein concentration first observed: 0.10 μM; measured IC₅₀: 41.9 nM.

(1R,3S)-3-(3-(3-(Pyridin-3-yl)propanamido)-1H-pyrazol-5-yl)-cyclopentyl (1-methylcyclopropyl)carbamate (20).

Following the general procedure using 5 as the amine (13.2 mg, 0.050 mmol), 3-(pyridin-3-yl)propanoic acid (16.6 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (35% formate salt, 13.5 mg, 0.033 mmol, 66%) as a white solid. The compound appears as a 4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.01 (s, br, 1H), 10.29 (s, 1H), 8.45 (s, 1H), 8.39 (dd, J = 4.9, 1.6 Hz, 1H), 7.64 (dt, J = 7.9, 2.0 Hz, 1H), 7.36 (s, 1H, Rot. A), 7.29 (dd, J = 7.7, 4.8 Hz, 1H), 7.09 (s, 1H, Rot. B), 6.33 (s, 1H, Rot. B), 6.26 (s, 1H, Rot. A), 5.02 (s, 1H, Rot. B), 4.97 (s, 1H, Rot. A), 3.06–2.95 (m, 1H), 2.88 (t,J = 7.4 Hz, 2H), 2.60 (t,J = 7.4 Hz, 2H), 2.49–2.38 (m, 1H), 2.08–1.92 (m, 1H), 1.92–1.80 (m, 1H), 1.80–1.59 (m, 2H), 1.59–1.46 (m, 1H), 1.23 (s, 3H), 0.59 (s, 2H), 0.46 (s, 2H); ¹³C NMR (126 MHz, DMSO-d₆) δ 169.0 (Rot. A), 164.9 (Rot. B), 155.7, 149.6, 147.3, 146.8, 146.7, 136.6, 135.8, 123.4, 93.2, 75.2 (Rot. B), 74.5 (Rot. A), 39.6 (under DMSO solvent peak), 36.5, 34.6, 32.4, 30.7, 28.7, 27.9, 22.9, 14.5 (Rot. B), 13.8 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₂₁H₂₈N₅O₃]⁺ expect 398.2187; found 398.2182; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 21.2 nM.

(1R,3S)-3-(3-(2-(4-Methoxyphenyl)acetamido)-1H-pyrazol-5-yl)-cyclopentyl isopropylcarbamate (21).

Following the General Procedure using 6 as the amine (12.6 mg, 0.050 mmol), 4-methoxyphenyl acetic acid (18.3 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (8.3 mg, 0.021 mmol, 41%) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.05 (s, 1H), 10.44 (s, 1H), 7.22 (d, J = 8.4 Hz, 2H), 6.93 (d, J = 7.1 Hz, 1H), 6.86 (d, J = 8.4 Hz, 2H), 6.26 (s, 1H), 4.97 (s, 1H), 3.72 (s, 3H), 3.61–3.52 (m, 1H), 3.49 (s, 2H), 3.08–2.94 (m, 1H), 2.47–2.38 (m, 1H), 2.05–1.92 (m, 1H), 1.92–1.80 (m, 1H), 1.76–1.60 (m, 2H), 1.59–1.48 (m, 1H), 1.01 (d, J = 4.8 Hz); ¹³C NMR (126 MHz, DMSO-d₆) δ 168.9, 158.4, 155.6, 147.7, 147.0, 130.5, 128.6, 114.1, 93.6, 75.0, 55.5, 42.6, 42.1, 39.5 (under DMSO solvent peak), 35.0, 32.8, 31.2, 23.0; HRMS: m/z: [M + H]⁺ calc’d for [C₂₁H₂₉N₄O₄]⁺ expect 401.2183; found 401.2151; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 2.8 nM.

(1R,3S)-3-(3-(2-(4-Methoxyphenyl)acetamido)-1H-pyrazol-5-yl)-cyclopentyl isopropyl(methyl)carbamate (22).

Following the general procedure using 7 as the amine (26.6 mg, 0.100 mmol), 4-methoxyphenyl acetic acid (36.6 mg, 0.220 mmol), HATU (83.7 mg, 0.220 mmol) as the coupling reagent, DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, and purifying via silica column chromatography (2–5% MeOH in DCM) and RP-HPLC (water/MeCN) yielded the target compound (24.8 mg, 0.060 mmol, 60%) as a pale orange solid. The compound appears as a 1.4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.04 (s, 1H), 10.41 (s, 1H), 7.22 (d,J = 8.4 Hz, 2H), 6.86 (d,J = 8.4 Hz, 2H), 6.27 (s, 1H), 5.01 (s, 1H), 4.22 (s, 1H, Rot. A), 4.06 (s, 1H, Rot. B), 3.71 (s, 3H), 3.49 (s, 2H), 3.07 (p,J = 7.9 Hz, 1H), 2.61 (s, 3H), 2.45–2.31 (m, 1H), 2.04–1.94 (m, 1H), 1.92–1.80 (m, 1H), 1.79–1.70 (m, 2H), 1.70–1.60 (m, 1H), 0.99 (d, J = 6.4 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 168.4, 158.0, 154.9, 147.3, 146.8, 130.0, 128.2, 113.7, 93.1, 75.9, 55.0, 45.8, 41.6, 39.1, 34.4, 32.5, 30.8, 27.0 (Rot. B), 26.7 (Rot. A), 19.7 (Rot. B), 19.3 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₃₁N₄O₄]⁺ expect 415.2340; found 415.2344; ASMS protein concentration first observed: 0.10 μM; measured IC₅₀: 1.2 nM.

(1R,3S)-3-(3-(2,4-Difluorobenzamido)-1H-pyrazol-5-yl)-cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (S4).

Following the general procedure using 10 as the amine (29.2 mg, 0.100 mmol), 2,4-difluorobenzoic acid (34.8 mg, 0.220 mmol), HATU (83.7 mg, 0.220 mmol) as the coupling reagent, DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, and purifying via silica column chromatography (2–5% MeOH in DCM) and RP-HPLC (water/MeCN) yielded the target compound (19.2 mg, 0.044 mmol, 44%) as a white solid. ¹H NMR (500 MHz, DMSO-d6) δ 12.21 (s, 1H), 10.67 (s, 1H), 7.69 (q, J = 7.3 Hz, 1H), 7.35 (td, J = 10.0, 2.1 Hz, 1H), 7.17 (td, J = 8.5, 2.1 Hz, 1H), 6.44 (s, 1H), 5.11–5.00 (m, 1H), 3.84 (p,J = 6.1 Hz, 1H), 3.11 (p,J = 8.1 Hz, 1H), 2.51–2.42 (m, 1H), 2.13–1.99 (m, 3H), 1.96–1.84 (m, 1H), 1.83–1.72 (m, 2H), 1.72–1.63 (m, 1H), 1.52–1.37 (m, 2H), 1.07 (d,J = 6.3 Hz, 3H), 1.04 (d,J = 6.3 Hz, 3H); ¹³C NMR (126 MHz, DMSO-d6) δ 163.3 (dd, ¹J_C-F = 250.0 Hz, ³J_C-F = 12.2 Hz), 160.9, 159.6 (dd, ¹J_C-F = 251.5 Hz, ³J_C-F = 12.9 Hz), 153.2, 146.9, 131.7 (dd, ³J_C-F = 10.4, 4.3 Hz), 121.2 (d, ²J_C-F = 14.1 Hz), 111.7 (dd, ²J_C-F = 21.6 Hz, ⁴J_C-F = 3.3 Hz), 104.5 (t, ²J_C-F = 26.2 Hz), 93.6, 75.2, 52.8, 52.5, 39.4 (under DMSO solvent peak), 34.5, 32.6, 30.9, 29.8, 28.8, 20.4, 19.2; ¹⁹F (471 MHz, DMSO-d6) δ −106.7, −109.5; HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₂₇N₄O₃F₂]⁺ expect 433.2046; found 433.2069; ASMS protein concentration first observed: 0.05 μM; measured IC₅₀: 3.8 nM.

(1R,3S)-3-(3-(2-(3-Fluorophenyl)acetamido)-1H-pyrazol-5-yl)-cyclopentyl 2,2-dimethylazetidine-1-carboxylate (S5).

Following the general procedure using 9 as the amine (13.9 mg, 0.050 mmol), 2-(3-fluorophenyl)acetic acid (17.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (2–10% MeOH in DCM) yielded the target compound (13.5 mg, 0.033 mmol, 65%) as a white solid. The compound appears as a 1.4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.08 (s, 1H), 10.53, 7.34 (q, J = 7.3 Hz, 1H), 7.17–7.11 (m, 2H), 7.06 (td, J = 8.7, 1.8 Hz, 1H), 6.28 (s, 1H), 5.05–4.97 (m, 1H, Rot. A), 4.97–4.90 (m, 1H, Rot. B), 3.75–3.66 (m, 1H), 3.63 (t,J = 7.5 Hz, 1H), 3.61 (s, 2H), 3.12–2.98 (m, 1H), 2.46–2.31 (m, 1H), 2.05–1.95 (m, 1H), 1.95–1.87 (m, 2H), 1.87–1.78 (m, 1H), 1.76–1.65 (m, 2H), 1.65–1.56 (m, 1H), 1.35 (s, 3H, Rot. B), 1.34 (s, 3H, Rot. B), 1.30 (s, 1.30, Rot. A), 1.27 (s, 1.30, Rot. A); ¹³C NMR (126 MHz, DMSO-d6) δ 167.5, 162.0 (d, ¹J_C-F = 243.1 Hz), 154.9 (Rot. A), 153.7 (Rot. B), 147.2 (Rot. A), 146.8 (Rot. B), 138.9 (d, ³J_C-F = 8.1 Hz), 130.1 (d, ³J_C-F = 8.4 Hz), 125.2, 115.8 (d, ²J_C-F = 21.4 Hz), 113.3 (d, ²J_C-F = 20.7 Hz), 93.0, 75.1 (Rot. A), 74.9 (Rot. B), 65.3 (Rot. A), 64.6 (Rot. B), 43.7 (Rot. B), 42.3 (Rot. A), 42.1, 39.5 (under DMSO solvent peak), 34.5 (Rot. A), 34.2 (Rot. B), 32.7 (Rot. A), 32. Four (Rot. B), 31.0 (Rot. A), 30.9 (Rot. B), 30.3 (Rot. B), 30.0 (Rot. A), 26.9 (Rot. A), 26.7 (Rot. A), 26.2 (Rot. B), 26.0 (Rot. B); ¹⁹F (471 MHz, DMSO-d6) δ −133.8; HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₂₈FN₄O₃]⁺ expect 415.2140; found 415.2145; ASMS protein concentration first observed: 0.05 μM; measured IC₅₀: 0.5 nM.

(1R,3S)-3-(3-(4-Methylbenzamido)-1H-pyrazol-5-yl)cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (S6).

Following the general procedure using 10 as the amine (14.6 mg, 0.050 mmol), 4-methylbenzoic acid (15.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (100% EtOAc) and RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (9.4 mg, 0.023 mmol, 46%) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.17, 10.60, 7.89 (d,J = 8.1 Hz, 2H), 7.28 (d, J = 8.1 Hz, 2H), 6.43 (s, 1H), 5.09–5.01 (m, 1H), 3.84 (p,J = 6.3 Hz, 2H), 3.10 (p,J = 8.1 Hz, 1H), 2.53–2.45 (m, 1H), 2.36 (s, 3H), 2.13–1.98 (m, 3H), 1.96–1.84 (m, 1H), 1.83–1.72 (m, 2H), 1.72–1.63 (m, 1H), 1.50–1.37 (m, 2H), 1.07 (d,J = 6.3 Hz, 3H), 1.04 (d,J = 6.3 Hz, 3H); ¹³C NMR (126 MHz, DMSO-d6) δ 164.2, 153.2, 147.0, 141.4, 131.4, 128.8, 127.7, 93.9, 75.2, 52.8, 52.5, 39.2 (under DMSO solvent peak), 34.8, 32.6, 30.9, 39.8, 28.8, 21.0, 20.4, 19.2; HRMS: m/z: [M + H]⁺ calc’d for [C₂₃H₃₁N₄O₃]⁺ expect 411.2391; found 411.2359; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 13.4 nM. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(3-(4-Sulfamoylbenzamido)-1H-pyrazol-5-yl)-cyclopentyl 2,2-dimethylazetidine-1-carboxylate (S7).

Following the general procedure using 9 as the amine (13.9 mg, 0.050 mmol), 4-sulfamoylbenzoic acid (22.1 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (5–10% MeOH in DCM) yielded the target compound (13.3 mg, 0.029 mmol, 58%) as a white solid. The compound appears as a 1.4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.26 (s, 1H), 10.93 (s, 1H), 8.11 (d,J = 8.3 Hz, 2H), 7.90 (d,J = 8.3 Hz, 2H), 7.50 (s, 2H), 6.49 (s, 1H), 5.07–5.01 (m, 1H, Rot. A), 5.01–4.95 (m, 1H, Rot. B), 3.76 (h,J = 7.7 Hz, 1H), 3.66 (t,J = 7.6 Hz, 1H), 3.11 (p,J = 8.2 Hz, 1H), 2.50–2.38 (m, 1H), 2.14–2.00 (m, 1H), 1.95 (t, J = 7.2 Hz, 2H), 1.92–1.83 (m, 1H), 1.81–1.72 (m, 2H), 1.72–1.63 (m, 1H), 1.38 (s, 6H, Rot. B), 1.35 (s, 3H, Rot. A), 1.34 (s, 3H, Rot. A); ¹³C NMR (126 MHz, DMSO-d₆) δ 163.3, 154.9 (Rot. A.), 153.7 (Rot. B), 147.2, 146.7, 146.4, 137.2, 128.4, 125.6, 94.0, 75.1 (Rot. A), 74.9 (Rot. B), 65.3 (Rot. A), 64.7 (Rot. B), 43.8 (Rot. B), 42.4 (Rot. A), 39.0 (under DMSO solvent peak), 34.5 (Rot. A), 34.3 (Rot. B), 32.7 (Rot. A), 32.4 (Rot. B), 31.0 (Rot. A), 31.0 (Rot. B), 30.4 (Rot. B), 30.1 (Rot. A), 27.0 (Rot. A), 26.8 (Rot. A), 26.2 (Rot. B), 26.1 (Rot. B); HRMS: m/z: [M + H]⁺ calc’d for [C₂₁H₂₈N₅O₅S]⁺ expect 462.1806; found 462.1798; ASMS protein concentration first observed: 0.50 μM; measured IC₅₀: 2.7 nM.

(1R,3S)-3-(3-(2-Cyclopropylacetamido)-1H-pyrazol-5-yl)-cyclopentyl isopropylcarbamate (S8).

Following a modified general procedure using 6 as the amine (12.6 mg, 0.050 mmol), 2-cyclopropylacetic acid (11.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, and DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, the compound was directly purified by RP-HPLC with no workup (water/MeCN + 0.1% formic acid). The target compound was isolated as a white solid (12.8 mg, 0.038 mmol, 77%). ¹H NMR (500 MHz, DMSO-d₆) δ 12.02 (s, 1H), 10.12 (s, 1H), 6.95 (d, J = 7.6 Hz, 1H), 6.32 (s, 1H), 4.98 (s, 1H), 3.62–3.50 (m, 1H), 3.09–2.96 (m, 1H), 2.53–2.38 (m, 1H), 2.13 (d,J = 7.0 Hz, 2H), 2.04–1.94 (m, 1H), 1.93–1.81 (m, 1H), 1.78–1.61 (m, 2H), 1.61–1.50 (m, 1H), 1.07–0.96 (m, 7H), 0.43 (q,J = 7.6 Hz, 2H), 0.14 (d,J = 4.5 Hz, 2H); ¹³C NMR (126 MHz, DMSO-d6) δ 169.7, 155.2, 147.3, 146.4, 93.3, 74.6, 42.1, 40.5, 39.1 (under DMSO solvent peak), 34.5, 32.4, 30.8, 22.6, 7.9, 4.1; HRMS: m/z: [M + H]⁺ calc’d for [C₁₇H₂₇N₄O₃]⁺ expect 335.2078; found 335.2065; ASMS protein concentration first observed: 1.00 μM; measured IC₅₀: 19.6 nM.

(1R,3S)-3-(3-(2-(Furan-2-yl)acetamido)-1H-pyrazol-5-yl)-cyclopentyl isopropyl(methyl)carbamate (S9).

Following a modified General Procedure using 7 as the amine (13.3 mg, 0.050 mmol), 2-furanacetic acid (13.9 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, and DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, the compound was directly worked up with no addition of LiOH·H₂O and no subsequent quench with AcOH. The compound was purified via silica column chromatography (8% MeOH in DCM), yielding the target compound (18.0 mg, 0.048 mmol, 96%) as a pale orange solid. The compound appears as a 4:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.12 (s, 1H), 10.48 (s, 1H), 7.55 (s, 1H, Rot. A), 7.53 (s, 1H, Rot. B), 6.41–6.38 (m, 1H, Rot. A), 6.38 (m, 1H, Rot. B), 6.29 (s, 1H), 6.23 (d,J = 3.0 Hz, 1H, Rot. A), 6.20 (d, J = 2.6 Hz, 1H, Rot. B), 5.02 (s, 1H), 4.23 (s, 1H, Rot. A), 4.09 (s, 1H, Rot. B), 3.67 (s, 2H, Rot. A), 3.57 (s, 2H, Rot. B), 3.09 (q, J = 7.9 Hz, 1H), 2.63 (s, 3H), 2.45–2.35 (m, 1H), 2.06–1.96 (m, 1H), 1.95–1.82 (m, 1H), 1.80–1.71 (m, 2H), 1.71–1.61 (m, 1H), 1.01 (d,J = 6.7 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d6) δ 170.6 (Rot. B), 165.6 (Rot. A), 154.9, 149.7, 147.1, 142.1 (Rot. B), 141.9 (Rot. A), 110.5, 107.8 (Rot. B), 107.7 (Rot. A), 93.1, 76.2 (Rot. B), 76.1 (Rot. A), 45.9, 39.1 (under DMSO solvent peak), 35.3 (Rot. A), 34.4, 33.6 (Rot. B), 32.5, 30.8, 27.0 (Rot. B), 26.7 (Rot. A), 19.7 (Rot. B), 19.4 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₁₉H₂₇N₄O₄]⁺ expect 375.2027; found 375.1994; ASMS protein concentration first observed: 1.00 μM; measured IC₅₀: 4.9 nM. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(3-(2-Hydroxy-2-(p-tolyl)acetamido)-1H-pyrazol-5-yl)-cyclopentyl ((R)-1-cyclopropylethyl)carbamate (S10).

Following the General Procedure using 8 as the amine (13.9 mg, 0.050 mmol), 2-hydroxy-2-(p-tolyl)acetic acid (18.3 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (1–10% MeOH in DCM) yielded the target compound (11.7 mg, 0.027 mmol, 55%) as a white solid. ¹H NMR (500 MHz, DMSO-d6) δ 12.09, 9.93, 7.34 (d, J = 8.0 Hz, 2H), 7.13 (d,J = 8.0 Hz, 2H), 6.98 (d, J = 8.0 Hz, 1H), 6.24 (s, 1H), 6.17 (s, br, overlapping, 1H, OH), 5.05 (s, 1H), 5.00–4.92 (m, 1H), 3.09–2.90 (2H), 2.47–2.38 (m, 1H), 2.27 (s, 3H), 2.03–1.93 (m, 1H), 1.92–1.79 (m, 1H), 1.76–1.60 (m, 2H), 1.59–1.48 (m, 1H), 1.06 (d,J = 6.7 Hz, 3H), 0.46–0.72 (m, 1H), 0.41–0.33 (m, 1H), 0.32–0.26 (m, 1H), 0.26–0.17 (m, 1H), 0.12–0.02 (m, 1H); ¹³C NMR (126 MHz, DMSO-d₆) δ 170.1, 155.4, 137.9, 136.7, 93.0, 74.6, 73.0, 49.9, 38.9 (under DMSO solvent peak), 34.7, 32.3, 30.7, 20.7, 20.5, 17.2, 2.7, 2.6; HRMS: m/z: [M + H]⁺ calc’d for [C₂₃H₃₁N₄O₄]⁺ expect 427.2340; found 427.2335; ASMS protein concentration first observed: 1.00 μM; measured IC₅₀: 11.2 nM. Two pyrazole ¹³C peaks could not be observed due to peak broadening.

(1R,3S)-3-(3-(2-Methylbenzamido)-1H-pyrazol-5-yl)cyclopentyl (2S,5S)-2,5-dimethylpyrrolidine-1-carboxylate (S11).

Following the General Procedure using 10 as the amine (29.2 mg, 0.100 mmol), 2-methylbenzoic acid (30.0 mg, 0.220 mmol), HATU (83.7 mg, 0.220 mmol) as the coupling reagent, DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, and purifying via silica column chromatography (2–4% MeOH in DCM) and RP-HPLC (water/MeCN) yielded the target compound (14.2 mg, 0.035 mmol, 35%) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.11 (s, 1H), 10.58 (s, 1H), 7.39 (d, J = 7.4 Hz, 1H), 7.34 (t,J = 7.4 Hz, 1H), 7.29–7.20 (m, 2H), 6.43 (s, 1H), 5.10–5.01 (m, 1H), 3.93–3.80 (m, 2H), 3.11 (p,J = 8.1 Hz, 1H), 2.50–2.43 (m, 1H), 2.36 (s, 3H), 2.13–1.99 (m, 3H), 1.96–1.85 (m, 1H), 1.84–1.73 (m, 2H), 1.73–1.64 (m, 1H), 1.51–1.38 (m, 2H), 1.07 (d,J = 6.3 Hz, 3H), 1.05 (d,J = 6.3 Hz, 3H); ¹³C NMR (126 MHz, DMSO-d₆) δ 166.8, 153.3, 147.0, 146.7, 136.7, 135.3, 130.4, 129.4, 127.3, 125.4, 93.6, 75.3, 52.8, 52.5, 39.6 (under DMSO solvent peak), 34.7, 32.6, 30.9, 29.8, 28.8, 20.4, 19.4, 19.2; HRMS: m/z: [M + H]⁺ calc’d for [C₂₃H₃₁N₄O₃]⁺ expect 411.2391; found 411.2359; ASMS protein concentration first observed: 1.00 μM; measured IC₅₀: 43.8 nM.

(1R,3S)-3-(3-(Tetrahydro-2H-pyran-4-carboxamido)-1H-pyrazol-5-yl)cyclopentyl 2,2-dimethylazetidine-1-carboxylate (S12).

Following the general procedure using 9 as the amine (13.9 mg, 0.050 mmol), tetrahydro-2H-pyran-4-carboxylic acid (14.3 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (2–7% MeOH in DCM) yielded the target compound (9.4 mg, 0.024 mmol, 48%) as a white solid. The compound appears as a 1.2:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.04 (s, 1H), 10.22 (s, 1H), 6.34 (s, 1H), 5.02 (s, 1H, Rot. A), 4.96 (s, 1H, Rot. B), 3.88 (d,J = 10.2 Hz, 2H), 3.74 (q,J = 7.0 Hz, 1H), 3.65 (t,J = 7.3 Hz, 1H), 3.32 (d,J = 10.2 Hz, 2H), 3.07 (s, 1H), 2.68–2.52 (m, 1H), 2.50–2.32 (m, 1H), 2.06–1.98 (m, 1H), 1.95 (t,J = 7.4 Hz, 2H), 1.92–1.80 (s, 1H), 1.80–1.67 (m, 2H), 1.63 (s, 5H), 1.38 (s, 6H, Rot. B), 1.34 (s, 3H, Rot. A), 1.32 (s, 3H, Rot. A); ¹³C NMR (126 MHz, DMSO-d₆) δ 171.9, 154.9 (Rot. A), 153.7 (Rot. B), 146.6, 93.3, 75.0 (Rot. A), 74.9 (Rot. B), 66.4, 65.3 (Rot. A), 64.6 (Rot. B), 43.7 (Rot. B), 42.3 (Rot. A), 39.1 (under DMSO solvent peak), 34.6 (Rot. A), 34.3 (Rot. B), 32.6 (Rot. A), 32.4 (Rot. B), 31.0, 30.4 (Rot. B), 30.0 (Rot. A), 28.9, 26.9 (Rot. A), 26.7 (Rot. A), 26.2 (Rot. B), 26.0 (Rot. B); HRMS: m/z: [M + H]⁺ calc’d for [C₂₀H₃₁N₄O₄]⁺ expect 391.2340; found 391.2344; ASMS protein concentration first observed: 1.00 μM; measured IC₅₀: 16.6 nM. One pyrazole ¹³C peak could not be observed due to peak broadening.

(1R,3S)-3-(3-(2-(4-Isobutylphenyl)propanamido)-1H-pyrazol-5-yl)cyclopentyl isopropylcarbamate (S13).

Following the general procedure using 6 as the amine (12.6 mg, 0.050 mmol), 2-(4-isobutylphenyl)propanoic acid (22.7 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (100% EtOAc) and RP-HPLC (water/MeCN) yielded the target compound (5.9 mg, 0.013 mmol, 27%) as a yellow solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.04 (s, 1H), 10.40 (s, 1H), 7.26 (d, J = 8.0 Hz, 2H), 7.08 (d, J = 8.0 Hz, 2H), 6.93 (d, J = 6.3 Hz, 1H), 6.29 (s, 1H), 4.97 (s, 1H), 3.80 (q, J = 7.0 Hz, 1H), 3.63–3.45 (m, 1H), 3.10–2.91 (m, 1H), 2.47–2.34 (m, 1H), 2.39 (d, J = 7.1 Hz, 2H), 2.03–1.91 (m, 1H), 1.91–1.82 (m, 1H), 1.82–1.74 (m, 1H), 1.73–1.60 (m, 2H), 1.60–1.49 (m, 1H), 1.34 (d, J = 7.0 Hz, 3H), 1.08–0.93 (m, 6H), 0.84 (d, J = 6.6 Hz, 6H); ¹³C NMR (126 MHz, DMSO-d₆) δ 171.4, 155.1, 147.2, 146.5, 139.4, 139.3, 128.8, 127.0, 93.2, 74.6, 44.5, 44.2, 42.1, 39.1 (under DMSO solvent peak), 34.4, 32.3, 30.7, 29.6, 22.6, 22.2, 18.4^’ HRMS: m/z: [M + H]⁺ calc’d for [C₂₅H₃₇N₄O₃]⁺ expect 441.2860; found 441.2857; ASMS protein concentration first observed: 5.00 μM; measured IC₅₀: 44.5 nM.

(1R,3S)-3-(3-(Picolinamido)-1H-pyrazol-5-yl)cyclopentyl (1-methylcyclopropyl)carbamate (S14).

Following the general procedure using 5 as the amine (26.4 mg, 0.100 mmol), picolinic acid (27.1 mg, 0.220 mmol), HATU (83.7 mg, 0.220 mmol) as the coupling reagent, DIPEA (76.6 μL, 56.9 mg, 0.440 mmol) as the base, and purifying via RP-HPLC (loaded in MeOH, water/MeCN eluent) yielded the target compound (24.8 mg, 0.067 mmol, 67%) as a white solid. The compound appears as a 4.0:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.18 (s, 1H), 10.45 (s, 1H), 8.72 (d, J = 4.4 Hz, 1H), 8.14 (d, J = 7.8 Hz, 1H), 8.07 (t, J = 7.7Hz, 1H), 7.68 (dd, J = 7.0, 5.4 Hz, 1H), 7.38 (s, 1H, Rot. A), 7.11 (s, 1H, Rot. B), 6.53 (s, 1H, Rot. B), 6.47 (s, 1H, Rot. A), 5.07 (s, 1H, Rot. B), 5.01 (s, 1H, Rot. A), 3.23–3.00 (m, 1H), 2.55–2.45 (m, 1H), 2.16–1.98 (m, 1H), 1.97–1.85 (m, 1H), 1.85–1.67 (m, 2H), 1.67–1.55 (m, 1H), 1.24 (s, 3H), 0.61 (s, 2H), 0.47 (s, 2H); ¹³C NMR (126 MHz, DMSO-d₆) δ 166.3 (Rot. B), 161.5 (Rot. A), 156.6 (Rot. B), 156.2 (Rot. A), 149.6, 149.1, 148.2, 145.8, 138.7, 127.5, 122.5, 93.6 (Rot. A), 93.3 (Rot. B), 75.7 (Rot. B), 75.0 (Rot. A), 39.6 (under DMSO solvent peak), 35.2, 33.4 (Rot. B), 32.8 (Rot. A), 31.6 (Rot. B), 31.2 (Rot. A), 29.1, 24.2 (Rot. B), 23.3 (Rot. A), 15.0 (Rot. B), 14.3 (Rot. A); HRMS: m/z: [M + H]⁺ calc’d for [C₁₉H₂₄N₅O₃]⁺ expect 370.1874; found 370.1862; ASMS protein concentration first observed: 10.00 μM; measured IC₅₀: 116.9 nM.

(1R,3S)-3-(3-(2-(Trifluoromethyl)benzamido)-1H-pyrazol-5-yl)-cyclopentyl 2,2-dimethylazetidine-1-carboxylate (S15).

Following a modified general procedure using 9 as the amine (27.8 mg, 0.100 mmol), 2-(trifluoromethyl)benzoic acid (83.7 mg, 0.440 mmol), HATU (167 mg, 0.440 mmol) as the coupling reagent, and DIPEA (115 μL, 85.3 mg, 0.660 mmol) as the base, Eighteen equiv. 1.5 M LiOH·H₂O and 80 °C were used for the hydrolysis step. The compound was directly injected onto RP-HPLC (water/MeCN) with no subsequent quench with AcOH or workup, yielding the target compound (8.6 mg, 0.019 mmol, 19%) as a white solid. The compound appears as a 1.5:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.16 (s, 1H), 10.87 (s, 1H), 7.79 (d, J = 7.8 Hz, 1H), 7.73 (t, J = 7.4 Hz, 1H), 7.66 (t, J = 7.6 Hz, 1H), 7.58 (d, J =7.4 Hz, 1H), 6.43 (s, 1H), 5.09–5.01 (m, 1H, Rot. A), 5.01–4.95 (m, 1H, Rot. B), 3.84–3.70 (m, 2H, Rot. B), 3.65 (t, J = 7.6 Hz, 2H, Rot. A), 3.19–3.04 (m, 1H), 2.50–2.35 (m, 1H), 2.13–2.00 (m, 1H), 1.95 (t,J = 7.4 Hz, 2H), 1.91–1.87 (m, 1H), 1.81–1.72 (m, 2H), 1.72–1.63 (m, 1H), 1.38 (s, 6H, Rot. B), 1.34 (d, J = 5.6, 6H, Rot. A); ¹³C NMR (126 MHz, DMSO-d₆) δ 164.9, 154.9 (Rot. A), 153.7 (Rot. B), 147.3 (Rot. B), 146.9 (Rot. A), 146.8, 136.0, 132.4, 129.8, 128.6, 126.2 (q, ³J_C–F = 4.5 Hz), 125.9 (q, ²J_C–F = 31.1 Hz), 123.8 (q, ¹J_C–F = 273.6), 93.6, 75.1 (Rot. A), 75.0 (Rot. B), 65.3 (Rot. A), 64.7 (Rot. B), 43.8 (Rot. B), 42.5 (Rot. A), 39.6 (under DMSO solvent peak), 34.4 (Rot. A), 34.2 (Rot. B), 32.7 (Rot. A), 32.4 (Rot. B), 31.0 (Rot. B), 31.0 (Rot. A), 30.4 (Rot. B), 30.1 (Rot. A), 26.9 (Rot. A), 26.7 (Rot. A), 26.2 (Rot. B), 26.0 (Rot. B); ¹⁹F (471 MHz, DMSO-d₆) δ −57.9 (Rot. A), −57.9 (Rot. B); ASMS protein concentration first observed: 20.00 μM; measured IC₅₀: 70.2 nM.

(1R,3S)-3-(3-(Quinoline-2-carboxamido)-1H-pyrazol-5-yl)-cyclopentyl isopropylcarbamate (S16).

Following the general procedure using 6 as the amine (12.6 mg, 0.050 mmol), quinoline-2-carboxylic acid (19.0 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, and purifying via silica column chromatography (2–5% MeOH in DCM) yielded the target compound (21.3 mg, 0.052 mmol, quantitative yield) as a white solid. ¹H NMR (500 MHz, DMSO-d₆) δ 12.34 (s, 1H), 10.53 (s, 1H), 8.64 (d, J = 8.5 Hz, 1H), 8.24 (d, J = 8.5 Hz, 2H), 8.12 (d, J = 8.2 Hz, 1H), 7.91 (t, J = 7.7 Hz, 1H), 7.76 (t, J = 7.7 Hz, 1H), 6.98 (d, J = 7.5 Hz, 1H), 6.55 (s, 1H), 5.02 (s, 1H), 3.59 (h,J = 6.9 Hz, 1H), 3.20–3.04 (m, 1H), 2.56–2.47 (m, 1H), 2.12–2.00 (m, 1H), 1.97–1.85 (m, 1H), 1.83–1.70 (m, 2H), 1.69–1.59 (m, 1H), 1.04 (s, 3H), 1.03 (s, 3H); ¹³C NMR (126 MHz, DMSO-d₆) δ 161.4, 155.3, 149.3, 147.4, 146.1, 146.0, 138.5, 130.9, 129.4, 129.1, 128.5, 128.2, 118.5, 93.3, 74.7, 42.2, 39.5 (under DMSO solvent peak), 35.0, 32.5, 30.7, 22.6; HRMS: m/z: [M + H]⁺ calc’d for [C₂₂H₂₆N₅O₃]⁺ expect 408.2030; found 408.201; ASMS protein concentration first observed: non binder; measured IC₅₀: 672.4 nM.

(1R,3S)-3-(3-((3S,5S,7S)-Adamantane-1-carboxamido)-1H-pyra-zol-5-yl)cyclopentyl (1-methylcyclopropyl)carbamate (S17).

Following the general procedure using 5 as the amine (13.2 mg, 0.050 mmol), 1-adamantanecarboxylic acid (19.8 mg, 0.110 mmol), HATU (41.8 mg, 0.110 mmol) as the coupling reagent, and DIPEA (38.3 μL, 28.4 mg, 0.220 mmol) as the base, purifying via silica column chromatography (75–100% EtOAc in hexanes) and RP-HPLC (water/MeCN + 0.1% formic acid) yielded the target compound (9.6 mg, 0.023 mmol, 45%) as a white solid. The compound appears as a 4.3:1 mixture of rotamers on the NMR time scale. ¹H NMR (500 MHz, DMSO-d₆) δ 12.02 (s, 1H), 9.59 (s, 1H), 7.35 (s, 1H, Rot. A), 7.09 (s, 1H, Rot. B), 6.25 (s, 1H), 5.02 (s, 1H, Rot. B), 4.97 (s, 1H, Rot. A), 3.10–2.92 (m, 1H), 2.46–2.39 (m, 1H), 2.07–1.92 (m, 4H), 1.92–1.82 (m, 7H), 1.75–1.61 (m, 8H), 1.60–1.49 (m, 1H), 1.23 (s, 3H), 0.59 (s, 2H), 0.46 (s, 2H); ¹³C NMR (126 MHz, DMSO-d₆) δ 175.1, 155.7, 147.0, 146.7, 93.8, 74.6, 40.4, 39.1 (under DMSO solvent peak), 38.2, 36.0, 34.7, 32.3 , 30.8, 28.6, 27.7, 22.7, 13.8; HRMS: m/z: [M + H]⁺ calc’d for [C₂₄H₃₅N₄O₃]⁺ expect 427.2704; found 427.2659 ASMS protein concentration first observed: breakthrough; measured IC₅₀: 871.2 nM.

Cell Painting

Cell Culture, Plating, and Treatment.

The HCT-116 human colon cancer cell line (ATCC Cat. No. CCL-247) was cultured in McCoy’s 5A Medium (Gibco: 16600082), and cells were plated into 384-well PhenoPlates (Revvity: 6057300) at 3000 cells/well in 50μL of media. The cells were allowed to attach overnight, and compounds were dispensed from freshly prepared 10 mM DMSO stock solutions using an HP D300e digital dispenser, with DMSO normalized to 1% of the total working volume. Compound treatment plating was randomized across the 384-well plate to minimize plate effects. Compounds were incubated for 24 h before cells were labeled with EdU using an EdU Cell Proliferation Kit for Imaging (Thermo Fisher C10337), where cells were incubated with EdU for 24 h before being fixed with 4% paraformaldehyde in PBS.

Immunostaining.

Cells were permeabilized (0.3% Triton x-100 in PBS) for 15 min before performing the Click-iT reaction as per the manufacturer guidelines. Cells were blocked for 1 h in 5% BSA (MilliporeSigma, A9647) and 0.01% Tween (FisherScientific, BP337–100) before incubation with primary Phospho-Histone H3 (Ser28) antibody (Life Technologies, 50912442) at 4 °C overnight in blocking buffer (antibody 1:500 in 25 μL per well). Cells were incubated with the secondary antibody (Thermo Fisher, A-31571) at 1:1000 in 25 μL per well, and counterstained with HCS CellMask Orange (H32713) at 1:100,00 and Hoechst 33342 (ThermoFisher Scientific, H3570) at 1:2000 in 25 μL per well for 1 h.

Image Acquisition and Analysis.

Sixteen fields per well were imaged using a Yokogawa CellVoyager 8000 high-content screening system with a 20X water immersion, 1.0 NA objective lens. Maximum intensity projection images were collected with a 50 μm pinhole, at 2 μm Z-spacing and 10 μm depth. Nuclei were detected in Hoechst 33342 channel images using Cellpose 2.0.5, and cell boundaries were identified by using these nuclei as a seed object and using the HCS CellMask Orange channel to delineate the cell body. Feature extraction was performed on the multichannel fluorescent images using CellProfiler 4.2.5, extracting >200 features that measured fluorescent intensity, intensity distribution, texture, size, and shape in each channel.

Cell viability was assessed across a 14-point, 2-fold compound dose range by normalizing the number of identified cells per condition to the DMSO vehicle control, which was set as 100% cell viability. Nonlinear regression analysis was performed, and CC50/EC50 values were calculated using GraphPad Prism 9.5.1. Cells positive for EdU and pHH3 were identified using random forest classifiers, one for each biomarker, resulting in >99% classification accuracy using CellProfiler Analyst 3.0.4.

Statistical Testing.

Cell viability was calculated by normalizing each compound (n = 3) to the vehicle control-treated wells (100% cell viability; n = 21). Nonlinear curve fitting of dose–response plots was performed using a three-parameter model for all compounds aside from compound S5, where the preferred model was a four-parameter model. Linear regressions were performed using automatic detection of outliers. All analyses were performed in GraphPad Prism 9.5.1.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.5c03614.

Experimental procedures, supplementary tables and images, and compound data and spectra (PDF) Molecular formula strings and associated biochemical and biological data (CSV)

ABBREVIATIONS

AI

artificial intelligence

ASMS

affinity-selection mass spectrometry

BT

breakthrough compound

C

collapsed

CDI

1,1’-carbonyldiimidazole

Cdk

cyclin-dependent kinase

conv.

conversion

CycE

cyclin E

DIPEA

N,N-diisopropylethylamine

D2B

direct-to-biology

DMT

design-make-test

E

extended

EdU

5-ethynyl-2’-deoxyuridine

ELSD

evaporative light scattering detection

ent-

enantiomer

HATU

O-(7-azabenzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexa-fluorophosphate

HTE

high-throughput experimentation

inhib.

inhibition

KDE

kernel density estimation

NB

no-binding

NMI

N-methylimidazole

NT

not tested

pHH3

phospho-Histone H3

Rot.

rotamer

t-SNE

t-distributed Stochastic Neighbor Embedding

UMAP

uniform manifold approximation and projection

UPLC-MS

ultrahigh pressure liquid chromatography–mass spectrometry

Data Availability Statement

All codes for molecular property computation and figure plotting are available at https://github.com/cernak-lab/Nanoscale-CDK2