General information

Yeast GFP clones, CIT2-GFP, AGP1-GFP, MEP2-GFP, LAP4-GFP and RPL19A-GFP were identified from an HTS campaign against the yeast GFP clone collection (Invitrogen) as the clones that respond to rapamycin. GFP clones were maintained on synthetic complete media (SCD) agar plates and grown in SCD media (Clontech, CA). S288C and BY4741 parental strains were obtained from stock cultures in our lab and are maintained and grown on SCD or YPD media (Fisher Scientific). The TORC1 bypass strain [31] was provided by Dr. Virgilio and was maintained in synthetic media lacking uracil (SD-Ura). The strain containing the CIT2-GFP reporter in rtg3 knockout background (BY4741, CIT2-GFP; rtg3::KmMX) was provided by Dr. Virgilio and was maintained on SCD or YPD plates. The CIT2-LacZ reporter strains in rtg2 or lst8 knockout background [9] were kindly provided by Dr. Liu and were maintained on YPD plates. Rtg3-GFP and Gln3-GFP strains were revived from yeast GFP clone collection and grown on YPD plate.

The fluorescent dyes Alexa Fluor® 405 and Alexa Fluor® 633 were purchased from Invitrogen™ (Eugene, OR). Rapamycin was purchased from TOCRIS Biosciences (Missouri, USA). Hyclone PBS buffer was purchased from ThermoFisher Scientific (Logan, Utah). BSA powder was purchased from Sigma-Aldrich (St. Louis, MO). Pluornic F-68 was purchased from Invitrogen.

Compounds ordered by UNM for SAR by commerce prior to involvement of the KU SCC were purchased from ChemDiv (San Diego, CA) and Ryan Scientific (Mt. Pleasant, SC). Unless otherwise indicated, all compound solutions were maintained and diluted in DMSO prior to addition to assay wells. Final DMSO concentrations were no more than 1% v/v. A Biomek® NX Multichannel (Beckman-Coulter, Fullerton, CA) was used for all cell and compound solution transfers for volumes greater than 1 μL. Low volume transfers (100 nL) were done via pintool (V&P Scientific, Inc., San Diego, CA). Compound dose response plates were generated with the Biomek® NX Span-8 (Beckman-Coulter, Fullerton, CA). The HyperCyt® high throughput flow cytometry platform (IntelliCyt™, Albuquerque, NM) was used to sequentially sample cells from 384-well microplates (2 μL/sample) for flow cytometer presentation at a rate of 40 samples per minute. Flow cytometric analysis was performed on a CyAn™ flow cytometer (Beckman-Coulter, Fullerton, CA). The resulting time-gated data files were analyzed with HyperView® software to determine compound activity in each well [32, 33]. Inhibition response curves were fitted by Prism® software (GraphPad Software, Inc., San Diego, CA) using nonlinear least-squares regression in a sigmoidal dose response model with variable slope, also known as the four-parameter logistic equation.

2.1. Assays

A. Primary Assay 1 (single point, 5-plex): multiplex HTS screen of TOR pathway GFP-fusion proteins in S. cerevisiae. AIDs: 1870, 1887, 1867,1873, 1862, 2066, 2029, 2025, 2023, 2016

This multiplex assay was conducted in a total volume of 10.1 μL in 384-well microtiter plates. The strains were grown separately overnight in SCD media in a shaking incubator at 30 ºC, and then stained for multiplexing with the Alexa Fluor 405 and Alexa Fluor 633 (see Figure 10 for staining scheme). Following staining, the yeast were combined and diluted into fresh SCD media at 0.5 OD₆₀₀. 5 μL SCD media supplemented with 0.03% pluronic were loaded into 384-well microtiter plates and 100 nL compounds were pintool-transferred. DMSO control and rapamycin (220 nM) controls were added to the first column and the second column respectively. A volume of 5 μL mixed cells were transferred into 384-well microtiter plates and incubated for 3 hours at 30 ºC with end-over-end rotation. The cells in the multiplex were interrogated for GFP expression levels using established high-throughput flow cytometric methodologies at the UNMCMD. Approximately 2 μL volumes from each well were collected at a rate of approximately 40 samples per minute. This results in analysis of approximately 1,000 cells of each cell type from each well. Flow cytometric data of light scatter and fluorescence emission at 530±20 nm (488 nm excitation, FL1), 665±10 nm (633 nm excitation, FL8) and 450±25 nm (405 nm excitation, FL6) were collected.

Figure 10

TORC1 bypass cell growth analysis on the probe ML231 and SID 96099781. 1:10 serial dilutions of BY4741 and TORC1 bypass cells were spotted onto the YPD plates supplemented with 7.5–60 μM compounds respectively

2.2. Probe Chemical Characterization

A. Probe Chemical Structure, Physical Parameters and Probe Properties

Figure 1Property summary of probe compound ML231 (SID 99300522, CID 3392161)

C. Solubility

Aqueous solubility of ML231 was measured in phosphate buffered saline (PBS) at room temperature (23 °C). PBS by definition is 137 mM NaCl, 2.7 mM KCl, 10 mM sodium phosphate dibasic, 2 mM potassium phosphate monobasic and a pH of 7.4. The solubility of probe ML231 (SID 99300522 CID 3392161) in this system was determined to be 0.05 μg/mL. [37]

The concentration of the probe (0.24 μM) is somewhat lower than the determined CIT2 EC₅₀ (3.8 μM) (Table 1). To test if ML231 and selected analogs were more soluble under the assay conditions, solubility was evaluated in yeast synthetic media containing 0.015% pluronic F-68 and 1% DMSO by counting solution particle formation by flow cytometry at the UNMCMD [38, 39] and independent results were obtained using the same media employing methods comparable to those used for the PBS buffer solubility determination. [37]

Table 1

Solubility experiments in various media with ML231 and analogs.

These experiments revealed that the aqueous solubility of this series, represented by 3–4 examples shown in Table 1 which included the probe ML231, is limited. Improved solubility in assay media as compared to PBS buffer was observed; however, the limitations persisted. Accumulation of ML231 in cells to higher levels than suggested by the solubility assays and solubility enhancement by reversible binding to biomolecules in the intracellular environment may account for the discrepancy between the EC₅₀ and apparent soluble concentration in media alone. Additional work is necessary to substantiate this hypothesis.

D. Stability

Aqueous stability was measured at room temperature (23 ºC) in PBS (no antioxidants or other protectants and DMSO concentration below 0.1%). The stability of probe compound ML231 (SID 99300522 CID 3392161), determined as the percent of compound remaining after 48 hours, was 89% [37]. Stability data are depicted as a graph showing the loss of compound with time over a 48 hour period with a minimum of 6 time points and provide the percent remaining compound at end of the 48 hours (Figure 2).

Figure 2

Aqueous stability of compound ML231 (SID 99300522, CID 3392161).

Notably, the duration of the assays used in characterizing ML231 and its analogs did not exceed 3 hours. As such it is not anticipated that the observed activity associated with ML231 is due to a reactive species. To further address if the scaffold was susceptible to nucleophiles and forming a reactive intermediate, ML231 was treated with glutathione. In each of four separate vials was added ML231 (5.0 mg, 0.017 mmol). L-Glutathione was added to each vial in the ratio of (ML231:glutathione) 1:1, 1:2, 1:3 and 1:4, respectively. DMSO (1 mL) was added to each vial and the reaction mixtures were stirred at room temperature for 48 h. Each reaction was monitored by LCMS over the 48 h period, and no change from starting material was observed in either the UV or the MS spectroscopy. After the 48 h experiment, the reaction containing the 1:1 ratio of ML231 to L-glutathione was doped with an additional 3 eq. of L-glutathione and heated at 80 °C for 2h. No change in UV was observed. An additional minor peak was noted (< 1%) in the MS spectrum besides starting material; however, the mass did not correlate to an identifiable displacement or addition product. The ML231 scaffold appears to be stable at room temperature when exposed to an excess (4 eq.) of glutathione and even under heating at high temperature.

E. Synthetic Route

The probe compound ML231 (SID 99300522 CID 3392161) was prepared in a one-pot procedure by taking advantage of the differential reactivity of the three chlorine atoms and sequential amination reactions (Figure 3).

Figure 3

General synthetic route for probe and associated analogues.

F. Submission of Five Related Analogues to the MLSMR

Five analogues have been fully characterized and prepared in preparation for submission to the MLSMR. The five selected analogs are shown in Figure 4.

Figure 4

Selected analogs to support probe ML231.

2.3. Probe Preparation

General experimental and analytical details:¹H and ¹³C NMR spectra were recorded on a Bruker AM 400 spectrometer (operating at 400 and 101 MHz respectively) or Bruker AM 500 spectrometer (operating at 500 and 125 MHz respectively) in CDCl₃ with 0.03% TMS as an internal standard or DMSO-d₆. The chemical shifts (δ) reported are given in parts per million (ppm) and the coupling constants (J) are in Hertz (Hz). The spin multiplicities are reported as s = singlet, br. s = broad singlet, d = doublet, t = triplet, q = quartet, dd = doublet of doublet and m = multiplet. The LCMS analysis was performed on an Agilent 1200 RRL chromatograph with photodiode array UV detection and an Agilent 6224 TOF mass spectrometer. The chromatographic method utilized the following parameters: a Waters Acquity BEH C-18 2.1 × 50mm, 1.7 μM column; UV detection wavelength = 214nm; flow rate = 0.4 mL/min; gradient = 5 – 100% acetonitrile over 3 minutes with a hold of 0.8 minutes at 100% acetonitrile; the aqueous mobile phase contained 0.15% ammonium hydroxide (v/v). The mass spectrometer utilized the following parameters: an Agilent multimode source which simultaneously acquires ESI+/APCI+; a reference mass solution consisting of purine and hexakis (1H, 1H, 3H-tetrafluoropropoxy) phosphazine; and a make-up solvent of 90:10:0.1 MeOH:Water: Formic Acid which was introduced to the LC flow prior to the source to assist ionization. Melting points were determined on a Stanford Research Systems OptiMelt apparatus.

The probe was synthesized by the route shown in Figure 3 (section 2.2E) and the following protocol: Probe ML231 (SID 99300522, CID 3392161): N⁴-(3-chlorophenyl)-N²-methyl-5-nitropyrimidine-2,4,6-triamine: To a solution of 2,4,6-trichloro-5-nitropyrimidine (1.0 eq, 0.657 mmol, 0.150 g) in EtOH (3.0 mL) at −40 °C, was added dropwise a solution of 3-chloroaniline (1.0 eq, 0.657 mmol, 0.084 g) and DIPEA (3.0 eq, 1.97 mmol, 0.255 g) in EtOH (1.5 mL). After stirring for 0.5 h at −40 °C, ammonia (2.0 M in MeOH, 4.0 eq, 1.3 mL) was added and the mixture was slowly warmed up to rt and stirred for 4h. Methylamine (33% wt. in absolute ethanol, 4.0 eq, 0.36 mL) was added and the mixture was kept on stirring for 16 h. The reaction was diluted with CHCl₃ (20 mL) and washed with brine (10 mL). The separated organic extracts were dried (MgSO₄) and concentrated. The residue was purified by chromatography (Biotage 50 g silica gel column, EtOAc/Hexane) to afford N⁴-(3-chlorophenyl)-N²-methyl-5-nitropyrimidine-2,4,6-triamine (0.077 g, 40%) as a yellow solid.

3.1. Summary of Screening Results

The primary screening results are summarized here with the flow chart (Figure 5) showing the overall workflow for the project and the AIDs uploaded to date. The primary screening was carried out in a 5-plex assay format in which CIT2, AGP1, MEP2, LAP4 and RPL19A GFP fusion clones were color-coded with Alexa Fluor dyes and evaluated simultaneously with respect to the alteration of GFP expression (Figure 6). Rapamycin was used as positive control to ensure the GFP clones have the appropriate response. With the treatment of rapamycin for 3 hrs, the GFP signal in CIT2, AGP1, MEP2 and LAP4 clones were increased to 1.5-fold or greater whereas that in RPL19A clone was decreased to 0.5 fold compared to the DMSO negative control. Compounds were scored active if the median channel fluorescence (MCF) of GFP in each clone were altered to the extent greater than 50% compared to the DMSO. The cut-off values for each clone were as follows: CIT2, AGP1 and LAP4 (increased to 1.5 fold), MEP2 (increased to 1.75 fold) and RPL19A (decreased to 0.66 fold). The primary screening results were uploaded as 10 AIDs: 1870, 2029-multiplex HTS screen of TOR pathway GFP-fusion protein in S. cerevisiae_specifically_CIT2; 1887, 2066- multiplex HTS screen of TOR pathway GFP-fusion protein in S. cerevisiae_specifically_AGP1; 1867, 2016- multiplex HTS screen of TOR pathway GFP-fusion protein in S. cerevisiae_specifically_MEP2; 1873, 2023- multiplex HTS screen of TOR pathway GFP-fusion protein in S. cerevisiae_specifically_LAP4, and 1862, 2025- multiplex HTS screen of TOR pathway GFP-fusion protein in S. cerevisiae_specifically_RPL19A. A total of 320,000 compounds were tested with 73, 225, 1059, 443 and 64 actives in CIT2, AGP1, MEP2, LAP4 and RPL19A clones respectively.

Figure 5

TOR compound triage and flow chart.

Figure 6

Schematic of screen strategy. A) Simplified yeast TORC1 signaling pathway showing the four signaling branches probed in this project. B) Dot plot showing the multiplex staining of 5 yeast GFP report clones using Alexa Flours 405 and 633.

Two subsequent SMR cherry picks resulted in single point confirmation of ~2400 compounds (AIDs: 2643, 2624, 2623, 2622, 2621, 2274, 2273, 2272, 2271 and 2270). The cherry pick single point confirmation screens resulted in 613 compounds for further dose response confirmation. The dose response confirmation dataset was uploaded as AIDs 2744, 2743, 2742, 2740, 2745 and resulted in 31, 28, 26, 228, 123 actives for CIT2, AGP1, MEP2, LAP4 and RPL19A branches respectively. A set of counter screens for 488/530 nm fluorescent compounds using S288c parental none-GFP yeast strain was also associated with the SMR cherry pick sets. These data were uploaded as AID 2757 where the 613 compounds were tested with a total of 21 compounds noted as green fluorescence. Autofluorescence was calculated on the basis of MCF detected in the green fluorescence channel (530 +/−20 nM) and the subtraction of the autofluorescence of the cells, as shown in the following equation: Fluorescence = MCF_CMPD - MCF_CELL where MCF_CMPD is the MCF of cells in the presence of test compound and MCF_CELL is the MCF of cells in the presence of the DMSO control. The compound was noted to be active if the fluorescence was greater than the average plus three standard deviations of the control wells on the plate. The activity score was calculated based on normalizing the fluorescence to the maximum measured.

A total of 255 compounds were confirmed. While one compound altered GFP expression in all five GFP clones as rapamycin does, some of the compounds were found to be selective on individual branches. The data was analyzed as follows: RawMCF values were converted to RawkMESF with the aid of calibration beads. SubkMESF were calculated by subtracting the baseline fluorescence measured from the S288c parental strain. Percent response was calculated by using the span of rapamycin response collected on the same plate. %Response= (SubkMESF@Conc–SubkMESF@DMSO)/( SubkMESF@Rap0.4μM– SubkMESF@DMSO). Note that for RPL19A, the absolute value of ( SubkMESF@Rap0.4μM– SubkMESF@DMSO) was used for calculation of the GFP decrease in RPL19A clone.

A total of 82 powder compounds of the 255 dose response confirmed hits were ordered by KU SCC on the basis of chemistry tractability and availability. These compounds were tested in the dose response confirmation (AIDs: 488792, 488795, 488801, 488808 and 488827). The 488/530 nm fluorescence profiling using the S288c parental non-GFP yeast strain was also performed in a dose response assay (AID 488829). The dose response confirmed the activity of 79 compounds. As one of the original goals of the project was to identify compounds with rapamycin-like activity with improved physiochemical behavior, the team was pleased to identify and confirm compound CID 3528206 (SIDs 96099781 and 26664184) that altered GFP expression in all five yeast clones, as seen with rapamycin (Figure 7A). This compound was further characterized in secondary assays, the outcomes of which are described in Figure 7.

Figure 7

Characterization of CID 3528206 (SID 96099781) as a TORC1 specific inhibitor. A) Dose response analysis of CID 3528206 (SID 96099781) on five GFP clones. B) Structure of CID 3528206 (SID 96099781). C) TORC1 bypass cell growth assay on CID 3528206 (SID (more...)

The first secondary assay was aimed at determining cytotoxicity of the compound in the TORC1 bypass cell growth assay (Figure 7C, AID: 504456). Compound CID 3528206 (SID 96099781) inhibited wild-type cell growth but not TORC1 bypass cell growth at 15–30 μM, suggesting that it may be not toxic to yeast. Otherwise, it would have inhibited both wild-type and TORC1 bypass cell growth.

The second assay was used to determine the impact on the TORC1 complex in the TORC1 bypass cell growth assay (Figure 7C, AID: 504456). Compound CID 3528206 (SID 96099781), behaved like rapamycin and inhibited wild-type cell growth but not TORC1 bypass cell growth at 15–30 μM, suggesting that it is functional at the level of TORC1 or upstream, as is rapamycin. This is in agreement with the dose response analysis by which only CID 3528206 (SID 96099781) was shown to be active in all 5 GFP clones in an analogous way as rapamycin.

The third assay of interest was used to determine the impact on inhibition of TORC1 activity by measuring Sch9p (a direct TORC1 substrate) phosphorylation (Figure 7D, AID: 504478). CID 3528206 (SID 96099781) inhibited Sch9p phosphorylation with an IC₅₀ of ~ 4 μM indicating that it inhibits TORC1 activity on the phosphorylation of Sch9p.

The fourth assay was used to determine the impact on inhibition of TORC2 activity by measuring phosphorylation of YPK1 (a direct TORC2 substrate (Figure 7E). CID 3528206 (SID 96099781) did not appear to inhibit YPK1 phosphorylation associated with TORC2.

Taken together, compound CID 3528206 (SIDs 96099781 and 26664184) may function as a specific small molecule inhibitor of TORC1 like rapamycin.

To gain preliminary SAR data for the scaffold on which CID 3528206 (SIDs 96099781 and 26664184) is based, a series of 30 structurally related compounds were picked from the DPI chemical library and included in the final round of cherry pick which contained 112 compounds. This final cherry pick compound set was tested in single-point conformation and dose response confirmation assay. Eleven AIDs were uploaded: 5 AIDs for single point confirmation (AIDs: 504321, 504334, 504336, 504338, 504340), 5 AIDs for dose response confirmation (AIDs: 488812, 488790, 488814, 488823 and 488825) and an AID for the fluorescent counter screen (AID 488818).

Of those compounds tested, one (SID 26671376, CID 3528206 – an early batch of the probe compound) stood out in its activity profile as an exclusive inhibitor of the CIT2 GFP clone even though the selective compound differed structurally from the rapamycin-like hit CID 3528206 by only the exchange of a 3-fluorophenyl group for a 3-chlorophenyl group at the same position (Figure 8). This result indicated that the nascent SAR on this single chemotype contained analogs exhibiting both rapamycin-like activity and branch selective activity. This interesting finding permitted the development of a branch selective probe and the determination of the structural and biological basis for the selectivity as compared to the rapamycin-like hit, CID 3528206.

Figure 8

Dose response analysis of probe compound ML231.

To date, 57 SAR compounds have been resupplied and assessed in dose response confirmation. The raw MCF data and % response data were plotted. Compounds were claimed as actives if two thresholds were met: 1) observation of the same fold-change of GFP expression cut-off value in the primary screen; for CIT2, AGP1 and LAP4 (increased to 1.5 fold or greater), for MEP2 (increased to 1.75 fold or greater) and for RPL19A (decreased to 0.66 fold or less), and 2) percent response (relative to rapamycin) at 100 μM was greater than 20%. A total of 9 compounds in this series were active in dose response (AIDs: 488792, 488795, 488801, 488808 and 488827) (Figure 9 and section 3.4 SAR tables). Compound (CID 3528206 SID 96099781) was the only compound that alters GFP expression in all five GFP clones.

Figure 9

Dose response curves for the active compounds. Dose response analysis was performed at least 3 times for all these compounds except SID 99300546 (CID 3739601). The data from one experiment are used as examples.

Compound (CID 50904403 SID 110923093) alters GFP expression in CIT2, AGP1, MEP2 and RPL19A clones, but not in the LAP4 clone. A total of seven compounds (SID 99300522, 99300541, 99300546, 99300549, 99300550, 110923089 and 110923090) are exclusively active on CIT2 expression. Among the 7 CIT2 branch selective compounds, CID 3392161 (SID 99300522) induces the second strongest response (91.7%) with the second best EC₅₀ (3.8 μM). CID 50904400 (SID 110923089) is the most potent compound (EC₅₀= 0.6 μM), but induces the weakest response (% response: 24.0%). CID 3449329 (SID 99300541) induces the strongest response (97.7%) with a lower EC₅₀ (8.8 μM). All 57 SAR compounds were assessed in the TORC1 cell growth secondary assays for evaluation of cytotoxicity and activity on TORC1 complex (AIDs: 504456). One compound (CID 46864179 SID 99300521) among these 57 compounds inhibits both wild-type and TORC1 bypass cell growth at 30 μM, suggesting it is toxic to yeast cells at this concentration, and only compound (CID 3528206 SID 96099781) was confirmed to inhibit wild-type cell growth but not the TORC1 bypass cell growth. The compound (CID 50904403 SID 110923093) targeting 4 GFP clones did not inhibit wild-type yeast cell. None of CIT2 selective compounds inhibited wild-type cell growth up to 60 μM (Figure 10 and data not shown). These data suggest that the CIT2 selective analogs may have a common mechanism of action but distinct from that of compound CID 3528206 (SID 96099781).

To investigate if the CIT2 selective compounds inhibit the phosphorylation of Sch9p as SID 96099781 (CID 3528206) does (Figure 7D), we evaluated the activity of SID 99300522 (CID 3392161), SID 99300541 (CID 3449329) and SID 99300549 (CID 4359574) in the Sch9p phosphorylation assay (Figure 11). All three compounds inhibited Sch9p phosphorylation. Compound SID 99300522 (CID 3392161) is shown to be the most potent inhibitor in this assay with IC₅₀ 2.3 μM. The IC₅₀s of SID 99300541 (CID 3449329) and SID 99300549 (CID 4359574) were 14.7 μM and 5.9 μM, respectively. However, compared to SID 96099781 (CID 3528206), which completely inhibited Sch9p phosphorylation, none of the CIT2 selective compounds fully inhibited Sch9p phosphorylation. SID 99300522 (CID 3392161) inhibited the Sch9p phosphorylation up to 89% while SID 99300541 (CID 3449329) and SID 99300549 (CID 4359574) only inhibited 70%–75% (Figure 11, and section 3.4 SAR tables). With respect to its selectivity, amplitude and potency, SID 99300522 (CID 3392161) was chosen as the probe in this probe report.

Figure 11

Sch9 phosphorylation assay on active compounds. A) Western blot showing the phosphorylated Sch9p (top band) in yeast cells treated with compounds at different concentrations (0.1–100μM). B) The intensity of top band was quantified to show (more...)

One key regulatory step in RTG pathway is the nuclear localization of Rtg1/3p [6, 8]. To investigate if the probe induces nuclear translocation of Rtg1/3p, we tested the probe SID 99300522 (CID 3392161), SID 96099781 (CID 3528206), SID 99300541 (CID 3449329) and rapamycin in the Rtg3p-GFP localization assay (Figure 12). Compared to the treatment with DMSO, Rtg3p-GFP was translocated into the nucleus after treatment with the probe SID 99300522 (CID 3392161), SID 96099781 (CID 3528206) and SID 99300541 (CID 3449329). Rapamycin also induced the nuclear translocation of Rtg3p-GFP. This result shows that the probe and its analogs can induce nuclear translocation of Rtg3p by which the CIT2 expression is induced.

Figure 12

Rtg3p-GFP translocation assay. Rtg3p-GFP cells were grown in SCD media and treated with 30μM compounds for 3hrs. The GFP localization was observed under the fluorescent microscope.

To further confirm the activity of the probe SID 99300522 (CID 3392161) on the RTG pathway and investigate its point of action, we carried out genetic analysis to measure CIT2 expression in the knockout background of RTG proteins and the Lst8p mutants, which are the key components of the RTG pathway (Figure 13A). First, we measured CIT2-GFP expression in the Rtg3p knock out mutant with the treatment of the probe SID 99300522 (CID 3392161), SID 96099781 (CID 3528206) and SID 99300541 (CID 3449329) as well as rapamycin and DMSO as controls (Figure 13B). In contrast to the CIT2-GFP expression in the wild-type background, the activity of the probe SID 99300522 (CID 3392161) and its analogs on CIT2-GFP expression is completely lost in Rtg3p mutant, as it is with rapamycin, suggesting that the activity of the chemotype is dependent on the Rtg3p protein and may function upstream of Rtg3p. Second, we measured CIT2-LacZ expression in both the wild-type background and the Rtg2p knockout background after treatment with the same set of compounds. The probe SID 99300522 (CID 3392161) and its analogs did not increase LacZ activity significantly in the Rtg2p mutant as rapamycin does (Figure 13C). These data suggest that the activity of the chemotype is dependent on the Rtg2p protein and may function upstream of Rtg2p. We also measured CIT2-LacZ expression in lst8-5 and lst8-1 mutants after treatment with the same set of compounds. None of these compounds could further increase LacZ expression in either lst8 mutants (data not shown), suggesting that the probe SID 99300522 (CID 3392161) and its analogs may function upstream of Lst8p. Taken together, this activity of the probe and its analogs on induction of CIT2 expression is dependent on RTG3 and RTG2 proteins, suggesting that the probe functions at the level of Rtg2p or upstream in the RTG pathway.

Figure 13

Chemical genetic analysis of active compounds on RTG pathway. A) RTG pathway in yeast (From Liu and Butow, 2006 [23]). B) Flow cytometry analysis of CIT2-GFP expression in the rtg3 knockout mutant with the treatment of 30μM compounds as well as (more...)

We have attempted to uncover the cellular targets of this chemotype by profiling their gene expression through the screening of yeast GFP clones. The probe SID 99300522 (CID 3392161), analog SID 96099781 (CID 3528206), rapamycin and DMSO were used to screen a set of 96 GFP clones (called RapaGFP clones) that were identified initially to be responsive to rapamycin from the yeast GFP collection (unpublished data). After the treatment of the GFP clones with the compounds for 3 hrs, the GFP signal of each clone was measured by flow cytometry. Compared to the GFP signal with the treatment of DMSO, the GFP clones treated with the compounds in which the GFP signal was increased to greater than 1.5-fold or decreased to 0.66-fold or less were claimed to be responsive. 53 GFP clones were confirmed to be responsive to rapamycin among which 41 clones were up-regulated and 12 were down-regulated (Figure 14). Note that 11 of the down-regulated clones encode ribosomal proteins, which is in agreement with the fact that rapamycin inhibits protein synthesis. The other 43 GFP clones did not respond to rapamycin or to any other compounds tested. In the presence of SID 96099781 (CID 3528206), 31 out of 41 rapamycin-up-regulated clones were up-regulated. Four of the rapamycin-down-regulated clones were down-regulated. Note that the rest of the 8 clones encoding ribosomal proteins also decreased ~20% by SID 96099781 (CID 3528206) (data not shown). This piece of data suggests that SID 96099781 (CID 3528206) exhibits a very similar pattern as rapamycin. This is consistent with the results from dose response confirmation and the secondary assays (Figure 7). Interestingly, the probe SID 99300522 (CID 3392161) just increased GFP expression in 6 clones and decreased GFP expression in ADE17 clone only (Figure 15). None of the clones encoding the ribosomal proteins were responsive to the probe SID 99300522 (CID 3392161) treatment, suggesting that the probe may not regulate the protein translation branch of TOR pathway. Among the 6 up-regulated clones, CIT2, DLD3 [40] and PYC1 [41, 42] are known to be regulated by the RTG pathway as well as nitrogen sources. PTR2 (Peptide Transport), BAT2 (Branched-chain Amino acid Transaminase) and DIP5 (DIcarboxylic amino acid Permease) play roles in maintaining amino acid homeostasis, suggesting that the probe SID 99300522 (CID 3392161) may regulate amino acid metabolism and/or affect the extracellular sensing of amino acid, which is controlled by the yeast plasma membrane SPS nutrient sensor. The SPS nutrient sensor has been reported to affect the RTG pathway upstream of Rtg2p [23, 43] via an unknown mechanism. Taken together with the observation that the probe SID 99300522 (CID 3392161) also acts upstream of Rtg2p, it will be very interesting to test if the probe targets the SPS pathway in the future.

Figure 14

RapaGFP clone screening. 96 RapaGFP clones were treated with 0.22μM rapamycin(Rapa), 30μM SID 96099781 (CID 3528206, TC1) – or 30μM SID 99300522 (CID 3392161, probe) for 3hrs. The genes with altered gene expression are (more...)

3.2. Dose Response Curves for Probe ML231

Figure 15Overlay of dose response curves for probe ML231 (SID 99300522, CID 3392161)

The mean percent of response of each GFP clone to the probe treatment in 6 independent experiments was used to fit by Prism 5.

3.3. Scaffold/Moiety Chemical Liabilities

The 2,4,6-triamino-5-nitropyrimidine scaffold and its derivatives have been easily handled in terms of stability to reaction conditions, exposure to acid or base, heating, and general manipulation. We have not observed decomposition nor have we experienced any chemical liability with these compounds. The structure does not contain moieties that are known generally to be reactive. Stability assessment was performed in 1× PBS buffer at pH 7.4 and room temperature. After 48 hours, it was determined that 89 percent of the parent probe compound remained, thus indicating reasonable stability under these conditions.

3.4. SAR Tables

Of the hits that were evaluated and validated, only one compound (SID 96099781, CID 3528206) resulted in an activity profile that altered GFP expression in all five GFP clones as seen with rapamycin. A chemistry effort was then undertaken to expand the underlying structure activity relationships associated with this scaffold (Figure 16). The preliminary SAR plan focused on four regions of the scaffold, predominately on the substitutions of the pyrimidine core. Those compounds that could be obtained commercially were purchased while unavailable analogs were synthesized in parallel. We synthesized 27 compounds and purchased 37 additional analogs that were purified and analyzed prior to assay. Interestingly, these compounds, all predicated on the validated hit, did not demonstrate an analogous effect on GFP expression as that seen with rapamycin and the hit itself. Instead, nearly all active analogs in this series exhibited pathway selectivity for induction of the CIT2-GFP expression. This finding, while unexpected for compounds so closely related to the hit structure, was rigorously validated and then supported with the identification of additional analogs in the series with analogous selectivity profiles.

Figure 16

Validated hit and SAR strategy.

The largest group of analogs that was studied resulted from systematic changes of the 3-fluorophenyl moiety of SID 96099781 (Table 2). Consistently, 4-halo or 4-alkoxy substituted phenyl rings in the R1 position resulted in loss of activity on all 5 clones (entries 2, 7, and 10–11). A similar story emerged for 2-substituted phenyl rings (entries 3 and 5). Some CIT2 inhibition registered for the 2-MeO-phenyl substituent (~4 μM, entry 8); however, the compound was not very effective, as a low response was associated with this result (33%).

Table 2

Summary of modifications to R₁.

Installation of 3-substituted phenyl rings revealed CIT2 selective potency for 3-chloro-, 3-bromo- and 3-MeO-phenyl substituents, although low responses plagued the latter two examples (entries 1, 4 and 9, respectively). The 3-chlorophenyl moiety (entry 1) imparted a reasonable CIT2 potency of 3.8 μM, response (92%), and lack of activity on the other four clones. The 3-fluorophenyl analog (entry 6, SID 96099781) was the original validated hit with effects observed on all 5 GFP clones. As the data in section 3.1 indicated, SID 96099781 was the only analog in the pyrimidine series that inhibited wild-type cell growth but not the TORC1 bypass cell growth, suggesting it acts at the level of TORC1 or upstream as rapamycin does. This is in contrast to the remaining analogs which appear to act downstream of TORC1. This differential site of action can account for the observation that SID 96099781 does not fit within the SAR profile surrounding this set of compounds. Compounds in this series were also evaluated for their effect on cell growth. All compounds in the series but the validated hit (entry 6) resulted in good wild-type cell growth, indicating lack of cytotoxicity.

The effect of slightly increasing the steric bulk or replacing the hydrogen of the methylamine appendage was investigated (Table 3). Introduction of an ethyl substituent in place of the methyl group at R2 (entry 2) resulted in enhanced potency on CIT2, AGP1, MEP2 and RPL19A as compared to the probe (entry 1). No activity on LAP4 was observed, but the potency across the remaining clones amounted to pan inhibition, though the response levels on AGP1 and MEP2 were low. Installation of an additional methyl group on the amine (entry 3) resulted in complete loss of activity across all clones, indicating the possible importance of the hydrogen in this position or as a participant in an active tautomeric structure.

Table 3

SAR summary for modification at R₂.

Attempts were made to modify the nitro group of the hit scaffold, as this particular functionality can contribute to poor solubility. One of the first structural amendments involved removal of the nitro group (entry 2, Table 4). In this case, the presence of the hydrogen was not a beneficial change, resulting in loss of activity on CIT2 as compared to the probe (entry 1) and suggesting that the electron withdrawing effects imparted by the nitro group might be important to activity. With this in mind, an electron-withdrawing nitrile group was introduced at the R3 position (entry 3) and an aryl ketone (entry 4) that was commercially available was also assessed. Unfortunately, these changes did not restore CIT2 activity, leading to an inactive profile across all five clones. These alterations represent a modest survey of changing the nitro group, and initial assessments indicate that the nitro group, a very polarizing functionality as compared to most complimentary replacements, is important to retaining CIT2 activity.

Table 4

SAR summary for modification of the nitro group at R₃.

The importance of the aniline portion of the scaffold was studied by introducing small alkyl substituents in place of the (N-H) hydrogen atoms (Table 5). Replacement of one hydrogen with a methyl group or both hydrogen atoms with methyl groups (entries 2 and 3, Table 5, respectively) resulted in no activity on any of the 5 clones, thus stressing the importance of the hydrogen atoms in any of several scenarios of an active analog: (a) participants in hydrogen bonding and/or (b) participants in an active tautomeric structure and/or (c) intolerance of steric bulk in this binding region.

Table 5

SAR summary for modification of the R₄ substituent.

3.5. Cellular Activity

The primary screen assay, TORC1 bypass cell growth assay and the chemical genetic assays in RTG mutants are all direct indication of cellular activity. The results of the primary screen assay and the chemical genetic assays indicate that the probe acts at the level of Rtg2p or upstream in the RTG pathway. The result from TORC1 bypass cell growth assay indicates that the probe is not toxic to the wild-type yeast cell growth at 60 μM. Note that the synthetic media used in the primary screen and the dose response analysis contains 0.015% pluronic F-68 that is required for the activity of rapamycin. But the probe showed the similar activity in the synthetic media without the addition of pluronic F-68, suggesting it is permeable to yeast cells.

3.6. Profiling Assays

The probe has been profiled against 50 kinases at a single concentration of 10 μM to assess promiscuity of the chemotype (Table 6). [44] Probe ML231 was dissolved in DMSO and tested at a final concentration of 10 μM. Prior to initiating a profiling campaign, the compound was evaluated for false positive against split-luciferase. Profiling was done in duplicate for ML231 against each kinase. The Percent Inhibition and Percent Activity Remaining were calculated using the following equation:

Table 6

Percent of activity remaining for various kinases when inhibited by ML231.

It is noteworthy that a PubChem search revealed that the probe ML231 (and analog SID 96099781) both are reported to inhibit mammalian GSK3β activity (AID 434954). This particular kinase was not offered as part of the profiling panel; however, the most potently inhibited kinase in this panel was GSK3a at 35.3%. Other kinases that were inhibited by the probe included CLK1 (28.3%), CLK2 (20.6%), AURKA (17.9%), RPS6KA1 (14.5%), STE (9.9%), SNARK (9.4%), and others less than 5%.

The team has also submitted the probe ML231 (SID 99300522 CID 3392161) to the NIH National Cancer Institute to elucidate the effect of the probe on the cancer cell line panel.

4.1. Comparison to Existing Art and How the New Probe is an Improvement

To our best knowledge, this is the first report of the RTG pathway selective probe. It will allow probing of the RTG branch without affecting other branches regulated by TORC1 signaling.

4.2. Mechanism of Action Studies

We have attempted to reveal its mechanism of action by searching for genes whose expression are changed upon the treatment of probe compound by screening 96 RapaGFP clones and microarray analysis.

We have performed two experiments to search for genes whose mRNA/protein expression is altered with the treatment of the probe. First, we screened 96 RapaGFP clones (see Figure 14) and identified not only the GFP clones regulated by the RTG pathway, as well as GFP clones encoding the genes involved in amino acid homeostasis, such as DIP5, BAT2 and PTR2 clones. These genes are up-regulated by the yeast membrane SPS (Ssy1-Ptr3-Ssy5) nutrient sensor for extracellular amino acids [48]. This suggests that the probe ML231 (SID 99300522, CID 3392161) may affect the SPS nutrient sensing pathway. Note that RTG pathway is regulated by glutamine, an amino acid that is also sensed by the SPS pathway (Figure 13A). The three RTG- regulated GFP clones (CIT2, DLD3 and PYC1) responsive to the probe may also be indirectly regulated by the SPS pathway. A key effector of the SPS pathway, Stp1p, is demonstrated to be regulated by the Tap42-Sit4p phosphatases complex, which is a well-known target of the TORC1 complex [49]. This finding suggests the connection between the TORC1 and the SPS signaling pathways [49]. Taken together, the potential cellular targets of the probe ML231 (SID 99300522, CID 3392161) may play roles in all the RTG, SPS and TORC1 pathways. It will be interesting to test whether the probe affects the Tap42-Sit4p phosphatases and regulates Stp1p. The screening assays for RapaGFP clones (Figure 14) allows us to evaluate several compounds with a limited amount (less than 1mg) at current stage, but it is a biased screen where the clones not responsive to rapamycin are missing. We plan to screen the whole GFP collection with the probe ML231 (SID 99300522, CID 3392161) in the future. This would generate more comprehensive information for the probe. Second, we have performed microarray analysis on the probe ML231 (SID 99300522, CID 3392161), rapamycin, SID 96099781 (CID 3528206) and 2 other CIT2 selective compounds. The data is being analyzed. The microarray data may provide very useful information on the possible cellular targets of the probe with the comprehensive comparison of the gene list of the probe and its analogs as well as rapamycin. We are going to do cluster/gene function analysis and promoter/transcription factor binding analysis to determine the possible biological pathway(s) that the probe candidate acts on. The data from the GFP collection screen and the microarray analysis would shed light on the cellular targets.

We also plan to test the probe in extended secondary assays to determine its mechanism of action (see details in section 4.3).

4.3. Planned Future Studies

In addition to the large-scale analysis of gene expression experiments, the probe ML231 will be further tested in extended secondary assays for revealing the mechanism of action and chemistry will be working to improve solubility.

1. Screen the whole yeast GFP collection with probe ML231, SID 96099781 (CID 3528206), rapamycin and DMSO. This screen will be done in 96-well plates at UNMCMD. The data will be analyzed as previously [50].
2. Analyze microarray data and confirm gene expression by RT-PCR. The microarray data will be analyzed by principle component analysis and two-dimensional hierarchical clustering. This two unsupervised data analysis will suggest relationship among the probe, SID 96099781 (CID 3528206), rapamycin and the active SAR analogs. The genes shared by these compounds will be used for pathway analysis with Genego systems to reveal the signaling pathways that may be regulated by the probe. The subsequent genetic and biochemical assays will be performed accordingly.
3. Stp1 localization assay. The GFP localization of Stp1p-GFP will be evaluated after treatment with the probe ML231 and its analogs as well as rapamycin. This experiment will test if ML231affects the SPS signaling pathway.
4. Test probe ML231 and active SAR compounds in the rich YPD media as well as in the synthetic media with different sole nitrogen sources, such as glutamate, glutamine, proline, urea and histidine. This experiment will suggest whether the probe candidate affects the regulation of glutamine and ammonia in the RTG pathway, which would aid in understanding the integration of the RTG, NCR and TORC1 pathways.
5. Extend research into mammalian cells. The active compounds, including the probe ML 231 and compounds SID 96099781, SID 99300541 and SID 99300549 as well as two inactive compounds SID 99300529 and SID 99300530 are being profiled in human gene expression chips in collaboration with the Broad Institute. Extended secondary assays in mammalian cells may be employed accordingly.
6. Chemistry will be done to improve the solubility and drug-likeness properties of the probe. Selective priorities include the identification of an isostere for the nitro group, the introduction of solubilizing groups on the R1 arene, more extensive revision of this component of the scaffold (aliphatic solubilizing groups) and modification of the central pyrimidine core.