Design, Synthesis, and Biological Application of Novel Photoaffinity Probes of Dihydropyridine Derivatives, BAY R3401

To explore the molecular mechanisms of BAY R3401, four types of novel photoaffinity probes bearing different secondary tags were synthesized. Their potency for glycogenolysis was evaluated in primary human liver HL-7702 cells and HepG2 cells. Probe 2d showed the best activity in primary human liver HL-7702 cells and HepG2 cells, with IC50 values of 4.45 μM and 28.49 μM, respectively. Likewise, probe 5d showed IC50 values of 6.46 μM in primary human liver HL-7702 cells and 15.29 μM in HepG2 cells, respectively. Photoaffinity labeling experiments were also performed and protein bands larger than 170 kDa were specifically tagged by probe 2d. The results suggest that the synthesized probe 2d might be a very promising tool for the isolation of the target proteins of BAY R3401.


Introduction
BAY R3401 is an orally bioavailable hypoglycemic agent for the treatment of type 2 diabetes, as reported by the Bayer Pharmaceutical Company [1]. This agent allows irreversible, nonselective suppression of hepatic glycogenolysis by inhibiting glycogen phosphorylase, which is the rate controlling enzyme of the glycogenolytic pathway [2]. The active metabolite, W1807, contributes significantly to its activity [3]. Nonetheless, much to the researchers' surprise, BAY R3401 inactivated glycogen phosphorylase by 63%, but glucose output dropped by 83% in the perfused liver [4]. It is difficult to explain the effects based only on the reported mechanism. Therefore, the exact mode of action of BAY R3401 has not been established.
Photoaffinity labeling is one of the major methods to directly capture small-molecule binding proteins [5]. The conventional approach, however, usually relies on the synthesis of photoaffinity probes and the identification of photolabeled fragments in proteins [6]. In general, a typical photoaffinity probe contains three functional groups. A bioactive scaffold ferries the probe to the enzyme active site, a photoreactive group generates a covalent and irreversible linkage between the probe and its target macromolecule after UV irradiation, and a tag (such as biotin or fluorophore) detects and/or visualizes the modified target enzymes [7].
Herein, we report the synthesis and biological application of four types of photoaffinity probes based on BAY R3401, which contains both a benzophenone photophore for covalent labeling of target proteins and a secondary handle for the subsequent detection or manipulation of labeled proteins (Figure 1). Probes bearing different secondary tags were exploited, either by direct attachment
The key intermediates, 20-22, were prepared individually via procedures similar to those reported previously, with some modifications (Scheme 2) [9]. Treatment of Lys(Boc)-OMe with 4-benzoylbenzoic acid using EDCI/DMAP afforded compound 16, bearing a benzophenone photophore. Hydrolysis of 16 with aqueous NaOH yielded carboxylic acid 17, followed by an esterification with propargyl bromide to produce the alkyne 18. Subsequent deprotection of 18 using TFA gave amide 19, then coupling 19 with D-biotin in DMF in the presence of EDCI, HOBt, and DIPEA as condensing agents produced biotin conjugate 20 with a biotin tag. Treatment of amino 19 with dansyl chloride (DNS-Cl) gave the desired fluorescent derivative 21. Amidation of amino 19 with ClCH 2 COCl was carried out to produce chloroacetyl compound 22 for the next step of azide displacement.
The preparation of a dual-labeled moiety was accomplished as follows (Scheme 3). The synthesis cycle began with deprotection of 16 using TFA, producing amide 23. Then, amidation with DNS-Cl gave compound 24. Hydrolysis of 24 with aqueous LiOH afforded carboxylic acid 25 with an 87% yield. Treatment of N-Cbz-N'-Boc-L-lysine with propargyl bromide gave alkyne 27. Subsequent deprotection of 27 using TFA, followed by coupling with D-biotin in DMF in the presence of isobutyl chloroformate and Et 3 N as condensing agents, produced biotin conjugate 28. The N-Cbz group was hydrolyzed off in 30% HBr in acetic acid at room temperature to give the amide compound 29. Reaction of 29 with carboxylic acid derivative 25 in the presence of HATU and DIPEA produced the dual label moiety with a 43% yield. The preparation of a dual-labeled moiety was accomplished as follows (Scheme 3). The synthesis cycle began with deprotection of 16 using TFA, producing amide 23. Then, amidation with DNS-Cl gave compound 24. Hydrolysis of 24 with aqueous LiOH afforded carboxylic acid 25 with an 87% yield. Treatment of N-Cbz-N'-Boc-L-lysine with propargyl bromide gave alkyne 27. Subsequent deprotection of 27 using TFA, followed by coupling with D-biotin in DMF in the presence of isobutyl chloroformate and Et3N as condensing agents, produced biotin conjugate 28. The N-Cbz group was hydrolyzed off in 30% HBr in acetic acid at room temperature to give the Scheme 4. Reagents and conditions: (a) CuSO4·5H2O, sodium ascorbate, CH2Cl2-H2O, r.t.; (b) NaN3, DMF, r.t.
The synthesis of control compounds 32 and 34, which lack the bioactive ligand BAY R3401 and only contain the photoaffinity and tag groups, is shown in Scheme 5. The cycle was started with the deprotection step, as above, followed by amidation with D-biotin and ClCH2COCl to give the control compound 32, and intermediate haloester 33. Then, treatment of 33 with NaN3 produced the control compound 34 with a 77% yield. To confirm the efficacy of designed probes 5a-5d in combining the tag after protein labeling, the reaction of the tag-free probe 5d with the alkyne-coupled biotin derivative in a simple model system was also studied (Scheme 6). The biotin derivative 35 was prepared based on a previously reported procedure described in [10]. Treatment of probe 5d with alkyne-biotin 35 under typical click conditions (CuSO4·5H2O, with sodium ascorbate as the reducing agent) in CH2Cl2-H2O, the conjugate product 36 could be isolated with a 13% yield as a white solid. These results demonstrated that the azide handle of the tag-free probe can be subsequently conjugated with an alkyne-tag through biocompatible copper-catalyzed azide-alkyne cycloaddition.  To confirm the efficacy of designed probes 5a-5d in combining the tag after protein labeling, the reaction of the tag-free probe 5d with the alkyne-coupled biotin derivative in a simple model system was also studied (Scheme 6). The biotin derivative 35 was prepared based on a previously reported procedure described in [10]. Treatment of probe 5d with alkyne-biotin 35 under typical click conditions (CuSO 4 ·5H 2 O, with sodium ascorbate as the reducing agent) in CH 2 Cl 2 -H 2 O, the conjugate product 36 could be isolated with a 13% yield as a white solid. These results demonstrated that the azide handle of the tag-free probe can be subsequently conjugated with an alkyne-tag through biocompatible copper-catalyzed azide-alkyne cycloaddition. The synthesis of control compounds 32 and 34, which lack the bioactive ligand BAY R3401 and only contain the photoaffinity and tag groups, is shown in Scheme 5. The cycle was started with the deprotection step, as above, followed by amidation with D-biotin and ClCH2COCl to give the control compound 32, and intermediate haloester 33. Then, treatment of 33 with NaN3 produced the control compound 34 with a 77% yield. To confirm the efficacy of designed probes 5a-5d in combining the tag after protein labeling, the reaction of the tag-free probe 5d with the alkyne-coupled biotin derivative in a simple model system was also studied (Scheme 6). The biotin derivative 35 was prepared based on a previously reported procedure described in [10]. Treatment of probe 5d with alkyne-biotin 35 under typical click conditions (CuSO4·5H2O, with sodium ascorbate as the reducing agent) in CH2Cl2-H2O, the conjugate product 36 could be isolated with a 13% yield as a white solid. These results demonstrated that the azide handle of the tag-free probe can be subsequently conjugated with an alkyne-tag through biocompatible copper-catalyzed azide-alkyne cycloaddition.

Cell Assay and SAR Analysis.
It is obviously important that the synthesized probes retain potency in bioassays. To evaluate the effects of all probes, the glycogenolysis assays were established in vitro, with primary human liver HL-7702 cells and HepG2 cells, based on the published method [11]. A well-known chloroindole inhibitor of glycogenolysis, CP-91149, was used as a positive control in these experiments.
The IC 50 values of the tested derivatives are listed in Table 1. Most of the newly synthesized probes maintained moderate inhibitory activity against glucagon-stimulated glycogenolysis in primary human liver HL-7702 cells and HepG2 cells. It is interesting to note that modification of dihydropyridine scaffold with bulky substituents resulted in a great increase in IC 50 both in primary human liver HL-7702 cells and HepG2 cells (e.g., BAY R3401 vs. 2b-2d, 3b-3d, 5a, 5c, 5d, 6a). The results are consistent with the SAR analysis of W1807, the active metabolite in cells of BAY R3401, which revealed that the N1-substituent may productive van der Waals interactions between the substitutions and its target proteins [8]. With different linkers, SAR analysis in HepG2 cells shows that the distance between the BAY R3401 moiety and secondary tags' moiety is important-a longer linker led to better inhibitory activity (e.g., 2a vs. 2d, 3a vs. 3d, and 5a vs. 5d). However, data analysis indicated no clear SAR for the distance in primary human liver HL-7702 cells. Within this series of compounds, probe 2d showed the best activity in primary human liver HL-7702 cells and HepG2 cells, with IC 50 values of 4.45 µM and 28.49 µM, respectively. Likewise, probe 5d showed an IC 50 value of 6.46 µM in primary human liver HL-7702 cells and 15.29 µM in HepG2 cells, respectively. Therefore, probe 2d and 5d may be very promising tools for isolation of the target proteins of BAY R3401.

Application of Activity-Based Profiling to Target Discovery
Based on the results of the potency in cell assay, probe 2d was selected for photolabeling studies to detect the binding proteins of BAY R3401 by PAGE and chemiluminescence [12]. The soluble proteomes prepared from HepG2 cells, were incubated with the 10 µM probe 2d, exposed to UV light for 30 min, followed by SDS-PAGE electrophoresis, and then transfer onto a PVDF membrane for detection with streptavidin-HRP [10]. Samples were prepared by incubating 2.0 mg/mL proteomes at different conditions: (Lane 1 and 6) marker; (Lane 2) with the 10 µM probe 2d and exposed to UV light for 30 min; (Lane 3) with the 10 µM probe 2d and BAY R3401, and then exposed to UV light for 30 min; (Lane 4) With the 10 µM control compound 32 and exposed to UV light for 30 min; and (Lane 5) with the 0 µM probe 2d and exposed to UV light for 30 min.
The results showed that a protein band larger than 170 kDa was seen in samples incubated with 2d. This labeling was specific since it was completed by BAY R3401, and no such labeling was seen when samples were incubated with control compound 32 rather than 2d ( Figure 3). These data demonstrate that the synthesized probe 2d can efficiently and specifically identify the binding protein(s) of BAY R3401 and suggest that 2d will be suitable for isolating the binding protein(s) of BAY R3401 by avidin-agarose chromatography.
Molecules 2018, 23, x FOR PEER REVIEW 11 of 28 The results showed that a protein band larger than 170 kDa was seen in samples incubated with 2d. This labeling was specific since it was completed by BAY R3401, and no such labeling was seen when samples were incubated with control compound 32 rather than 2d ( Figure 3). These data demonstrate that the synthesized probe 2d can efficiently and specifically identify the binding protein(s) of BAY R3401 and suggest that 2d will be suitable for isolating the binding protein(s) of BAY R3401 by avidin-agarose chromatography. Then, we used a more detailed proteomic analysis, with a mass spectrometric analysis method, to detect the protein band larger than 170 kDa. Several proteins have been identified, and some of the results are summarized in Table 2. We hypothesise that the PH-interacting protein, histone H4, hexokinase-2 and solute carrier family 2 might be the target proteins of BAY R3401 based on the protein probability scores in MaxQuant. We are now in the process of verifying these proteins. A comparison between the streptavidin blot analysis and Coomassie brilliant blue (CBB) staining was carried out. The detailed information of CBB is in Supplementary information. The results also showed that the streptavidin blot analysis was specific since the protein labeled by probe 2d is clearly shown after the streptavidin blot analysis while CBB staining had a strong background with the gel image under identical staining conditions.  Then, we used a more detailed proteomic analysis, with a mass spectrometric analysis method, to detect the protein band larger than 170 kDa. Several proteins have been identified, and some of the results are summarized in Table 2. We hypothesise that the PH-interacting protein, histone H4, hexokinase-2 and solute carrier family 2 might be the target proteins of BAY R3401 based on the protein probability scores in MaxQuant. We are now in the process of verifying these proteins. A comparison between the streptavidin blot analysis and Coomassie brilliant blue (CBB) staining was carried out. The detailed information of CBB is in Supplementary information. The results also showed that the streptavidin blot analysis was specific since the protein labeled by probe 2d is clearly shown after the streptavidin blot analysis while CBB staining had a strong background with the gel image under identical staining conditions.

Chemistry
NMR experiments were performed on a Bruker Avance III 400 MHz and a Bruker Fourier 300 MHz. The spectra are referenced internally according to the residual solvent signals of TMS (δ = 0.00 ppm). Positive or negative ion LCMS data were obtained at 303 K by a quadrupole Mass Spectrometer on Agilent LC/MSD 1200 Series using a 50 × 4.6 mm (5 µm) ODS column. Prep-HPLC experiments were performed by flash welchrom C18 column (150 × 20 mm) chromatography.

General Procedure for Synthesis of compounds 13a-13d
To a solution of 12 (12.20 g, 0.03 mol) in anhydrous acetonitrile (120 mL) was added K 2 CO 3 (12.50 g, 0.09 mol) in groups, and then Bromo-PEG(n)-bromide (0.15 mol) was added to the reaction mixture slowly under an ice bath, and the temperature was raised to 80 • C for 1.5 h. After cooling to r.t., the mixture was diluted with water (150 mL) and extracted by EtOAc (100 mL × 2). The combined organic phase was washed with brine (100 mL × 2), dried over anhydrous Na 2 SO 4 , and evaporated.

Glycogenolysis Inhibition in Liver HL-7702 Cells and HepG2 Cells
The inhibition of hepatic glycogenolysis was monitored by the measurement of liver glycogen using a microplate reader (BIO-RAD, Bio-Rad Laboratories, Hercules, CA, USA), which was done quantitatively by the anthrone reagent (Sigma, Saint Louis, MO, USA) colorimetric method based on the published method [11]. Primary human liver HL-7702 cells or HepG2 cells (Sigma) were treated with the test compound or DMSO solvent (final concentration, 0.10%), followed by 60 min incubation with 0.3 nM glucagon (GGN). Assays were terminated by centrifugation, and cells were digested with 30% KOH followed by glycogen determination. The IC 50 values were estimated by fitting the inhibition data to a dose-dependent curve using a logistic derivative equation.