Synthesis of New Brassinosteroid 24-Norcholane Type Analogs Conjugated in C-3 with Benzoate Groups

The metabolism of brassinosteroid leads to structural modifications in the ring skeleton or the side alkyl chain. The esterification and glycosylation at C-3 are the most common metabolic pathways, and it has been suggested that conjugate brassinosteroids are less active or inactive. In this way, plants regulate the content of active brassinosteroids. In this work, the synthesis of brassinosteroid 24-norcholane type analogs conjugated at C-3 with benzoate groups, carrying electron donor and electron attractant substituents on the aromatic ring, is described. Additionally, their growth-promoting activities were evaluated using the Rice Lamina Inclination Test (RLIT) and compared with that exhibited by brassinolide (used as positive control) and non-conjugated analogs. The results indicate that at the lowest tested concentrations (10−8–10−7 M), all analogs conjugated at C-3 exhibit similar or higher activities than brassinolide, and the diasteroisomers with S configuration at C-22 are the more active ones. Increasing concentration (10−6 M) reduces the biological activities of analogs as compared to brassinolide.


Introduction
Brassinosteroids (BRs) are an important group of polyhydroxylated sterol plant growth regulators in multiple developmental processes, at nanomolar to micromolar concentration, including cell division, cell elongation, vascular differentiation, reproductive development, and modulation of gene expression [1]. BRs also influence various other developmental processes such as the germination of seeds, rhizogenesis, flowering, senescence, abscission, and maturation. They also confer resistance to plants against various abiotic and biotic stresses [2][3][4][5].
On the other hand, a series of C-3 esterified derivatives of 24-epibrassinolide (13-15) and synthetic BRs analogs (16)(17)(18)(19) (Figure 2) have been reported [21,22]. However, biological evaluations in the Bean Second-Internode Bioassay (BSIB) for compounds 16 and 17 indicated that these analogs are less active than 24-epibrassinolide [22]. These results are in line with previously established structure-activity relationships ob-tained for natural BRs. These structure-activity relationship (SAR) studies have been made using BSIB and the Rice Lamina Inclination Test (RLIT) [23][24][25], and their main goal is to define general structural requirements for the growth-promoting activity of BRs [24,[26][27][28][29]. These results have been used to guide the synthesis of BRs analogs with a variety of structural modifications but keeping those considered essential for biological activity.
Several studies have proved that synthetic BRs analogs with significant structural changes and different substituents, both in the ring and the alkyl chain, can induce similar or even higher biological effects in plants as compared to natural BRs [30][31][32][33][34][35][36]. Some recent reviews of the growth-promoting activity of BRs and their analogs have established novel structural requirements for the existence of biological activity [23,[37][38][39]. For example, it has been shown that methyl ethers at C-3 are more active than 1 in the RLIT [40], whereas benzoate esters in the C-3 position were found to be less active than 24-epibrassinolide in the BSIB test [22].
In a previous in silico study, we have assessed the effect on activity of different groups attached to position C-3 of BRs analogs. The results suggest that bulky groups reduce the activity, whereas functionalization with electronegative and hydrophobic groups would increase it [29]. Thus, in this work, we present the synthesis of four new BR 24-norcholane type analogs conjugated with benzoate groups in C-3 ( Figure 2, compounds 18a, 18b, 19a, and 19b). The aromatic ring of the benzoate group contains electron-donor and electronwithdrawing substituents. Their growth-promoting activities were evaluated using RLIT, and the results were compared with those reported for other structurally similar analogs ( Figure 2, compounds 20a and 20b) [38,41,42].
The synthesis and evaluation of biological activity of these BRs analogs, conjugated in C-3 with benzylic esters, are studied either to get new active molecules or to elucidate if esterification could be a metabolic path for exogenous BRs.

Chemistry
To obtain the new BR analogs conjugated in C-3 (18a, 18b, 19a, and 19b, Figure 2), the synthetic strategy shown in Scheme 1 was developed. The synthesis of the key intermediate alkene 28 has been previously reported [43], but herein, we have introduced some modifications in the synthesis steps to increase the yields of reactions. In addition, more clear spectroscopic evidence ( 1 H-and 13 C-NMR) is provided [43][44][45].
Oxidative decarboxylation of the side chain of compound 22, with the PhI(OAc) 2 / Cu(OAc) 2 system [44,45], leads to olefin 23 in 99.6% yield (yield data were not reported by other authors). In the 1 H-NMR of compound 23 ( Figure S2 [44,45]. Meanwhile, in the 13 C-NMR ( Figure S2, Supplementary Materials), the carbons C-22 and C-23 appear at δ C = 145.06 and 111.69 ppm, respectively. These signals confirm the presence of terminal alkene. The saponification of diacetate 23 with the system K 2 CO 3 /acetone/methanol/reflux leads to diol 24 in 97.1% yield (ref. 98% yield, [43]). Although compound 24 was previously reported, no NMR spectroscopic data were reported [43,45]. So, the observed signals in the 1 H-NMR spectrum ( Figure S3, Supplementary Materials) at δ H = 4.02-3.96 ppm (1H, m) and 3.48-3.42 ppm (1H, m) were assigned to carbinolic hydrogens H-6 and H-3, respectively (Table 1). While in the 13 C-NMR ( Figure S3, Supplementary Materials), the carbons C-6 and C-3 appear at δ C = 67.63 and 71.72 ppm, respectively ( Table 1). The assignments for the H-6 and H-3 signals were confirmed by the 2D HSQC spectrum of compound 24. The subsequent oxidation of compound 24 with the PCC/CH 2 Cl 2 system produces a mixture of three oxidation products (Scheme 1), which were efficiently separated by flash chromatographic column. So, the least polar product was identified as diketone 26 (19.1% yield), a product of intermediate polarity identified as monoketone 25 (2.4% yield), and the most polar product identified as the desired monoketone 27 (40.2% yield). Diketone 26 was previously obtained by the oxidation of glycol with Jones reagent in 95% yield [45]. Meanwhile, diketone 26 and monoketone 27 were obtained by oxidation with the PDC/CH 2 Cl 2 system, with 21% and 61.7% yields, respectively [43]. The IR and 1 H-NMR spectroscopic data for compounds 26 and 27 were consistent with those reported ( Figures S5 and S6, Supplementary Materials) [43,45]. However, none of these previous works reported obtaining monoketone 25. In the 1 H-NMR spectrum of this compound ( Figure S4, Supplementary Materials), the observed signal at δ H = 4.15-4.10 ppm (1H, m) was assigned to carbinolic hydrogen H-6, whereas in the 13 C-NMR spectrum ( Figure S4, Supplementary Materials), the observed signal at δ C = 67.73 ppm corresponds to C-6. Table 1 shows the differences detected for the main signals observed in the 1 H-and 13 C-NMR spectra of compounds 24 to 27. All this information was confirmed by the 2D HSQC correlation spectra of compounds 25-27.
Diketone 26 was conveniently converted to the desired monoketone 27 by selective reduction with NaBH 4 /MeOH [43] at low temperature (0-5 • C) with 76.3% yield (step e, Scheme 1). The spectroscopic data of this compound and 27, which was obtained by direct oxidation from 24 (step d, Scheme 1), were identical.
Then, compound 27 was easily isomerized under acid condition (2.5% v/v HCl/MeOH) to give the derivative 28 possessing 5α-cholestan-6-one skeleton (74.8% yield) [43,44,[46][47][48][49]. The IR, 1 H-and 13 C-NMR spectroscopic data registered for compound 28 were consistent with those reported ( Figure S7, Supplementary Materials) [43,46]. The C-3 benzoylation reactions of 28 were carried out according to the methodology reported for other steroidal nuclei [41,50,51]. So, treatment of 28 with 4-methylbenzoyl chloride/DMAP in CH 2 Cl 2 and pyridine led to 4-metylbenzoate derivative 29 with 87.9%. Similarly, the reaction of 28 with 2-fluorobenzoyl chloride led to 2-fluorobenzoate derivative 30 with 56.0%. The structures of both derivatives were mainly characterized by 1 H and 13 C spectroscopy. For derivative 29, the presence of aromatic signals at δ H = 8.00 ppm (2H, d, J = 9.0 Hz) and 6.93 ppm (2H, d, J = 9.0 Hz) were assigned to the hydrogens HAr-2' and HAr-3 , respectively, whereas the signals appearing at δ C = 163. 36  Recently, the synthesis of glycols C22/C23 in steroids with the shortest side chain of 24-nor-5α-cholane type by a Sharpless dihydroxylation reaction has been reported [42]. The results showed that this type of hydroxylation leads to a mixture of C-22 glycols (R/S) with an approximate 1:1 ratio of both diastereomers [42]. Thus, both olefins 29 and 30 were dihydroxylated following this method and using dihydroquinidine p-chlorobenzoate (DHQD-CLB) as a chiral ligand (Scheme 1) [32,42]. The Sharpless dihydroxylation of derivative 29 produced the 18a/18b diastereoisomer mixture with a total 91.6% yield. The diastereomeric ratio of each glycol in the mixture can be established by the integration of 1 H-NMR signals assigned to the C-21 methyl group, which appear at δ H = 0.921 and 0.953 ppm in 18a and 18b diastereoisomers, respectively. Based on these NMR measurements, the relative ratio of 18a:18b was determined as 1.0:1.0. Subsequently, the diastereoisomers mixture was separated by a semi-preparative HPLC system, allowing obtaining the analogs 18a and 18b.
The structure and stereochemistry at C-22 of compounds 18a and 18b was established by a simple comparison of 1 H-and 13 C-NMR spectra obtained for derivatives 20a and 20b, which were previously reported [41,42]. These comparisons considered chemical shifts (δ), coupling constants (J), and multiplicities of signals corresponding to H-22, H-23a, H-23b, and CH 3 -21 ( 1 H-NMR) and chemical shifts (δ) in 13 C-NMR of both epimers. The main differences in these spectroscopic parameters are listed in Table 2. Similarly, a Sharpless dihydroxylation of derivative 30 produced the 19a/19b diastereoisomers mixture with a total 80.8% yield. The diastereomeric ratio of each glycol in the mixture was 1.0:1.0 (established by the integration of 1 H-NMR signals assigned to the C-21 methyl group, which appear at δ H = 0.917 and 0.954 ppm in 19a and 19b diastereoisomers, respectively). The diastereoisomers mixture was separated by semi-preparative HPLC system, allowing obtaining analogs 19a and 19b. Similar to the above, the main differences in spectroscopic parameters of epimers are listed in Table 3. In summary, four new BRs 24-norcholane type analogs conjugated at the C-3 position with benzoate groups substituted with electron donor and electron-withdrawing groups in the p-position (compounds 18a, 18b, 19a and 19b) have been synthesized and characterized.

Biological
In this work, the activity of new BR 24-norcholane type analogs conjugated at the C-3 position was evaluated using the Rice Lamina Inclination Test. The results of this test were compared with those obtained for other free analogs of 24-norcholane type (analogs 20a and 20b [38]) and with brassinolide. This assay was used because of its specificity and high sensitivity for 1 and their analogs [31,52,53]. The bending angles were measured as the difference between the induced angle produced by treatment with each compound and that found for the negative control. Results obtained for 1, which was used as positive control, and BR analogs 18a, 18b, 19a, 19b, 20a, and 20b are listed in Table 4. Table 4. Comparison between BRs C-3 conjugated 24-norcholane and free 24-norcholane type analogs on lamina inclination of rice seedlings.

Bending Angle between Laminae and Sheaths
(Degrees ± SD) .0 † Data previously obtained and reported in reference [38]. Brassinolide (1) was used as positive control. The negative control only contained sterile distilled water. These values represent the mean ± standard deviation of two independent experiments with at least six replicates each (n = 12). (*) Represents experiments with a significant difference between positive control (1) and analog treatments at p < 0.05 significance level (least square differences (LSD) t-test).
To simplify the data analysis, we will consider the data obtained at 1 × 10 −8 and 1 × 10 −7 M to analyze the correlation between chemical structure and biological activity. The results indicate that at these concentrations, 18b and 19b are the most active in the series of conjugated analogs (19b was the most active at the concentration of 1 × 10 −8 M, whereas 18a was the most active at the concentration of 1 × 10 −7 M) (see Table 3), and they were more active than the free analogs 20a and 20b. Another important effect to consider is related to the configuration on the C-22 carbon of the side chain. Thus, at the concentration of 1 × 10 −8 , analogs 18a and 19a with C-22(R) configuration are less active than analogs 18b and 19b with C-22(S) configuration. However, an opposite effect is observed for analogs 20a and 20b. Similarly, at the concentration of 1 × 10 −7 M, the analogs 19b and 20b with C-22(S) configuration are more active that analogs 19a and 20a with C-22(R) configuration. However, an opposite effect is observed for analogs 18a and 18b. The results observed for the pairs 18a/18b (at 1 × 10 −7 M) and 20a/20b (at 1 × 10 −8 M) would be aligned with those reported for natural occurring BRs with an intact side chain, which indicates that glycol function with C-22(R) and C-23(R) configuration appears essential for a high biological activity and are more active than those with C-22(S) and C-23(S) configuration [3,54]. However, these apparently contradictory structural effects of BRs analogs could be explained in attributed to the shorter side chains. This structural feature could give a greater rotational freedom degree.

Chemistry
All reagents were purchased from commercial suppliers and used without further purification. Melting points were measured on a SMP3 apparatus (Stuart-Scientific, now Merck KGaA, Darmstadt, Germany) and are uncorrected. 1 H-, 13 C-, 13 C-DEPT-135, gs 2D HSQC, and gs 2D HMBC NMR spectra were recorded in CDCl 3 and MeOD solutions, and they are referenced to the residual peaks of CHCl 3 at δ = 7.26 ppm and δ = 77.00 ppm for 1 H and 13 C, respectively and CD 3 OD at δ = 3.30 ppm and δ = 49.00 ppm for 1 H and 13 C, on an Avance 400 Digital NMR spectrometer (Bruker, Rheinstetten, Germany) operating at 400.1 MHz for 1 H and 100.6 MHz for 13 C, and JEOL JNM-ECA 500 NMR spectrometer (JEOL, Tokyo, Japan) operating at 500. 16 MHz for 1 H, 125.77 MHz for 13 C, and 470.62 MHz for 19 F. Chemical shifts are reported in ppm and coupling constants (J) are given in Hz; multiplicities are reported as follows: singlet (s), doublet (d), doublet of doublets (dd), doublet of triplets (dt), triplet (t), quartet (q), multiplet (m), and broad singlet (bs). IR spectra were recorded as KBr disks in a Fourier Transform Infrared (FT-IR) 6700 spectrometer (Nicolet, Thermo Scientific, San Jose, CA, USA) and frequencies are reported in cm −1 . High-resolution mass spectra (HRMS) were recorded in an API HRMS instrument, and the samples were dissolved in chloroform (or chloroform: methanol; 1:1; v/v, in the case of hydroxylated compounds) to a concentration of 10 µg mL −1 . The ASAP (Atmospheric Solids Analysis Probe) was dipped into the sample solution, placed into the ion source, and analyzed in full scan mode. The source of the Synapt G2-Si mass spectrometer (Waters, Manchester, UK) was operated in positive ionization mode (ASAP + ), if not stated otherwise, at a source temperature of 120 • C. The corona needle current was kept at 5 µA and the collision energy was kept at 4 V. The probe temperature was ramped up from 50 to 600 • C in 3 min. Data were acquired from 50 to 1000 Da with 1.0 s scan time in high-resolution mode. The data were processed using the Masslynx 4.1 software (Waters, Milford, MA, USA). A mass accuracy of 1 ppm or less was achieved with the described instrumentation for all compounds. For analytical TLC, silica gel 60 in a 0.25 mm layer was used, and TLC spots were detected by heating after spraying with 10% H 2 SO 4 in H 2 O. Chromatographic separations were carried out by conventional column on silica gel 60 (230-400 mesh) using EtOAc-hexane gradients of increasing polarity. All organic extracts were dried over anhydrous magnesium sulfate and evaporated under reduced pressure, below 40 • C. The HPLC system consisted of a Waters semi-preparative HPLC system including a quaternary pump, a liquid handler, and UV-Vis and Evaporative Light Scattering Detector (ELSD) detectors. The semi-preparative column was filled with silica gel. 3α-Hydroxy-24-nor-5α-chol-22-en-6-one (28) Compound 27 (6.08 g, 17.65 mmol) was dissolved in 100 mL of 2.5% v/v HCl-MeOH at room temperature and constant agitation for 48 h. The end of the reaction was verified by TLC. The solvent was evaporated under reduced pressure, and the crude was re-dissolved in 60 mL of EtOAc. The organic layer was washed with saturated solution of NaHCO 3 (2 × 15 mL) and water (2 × 30 mL), dried over MgSO 4 , and filtered. The solvent was evaporated under reduced pressure. The crude was re-dissolved in CH 2 Cl 2 (5 mL) and chromatographed on silica gel with PE/EtOAc mixtures of increasing polarity 6-Oxo-24-nor-5α-chol-22-en-3α-yl 4-methylbenzoate (29) A solution of compound 28 (150 mg, 0.44 mmol) and DMAP (44 mg, 0.36 mmol) was prepared in 3 mL of anhydrous pyridine. To this solution, 4-methylbenzoyl chloride 174 µL (1.32 mmol) was added by slow dripping, and the reaction was maintained at room temperature with constant stirring for 2 h. The end of the reaction was verified by TLC. After completion of the reaction, 3 mL of hot water was added. After an additional 20 min of stirring, the mixture was extracted with EtOAc (20 mL) and washed successively with saturated NaHCO 3 solution (2 × 10 mL) and water (2 × 10 mL), dried over anhydrous MgSO 4 , and filtered. The solvent was evaporated under reduced pressure, and the crude was redissolved in CH 2 Cl 2 (5 mL) and chromatographed on silica gel with EtOAc/cyclohexane A solution of compound 29 (120 mg, 0.348 mmol) and DMAP (35 mg, 0.286 mmol) was prepared in 3 mL of anhydrous pyridine. To this solution, 2-fluorobenzoyl chloride 123.5 µL (1.05 mmol) was added by slow dripping, and the reaction was maintained at room temperature with constant stirring for 3 h. The end of the reaction was verified by TLC. After completion of the reaction, 3 mL of hot water was added. After an additional 20 min of stirring, the mixture was extracted with EtOAc (20 mL) and washed successively with saturated NaHCO 3 solution (2 × 10 mL) and water (2 × 10 mL), dried over anhydrous MgSO 4 , and filtered. The solvent was evaporated under reduced pressure, and the crude was re-dissolved in CH 2 Cl 2 (5 mL) and chromatographed on silica gel with EtOAc/cyclohexane (1.0:19) mixture. Compound 30 (91 mg, 56.0% yield) was obtained as a colorless solid, m.p. = 120.6-120. 562 mmol in 20 mL of t-butanol) were added, and the mixture reaction was stirred at room temperature for 5 h. The end of the reaction was verified by TLC; then, H 2 O (10 mL) and a saturated solution of Na 2 S 2 O 3 . 5H 2 O (2 mL) were added. The mixture was stirred for another 20 min. Later, it was extracted with EtOAc (2 × 35 mL) and washed with water (2 × 35 mL), and both organic phases were combined, dried over MgSO 4 , and filtered. The solvent was evaporated under reduced pressure. The crude was re-dissolved in CH 2 Cl 2 (1.0 mL) and chromatographed on silica gel with an EtOAc/cyclohexane (16:4) mixture. A mixture of 19a/19b = 1.0/1.0 was obtained (65 mg: 80.8% yield). The separation by HPLC of an analytical sample allowed the separation and obtaining of the pure compounds 19a and 19b.

Rice Lamina Inclination Test (RLIT)
The biological activity of the growth of the compounds was evaluated by the rice lamina inclination test [55,56], according to a previously described procedure [38], and using the same a Zafiro cultivar (Oryza sativa) provided by the Institute of Agricultural Research (INIA-Quilamapu-Chile) as previous studies.
The seeds were sown and cultivated until the seedlings presenting the second internode of the rice blade were selected for cutting. Six segments per treatment were incubated in Petri dishes containing 60 mL of distilled water, and the amount of test compound (BRs  analogs 18a, 18b, 19a, 19b, 20a, and 20b and positive control (1)) needed to reach final concentrations equal to 1 × 10 −8 M; 1 × 10 −7 M; and 1 × 10 −6 M. The negative control only contained sterile distilled water. All treatments were incubated by 48 h at 25 • C in darkness, and the angles developed between the blade and the sheath were measured. Each experiment was performed by duplicate.
Results were expressed as mean ± standard deviation (SD) using twelve angle measurements. Statistical analysis was done using a statistical package Excel by applying mean values using one-way ANOVA with the post-hoc least square differences (LSD) test to determine if there was a significant difference between the positive control and the treatments. A P value of less than 0.05 was considered significant.

Conclusions
Brassinosteroid 24-norcholane type analogs conjugated at C-3 and configurations S and R on the C-22 carbon of the side chain have been synthesized and characterized. The synthesis uses hyodeoxycholic acid as the starting material, and epimers with different configuration at C-22 are obtained. These epimers have been separated, and their growthpromoting activity was measured using RLIT. The results show that the esterification of BRs analog at C-3 has no effect on the biological activity of synthetic analogs. This suggest that reducing activity by esterification at C-3 requires a long chain carboxylic acid. In addition, the presence of a hydroxyl group at C-3 is not an essential structural feature for activity. This result confirms previous SAR where it has been proposed that activity is not determined by the presence or absence of specific groups in the BR structure.

Conflicts of Interest:
The authors declare no conflict of interest.