Synthesis and Biological Activity of Brassinosteroid Analogues with a Nitrogen-Containing Side Chain

Brassinosteroids are a class of plant hormones that regulate a broad range of physiological processes such as plant growth, development and immunity, including the suppression of biotic and abiotic stresses. In this paper, we report the synthesis of new brassinosteroid analogues with a nitrogen-containing side chain and their biological activity on Arabidopis thaliana. Based on molecular docking experiments, two groups of brassinosteroid analogues were prepared with short and long side chains in order to study the impact of side chain length on plants. The derivatives with a short side chain were prepared with amide, amine and ammonium functional groups. The derivatives with a long side chain were synthesized using amide and ammonium functional groups. A total of 25 new brassinosteroid analogues were prepared. All 25 compounds were tested in an Arabidopsis root sensitivity bioassay and cytotoxicity screening. The synthesized substances showed no significant inhibitory activity compared to natural 24-epibrassinolide. In contrast, in low concentration, several compounds (8a, 8b, 8e, 16e, 22a and 22e) showed interesting growth-promoting activity. The cytotoxicity assay showed no toxicity of the prepared compounds on cancer and normal cell lines.


Introduction
Brassinosteroids (BRs, Figure 1) are a group of plant steroid hormones that regulate many important aspects of plant growth and development, such as cell division, elongation and differentiation, pollen tube growth, seed germination, regulation of gene expression, enzyme activation and photosynthesis [1][2][3]. They also induce tolerance against a wide range of biotic and abiotic stresses, such as water or drought, temperature, oxidative stresses, high salinity and different environmental pollutants [4,5]. At the molecular level, BRs modify the pathway of enzymes and change the gene expression when plants are exposed to stress. Plants perceive brassinosteroids at the cell membrane, using the membrane-integral receptor kinase brassinosteroid insensitive 1 (BRI1) and its co-receptor BAK1 [6][7][8][9]. the encoded protein, BRI1, is a member of a large family of plant LRR (leucine-rich repeat) receptor-like kinases, characterized by an extracellular LRR domain, a single-pass transmembrane segment and a cytoplasmic kinase domain. BRI1 has been established as an authentic brassinosteroid receptor through genetic and biochemical investigations [10]. Moreover, recent studies have shown that natural BRs and their synthetic analogues have potential application not only in agriculture, but also in medicine due to their antiviral [11,12], immunomodulatory and neuroprotective activities [13][14][15], and anti-proliferative effects on animal cells in vitro [16][17][18][19][20].
ol. Sci. 2021, 22, x FOR PEER REVIEW 2 of 28 their antiviral [11,12], immunomodulatory and neuroprotective activities [13][14][15], and anti-proliferative effects on animal cells in vitro [16][17][18][19][20]. Since the first BR isolation and structural identification, chemists began to synthesize new brassinosteroid analogues with different modifications of the sterol skeleton and tested them on various biological activities, especially in plant bioassays. Modification of the side chain is one of the main aims of synthesis. Such modifications include shortening or prolonging the side chain, oxidation or reduction of oxygen functional groups, incorporation of heteroatoms other than oxygen (mostly nitrogen and halogens) and modification of newly synthesized functional groups (e.g., acetylation, methylation, etc.) [21][22][23][24][25][26][27][28]. Preparation of derivatives with a nitrogen-containing side chain has not been at the forefront of the interest of steroidal chemists. Few examples have been published so far. Back et al. described the synthesis and biological activities of 25-azabrassinolide. Unfortunately, the rice lamina inclination biotest showed no biological activity [29]. Several BR analogues with nitrogen-containing heterocycles (e.g., isoxazoline, oxazole, thiazole and pyridine) were synthesized by Wendeborn's group [25]. Other analogues with nitrogen groups in the side chain were prepared mostly for other chemical modifications, e.g., azide derivatives for photoaffinity labeling [30] or activated esters for conjugation reactions [31,32].
The aim of this study was to synthesize new BRs analogues with nitrogen functional groups (amides, amines and ammonium salts) in the side chain and study their biological properties. The synthesis was also based on molecular docking. It uses computational methods for prediction of interaction between receptors and ligands. It is a tool for predicting the first results during the preparation of new ligands. In our study, we were interested mostly in tertiary amides and amines to avoid the presence of hydrogen directly attached to nitrogen (with the exception of hydrochloride salts of tertiary amines whose existence is pH-dependent) and because they have an additional hydrogen bond acceptor than standard BRs hydroxy groups. The plant growth-promoting activity of synthetic analogues was assayed using the Arabidopsis root sensitivity bioassay. Further, the cytotoxic activity of all compounds was tested on one normal and three cancer cell lines.

Chemistry
The synthesis of all nitrogen-containing compounds was divided into three groupsshort chain amides (Scheme 1), short chain amines and ammonium chlorides (Scheme 2) and long chain derivatives (Scheme 3). The synthesis of amides with a short side chain started with the known tosylate 1 [33]. Other reaction steps correspond to standard brassinosteroid synthesis [3]. Firstly, hydroxy ester 2 is formed from tosylate, followed by oxidation with Jones reagent to oxo-ester 3. This ketone undergoes rearrangement to olefin 4. Hydrolysis of the ester group afforded carboxylic acid 5 that is suitable for amide formation. Five different secondary amines were used for the amidation reaction-dimethylamine, diethylamine, pyrrolidine, piperidine and morpholine. In all five cases, olefinic amides 7a-e were obtained in very good yield. However, activated ester 5 can also be Since the first BR isolation and structural identification, chemists began to synthesize new brassinosteroid analogues with different modifications of the sterol skeleton and tested them on various biological activities, especially in plant bioassays. Modification of the side chain is one of the main aims of synthesis. Such modifications include shortening or prolonging the side chain, oxidation or reduction of oxygen functional groups, incorporation of heteroatoms other than oxygen (mostly nitrogen and halogens) and modification of newly synthesized functional groups (e.g., acetylation, methylation, etc.) [21][22][23][24][25][26][27][28]. Preparation of derivatives with a nitrogen-containing side chain has not been at the forefront of the interest of steroidal chemists. Few examples have been published so far. Back et al. described the synthesis and biological activities of 25-azabrassinolide. Unfortunately, the rice lamina inclination biotest showed no biological activity [29]. Several BR analogues with nitrogencontaining heterocycles (e.g., isoxazoline, oxazole, thiazole and pyridine) were synthesized by Wendeborn's group [25]. Other analogues with nitrogen groups in the side chain were prepared mostly for other chemical modifications, e.g., azide derivatives for photoaffinity labeling [30] or activated esters for conjugation reactions [31,32].
the aim of this study was to synthesize new BRs analogues with nitrogen functional groups (amides, amines and ammonium salts) in the side chain and study their biological properties. the synthesis was also based on molecular docking. It uses computational methods for prediction of interaction between receptors and ligands. It is a tool for predicting the first results during the preparation of new ligands. In our study, we were interested mostly in tertiary amides and amines to avoid the presence of hydrogen directly attached to nitrogen (with the exception of hydrochloride salts of tertiary amines whose existence is pH-dependent) and because they have an additional hydrogen bond acceptor than standard BRs hydroxy groups. the plant growth-promoting activity of synthetic analogues was assayed using the Arabidopsis root sensitivity bioassay. Further, the cytotoxic activity of all compounds was tested on one normal and three cancer cell lines.

Chemistry
the synthesis of all nitrogen-containing compounds was divided into three groupsshort chain amides (Scheme 1), short chain amines and ammonium chlorides (Scheme 2) and long chain derivatives (Scheme 3). the synthesis of amides with a short side chain started with the known tosylate 1 [33]. Other reaction steps correspond to standard brassinosteroid synthesis [3]. Firstly, hydroxy ester 2 is formed from tosylate, followed by oxidation with Jones reagent to oxo-ester 3. This ketone undergoes rearrangement to olefin 4. Hydrolysis of the ester group afforded carboxylic acid 5 that is suitable for amide formation. Five different secondary amines were used for the amidation reactiondimethylamine, diethylamine, pyrrolidine, piperidine and morpholine. In all five cases, olefinic amides 7a-e were obtained in very good yield. However, activated ester 5 can also be isolated after 4-5 h of reaction. As the last reaction, we used Upjohn dihydroxylation to prepare five dihydroxy amides, 8a-e. Due to the reduction step in the synthesis of amine analogues, the previous reaction strategy cannot be used here. It is necessary to protect both the hydroxy groups and ketone. Thus, we started with the known dihydroxy ester 9 [34]. It was firstly protected on hydroxy groups with acetone (compound 10), followed by protection of the ketone with ethylene glycol (compound 11). Fully protected ester was then hydrolyzed to acid 12 which is used for amide formation. Amides 13a-e were then subjected to reduction with lithium aluminum hydride to obtain amines 14a-e. Subsequent deprotection with basic work-up afforded dihydroxy amines 15a-e, whose treatment with hydrochloric acid formed ammonium hydrochloride 16a-e. Scheme  For the preparation of long side chain analogues, we used the known fully protected ester 17 [35]. Hydrolysis of the ester and condensation of acid 18 with amines afforded five fully protected amides, 19a-e. Firstly, all amides were deprotected to form tetrahydroxy amides 20a-e. Another portion of amides was used in lithium aluminum hydride reduction to prepare amines 21a-e. After their acidic deprotection, we obtained ammonium chlorides 22a-e. the deprotection on the side chain must be carried out at 45 • C. Lower temperature (below 30 • C) led only to partial deprotection on the A-ring and isolation of amide 23 and ammonium chloride 24 with the still protected side chain.

Molecular Docking
Molecular docking is a useful tool to understand the pose and energetics of a proteinligand complex. the binding site of BRI1 is located on the surface of the receptor ectodomain as a nonpolar cavity lined by nonpolar amino acids. Brassinolide fits into the cavity via its nonpolar side and displays its hydroxyl groups towards the solvent and protein partners [26]. Due to this reason, we used only tertiary amides and amines. Especially in the case of long side chain derivatives, the presence of a new hydrogen bond donor can affect their pose in the protein cavity. Molecular docking predicted similar or better binding energies than for brassinolide for compounds 8c, 8d, 8e, 15d, 20c, 22c and 22d (see Table S1). This implies that derivatives with nitrogen groups in the side chain should bind within the BRI1 cavity at least as easily as brassinolide itself. Based on molecular docking, these were also candidates for showing similar binding experimentally. For poses of all nitrogen-containing derivatives used for the plant biotest, see Figures S1-S20 in the Supplementary Material.

Biology
Several bioassays such as the first bean internode, root growth and rice lamina inclination have been developed to evaluate the growth-promoting activity of BR derivatives [36]. In this work, the activity of new BR analogues was evaluated using the Arabidopsis root sensitivity test because of its simplicity and high sensitivity for BRs. [27,28]. the characteristic effect of exogenously applied BRs on light-grown Arabidopsis is dose-dependent-BRs in higher concentrations cause inhibition of root growth; vice versa, low concentrations are stimulating [37][38][39]. the results obtained for 24-epiBL, which was used as a positive control, and the BR derivatives are shown in Figure 2, Figure 3 and Figure S21. the most active BR (24-epiBL) inhibited Arabidopsis root growth in 1 nM and higher concentrations. Most of the newly synthesized BR derivatives were not active, and only the long side chain derivatives 20a and 20c show a weak inhibitory effect in the concentration range, which was not confirmed using the higher tested concentration of 100 nM ( Figure 3 and Figure S21). Compound 22b shows the same effect only at low concentrations; on the contrary, compound 8a shows weak inhibition up to the highest tested concentration of 100 nM (see Figure S21). On the other hand, in low concentrations (0.1 and 1 nM), a slight elongation of roots was observed for derivatives with a short side chain (8a, 8b, 8e and 16e) compared to the control. A similar trend was shown for dimethylammonium hydrochloride 22a and morpholinium hydrochloride 22e with a long side chain. For numerical values of the average, SD and p-value, see Table S2.
the structure-activity relationship of brassinosteroids has been studied in detail in recent years and it has been postulated that the (2α,3α)-and (22R,23R)-vicinal diol moieties are required for optimum bioactivity and 7-oxalactone BRs have stronger biological activity than 6-oxo types, whereas B-ring non-oxidized BRs reveal no activity in biological tests [25,40,41]. Recently, several research groups prepared BR synthetic analogues with various nitrogen modifications in the side chain. Wendeborn and co-workers [25] synthesized BR derivatives where the isoxazoline ring was introduced to replace the (22,23)vicinal diols and showed no statistically significant activity in the bean second internode elongation bioassay. As well as 25-azabrassinolide, it proved to be completely inactive at all doses studied in the rice leaf lamina inclination assay [29]. the low biological activity of BR analogues with the nitrogen-containing side chain is probably due to the lower BRI-1 receptor affinity.
the effect of the prepared BRs derivatives on the viability (in Calcein AM assays) of BJ human fibroblasts (as an example of a "normal" cell line) and human cancer cell lines of various histopathological origins, including: T-lymphoblastic leukemia CEM, breast carcinoma MCF7 and cervical carcinoma cell line HeLa, was also studied. Cells of all these lines were exposed to six 3-fold dilutions of each drug for 72 h prior to determination of cell survival. the IC 50 (concentration leading to 50% inhibition of viability) values obtained from the Calcein AM cytotoxicity assay were calculated. 28-Homocastasterone was used as a positive control, which is the most potent natural brassinosteroid towards CEM cells (IC 50 13 µM [42]). All tested BR nitrogen-containing derivatives had no detectable activity, even when tested in concentrations of up to 50 µM. No BRs derivative-mediated loss of viability was observed in the BJ fibroblasts (see Table S3). Seven-day-old Arabidopsis thaliana seedlings (Columbia ecotype, Col-0) were treated with DMSO as control/24epiBL/BR short side chain derivatives; (A)-amides, (B)-amines, (C)-ammonium hydrochlorides. For each treatment, more than 15 seedlings were analyzed in two biological repeats. Error bars represent SD. Asterisks represent significant changes (t-test); * represents p-value < 0.05, in comparison to control.  For each treatment, more than 15 seedlings were analyzed in two biological repeats. Error bars represent SD. Asterisks represent significant changes (t-test); * represents p-value < 0.05, in comparison to control.

Materials and Instruments
the melting points (Mp) were determined on a Stuart SMP30 instrument (Bibby Scientific Ltd., Staffordshire, UK). Elemental analyses were performed using an EA 1108 elemental analyzer (Fison Instruments, Glasgow, UK); the values (C, H, N) agreed with the calculated values within acceptable limits. the NMR spectra were taken on a JEOL JNM-ECA 500 (JEOL, Tokyo, Japan; 1 H, 500 MHz; 13 C, 125 MHz) spectrometer equipped with a 5 mm JEOL Royal probe. 1 H NMR and 13 C NMR chemical shifts (δ) were calibrated using tetramethylsilane (TMS, 1 H δ = 0 ppm) or solvents: CDCl 3 ( 1 H δ = 7.26 ppm, 13 C δ = 77.00 ppm), CD 3 OD ( 1 H δ = 3.31 ppm, 13 C δ = 49.00 ppm), DMSO-d 6 ( 1 H δ = 2.46 ppm, 13 C δ = 40.00 ppm) or D 2 O ( 1 H δ = 4.79 ppm). Chemical shifts are given in ppm (δ-scale) and coupling constants (J) are given in Hz. All values were obtained by first-order analysis. the most common abbreviations for NMR signals were used: s-singlet, d-doublet, t-triplet, q-quartet, m-multiplet, b-broad (and combinations thereof, e.g., dd-doublet of doublet, bs-broad singlet, etc.). For API HRMS analysis, the samples were dissolved in chloroform (or chlo-roform/methanol; 1:1; v/v, in the case of hydroxylated compounds) to a concentration 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+) and, 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 a value of 4. 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 and processed using the Masslynx 4.1 software (Waters, Manchester, UK). Mass accuracy of 1 ppm or less was achieved with the described instrumentation for all compounds. Merck silica gel Kieselgel 60 (230-400 mesh) was used for column chromatography. Reagents and solvents were purchased from Sigma-Aldrich (St. Louis, MO, USA) and were not purified.

Chemistry
Ethyl (20S)-6β-hydroxy-3α,5-cyclo-5α-pregnane-20-carboxylate (2) Potassium acetate (5.4 g; 55 mmol) was added to a solution of tosylate 1 (3.0 g; 5.54 mmol) in acetone (140 mL) and water (40 mL). the reaction was stirred under reflux for 7 h. Then, excess of acetone was removed under reduced pressure and the residual suspension was dissolved in diethyl ether and extracted twice with water. Organic solvents were dried with sodium sulfate and removed under reduced pressure. the crude product was purified on silica gel (mobile phase-30% ethyl acetate in cyclohexane) to afford hydroxyl ester 2 (2.1 g; 97%) as a white solid.
Mp 118-120 • C. 1  (20S)-6-oxo-5α-pregn-2-ene-20-carboxylic acid (5) Potassium hydroxide (1.0 g; 18 mmol) was added to a solution of ester 4 (1.4 g; 3.63 mmol) in ethanol (100 mL) and water (5 mL). the reaction mixture was stirred under reflux for 8 h. the excess ethanol was then removed under reduced pressure. the suspension was diluted with ethyl acetate and extracted with 5% aqueous hydrochloric acid and twice with water. Organic solvents were dried with sodium sulfate and evaporated under reduced pressure. the crude product was purified on silica gel (mobile phase-50-80% ethyl acetate in cyclohexane) to afford acid 5 (1.18 g; 91%) as a white solid. All data correspond to the data published [34].

Conclusions
In this study, 25 brassinosteroid analogues with a nitrogen-containing side chain were synthesized based on molecular docking and tested using an Arabidopsis sensitivity bioassay and Calcein AM cytotoxicity assay. New analogues were divided into two groups-short side chain BRs and long side chain BRs. As a nitrogen-containing functional group, tertiary amides, tertiary amines and ammonium chlorides were prepared. the synthesized substances showed no significant inhibitory activity compared to natural 24-epibrassinolide. In contrast, several compounds showed interesting growth-promoting activity. the cytotoxicity assay showed no toxicity of the prepared compounds on cancer or normal cell lines. Unfortunately, the results of the molecular docking, which proved the good binding affinity of the tested compounds to the BRI1 receptor, were not confirmed by plant experiments. However, the very low or no biological activity of nitrogen derivatives can also be caused by different reasons other than low affinity to the receptor, e.g., by an unknown mechanism that prevents transport of these nitrogen compounds to the receptor. Based on these observations, it is possible to state that the presence of an amide or amine in the side chain of BRs analogues significantly reduces the biological activity in plants.