Scalable Synthesis and Cancer Cell Cytotoxicity of Rooperol and Analogues

Plant polyphenols, such as the African potato (Hypoxis hemerocallidea)-derived bis-catechol rooperol, can display promising anticancer activity yet suffer from rapid metabolism. Embarking upon a program to systematically examine potentially more metabolically stable replacements for the catechol rings in rooperol, we report here a general, scalable synthesis of rooperol and analogues that builds on our previous synthetic approach incorporating a key Pd-catalyzed decarboxylative coupling strategy. Using this approach, we have prepared and evaluated the cancer cell cytotoxicity of rooperol and a series of analogues. While none of the analogues examined here were superior to rooperol in preventing the growth of cancer cells, analogues containing phenol or methylenedioxyphenyl replacements for one or both catechol rings were nearly as effective as rooperol.


Introduction
African potato (Hypoxis hemerocallidea) is a widely used medicinal plant in southern Africa [1]. Ethanolic extracts of the corms of H. hemerocallidea contain 10-15% of hypoxoside, a bis-glucoside of the aglycone rooperol, 1 (Figure 1) [1]. Rooperol has demonstrated a cytotoxicity against cancer cell lines [2][3][4], and an orally administered extract of H. hemercallidea has been the subject of Phase I clinical trial in advanced lung cancer patients [5,6]. Interestingly, in this Phase I trial, no dose-limiting toxicity was identified, and 5 of the 24 patients enrolled showed some signs of response, including one complete response (>5 years) [6]. Despite this hint of clinical promise, pharmacokinetic studies demonstrated the very rapid metabolism of hypoxoside and rooperol, with only Phase II metabolites of rooperol detected in the blood of these patients [5].
Rooperol is one of a number of plant polyphenols that have demonstrated promise as anticancer agents [7][8][9]. However, a common issue with these anticancer polyphenols is metabolic instability [10]. Our work has focused on the design and evaluation of analogues of biologically interesting plant polyphenols with the goal of improved metabolic stability [11,12]. In the present case, our efforts focused on rooperol analogues with more metabolically stable replacements for this natural product's catechol moieties. A number of synthetic studies of rooperol and analogues have been reported. The first total synthesis by Drewes and co-workers employed a problematic Csp-Csp3 coupling step [13]. We subsequently reported a very concise synthesis of rooperol and analogues, which incorporated two key strategies: a Friedel-Crafts reaction involving tetrachlorocyclopropene and electron-rich aromatics and a Pd-catalyzed decarboxylative coupling of the resulting 3-arylprop-2-ene-1-yl arylpropiolates 2 ( Figure 1) [14]. Although this route could afford rooperol analogues in as few as three steps, it suffers from two limitations. First, the route shown in Figure 1 lacks generality in that only very electron-rich aromatic compounds participate in the initial Friedel-Crafts reaction, and only certain arylprop-2-ene-1-ols directly afford the esters 2 upon alcoholysis of the initially formed Friedel-Crafts product. Second, the Friedel-Crafts reaction did not proceed well above 1 mmole scale, limiting the amount of material that could be prepared for biological assay. Thus, in order to continue our efforts to prepare and evaluate metabolically stable rooperol analogues, we first had to redesign this synthesis. Here, we report a more general, scalable synthesis of rooperol and analogs that retain the key Pd-catalyzed decarboxylative coupling from our earlier work. Using this improved synthesis, we prepared a number of rooperol analogs and report here their activity against HeLa cancer cells.

Results
We set out to evaluate a variety of potential replacements for the catechol moieties of rooperol. As shown in Figure 2, these replacements included variations in the number (B) and placement (C) of the catechol hydroxyl groups and substitution of these groups with fluorine (D) or methylenedioxy groups (E). In addition, we also explored the 4Hbenzo[d] [1,3]dioxine group (F) as a replacement for the catechol groups in rooperol. Given the limitations of our previous route to rooperol and analogues (Figure 1), the preparation of the analogues contemplated in Figure 2 required an alternative synthetic scheme. Our revised synthesis of protected rooperol and analogs is shown in Scheme 1. A number of synthetic studies of rooperol and analogues have been reported. The first total synthesis by Drewes and co-workers employed a problematic Csp-Csp3 coupling step [13]. We subsequently reported a very concise synthesis of rooperol and analogues, which incorporated two key strategies: a Friedel-Crafts reaction involving tetrachlorocyclopropene and electron-rich aromatics and a Pd-catalyzed decarboxylative coupling of the resulting 3-arylprop-2-ene-1-yl arylpropiolates 2 ( Figure 1) [14]. Although this route could afford rooperol analogues in as few as three steps, it suffers from two limitations. First, the route shown in Figure 1 lacks generality in that only very electron-rich aromatic compounds participate in the initial Friedel-Crafts reaction, and only certain arylprop-2ene-1-ols directly afford the esters 2 upon alcoholysis of the initially formed Friedel-Crafts product. Second, the Friedel-Crafts reaction did not proceed well above 1 mmole scale, limiting the amount of material that could be prepared for biological assay. Thus, in order to continue our efforts to prepare and evaluate metabolically stable rooperol analogues, we first had to redesign this synthesis. Here, we report a more general, scalable synthesis of rooperol and analogs that retain the key Pd-catalyzed decarboxylative coupling from our earlier work. Using this improved synthesis, we prepared a number of rooperol analogs and report here their activity against HeLa cancer cells.

Results
We set out to evaluate a variety of potential replacements for the catechol moieties of rooperol. As shown in Figure 2, these replacements included variations in the number (B) and placement (C) of the catechol hydroxyl groups and substitution of these groups with fluorine (D) or methylenedioxy groups (E). In addition, we also explored the 4Hbenzo[d] [1,3]dioxine group (F) as a replacement for the catechol groups in rooperol. A number of synthetic studies of rooperol and analogues have been reported. The first total synthesis by Drewes and co-workers employed a problematic Csp-Csp3 coupling step [13]. We subsequently reported a very concise synthesis of rooperol and analogues, which incorporated two key strategies: a Friedel-Crafts reaction involving tetrachlorocyclopropene and electron-rich aromatics and a Pd-catalyzed decarboxylative coupling of the resulting 3-arylprop-2-ene-1-yl arylpropiolates 2 ( Figure 1) [14]. Although this route could afford rooperol analogues in as few as three steps, it suffers from two limitations. First, the route shown in Figure 1 lacks generality in that only very electron-rich aromatic compounds participate in the initial Friedel-Crafts reaction, and only certain arylprop-2-ene-1-ols directly afford the esters 2 upon alcoholysis of the initially formed Friedel-Crafts product. Second, the Friedel-Crafts reaction did not proceed well above 1 mmole scale, limiting the amount of material that could be prepared for biological assay. Thus, in order to continue our efforts to prepare and evaluate metabolically stable rooperol analogues, we first had to redesign this synthesis. Here, we report a more general, scalable synthesis of rooperol and analogs that retain the key Pd-catalyzed decarboxylative coupling from our earlier work. Using this improved synthesis, we prepared a number of rooperol analogs and report here their activity against HeLa cancer cells.

Results
We set out to evaluate a variety of potential replacements for the catechol moieties of rooperol. As shown in Figure 2, these replacements included variations in the number (B) and placement (C) of the catechol hydroxyl groups and substitution of these groups with fluorine (D) or methylenedioxy groups (E). In addition, we also explored the 4Hbenzo[d] [1,3]dioxine group (F) as a replacement for the catechol groups in rooperol. Given the limitations of our previous route to rooperol and analogues (Figure 1), the preparation of the analogues contemplated in Figure 2 required an alternative synthetic scheme. Our revised synthesis of protected rooperol and analogs is shown in Scheme 1. Given the limitations of our previous route to rooperol and analogues (Figure 1), the preparation of the analogues contemplated in Figure 2 required an alternative synthetic scheme. Our revised synthesis of protected rooperol and analogs is shown in Scheme 1. We prepared arylprop-2-ene-1-ols 5 from the corresponding aldehydes 8 via reduction of the methyl esters 4, obtained via Wittig reaction. Separately, the same aldehydes 8 were subjected to Corey-Fuchs alkynylation [15] via the vinyl dibromides 6, which were subjected to elimination with nBuLi, followed by trapping of the resulting alkynyl anions with CO 2 to afford the arylpropiolic acids 7. All of these transformations occurred without incident to afford good yields of the products 4A-F, 5A-F, 6A-E, and 7A-E with the exception of the nBuLi elimination/trapping of the vinyl dibromide 6D derived from 3,4-difluorobenzaldehyde. This reaction proved somewhat capricious, and at best, only modest yields of the corresponding acid 7D were obtained. We prepared arylprop-2-ene-1-ols 5 from the corresponding aldehydes 8 via reduction of the methyl esters 4, obtained via Wittig reaction. Separately, the same aldehydes 8 were subjected to Corey-Fuchs alkynylation [15] via the vinyl dibromides 6, which were subjected to elimination with nBuLi, followed by trapping of the resulting alkynyl anions with CO2 to afford the arylpropiolic acids 7. All of these transformations occurred without incident to afford good yields of the products 4A-F, 5A-F, 6A-E, and 7A-E with the exception of the nBuLi elimination/trapping of the vinyl dibromide 6D derived from 3,4difluorobenzaldehyde. This reaction proved somewhat capricious, and at best, only modest yields of the corresponding acid 7D were obtained. Scheme 1. Scalable and general synthesis of protected rooperol (1′AA) and analogues.
The preparation of the carboxylic acid 7F followed a different route, as shown in Scheme 2. The previously reported alkyne 9F [16] was deprotonated with nBuLi, and the resulting anion was trapped with CO2 to afford 7F in good yield. While the aldehydes 8D and 8E are commercially available, the tert-butyldimethylsilyl-protected aldehydes 8A-C were prepared by reacting the commercially available hydroxy-substituted benzaldehydes 10A-C with tert-butyldimethylsilyl chloride in the presence of imidazole (Scheme 2). The previously reported aldehyde 8F was prepared from the corresponding bromide 11F following literature precedent [16] (Scheme 2). Scheme 1. Scalable and general synthesis of protected rooperol (1 AA) and analogues.
The preparation of the carboxylic acid 7F followed a different route, as shown in Scheme 2. The previously reported alkyne 9F [16] was deprotonated with nBuLi, and the resulting anion was trapped with CO 2 to afford 7F in good yield. While the aldehydes 8D and 8E are commercially available, the tert-butyldimethylsilyl-protected aldehydes 8A-C were prepared by reacting the commercially available hydroxy-substituted benzaldehydes 10A-C with tert-butyldimethylsilyl chloride in the presence of imidazole (Scheme 2). The previously reported aldehyde 8F was prepared from the corresponding bromide 11F following literature precedent [16] (Scheme 2).
We prepared a number of symmetrical esters 2 in which the catechol ring or replacement was the same on both the alcohol and acid moieties via DCC coupling of the acids 7 and alcohols 5 (Scheme 1). Each of these esters was then subjected to Pd-catalyzed decarboxylative coupling to afford protected rooperol 1′AA and the protected symmetrical rooperol derivatives 1′BB and 1′CC as well as the rooperol analogues 1DD, 1EE, and 1FF (Scheme 1).
In addition to these symmetrical rooperol analogues, two analogues with two different replacements for the catechol moieties were also prepared (Scheme 3). DCC coupling of the alcohol 4E with phenylpropiolic acid afforded the ester 2EG, while coupling with the acid 7A afforded the ester 2EA. Each of these was subjected to Pd-catalyzed decarboxylative coupling to afford the rooperol analog 1EG and the protected rooperol analog 1′EA (Scheme 3).
We prepared a number of symmetrical esters 2 in which the catechol ring or replacement was the same on both the alcohol and acid moieties via DCC coupling of the acids 7 and alcohols 5 (Scheme 1). Each of these esters was then subjected to Pd-catalyzed decarboxylative coupling to afford protected rooperol 1 AA and the protected symmetrical rooperol derivatives 1 BB and 1 CC as well as the rooperol analogues 1DD, 1EE, and 1FF (Scheme 1).
In addition to these symmetrical rooperol analogues, two analogues with two different replacements for the catechol moieties were also prepared (Scheme 3). DCC coupling of the alcohol 4E with phenylpropiolic acid afforded the ester 2EG, while coupling with the acid 7A afforded the ester 2EA. Each of these was subjected to Pd-catalyzed decarboxylative coupling to afford the rooperol analog 1EG and the protected rooperol analog 1 EA (Scheme 3).
The protected rooperol analogues 1 were deprotected by two different methods. For the analogs containing a catechol group, we used our previously described silyl deprotection method [14] using HBr and KF, which afforded rooperol (1AA) and analogue 1EA (Scheme 3). The other analogues were prepared by AcOH-buffered TBAF deprotection to afford 1BB and 1CC (Scheme 3).
With these rooperol analogues in hand, we evaluated the utility of the various catechol replacements by determining the cancer cell cytotoxicity of these compounds versus rooperol using a MTT assay (Table 1). Rooperol displays cytotoxicity against all three cell lines examined: HeLa (cervical adenocarcinoma), H460 (lung carcinoma), and A549 (lung carcinoma). All of the rooperol analogues examined here were also tested against HeLa cells, which were slightly more sensitive to rooperol compared to the other cell lines. All of the analogues, with the exception of 1DD, show some activity against HeLa cells, with analogues 1CC and 1EA displaying activity close to that of rooperol. Interestingly, while the symmetrical analogue 1EE lacks good activity against HeLa cells, the two asymmetrical analogues containing the same methylenedioxyphenyl catechol replacement, 1EA and 1EB, show better cytotoxicity. It is also interesting that 1EE has nearly the same activity as rooperol against A549 cells, while 1DD is much less active against both A549 and H460 cells compared to rooperol. The protected rooperol analogues 1′ were deprotected by two different methods. For the analogs containing a catechol group, we used our previously described silyl deprotection method [14] using HBr and KF, which afforded rooperol (1AA) and analogue 1EA (Scheme 3). The other analogues were prepared by AcOH-buffered TBAF deprotection to afford 1BB and 1CC (Scheme 3).
With these rooperol analogues in hand, we evaluated the utility of the various catechol replacements by determining the cancer cell cytotoxicity of these compounds versus rooperol using a MTT assay (Table 1). Rooperol displays cytotoxicity against all three cell lines examined: HeLa (cervical adenocarcinoma), H460 (lung carcinoma), and A549 (lung carcinoma). All of the rooperol analogues examined here were also tested against HeLa cells, which were slightly more sensitive to rooperol compared to the other cell lines. All of the analogues, with the exception of 1DD, show some activity against HeLa cells, with analogues 1CC and 1EA displaying activity close to that of rooperol. Interestingly, while the symmetrical analogue 1EE lacks good activity against HeLa cells, the two asymmetrical analogues containing the same methylenedioxyphenyl catechol replacement, 1EA and 1EB, show better cytotoxicity. It is also interesting that 1EE has nearly the same Scheme 3. Synthesis of unsymmetrical rooperol analogs and deprotection reactions.

Discussion
A scalable and general synthesis of rooperol and analogues has been developed and used to evaluate the cancer cell cytotoxicity of a number of analogues that may have improved metabolic stability compared to rooperol. The synthesis of rooperol presented here is longer (seven total steps, longest linear sequence of five steps) but higher yielding (27% vs. 17%) compared to our previous total synthesis [14]. In addition to preparing rooperol, this route was employed to prepare a series of symmetrical and unsymmetrical analogues that were evaluated for cancer cell cytotoxicity compared to rooperol. The symmetrical compound 1BB, bearing 4-hydroxyl substituents on each aromatic ring, displays activity against HeLa cells that is approximately one-half that of rooperol (GI 50 = 33.2 vs. 18 µM) and similar to an unsymmetrical analog bearing a methylenedioxyphenyl group (1EA) in place of one catechol group of rooperol. Interestingly, while the symmetrical analogue bearing two methylenedioxyphenyl groups, 1EE, is much less active against HeLa cells compared to rooperol, 1EE is very similar to rooperol in activity against A459 cells (GI 50 = 28.4 vs. 26.0 µM). Notably, because 1EE lacks the catechol moieties of rooperol, it is not as prone to redox cycling and therefore chemically more stable than rooperol. Together, these results indicate that a more extensive search for symmetrical and asymmetrical rooperol analogues is warranted. In addition, the results obtained here indicate that the phenol and methylenedioxyphenyl catechol replacements are promising, the latter particularly against lung cancer cells. Work establishing the metabolic stability of these analogues versus rooperol is on-going.

Materials and Methods
All reactions were carried out under argon in oven-dried glassware with magnetic stirring. Unless otherwise noted, all materials were obtained from commercial suppliers and were used without further purification. THF was distilled from sodium/benzophenone prior to use. Flash chromatography was performed with EM Reagent silica gel (230-400 mesh) using the mobile phase indicated. Melting points (open capillary) are uncorrected. Unless otherwise noted, 1 H and 13 C NMR spectra were determined in CDCl 3 on a spectrometer operating at 400 and 100 MHz, respectively, and are reported in ppm using solvent as internal standard (7.26 ppm for 1 H and 77.0 ppm for 13 C in CDCl 3 ). Mass spectra were obtained by atmospheric pressure chemical ionization (APCI), chemical ionization using methane as the ionizing gas (CI), or by electrospray ionization (ESI). Copies of all NMR and MS spectra are available in the Supporting Information.
General Procedure-Silyl Protection of Benzaldehydes: 1,2-bis-((tert-butyldimethylsilyl)oxy)benzaldehyde (8A) [17]: To solution of 3,4-dihydroxybenzaldehyde (4 g, 29 mmol) in dichloromethane (120 mL) under argon was added imidazole (7.9 g, 4 equivalent). The reaction mixture was cooled in an ice bath and tert-butyldimethyl-silyl chloride (12 g, 2.5 equivalent) was added. The reaction mixture was allowed to stir overnight, then diluted with dichloromethane and washed twice with 1 N HCl, once with saturated aqueous NaHCO 3 , once with brine, and then, the organic layer was dried over Na 2 SO 4 . After filtration, the solution was evaporated under vacuum and the residue subjected to flash chromatography purification (SiO 2 , 10% EtOAc/ hexanes) to yield the product (9.7 g, 91%) as a viscous oil, which crystallized to a pale yellow solid. Mp