Progress in Studies on Rutaecarpine. II.—Synthesis and Structure-Biological Activity Relationships

Rutaecarpine is a pentacyclic indolopyridoquinazolinone alkaloid found in Evodia rutaecarpa and other related herbs. It has a variety of intriguing biological properties, which continue to attract the academic and industrial interest. Studies on rutaecarpine have included isolation from new natural sources, development of new synthetic methods for its total synthesis, the discovery of new biological activities, metabolism, toxicology, and establishment of analytical methods for determining rutaecarpine content. The present review focuses on the synthesis, biological activities, and structure-activity relationships of rutaecarpine derivatives, with respect to their antiplatelet, vasodilatory, cytotoxic, and anticholinesterase activities.


Introduction
Rutaceous plants, especially Evodia rutaecarpa (its dried fruit is called 'Wu-Chu-Yu' in China), have long been used to treat gastrointestinal disorders, headache, amenorrhea, and postpartum hemorrhage in traditional oriental medicine [1,2]. The alkaloid, rutaecarpine (8,13-dihydroindolo-[2′,3′:3,4]pyrido [2,1b]quinazolin-5(7H)-one, 1a, Figure 1) was first isolated in 1915 by Asahina and Kashiwaki from an acetone extract of base-treated Evodia rutaecarpa [3][4][5] and later from 'Wu-Chu-Yu' [6]. Interest in the OPEN ACCESS molecule has since been growing, presumably due to its characteristic structure and intriguing biological properties (733 references were found in the SciFinder database provided by the American Chemical Society). In addition, 55 patents have been issued regarding its isolation, biological activity, synthesis, metabolism, and toxicology. Numbers of papers covering rutaecarpine are summarized in Table 1.  Of the 17 review papers written to date, eight have focused on the synthesis of rutaecarpine [7][8][9][10][11][12][13][14], seven on pharmacology [15][16][17][18][19][20][21], one on the modulation of cytochrome P450 [22], and one on detection methods [23]. A review published in 1983 by Bergman [7] covers the nomenclature, structure, synthesis and pharmacological properties of rutaecarpine and of related quinazolinone alkaloids. The review of Wang et al., written in Chinese in 2006, provides details of the synthesis of rutaecarpine based on construction patterns of the five-ring system [10]. Shakhidoyatov and Elmurado's review covered the most recent view on the general point of view for tricyclic quinazoline alkaloids [14]. A review written in 1999 by Sheu addressed the in vitro and in vivo pharmacology of rutaecarpine [15] and later described the cardiovascular pharmacological actions of rutaecarpine in his recent review [20]. More recently, in 2010, Jia and Hu reviewed its cardiovascular protective effects [19]. The present work focuses on the synthesis, biological activities, and structure-activity relationships, with respect to the antiplatelet, vasodilatory, cytotoxic, and anticholinesterase activities, of rutaecarpine derivatives, and complements our first review published in 2008 [24].

Synthesis of Rutaecarpine
A simple retrosynthetic analysis leads to tryptamine (2) and its equivalents for the indole moiety, and anthranilic acid (3a) and its equivalents for the quinazolinone moiety, which leaves an additional one-carbon unit needed for the C13b atom in rutaecarpine (Scheme 1).
Tryptamine (2) has been one of most popular starting materials [1,11] and the compounds 4, 5, and 6 ( Figure 2) have been used as alternative starting materials which provide the A,B,C-ring system and the one-carbon unit at C13b. On the other hand, a series of benzoic acid derivatives 3b-i with nitrogen at the ortho-position ( Figure 3) were employed as an equivalent for 3a as the counterparts for tryptamine. In fact, in 1927 Asahina et al. reported two synthetic procedures for the synthesis of rutaecarpine using these equivalents as a starting materials-one procedure involved a three-step synthesis from 3-(2-aminoethyl)indole-2-carboxylic acid (4) and 2-nitrobenzoyl chloride (3e) (yield not given) [25] and the other a one-pot synthesis (in 24% yield) from 1,2,3,4-tetrahydro-1-oxo-β-carboline (5b) and methyl anthranilate (3d) in the presence of PCl3 [26] (Scheme 2).   [25,26].
Since the classification of syntheses in our previous review [1] was based on the structures of starting materials, we kept the same classification in the present review, that is: (1) tryptamine-derived syntheses; (2) tetrahydro-β-carboline-derived syntheses; and (3) miscellaneous.

Synthesis Using Tryptamine
Lee et al. [27,28] and Kamal et al. [29] used a one-pot reductive-cyclization of nitro (9a) and azide compounds (9b), respectively, to construct the quinazolinone skeleton. Tryptamine was subjected to a Bischler-Napieralsky reaction to afford starting compound 7, which was then condensed with 3e and 2-azidobenzoyl chloride (3f) to afford 8a and 8b, respectively. Cleavage of the exocyclic double bond led to the corresponding ketone 9. It is worth noting that cleavage of the exocyclic double bond on 8 by ozonolysis failed, whereas oxidative cleavage with KMnO4 lead to ketones 9a and 9b in 32% and 67% yield, respectively (Scheme 3). Scheme 3. Synthesis of rutaecarpine by Lee et al. [27,28] and Kamal et al. [29].
The reduction of the nitro group in 9a by tin chloride resulted in subsequent cyclization giving 1a in 94% yield [27,28]. On the other hand, the related 2-azobenzamide (9b) undergoes an aza-Wittig reductive cyclization in the presence of Ph3P and NH4OH or Ni2B in HCl-MeOH under microwave irradiation [29].
More recently, base-initiated intramolecular anionic cascade cyclization [32] of the 2-cyano compound 14 was applied to rutaecarpine synthesis by Liang, et al. [33]. The authors optimized the reaction conditions and found DBU was the reagent of choice for the conversion. The prerequisite 2-cyanoindole compound 14 was prepared in two-steps from tryptamine and 2-fluorobenzoic acid via a 2-chloroindolenine, generated by an electrophilic aromatic substitution reaction at the C2 position in the indole moiety by t-butyl hypochlorite, followed by nucleophilic substitution of the 2-chloro group by cyanide anion in the presence of BF3 (Scheme 6). Scheme 6. Synthesis of rutaecarpine by Liang et al. [33].

Miscellaneous
Synthetic methods not employing anthranilic acid, tryptamine, or their equivalents are very rare. Recently, Pan and Bannister employed a sequential Sonogashira reaction and Larock indole synthesis, whereby Sonogashira's Pd(0)-catalyzed ethynylation of 19 to 18 led to 20, which subsequently underwent an intramolecular Pd(0)-catalyzed indole formation [40] to produce 21. The acid-catalyzed cyclization of 21 led to 1a in 81% yield with a trace of isomeric compound 22 [41] (Scheme 9). To the best of our knowledge, this procedure is the first example of construction of the C-ring to synthesize rutaecarpine via N6-C7 bond formation.
The same group [49] reported a hybrid between the alkaloids rutaecarpine and luotonin A [50,51]. Vilsmeier-Haack formylation of 2-(1H-indol-2-yl)quinazolin-4(3H)-one (25a) gave 26 which was subjected to either direct or indirect reduction to alcohol followed by acid catalyzed cyclization to produce 27a,b. On the other hand, a direct cyclization followed by chlorination under Vilsmeier-Haack conditions led to 27c [52] (Scheme 11). Such a chloro-compound represents a good substrate for introducing substituents by nucleophilic substitution [52] for the synthesis of related series of compounds. Scheme 11. Synthesis of a hybrid between rutaecarpine and luotonin A [50,52].

Biological Properties
A review written in 2003 by Hu and Li, comprehensively described the in vitro and in vivo pharmacology of rutaecarpine [18], in which pharmacological actions were classified as; cardiovascular effects, antiplatelet activity, antithrombotic activity, anticancer activity, anti-inflammatory and analgesic effects, effects on the endocrine system, anti-obesity and thermoregulatory effects, effects on smooth muscle (except cardiovascular), and others. In addition, rutaecarpine ameliorated body weight gain by inhibiting orexigenic neuropeptides NPY and AgRP in mice [53] and reducing lipid accumulation by AMPK (AMP activated protein kinase) activation and UPR (unfolded protein response) suppression [52]. Recently, Xu et al. reported the anti-atherosclerosis activity (EC50 = 0.27 μM) by up-regulating ATP-binding cassette transporter A1 (ABCA1) [54,55]. The present review addresses structure-activity relationships with respect to antiplatelet activity, vasodilator activity, cytotoxicity, and anticholinesterase activity.

Vasodilator Activity
An early study showed that phenylephrine-induced contraction of isolated rat mesenteric arterial segments with intact endothelium was relaxed by 90% by 0.1 mM rutaecarpine and that such relaxation was concentration-dependent in the 0.1 μM-0.1 mM range [59]. Further study revealed that NO-dependent vasodilation is primarily responsible for the vasodilatory activity of rutaecarpine [60]. Subsequently, the vasodilatory effect of rutaecarpine was also related to the stimulation of endogeneous calcitonin gene-related peptide (CGRP) release via the activation of transient receptor potential vanilloid subfamily, member 1 (TRPV1) [61,62]. Chen et al. synthesized 12 rutaecarpine derivatives and 11 analogues, and then evaluated their vasodilator activities (data not shown). These authors found two important trends regarding the vasodilator activities of rutaecarpine-related anti-hypertensives: (1) the N14 atom of rutaecarpine might be the key site, and (2) the 5-carbonyl probably makes a lower contribution, while simple substitution on the indole or quinazoline rings does not enhance vasodilatory effects [62]. Although the prepared compounds showed better activity than rutaecarpine (EC50 = 1.33 μM), such a finding would suggest a new direction for the discovery of valuable TRPV1 agonists as anti-hypertensive drugs if rutaecarpine had proper substituent (s) at the proper position (s).
Regarding the mechanisms responsible for cytotoxicity, inhibitory activities against topoisomerase (topo) I and II have been studied. These inhibitory activities appeared to be affected by substitution on the E-ring but not by substitutions on rings A and/or C [64,66,67], except for 11-bromorutaecarpine. In fact, 10-bromorutaecarpine (1g) and 3-chlororutaecarpine (1c) showed strong inhibitory activities (79.54% and 84.35%, respectively) against topo I and were comparable to camptothecin (82.62%) at 100 μM against 0.2 unit topo I and similar to that against 0.2 units of topo II [67] (see Table 5). In addition, rutaecarpine inhibited tumor cell migration by approximately 30%-40% at 100 μg·mL −1 [68], which would open a new study-window on the use of rutaecarpine as an antitumor agent.
On the other hand, dihydrorutaecarpines (29) showed increased activity and selectivity toward a CNS (U251) cancer cell line in the concentration range 0.02-7 μM and toward a renal cancer cell line at 0.08-20 μM ( Table 4). The parent 30a showed strong activity (GI50 = 0.02 μM) and selectivity toward a CNS cancer cell line and the 10-methylthio compound (29b) showed strong activity (GI50 = 0.08 μM) and selectivity toward a renal (ACHN) cancer cell line.
It is worth noting that evodiamine (29h) showed strong in vitro cytotoxicity against the following cell lines; human leukemia (HL-60, GI50 = 0.   (11, [75] 0  Most of the values were given with standard error of the mean (SEM), but intentionally omitted SEM for clarity. Tumor cell lines: CNS cancer (U251), lung cancer (H522), melanoma cancer (UACC62), ovarian cancer (SKOV3), prostate cancer (DU145), and renal cancer (ACHN). a The cytotoxicity GI 50 values are the concentrations corresponding to 50% growth inhibition, and are the averages of at least two determinations.
Although the cytotoxicity of 13b,14-dihydrorutaecarpine (29a) is somewhat more potent than that of rutaecarpine (1a), it is not easy to establish any possible structure-activity relationship between 13b,14-dihydrorutaecarpine derivatives 29 and the corresponding rutaecarpine derivatives 1, not only because the substitution patterns are different but also because tested cancer cell lines are different.
The poor solubility of rutaecarpine and its derivatives in common organic solvents and in water results in poor bioavailability, which is the main obstacle that needs to be resolved for the further development of rutaecarpine and its derivatives as a drug. In fact, the excellent in vitro activities (GI50 = 1-8 μM) of compounds 29e, 29f, and 29i were not reflected by xenograft model results, presumably because of their poor bioavailability [69].
Generally, benzo-annulation increases electronic dispersion and planar dimensions thus may play an important role in interactions with receptor sites [76,77]. A series of benzo-annulated rutaecarpines were prepared using Fischer indole synthesis and their cytotoxicities against selected human cancer cell lines and their inhibitory effects on topo I and II were evaluated [78] (Table 5). However, currently available data are not sufficient to indicate any clear structure-activity relationships.
It should be noted that an isostere 24 with a sulfone moiety was not as active as rutaecarpine. The percentage of apoptotic cells corresponding to the sub-GI phase of 5-sulfarutaecarpine was found to be 29.2%, which compares with the 14.5% of rutaecarpine [48]. In addition, the hybrids 27a showed an increase of cytotoxic activities against HeLa cells and apoptosis inducing effects at a concentration comparable to that of etoposide. The percentages of apoptotic cells corresponding to the sub-G1 phase of 26 and 27a at the 10 −6 mol·L −1 were 38.6% and 24.1%, respectively, which are comparable to the 14.5% of rutaecarpine while a positive control (etoposide) gave 15.4%.

Inhibitory Activity on Acetylcholinesterase
The early studies on the strong inhibitory activity (64% inhibition at 100 μg·mL −1 ) of fruits extract of Evodia officinalis against acetylcholinesterase [79] and its strong in vivo anti-amnesic activity (IC50 = 6.3 μM) led to the finding that dehydroevodiamine (31) was the origin of such biological activities [80] ( Tables 6 and 7). These results led to more systematic studies on the anti-cholinesterase activity of rutaecarpine [81].
Wang, et al. prepared a series of rutaecarpine (compounds 1q-w) ( Table 6) and 7,8-dehydrorutaecarpine (11, vide infra) derivatives ( Table 7). Most of the rutaecarpine derivatives showed strong inhibitory activity against acetylcholinesterase (eeAChE) from electric eel and butylcholinesterase (eqBuChE) from equine serum with a selectivity on AChE over BuChE in the range 0.1-297.5 [81]. Additional structural modifications of 11 lead a dramatic increase in anticholinesterase activity up to 0.61 nM (11h) and selectivity on AChE up to over 3000 [81,82]. The original data were given as mean +/− standard error of the mean (SEM), but intentionally omitted SEM for clarity. a 50% inhibitory concentration (means of at least four independent experiments) of eeAChE from electric eel; b 50% inhibitory concentration (means of at least four independent experiments) of eqBuChE from equine serum; c Selectivity Index for AChE = IC 50 (BuChE)/IC 50 (AChE).
Interestingly, 5b, not only natural product but also one of the favored starting materials for rutaecarpine synthesis, showed promising inhibitory activity on AChE (IC50 = 83.38 μM) [84], implying a new vista towards designing new analogues of rutaecarpine for the treatment of Alzheimer's disease. In addition to 5b, studies on truncated rutaecarpines have led to new promising lead compounds, such as 25 [85] ( Table 8), 32 [86], and 33 [87] (Figure 4).  The original data were given as mean +/− standard error of the mean (SEM), but intentionally omitted SEM for clarity. a 50% inhibitory concentration (means of at least four independent experiments) of eeAChE from electric eel; b 50% inhibitory concentration (means of at least four independent experiments) of eqBuChE from equine serum; c Selectivity Index for AChE = IC 50 (BuChE)/IC 50 (AChE).
The anticholinesterase activity and the selectivity on AChE were somewhat related to the back-bone structure and the length of the side chain: The derivatives with a backbone with an aromatic C ring (11) showed better activity and selectivity than non-aromatic (1) Table 8). On the other hand, an increase of the length of the side chain would increase the activity (1q vs. 1t, 1r vs. 1u, and 1s vs. 1v; 11c vs. 11e; and 25c vs. 25d) and selectivity except in the case of 1s vs. 1v.
The selectivity indexes (SI, calculated by IC50 for BuChE/IC50 for AChE) on AChE of all the rutaecarpine derivatives ranged from 5.3-3225. Although the selectivity on AChE over BuChE is a concern for curing Alzheimer's disease, clinically useful physostigmine (SI = 3.47) [84], galanthamine HBr (SI = 13.1 [86]) and donepezil (SI = 1252 [88]) show selectivity for AChE over BuChE while rivastigmine (SI ≤ 0.008 [86]) and neostigmine (SI = 0.58 [89]) show selectivity for BuChE over AChE. These results imply that it may not be an advantage for a cholinesterase inhibitor to be selective for AChE or BuChE, but instead suggest that higher efficacy requires a good balance between AChE and BuChE. The original data were given as mean +/− standard error of the mean (SEM), but intentionally omitted SEM for clarity. a 50% inhibitory concentration (means of at least 3 independent experiments) of eeAChE from electric eel; b 50% inhibitory concentration (means of at least 3 independent experiments) of eqBuChE from equine serum; c Selectivity Index for AChE = IC 50 (BuChE)/IC 50 (AChE).

Conclusions
Rutaecarpine is one of the important alkaloids isolated from the Rutaceae and related plants, and it exhibits various interesting biological properties. Recent years have witnessed steady progress in understanding the chemistry and biology of rutaecarpine. Furthermore, it should be noted that reports have been issued on the beneficial effects of rutaecarpine analogues on controlling lipid accumulation [52], obesity [53], and atherosclerosis [54,55], and. The present review focuses on the synthesis of rutaecarpine derivatives and on their biological activities, especially on structure-activity relationships and their antiplatelet, vasodilatory, cytotoxic, and anticholinesterase activities. More efficient and/or practical methods are needed for the synthesis of rutaecarpine derivatives, not only to pursue structure-activity relationship studies but also to identify novel potent lead compounds for drug development.

Author Contributions
J.K.S. and Y.J. wrote the manuscript. H.W.C. critically revised the manuscript typically section for biological properties.

Conflicts of Interest
The authors declare no conflict of interest.