Chemistry of Renieramycins. Part 14: Total Synthesis of Renieramycin I and Practical Synthesis of Cribrostatin 4 (Renieramycin H)

The first total synthesis of (±)-renieramycin I, which was isolated from the Indian bright blue sponge Haliclona cribricutis, is described. The key step is the selenium oxide oxidation of pentacyclic bis-p-quinone derivative (3) stereo- and regioselectively. We also report a large-scale synthesis of cribrostatin 4 (renieramycin H) via the C3-C4 double bond formation in an early stage based on the Avendaño’s protocol, from readily available 1-acetyl-3-(3-methyl-2,4,5-trimethylphenyl)methyl-piperazine-2,5-dione (8) in 18 steps (8.3% overall yield). The synthesis provides unambiguous evidence supporting the original structure of renieramycin I.


Introduction
Many tetrahydroisoquinoline antitumor natural products, such as renieramycins, saframycins, and ecteinascidins, have attracted considerable interest due to their extraordinary structures and meager availability in nature, as well as their potent antitumor activity [1,2]. Among them, renieramycins H (1h) and I (1i) were isolated from the methanol extract of the Indian bright blue sponge Haliclona OPEN ACCESS cribricutis collected from the intertidal region of Okha, Gujarat, in 1988 [3]. Original structures 1h and 1i were given the names renieramycins H and I, respectively. Thereafter, we revised the structure of renieramycin H to that of cribrostatin 4 (2) [4][5][6], which was independently isolated from the blue sponge Cribrochalina sp. collected from reef passages in the Republic of Maldives, based on 13 C NMR studies of several semi-synthetic models ( Figure 1) [7,8]. Cribrostatin 4 (2) has attracted the interest of several medicinal chemistry experts because of its unique structure and cytotoxicity despite the lack of the hemiaminal or aminonitrile function at C-21. Three total syntheses of 2 have been reported [9][10][11]. Recently, we completed a 21-step stereocontrolled total synthesis of (±)-2 from 1-acetyl-3-(3-methyl-2,4,5-trimethylphenyl)methyl-piperazine-2,5-dione (8) in 3.4% overall yield [12,13]. Furthermore, we have accomplished the total synthesis of renieramycin G (1g) [14,15]. The availability of 1g and 2 has enabled us to prepare several renieramycin derivatives having a lactam carbonyl to understand the molecular basis of their impressive cytotoxicity profiles. We present herein an alternative large-scale approach for the total synthesis of 2. This approach might yield a variety of novel analogs of cribrostatin 4 (2), as well as C3-C4 unsaturated bis-p-quinone derivatives, such as renieramycin I (1i), for detailed studies of structure activity relationships (SARs) of these classes of antitumor marine natural products.

Results
The most serious problem in our previous cribrostatin 4 (2) synthesis was that 1-epi-pentacyclic alcohol (4) (Chart 1) might be formed, and the undesired stereochemistry had to be converted into the natural one at C-1 position via enolate formation through several cycles. Avendaño et al. reported that the stereocenter at C-3 of 1,3-trans-compound 5 [16] could be transformed into corresponding 1,3-cis-compound 7 via unsaturated compound 6 through regioselective radical bromination, followed by hydrogenation from the less hindered α-face in good yield [17]. They applied this protocol to the preparation of pentacyclic phthalascidin analogs [18]. We were very interested in this procedure for constructing the 1,3-cis relationships of renieramycins (Scheme 1) [19]. Based on Avendaño's protocol, we designed an alternative synthetic plan that involves the key transformations outlined in Scheme 2: (1) construction of tricyclic compound 9 having an α,β-unsaturated amide carbonyl from readily available compound 8 [20]; (2) condensation of 9 with benzaldehyde derivative and subsequent regio-and stereospecific hydrogenation leading to compound 11; (3) construction of pentacyclic framework and conversion of ester into our intermediate 12, which can be transformed into cribrostatin 4 intermediate 3 [13] (Scheme 2).

Scheme 2.
Strategy for practical synthesis of compound 3, which will be converted into 1i and 2.
According to the results of our previous studies [21,22], treatment of 8 with trimethylsilyl chloride (TMSCl) in the presence of triethylamine (TEA) in CH2Cl2 gave O-trimethylsilyl lactim intermediate 14, which was treated with 2,2-diethoxyethyl benzoate in the presence of trimethylsilyl triflate (TMSOTf) and acetic anhydride to give 15 as an inseparable mixture of diastereomers (15a:15b = 10:3) in 92% yield. After exerting a great deal of effort to separate this mixture by column chromatography several times, we obtained both isomers in their pure forms, and detailed 2D NMR studies confirmed the structures of 15a (minor) and 15b (major). The NMR spectrum of 15a displayed H-1 and H-3 proton signals at δ 6.20 and δ 4.70, respectively, whereas the NMR spectrum of 15b showed H-1 and H-3 proton signals appearing at δ 6.15 and δ 4.07, respectively. An observable nuclear Overhauser enhancement (NOE) between H-3 and H-22 revealed that compound 15a has the trans form (Scheme 3). We then studied the conversion of 15 into unsaturated compound 9, which is the first key step of our synthesis. A preliminary experiment was carried out using major isomer 15a. According to the typical conditions of Avendaño et al. [17], 15a was treated with N-bromosuccinimide (NBS: 1.0 equiv.) and 2,2′-azobisisobutyronitrile (AIBN: 0.1 equiv.) in CCl4 at 80 °C for 6 h to generate 9 (55%) and 16 (4%) plus unreacted 15a (33%). The 1 H NMR spectrum of 9 showed an H-4 olefinic proton signal that appeared as a singlet at δ 7.47. The 1 H NMR spectrum of 16 showed characteristic AB type doublet proton signals at δ 4.60 and δ 4.57 along with the H-4 olefinic singlet proton signal at δ 7.41. Accordingly, 16 might be a product of over-reaction product at C-6 aromatic methyl group. After extensive investigation of the reaction conditions, we found that the yield of our target 9 could be improved by slightly lowering the reaction temperature (60 °C) and excluding AIBN. Thus, the reaction of 15a with NBS (2 equiv.) in CCl4 at 60 °C for 6 h gave 9 (69%) and 16 (13%). It was extremely difficult to separate 9 and 16 in a large scale using silica gel column chromatography. However, catalytic reduction of the above mixture using 10% Pd/C in 2-propanol and DMF at 25 °C for 11 h gave 9 as the sole product in 82% overall yield. Accordingly, the transformation of 15a into 9 without any purification of the intermediates was found to be the best choice in terms of overall yield (9 in 71% yield in four steps).
With key intermediate 9 in hand, we next looked into ways to design a practical transformation of 9 into 12, which was the key intermediate in our previous total synthesis of cribrostatin 4 (2) (Schemes 4 and 5). Condensation of 9 with benzaldehyde derivative 17 [20] in the presence of potassium tert-butoxide gave (Z)-arylidenepiperazinedione 10 in 70% yield. Catalytic hydrogenation of the trisubstituted double bond of 10 over 10% Pd on carbon in MeOH at 25 °C proceeded chemoselectively to give desired 11a (72%) along with 11b (21%). Detailed 2D NMR studies were performed to confirm the structures of 11a and 11b. The NMR spectrum of 11a displayed H-1 and H-13 proton signals at δ 6.53 and δ 4.34, respectively, whereas the NMR spectrum of 11b had H-1 and H-13 proton signals appearing δ 6.43 and δ 4.45, respectively. An NOE between H-1 and H-13 proton signals was observed in 11a but not 11b. Thus, the hydrogenation of 10 obviously occurred stereoselectively from the α-face to generate H-1 and H-13 cis isomer 11a. It is proposed that the steric hindrance due to the C-8 methoxy group and the C-21 carbonyl group was responsible for the β-axial orientation of C-1 substituent as shown in conformer X of 10.

Scheme 4. Preparation of key intermediate 11a.
The piperazinedione ring of 11a was activated by introducing a 2-propyloxycarbonyl group to give imide 18 in 96% yield. Chemoselective reduction of 18 in the conventional manner afforded a hemiaminal, which was treated with formic acid at 25 °C for 0.5 h to afford 19 [23] in 82% yield. Deprotection of 19 with TFA and H2SO4 gave secondary amine 20, which was transformed into 21 by reductive methylation in high yield. Hydrolysis of 21 with 10 N aqueous LiOH in THF/MeOH at 25 °C for 8 h gave primary alcohol 12 in 97% yield, which is identical to the intermediate in our previous total synthesis [13].
The conversion of 12 into 22 was accomplished by partial demethylation with boron tribromide (BBr3), followed by oxidative demethylation to give bis-p-quinone 22 in 67% yield (Scheme 6). Acylation of 22 with in situ prepared angeloyl chloride in dichloromethane gave common intermediate 23 in 84% yield. Encouraged by the results of our extensive model studies, including the transformation of several natural products [24][25][26], the introduction of a methoxy group to the C-14 position of 23 was achieved using 10 equiv. of SeO2 in a mixture of methanol and dioxane at 100 °C for six days to give 1i in 43% yield along with secondary alcohol 24 in 29% yield. The orientation of the methoxy group of 1i was assigned on the basis of the signal of 14-H (δ 4.34, d, J = 1. 4 Hz). The spectroscopic properties of synthetic 1i were in complete accord with those of natural renieramycin I (1i) [27]. Scheme 6. Transformation of compound 12 into renieramycin I (1i) and cribrostatin 4 (2) through compound 23.

Experimental Section
IR spectra were obtained with a Shimadzu Prestige 21/IRAffinity-1 FT-IR spectrometer. 1 H-and 13 C-NMR spectra were recorded on a JEOL JNM-ECA 500 FT NMR spectrometer at 500 MHz for 1 H and 125 MHz for 13 C; a JEOL JNM-AL 400 NMR spectrometer at 400 MHz for 1 H and 100 MHz for 13 C; and a JEOL JNM-AL 300 NMR spectrometer at 300 MHz for 1 H and 75 MHz for 13 C (ppm, J in Hz with TMS as internal standard). All proton and carbon signals were assigned by extensive NMR measurements using COSY, HMBC, and HMQC techniques. Mass spectra were recorded on a JEOL JMS 700 instrument with a direct inlet system operating at 70 eV. Elemental analyses were conducted on a YANACO MT-6 CHN CORDER elemental analyzer.