|Home | About | Journals | Submit | Contact Us | Français|
The first total synthesis of (+)-calphostin D and the total synthesis of (+)-phleichrome are outlined. The convergent syntheses utilize an enantiopure biaryl common intermediate, which is formed via an enantioselective catalytic biaryl coupling. The established axial chirality is transferred to the perylenequinone helical stereochemistry with good fidelity. Additionally, efforts focus on the installation of the stereogenic C7,C7′-2-hydroxypropyl groups. Three routes were evaluated to establish the C7,C7′-stereochemistry, in which the successful route involved a double epoxide alkylation with a complex axial chiral biscuprate. This strategy not only allowed the synthesis of the unnatural isomers of calphostin D and phleichrome for assessment in biological systems, but also provided valuable information for the syntheses of the more complex cercosporin and hypocrellin A.
The natural products 1–4 (Figure 1) are representative members of the mold perylenequinones, one of the three major classes of naturally occurring perylenequinones.1 They are characterized by a helical chiral oxidized pentacyclic core in conjunction with C7,C7′-substitution containing centrochiral stereocenters. Calphostin D (R1 = R2 = H, 1d) and phleichrome (2) are the simplest of this growing class of natural products. Calphostin D (1d) is accompanied in nature by its substituted counterparts: calphostin A (R1 = R2 = COPh, 1a), calphostin B (R1 = H, R2 = COPh; 1b) and calphostin C (R1 = COPh, R2 = CO2(p-OH-Ph); 1c). The pigments 1a–d and 2 are isolates of the Cladosporium fungi, specifically Cladosporium cladosporioides (1a–d) and Cladosporium phlei (2).2,3 The more architecturally complex perylenequinones, cercosporin (3)4 and hypocrellin A (4)5 have also been isolated from different fungi.
From a synthetic viewpoint, 1–3 contain the same stereochemical elements – helical chirality and stereogenic C7,C7′-2-hydroxypropyl groups. Prior to our investigations, the total syntheses of the calphostins and phleichrome were reported involving diastereoselective biaryl couplings.6 Unfortunately, the chiral naphthalenes provided only modest stereocontrol during the dimerizations (Scheme 1).7 Furthermore, 3 with the opposite helical chirality and an additional seven-membered ring remained a target that would be difficult to access via the reported approaches. The synthetic challenge of 3 centers on the bridging seven-membered ring, which lowers the atropisomerization barrier such that significant atropisomerization occurs at 37 °C.8
As such, we pursued a flexible strategy that would permit stereoselective synthesis of any stereoisomer of the calphostin/phleichrome framework for biological evaluation and provide an entry into the more complex 3. In this paper, we describe the evolution of the total synthesis of ent-2 and ent-1, with the main focus being the stereoselective installation of the C7,C7′-2-hydroxypropyl groups.
We sought a synthetic strategy to permit a flexible approach to all the perylenequinone natural products 1–4 (Figure 1). In the previous installment of this series, we described the synthesis of helical chiral perylenequinones absent any centrochiral stereocenters. The synthesis employed an enantioselective biaryl coupling to establish the axial chirality from which the corresponding helical stereochemistry was generated with complete stereocontrol. Having shown with this work that compounds such as helical chiral 9 are configurationally stable, we proposed that such an intermediate could be employed in a biomimetic synthesis of ent-1, ent-2, 3, and 4. Specifically, reduction of intermediate 9 directed by the helical axis would furnish ent-2 and 3 (path a) while chelate-controlled reduction would provide ent-1d (path a′). Furthermore, a transannular aldol reaction would access hypocrellin ent-4 (path b). This proposal would circumvent the moderate selectivities encountered in the establishing the axial/helical stereochemistry in prior approaches (Scheme 1, Scheme 2 path c).9 Furthermore, there is no direct means to generate hypocrellin (ent-4) via the strategy in path c; oxidation of the alcohols would be required resulting in loss of their stereochemical information.
Thus, we began our efforts toward (+)-calphostin D (ent-1d) and (+)-phleichrome (ent-2) by devising a synthesis of chiral binaphthalene 10 in order to prepare 9 (Scheme 2). Previously, we had discovered that diaza-cis-decalin catalyst 28 (Table 1) was effective in the coupling of functionalized 2-naphthols and determined the optimal substitution patterns to generate products with high yield and selectivity.10 With these constraints in mind, 15–17 illustrated in Scheme 3 were anticipated to be suitable substrates. Binaphthol coupling as late as possible (i.e. 15) was desirable as the monomers were simpler to manipulate than the corresponding dimers. To determine the optimal pathway in Scheme 3, a series of substrates was synthesized and examined in the asymmetric catalytic naphthol coupling.
Commercially available para-methoxyphenylacetic acid (18) was subjected to regioselective bromination, acid chloride formation, condensation with dimethylmalonate, and Friedel-Crafts cyclization to afford 19 (Scheme 4).11 The bromonaphthalenediol 19 was protected as the bisacetate 20a, using acetic anhydride and pyridine, and as the bispivalate 20b, using pivaloyl chloride and triethylamine. In the first branch point of the naphthol syntheses, selective removal of the less sterically encumbered C2-acyl groups supplied the coupling substrates 21a and 21b.
In the previous paper of this series, the failed synthesis of 22a via cyclization of C3-allylated-18 was attributed to the instability of the allyl group to the Friedel-Crafts cyclization conditions. However, installation of the allyl group after formation of the naphthalene proceeded readily. While the bromo-substrates 20a–b performed poorly in Suzuki couplings, Stille coupling with allyl tri-n-butyltin provided the desired C7-allyl naphthalenes 22a and 22b in suitable yields (Scheme 4). An efficient selective deprotection of the C2-acetate yielded the allyl coupling monomer 24. A Wacker oxidation of 22b afforded the ketone in 63% yield, and after removal of the C2-pivalate, the crucial ketone naphthol 23 was provided.
Because substrates with C7,C7′-halogens provide a readily-convertible scaffold, the C7-iodonaphthol12 was examined in addition to the C7-bromo variant. Beginning with the use of iodine monochloride as the halogen source (Scheme 5), the route mirrors that of bromonaphthol 20a (Scheme 4). Though the syntheses are comparable, significant differences were noted in the Friedel-Crafts cyclization step. The use of sulfuric acid with the iodo analog 25 provided none of the desired product, whereas these conditions efficiently afforded bromo analog 29. For the more sensitive iodo analog, P2O5 in methanesulfonic acid proved essential in the cyclization, although small amounts of by-products were still observed. The overall yields to the diacetates are slightly less for the iodo (40%) vs the bromo (53%) series; however, the iodo substrate provided advantages at later stages (see below).
Employing the optimized conditions (40 °C under oxygen for 48 h), the biaryl coupling screening results for the above substrates are collected in Table 1. As expected, substrates with the electron withdrawing bromo substituent reacted more slowly compared to reported substrate 29a (entry 1).10d Rates improved with acetonitrile as a solvent in place of dichloroethane such that bromo analog 29b was obtained with good yield (67%) and selectivity (81% ee) (entry 2). Replacement of the C4-acetoxy group with the more stable pivaloyl group was promising, but would require further optimization due to lower yield and selectivity (entry 3). With iodo substrate 27, the standard coupling conditions again provided low conversion (entry 4). Use of a mixed solvent system (1:1 dichloroethane:acetonitrile) result in some improvement in yield and good enantioselection (entry 5). Due to substrate solubility acetonitrile alone was initially not considered, but reaction at lower concentrations for a longer period of time (2 days) afforded 29d in 81% yield and 80% enantioselectivity (entry 6). Further attempts to increase selectivity were halted when it was discovered that the high crystallinity of 29d, could provide >99% ee material in one trituration with excellent mass recovery. In a final comparison of the bromo- and iodo-derivatives, this property rendered the synthesis of highly enantioenriched material with iodobiaryl 29d more efficient even though the yields in the synthesis of iodonaphthol substrate 27 was somewhat lower yielding than that of the corresponding bromonaphthol 21a.
The best results for the asymmetric biaryl coupling were obtained with the allylated naphthol 24 to provide 29e in good yield (94%) and enantioselectivity (85%) (entry 7, Table 1). Enantiopure material could be obtained, but unlike the iodo analog 29d, multiple triturations had to be employed. The survival of the oxidatively sensitive allyl group speaks to the compatibility of the coupling conditions with most functionality. Unfortunately, this generalization did not hold true with the corresponding ketone [2-(oxo)-propyl] substrate (entry 8), where formation of 29f was slow and accompanied by significant decomposition. Both enol formation and oxidative enol coupling could compete with the coupling reaction, accounting for these observations.
To install the C7,C7′-stereochemistry we desired a synthetic strategy that was both direct and flexible, utilizing intermediates that could lead to all of the natural products 1–4. In the prior syntheses of the calphostins and phleichrome, formation of the stereogenic C7,C7′ 2-hydroxypropyl groups preceded dimerization (see Scheme 1). Since our strategy (Scheme 2) introduces this stereochemical array after coupling, maximum diversity is possible permitting the synthesis of multiple natural products.
Two distinct paths were envisioned (Scheme 6). In path a, the axial stereochemistry could be utilized as a relay to control the C7,C7′-stereochemistry. We demonstrated the fidelity of this approach in our synthesis of hypocrellin A,12 which is detailed in the fourth paper in this series. In applying the strategy to phleichrome, the alcohol stereocenters could be generated in a diastereoselective fashion by either the reduction of diketone 31 (path a, R = Me) or the methyl addition to dialdehyde 32 (path a, R = H). Alternatively, the stereogenic C7,C7′-substitution could be introduced from an external source, as seen in the epoxide opening of path b. The latter approach is attractive in that independent introduction of the C7,C7′-stereochemistry would permit a more facile synthesis of the complete diastereomeric series of 1, 2, and 3. Though the flexibility of path b was attractive, we initially investigated the likely biomimetic path a (R =Me). Specifically, the breadth of precedent for stereoselective ketone reduction and the interception of a late-stage intermediate (9) from the hypocrellin A synthesis, led us to investigate the diastereoselective reduction of 31.
The asymmetric reduction of ketones to form chiral secondary alcohols13 is a common transformation in synthesis that is widely regarded as a “solved problem”. However, consideration of the reduction of a diketone 31 to form 30a reveals problematic aspects still encountered in this methodology. First, the groups flanking the ketone, methyl and benzyl are sterically similar such that facial bias in asymmetric reductions is difficult. Second, the enolization of benzylic ketones results in low yields. The challenges of the motif can be seen in the approaches to these centers in the prior calphostins syntheses.6
To test the viability of the ketone reduction approach, bisketone 31 was employed initially. The allyl biaryl 29e offered the most direct means to the desired bisketone 31. Following an one-pot deacetylation/methylation, a Wacker reaction provided substrate 31 (Scheme 7).
Molecular models (Scheme 8) suggested that the axial biaryl stereochemistry would block one prochiral face of the ketone to allow for a stereoselective approach of an achiral reducing agent. The reductions could provide three diastereomers: (M,R,R)-diastereomer 30a, (M,S,S)-diastereomer 30b, and meso-(M,R,S)-diastereomer 30c. Unfortunately, statistical mixtures of the isomers (dr = 1:1:2, 30a:b:c; Scheme 8) were obtained using NaBH4 or the larger L-Selectride. Apparently, the biaryl axis is too distant to provide any stereocontrol over approach to the C7,C7′-ketone functionality. This outcome is reminiscent of the prior perylenequinone syntheses,6 where poor diastereoselectivity was observed due to the distance of the C7,C7′-stereochemistry from the forming biaryl bond (Scheme 1). Furthermore, the free rotation about the aryl-methylene bond would result in conformers where the opposite faces are blocked by the axial chiral biaryl.
At this point, external asymmetric reducing agents were examined since internal diastereocontrol was absent. In a parallel synthesis to the biaryl counterpart, a deacetylation/methylation followed by a Wacker reaction yielded model naphthalene 34 (Scheme 9) for this study. Due to the similar steric demands of the methyl and benzyl ketone substituents, the development of asymmetric reductions of arylacetones has been limited. Although 3714 and 3815 were well precedented for this transformation (eq 1), they were ineffectual with naphthalene 34 (17% and 36% ee), respectively (entry 1–2; Table 2). Even the reliable CBS catalyst16 only provided a modest 50% ee (entry 3). Commercially available amino alcohol 41 was also evaluated but with no improvement to selectivity (entry 4).
Other reduction methods including α-pinene borane reagent 42,17 (+)-TADDOL/Ti(Oi-Pr)4/catecholborane,18 pyrrolidinyl-proline 43/DIBAL/SnCl2,19 BINAL 44,20 and RuCl2(BINAP)/diphenylethylenediamine/1000 psi H221 provided little or no enantioselectivity with 34 (entry 5–9, Table 2) in spite of strong precedents with related systems. Most surprising was the complete absence of hydrosilylation of ketone 34 with a rhodium-pybox catalyst (entry 10), even though the same catalyst performed well in our hands with the closely related 2-methoxy-phenylacetone (82% ee).22 Both immobilized Geotrichum candidum and Baker’s yeast have been known to reduce phenylacetone in >99% ee.23,24 Unfortunately, Baker’s yeast had no effect on our model ketone 34 (entry 11, Table 2), although we successfully reduced acetophenone under the same conditions. Apparently, the additional steric hindrance from the ortho-methoxy group has a profound effect on reactivity with enzymatic catalysts.
Since the best result was obtained with the CBS catalyst 40 (entry 3, Table 2), these conditions were applied to the chiral biaryl 31 (Scheme 8). Unfortunately, only moderate selectivity (dr = 1.3:1.0:2.0, 30a:b:c) was observed regardless of which axial antipode (M or P) was used. These results highlight that benzyl methyl ketone substrates remain a problematic asymmetric reduction class compared to aryl alkyl ketones or even many dialkyl ketones.13,16,23 At this point, attention was turned to the second strategy in path a: stereoselective methyl addition to bisaldehyde 32 (Scheme 6).
We initially proposed to use the biaryl axis to direct a diastereoselective methyl addition to 32 (path a R=H, Scheme 6), but the lack of stereocontrol observed in the reduction of 31 would likely be problematic in this route as well. Thus, chiral catalysts were surveyed in the enantioselective addition of Me2Zn to model phenylacetaldehyde 45. Although the asymmetric addition of dialkylzinc reagents to aldehydes is quite common,25 few examples have been reported with α-arylacetaldehydes and Me2Zn26 likely due to the challenges surrounding the reaction: 1) the acidity of the aldehyde 45, making aldol by-products likely and 2) the use of the less reactive, more basic ZnMe2 relative to the more common ZnEt2. A survey of several promising catalyst systems from the literature including MIB,27 more reactive amino alcohol 47,28 BINOL titanium complexes,29 and titanium salen complexes30,31 (entries 1–4, Table 3) was not promising as aldol byproducts predominated.32
Since substrate activation seemed to be crucial, the highly reactive bis(sulfonamide) catalysts,25,33 which catalyze additions even to less-reactive ketones,34 were assessed. Encouragingly, the bis(sulfonamides) 49a and 49b were capable of catalyzing the methyl addition to generate the desired alcohol 46 (41–65% yield, 10–24% ee; entry 5, Table 3). Reducing the amount of Ti(Oi-Pr)4 and ZnMe2 did increase the selectivity up to 70% ee, but also increased the aldol by-adducts. Prior to further optimization, the bis(sulfonamide) catalysts were applied to model naphthalene 52.
Synthesis of naphthalene aldehyde 52 commenced from intermediate 20a, which was subjected to a Sonagashira coupling to install generate alkyne 50 (Scheme 10). Subsequent tetra-n-butylammonium fluoride treatment furnished the terminal alkyne. A one-pot deacetylation/methylation of the C2,C4 phenols was achieved using NaH (60%) and MeI in wet DMF to supply 51 with in high yield. Upon screening a range of hydroboration reagents [bis-sec-isoamylborane (Sia2BH), BH3 THF, catecholborane, and Cy2BH], Cy2BH was found to be the most successful providing 52 after hydrogen peroxide oxidation in low yield (30%). Before further optimization, the alkylation conditions were examined on this substrate. Unfortunately, when the methyl addition with 49b was attempted only a mixture of aldol adducts were produced. Steric interactions from the C6-methyl ether that is not present in model 45 could account for the inability of the catalyst to activate the substrate, allowing deprotonation of the α-center as the only viable pathway. At this juncture, the difficulties encountered in both routes of path a from Scheme 6 stimulated us to investigate the alternate path b utilizing an external chiral reagent as a source for the C7,C7′-stereochemical array.
Although we had initially examined biomimetic diastereoselective approaches, the introduction of an independent chiral fragment (path b, Scheme 6) presents distinct advantages with respect to convergency. As discussed earlier, the use of separate fragments allows facile entry to all diastereomeric combinations of the natural products 1–3 including the unnatural isomers. We elected to investigate copper-mediated epoxide openings to achieve this goal. Since the biaryl axis seems to exert minimal stereocontrol over reactions at the C7,C7′-position, the epoxide opening should not be limited to a matched case (double diastereocontrol), meaning both the (R)- and (S)-epoxide can be utilized with equal facility. Prior to this series of papers, we published an overview of this work;9 Table 4 and the discussion below provide a full report of this chemistry in the calphostin/phleichrome system.35
While Grignard-derived cuprates enjoy considerable precedent in epoxide openings,36 there are few examples of biscuprates being employed in this alkylation. While complex cuprates have been used successfully (Eq 2)37 and simple biscuprates have been employed in epoxide alkylation (Eq 3),38 we could locate no reports of a highly functionalized dianion effecting two ring-openings. Our primary concerns were: 1) the metal-halogen exchange on an electron-rich system and in the presence of the C3-methyl esters; 2) the stability of the electron-rich bisarylcuprate, and 3) the use of stoichiometric biscuprate rather than the excess that is typical for cuprate additions. These concerns were assessed by means of several naphthalene systems, including bismethyl ether 61 and iso-propyl ether 60.39
Starting from iodo-intermediates, the two naphthalene substrates were synthesized as shown in Scheme 11. Mitsunobu reaction with 27 was used to install the C2-iso-propyl ether and was followed by an one-pot deacetylation/methylation to generate 60 in high yield. Methylation of the bisnaphthol 59 with55 dimethylsulfate and potassium hydroxide provided the bismethyl ether 61 directly. The low yield (33%) was attributed to a by-product, resulting from the electrophilic methylation of the C1-position.
To examine the functional group compatibility of the epoxide-opening reaction a variety of substrates were evaluated. Bromonaphthalene 20a formed the Grignard reagent, but only at elevated temperatures, which caused removal of the C2,C4-acetate groups (entry 1, Table 4). While organolithium formation from the iodonaphthalene 26 with t-BuLi was unsuccessful due to addition to the C3-ester, the corresponding Grignard reagent was formed readily at low temperatures (−40 °C) using i-PrMgBr40 (entry 2). The attempted cuprate formation and epoxide-opening from this Grignard only resulted in cleavage of the C2,C4-acetates (entry 3). Pleasingly, when the C2,C4-hydroxyl groups were masked as methyl ethers in 61, the epoxide-opening could be completed in 65% yield with complete regioselection (entry 4).
Since we plan to undertake this same transformation twice on one substrate (65% yield with the monomer would correspond to 43% yield with the dimer), the conditions were optimized further with iodonaphthalene 60. The initial epoxide-opening with 60 yielded modest amounts of the desired product (40%). Fortunately, the use of rigorously oxygen- and water-free conditions drastically decreased the amount of arene formed, improving the yield of 64 to 77% (entry 5a, Table 4). The use of Et2O as a solvent (entry 5b) had no effect on the reaction. The use of other copper reagents (CuCN, entry 5c; CuBr, entry 5d) and additives such as TMSCl or HMPA (entry 5e–f) provided lower yields of the desired 64 due to more protodemetallation of the organocopper reagent. Ultimately, careful purification of the CuI by recrystallization proved to be the most important finding, providing 64 in 87% yield (entry 5g). Interestingly, the use of lower temperatures (−78 °C) did not improve the outcome (entry 5h) indicating that the cuprate and product are fairly robust. With isolated yields in the 85% range (entry 5g), acceptable yields (85% × 85% = 72%) were anticipated in the dimeric systems.
The fact that acetate protecting groups were unacceptable in the epoxide opening (entry 3, Table 4), indicated that this key transformation must be conducted after biaryl formation (a C4-acetate is necessary for high selectivity in the biaryl coupling)10d in order to reduce the number of protecting group steps. Conveniently, this sequence permits more flexibility in the strategy, allowing the syntheses of 1–3 to diverge at a late biaryl intermediate (33, Scheme 6). In spite of these advantages, the formation of a functionalized dianionic organocuprate and two epoxide alkylations remained speculative. The success of the transformation relies on the derived biscuprate behaving as two independent cuprates (Scheme 12). If the two entities interact significantly, side products resulting from an intramolecular reaction of intermediate 66 would occur. In spite of these reservations, enantiopure M-29d was deacetylated and methylated in a one-pot protocol using the previously described conditions (Scheme 6, ,7,7, ,8)8) to yield 33 in 94% yield (Scheme 12). With the optimized epoxide-opening conditions, we discovered that (R)-propylene oxide reacted smoothly with the biscuprate of 33 providing the target structure 30a with two new stereocenters in 75% yield as a single diastereomer (2 couplings, 81% yield each). Interestingly, the cuprates formed from 65 do appear to act independently since the only by-products isolated arise from protodemetallation.
With the stereochemical issue resolved, synthesis of one of the simplest mold perylenequinone natural products, (+)-phleichrome (ent-2), was undertaken. Our studies commenced with application of the PhI(OCOCF3)2-induced oxidation41 of the C5,C5′-positions described in the previous paper of this series to epoxide opening product 30a (Scheme 12). A survey of protecting groups (Me, TBS, Ac, and Bz) on the newly installed C7,C7′-hydroxyl stereocenters revealed that only the benzoate group was able to withstand the reaction conditions to provide 68d without substantial decomposition (Scheme 13).
With the necessary oxygenation pattern established, our next task was removal of the C3,C3′-ester groups via decarboxylation of the respective C3,C3′-diacid. Significantly, the C3,C3′-ester groups served four distinct purposes: 1) coordination to the catalyst during biaryl coupling (Table 1) to enable a highly enantioselective process; 2) stabilization of the highly electron-rich biaryl in the biscuprate epoxide alkylations (Scheme 12); 3) blocking the C3,C3′-position during the C5,C5′-oxidation (Scheme 13); 4) providing an avenue for C3,C3′-derivatization.9 As outlined in the previous paper of this series, the lack of success of conventional decarboxylation protocols42 led us to develop a palladium-catalyzed decarboxylation protocol.12,43 Following benzylation of the bisphenol 68d, a deprotection/reprotection sequence of the C7,C7′-hydroxypropyl groups was needed to withstand the decarboxylation protocol. Thus, the bisbenzoate was cleaved with K2CO3 and MeOH to afford 69, which was subjected to TBSOTf and 2,6-lutidine to provide the bissilyl ether (Scheme 14). The high temperatures that were needed to saponify the sterically encumbered C3,C3′-ester groups resulted in atropisomerization of the biaryl axis. Consequently, a three-step (reduction/oxidation/oxidation) protocol was used to synthesize diacid 70 in high yield (85% over three steps). The novel palladium-mediated decarboxylation of 70 proceeded smoothly to provide the key intermediate 71 in moderate yield and with no loss of enantioenrichment.43
After cleavage of the benzyl ethers, the bisphenol was oxidized by MnO2 to afford perylenequinone (Scheme 15).6c,d The use of MgI2 allowed for the selective removal of the C4,C4′-methyl ethers yielding 72.6b,e,f However, all attempts to remove the C7,C7′-silyl groups resulted in no reaction or significant decomposition, providing none of the desired ent-2.44 Analysis of our initial foray revealed many protection/deprotection steps in addition to the final protecting group problem. Central to these problems was the PhI(OCOCF3)2 (PIFA) oxidation. While the method enabled phenol formation, its incompatibility with most protecting groups (Scheme 13) ultimately restricted and lengthened the synthesis.
Though our initial goal was the synthesis of (+)-phleichrome (ent-2), the use of an external chiral source in the epoxide alkylations (Scheme 12) made a convergent synthesis of the epimers, ent-1 and ent-2,9 straightforward (Scheme 16). For the purposes of this discussion, the total syntheses illustrated in Scheme 16 represent the culmination of our synthetic studies and will be used to draw a comparison between the first and second approaches.45
The first and second generation strategies diverge after the epoxide alkylation of intermediate 33 (Scheme 16). Notably, both (R)- and (S)-propylene oxide were used to provide the diastereomers (M,R,R)-73 and (M,S,S)-73, respectively, after benzylation of the newly formed alcohol stereocenters. While different rates might be expected due to double diastereodifferentiation (matched and mismatched cases), no difference in the reaction rate was seen here. The benzyl protection was chosen to minimize protecting group manipulations, since a global debenzylation would be undertaken prior to perylenequinone formation. While such benzyl ethers are compatible with the latter stages of chemistry described above (Scheme 14-Scheme 15), they are not compatible with the key PIFA oxidation (Scheme 13).46 The only suitable protecting group of the C7,C7′-hydroxypropyl groups for the PIFA oxidation was benzoate which was not viable in the remainder of the synthesis, requiring protecting group exchanges. For these reasons, a new C5,C5′-oxidation route was investigated with benzyl ethers (M,R,R)-73 and (M,S,S)-73 (Scheme 16).
In initial work, we had shown that halogenation and lithiation of the C5-position was facile; however, oxygenation did not proceed.47 In the intervening time, important advances were made in the palladium-catalyzed couplings of aryl halides with oxygen nucleophiles.48,49 Even though these methods had not been utilized in highly hindered systems as encountered here, we aimed to test their feasibility with our challenging highly functionalized, electron-rich system. To this end, chlorination using sulfuryl chloride readily afforded the aryl chloride substrate (Scheme 16).45 Optimization of Buchwald’s protocol,48c involving the catalyst system derived from Pd2dba3 and the X-phos(t-Bu) ligand, enabled the coupling of the bisaryl chloride with KOH to provide the desired bisphenols. Unfortunately, the same reactions with alkoxides such as benzyl alkoxide were poor. Presumably, the steric hindrance of the reacting position combined with that of the alkoxide disfavors O-arylation. However, immediate protection of the unstable bisphenols with benzyl bromide supplied the tetrabenzyl ethers (M,R,R)-74 and (M,S,S)-74 in high yield. In comparison to the first generation approach, the new oxidation procedure allowed a more direct and higher-yielding route to 74 (Scheme 16). Whereas the electronics of the naphthalenes played a large role in the PIFA oxidation chemistry such that each substrate required optimization, the palladium-catalyzed coupling was surprising general.
The final improvement in the second generation strategy was the rhodium-mediated decarbonylation protocol utilized in the removal of the C3,C3′-ester groups of (M,R,R)-74 and (M,S,S)-74 (Scheme 16). Though the palladium-mediated decarboxylation43 of diacid 70 (Scheme 14) was an important contribution employed in the synthesis of perylenequinone analogs,9 an unexpected reaction occurred in this transformation during the cercosporin synthesis.45 As such, a rhodium-mediated decarbonylation approach was examined. After optimization on several model systems,45 the (+)-calphostin D (ent-1d) and (+)-phleichrome (ent-2) syntheses were used as the penultimate test of the decarbonylation protocol.
The requisite bisaldehydes were generated by reduction to the alcohol using DIBALH and then oxidation with o-iodoxybenzoic acid (IBX) (Scheme 16). Pleasingly, treatment with RhCl(PPh3)3 in diglyme at 95 °C supplied a smooth decarbonylation provided that rigorously oxygen-free conditions were employed. The desired (M,R,R)-75 and (M,S,S)-75 were provided with no loss in enantioenrichment and with increased yields of 75% and 90% (cf. 71, Scheme 14). After global removal of all four benzyl ethers of 75, the total syntheses of ent-2 and ent-1 culminated with perylenequinone formation and a selective cleavage of the C4,C4′-methyl ethers.
The first total synthesis of (+)-calphostin D and the total synthesis of (+)-phleichrome from commercially available 18 have been developed. The products were generated in 17 steps with overall yields of 5.3% (average of 87% per step) for ent-2 and 5.2% (average of 87% per step) for ent-1. The syntheses diverge after the first seven steps from enantiopure biaryl (M)-29d, which is formed via an enantioselective catalytic biaryl coupling. While our initial biomimetic route to the stereogenic C7,C7′-2-hydroxypropyl groups was unsuccessful, invaluable information was gained concerning the challenges surrounding this substitution pattern. Furthermore, weaknesses in current asymmetric ketone reduction and aldehyde alkylation methods have been highlighted providing impetus for further study. Ultimately, a three-component coupling reaction was developed involving the union of a complex axial chiral biscuprate with two equivalents of a centrochiral epoxide. This strategy permitted stereoselective access to both ent-2 and ent-1. With the centrochiral centers established, the C5,C5′-oxidation evolved from a capricious PIFA reaction to a remarkably robust palladium-catalyzed O-arylation. Two strategies, a palladium-catalyzed decarboxylation and rhodium-mediated decarbonylation, were found viable for removal of the C3,C3′-ester functionality to establish the perylenequinone substitution pattern. This investigation not only provided the unnatural isomers of calphostin D (1) and phleichrome (2) for assessment in biological systems, but also provided valuable information for the syntheses of the more complex cercosporin (3)9 and hypocrellin (ent-4)12 which are detailed in the subsequent papers in this series.
A MeOH (275 mL × 2) and K2CO3 (1.4 g × 2, 10.5 mmol) mixture is heated and sonicated to promote salt dissolution and then cooled to 0 °C. To a chilled (0 °C/ice bath) solution of diacetate 26 (6.0 g × 2, 13.1 mmol) in CH2Cl2 (110 mL × 2) was added the MeOH/K2CO3 mixture. The mixture was stirred at 0 °C for 0.5 h under argon. After quenching with 1 N HCl, the aqueous phase was extracted with CH2Cl2. The organics were washed with brine and dried (Na2SO4). After the solvent was evaporated, the residue was recrystallized from hexanes/CH2Cl2 to yield 27. Subsequent reacylation of the filtrate (containing C4-naphthol and C2,C4-diol) and application of the above procedure afforded 9.5 g of 27 in an 84% overall yield: 1H NMR (360 MHz, CDCl3) δ 2.49 (s, 6H), 3.96 (s, 6H), 4.04 (s, 6H), 7.03 (s, 2H), 7.15 (s, 2H), 8.22 (s, 2H), 10.50 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 20.8, 53.1, 56.2, 94.8, 99.5, 108.2, 109.1, 122.3, 134.1, 137.9, 154.4, 155.3, 168.9; IR (thin film) 2926, 1772, 1729, 1440 cm−1; HRMS (ESI) calcd for C15H13IO6Na (MNa+) 438.9654, found 438.9643.
To a solution of 27 (1.7 g, 4.1 mmol) in MeCN (550 mL) was added 20 mol% CuI.(S,S)-1,5-diaza-cis-decalin catalyst 28 (292 mg, 0.84 mmol). After stirring for 3 d under oxygen, the solution was quenched with 1 N HCl. The aqueous phase was extracted with EtOAc and the organics were washed with brine, dried (Na2SO4), and concentrated. The resultant resin was chromatographed (50% EtOAc/hexanes) to give 29d in 81% ee. Trituration from CH2Cl2 and hexane (1:5) afforded 29d in >99% ee as a yellow solid (1.3 g, 80%): +26.5 (c 0.5, CH2Cl2, >99% ee (M)); 1H NMR (360 MHz, CDCl3) δ 2.54 (s, 6H), 3.98 (s, 6H), 4.04 (s, 6H), 7.07 (s, 2H), 7.65 (s, 2H), 10.72 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 21.2, 53.6, 56.6, 96.2, 100.5, 108.7, 113.9, 122.9, 133.8, 136.3, 148.2, 153.3, 154.8, 169.1, 169.6; IR (film) 3096, 2957, 1768, 1671, 1613, 1563, 1478, 1440 cm−1; HRMS (ES) calcd for C30H24I2O12Na (MNa+) 852.9300, found 852.9250; CSP HPLC (Chiralpak AD, 1.0 mL/min, 80:20 hexanes:i-PrOH) tR (M)= 22.0 min, tR (P)=30.2 min.
To a solution of 29d (725 mg, 0.87 mmol) in DMF (25 mL) was added NaH (60% in oil, 1.0 g, 26 mmol), and MeI (1.6 mL, 26 mmol). After stirring for 4 h at room temperature under argon, the mixture was quenched with 1 N HCl. The aqueous phase was extracted with EtOAc and the combined organic fractions were washed with 1 N HCl (3X) and brine (2X). After drying (Na2SO4) and concentration, the residue was chromatographed (25% EtOAc/hexanes) to yield 33 as a white foam (660 mg, 94%): −57.4 (c 0.5, CH2Cl2, >99% ee (M)); 1H NMR (360 MHz, CDCl3) δ 3.36 (s, 6H), 4.00 (s, 6H), 4.02 (s, 6H), 4.14 (s, 6H), 7.42 (s, 2H), 7.63 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 52.7, 56.5, 61.9, 62.6, 91.9, 100.7, 118.3, 120.9, 125.9, 131.4, 136.9, 152.0, 153.3, 155.2, 166.9; IR (film) 2945, 1733, 1579, 1463, 1436 cm−1; HRMS (ESI) calcd for C30H28I2O10Na (MNa+) 824.9669, found 824.9638.
A flame-dried Schlenk flask was charged with the aryl iodide and the system was vacuum purged with argon (3x). After dissolution in anhydrous THF the solution was cooled to −40 °C and i-PrMgBr (1 M in THF, 1.25 equiv) was added, dropwise along the sides of the flask. The reaction mixture was stirred at −40 °C for 40 min, under argon. CuI (recrystallized from aqueous NaI and stored in an inert atmosphere box, 0.5 equiv) was introduced to a separate flame-dried Schlenk flask, and the system was vacuum purged with argon (3x). After addition of anhydrous THF, the mixture was cooled to −40 °C. The contents of the first flask (Grignard solution) were added dropwise to the second flask (CuI mixture) via cannula. After stirring for 30 min at −40 °C under argon, a solution of (R)-propylene oxide (2.5 equiv) was added dropwise over 5 min. The mixture was stirred at −40 °C for 30 min and was then allowed to slowly warm to 0 °C over 1 h. The reaction was quenched with 1 N HCl, and then extracted with EtOAc. The combined organic fractions were washed with 1 N HCl and brine, dried (Na2SO4), and concentrated in vacuo. Purification was then accomplished by SiO2 chromatography.
The epoxide-opening was carried out according to the General Procedure with 33 (500 mg, 0.623 mmol) and i-PrMgBr (1 M in THF, 1.87 mL, 1.87 mmol) in THF (7.0 mL); CuI (119 mg, 0.623 mmol) in THF (4 mL) and (R)-propylene oxide (175 μL, 2.49 mmol). The material was chromatographed (SiO2, 50% EtOAc/hexanes) and the product 30a was obtained diastereomerically pure as a white foam (312 mg, 75%): −120.2 (c 0.45, CH2Cl2); 1H NMR (300 MHz, CDCl3) δ 1.10 (d, J = 6.2 Hz, 6H), 1.99 (br s, 2H), 2.37 (dd, J = 8.7, 13.2 Hz, 2H), 2.89 (dd, J = 3.1, 13.2 Hz, 2H), 3.32 (s, 6H), 3.95 (m, 2H), 3.96 (s, 6H), 3.98 (s, 6H), 4.14 (s, 6H), 6.95 (s, 2H), 7.44 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 23.6, 41.2, 52.8, 55.7, 62.1, 62.9, 67.4, 100.6, 119.9, 120.5, 125.2, 128.5, 130.4, 131.6, 151.7, 153.5, 156.3, 167.6; IR (film) 3313, 2950, 1730, 1591, 1498, 1444 cm−1; HRMS (ES) calcd for C36H42O12Na (MNa+) 689.2572, found 689.2563.
The epoxide-opening was carried out according to the General Procedure with iodo-substrate 33 (475 mg, 0.592 mmol) and i-PrMgBr (1 M in THF, 1.8 mL, 1.8 mmol) in THF (7.0 mL) at −78 °C; CuI (113 mg, 0.592 mmol) in THF (4.0 mL) and (S)-propylene oxide (166 μL, 2.37 mmol). The material was chromatographed (SiO2, 50% EtOAc/hexanes). Product 30b was obtained diastereomerically pure as a white foam (292 mg, 74%): −61.4 (c 0.35, CH2Cl2); 1H NMR (500 MHz, CDCl3) δ 1.04 (d, J = 6.2 Hz, 6H), 2.04 (br s, 2H), 2.66 (dd, J = 7.5, 13.5 Hz, 2H), 2.71 (dd, J = 4.6, 13.5 Hz, 2H), 3.32 (s, 6H), 3.88 (m, 2H), 3.97 (s, 6H), 3.98 (s, 6H), 4.13 (s, 6H), 6.97 (s, 2H), 7.45 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 23.1, 40.4, 52.8, 55.7, 62.1, 62.9, 67.8, 100.6, 119.8, 120.6, 125.2, 128.4, 130.4, 131.3, 151.9, 153.5, 156.3, 167.5; IR (film) 3414, 2950, 1730, 1591, 1498, 1444 cm−1; HRMS (ESI) calcd for C36H42O12Na (MNa+) 689.2572, found 689.2565.
To a solution of 30a (300 mg, 0.45 mmol) in DMF (12 mL) was added benzyl bromide (1.1 mL, 9.0 mmol) and n-Bu4NI (33 mg, 0.090 mmol). NaH (60% in oil, 270 mg, 6.8 mmol) was added and the reaction stirred under argon. After completion as judged by TLC, the mixture was acidified with 1 M HCl, diluted and washed with EtOAc (2X). The combined organic portions were washed with NH4Cl (aq, 2X), dried (Na2SO4) and concentrated. Purification by column chromatography (10–50% EtOAc/hexanes) afforded (M,R,R)-73 as a yellow resin (337 mg, 88%): −103.3 (c 0.3, CH2Cl2, >99% ee); 1H NMR (300 MHz, CDCl3) δ 0.97 (d, J = 6.1 Hz, 6H), 2.48 (dd, J = 7.1, 13.2 Hz, 2H), 3.00 (dd, J = 5.7, 13.2 Hz, 2H), 3.33 (s, 6H), 3.64 (m, 2H), 3.91 (s, 6H), 3.97 (s, 6H), 4.13 (s, 6H), 4.34 (br m, 4H), 7.03 (s, 2H), 7.11 (m, 4H), 7.22 (m, 6H), 7.40 (s, 2H); 13C NMR (125 MHz, (CD3)2CO) δ 19.9, 38.4, 52.6, 55.8, 61.9, 63.0, 70.5, 74.5, 100.8, 120.7, 121.5, 125.7, 127.7, 128.0, 128.8, 128.9, 130.9, 132.3, 140.3, 152.2, 154.1, 157.3, 167.5; IR (film) 2943, 1738, 1591, 1498, 1452 cm−1; HRMS (ES) calcd for C50H54O12Na (MNa+) 869.3513, found 869.3480.
Bisbenzyl ether (M,S,S)-73 was prepared in the same manner as (M,R,R)-73 and was obtained as a yellow resin (285 mg, 79%): −68.1 (c 0.3, CH2Cl2, >99% ee); 1H NMR (500 MHz, (CD3)2CO) δ 0.96 (d, J = 6.1 Hz, 6H), 2.61 (dd, J = 5.6, 13.6 Hz, 2H), 2.76 (dd, J = 7.0, 13.6 Hz, 2H), 3.30 (s, 6H), 3.71 (m, 2H), 3.93 (s, 6H), 3.94 (s, 6H), 4.09 (s, 6H), 4.19 (d, J = 12.2 Hz, 2H), 4.32 (d, J = 12.2 Hz, 2H), 7.00 (m, 4H), 7.06 (s, 2H), 7.16 (m, 6H), 7.46 (s, 2H); 13C NMR (125 MHz, (CD3)2CO) δ 20.0, 39.0, 52.6, 55.8, 61.9, 63.0, 70.6, 74.3, 100.8, 120.7, 121.5, 125.7, 127.7, 127.9, 128.7, 128.9, 130.9, 132.4, 140.2, 152.3, 154.0, 157.2, 167.5; IR (film) 2943, 1738, 1591, 1498, 1452 cm−1; HRMS (ES) calcd for C50H54O12Na (MNa+) 869.3513, found 869.3517.
Bisbenzyl ether (M,R,R)-73 (30 mg, 0.036 mmol) in anhydrous CH2Cl2 (0.7 mL) was treated with SO2Cl2 (7.0 μL, 0.089 mmol) and was allowed to stir at room temperature under argon until the reaction was complete, as determined by TLC. The mixture was quenched with H2O, extracted with CH2Cl2, washed with brine, dried (Na2SO4), and concentrated. Purification was accomplished by chromatography (25% EtOAc/hexanes) to yield (M,R,R)-bischloride as a yellow resin (30 mg, 94%): −49.0 (c 0.15, CH2Cl2, >99% ee); 1H NMR (500 MHz, CDCl3) δ 1.00 (d, J = 6.1 Hz, 6H), 2.50 (dd, J = 6.6, 13.6 Hz, 2H), 2.97 (dd, J = 6.4, 13.6 Hz, 2H), 3.35 (s, 6H), 3.59 (m, 2H), 3.84 (s, 6H), 3.97 (s, 6H), 4.01 (s, 6H), 4.25 (d, J = 12.1 Hz, 2H), 4.34 (d, J = 12.1 Hz, 2H), 7.02 (s, 2H), 7.06 (m, 4H), 7.20 (m, 6H); 13C NMR (125 MHz, CDCl3) δ 19.7, 38.3, 52.9, 60.9, 61.9, 64.8, 70.7, 75.0, 120.7, 122.5, 122.7, 124.4, 127.2, 127.5, 127.6, 128.4, 133.7, 135.2, 138.9, 152.9, 154.6, 154.8, 167.0; IR (film) 2935, 1738, 1591, 1552, 1452 cm−1; HRMS (ESI) calcd for C50H52Cl2O12Na (MNa+) 937.2734, found 937.2750.
An oven-dried microwave tube with a crimp top and Teflon septa containing a stirbar was charged with aryl halide (M,R,R)-bischloride (120 mg, 0.13 mmol) and KOH (44 mg, 0.79 mmol). In an inert atmosphere box, the substrate-containing microwave tube was charged with Pd2dba3 (18 mg, 0.020 mmol) and x-phos(t-Bu) ligand (33 mg, 0.79 mmol), and the reaction tube was crimped in the inert atmosphere box to avoid exposure to oxygen. The tube was further evacuated and backfilled with argon (2X). A solution of 1,4-dioxane (1.7 mL) and de-ionized water (1.2 mL) was vigorously purged with argon for 1 h prior to use. At this time, the solvent mixture was added to the reaction tube and the mixture was stirred in a preheated oil bath (90 °C) until the aryl halide was consumed as judged by TLC. The reaction mixture was cooled to 0 °C, carefully acidified with aqueous HCl (0.5 N), and the resulting mixture was extracted with EtOAc (2X). The organic layer was dried (Na2SO4) and concentrated to yield an orange oil. This unstable oil was immediately dissolved in anhydrous DMF (2.5 mL) and treated with BnBr (300 μL, 2.6 mmol) and NaH (95%, 70 mg, 2.6 mmol) under argon and allowed to stir at room temperature for 1 h. The reaction was quenched with NH4Cl (aq) and washed with EtOAc (2X). The organic phase was washed with NH4Cl (aq, 2X), dried (Na2SO4), and the solvent was evaporated. Purification was accomplished by chromatography (25% EtOAc/hexanes) to yield (M,R,R)-74 as a yellow resin (120 mg, 86 % yield): −29.5 (c 0.3, CH2Cl2, >99% ee); 1H NMR (300 MHz, CDCl3) δ 1.02 (d, J = 6.1 Hz, 6H), 2.50 (dd, J = 6.6, 13.3 Hz, 2H), 2.99 (dd, J = 6.4, 13.2 Hz, 2H), 3.37 (s, 6H), 3.60 (m, 2H), 3.85 (s, 6H), 3.96 (s, 6H), 3.99 (s, 6H), 4.27 (d, J = 12.1 Hz, 2H), 4.36 (d, J = 12.1 Hz, 2H), 5.02 (d, J = 10.0 Hz, 2H), 5.06 (d, J = 10.0 Hz, 2H), 6.93 (s, 2H), 7.07 (m, 4H), 7.18 (m, 6H), 7.40 (m, 6H), 7.61 (m, 4H); 13C NMR (125 MHz, CDCl3) δ 19.9, 38.2, 52.7, 61.4, 61.9, 64.5, 70.7, 75.4, 76.9, 120.5, 120.6, 123.1, 124.0, 127.5, 127.6, 128.1, 128.4, 128.6, 129.0, 133.4, 135.4, 138.0, 139.1, 146.6, 150.5, 152.3, 154.3, 167.5; IR (film) 2935, 1738, 1591, 1452 cm−1; HRMS (ESI) calcd for C64H66O14Na (MNa+) 1081.4350, found 1081.4380.
(M,S,S)-bischloride was prepared in the same manner as diastereomer (M,R,R)-bischloride and was obtained as a yellow resin (215 mg, 99%): −20.0 (c 0.15, CH2Cl2, >99% ee); 1H NMR (500 MHz, CDCl3) δ 0.94 (d, J = 6.1 Hz, 6H), 2.57 (dd, J = 6.3, 13.6 Hz, 2H), 2.81 (dd, J = 6.8, 13.6 Hz, 2H), 3.31 (s, 6H), 3.68 (m, 2H), 3.84 (s, 6H), 3.98 (s, 6H), 4.02 (s, 6H), 4.22 (d, J = 12.0 Hz, 2H), 4.35 (d, J = 12.0 Hz, 2H), 6.95 (s, 2H), 7.03 (m, 4H), 7.19 (m, 6H); 13C NMR (125 MHz, CDCl3) δ 19.8, 38.9, 52.9, 60.9, 61.9, 64.7, 70.7, 74.7, 120.6, 122.5, 122.7, 124.5, 127.4, 127.5, 127.6, 128.4, 133.8, 135.1, 138.8, 153.1, 154.4, 154.8, 166.9; IR (film) 2935, 1738, 1591, 1552, 1452 cm−1; HRMS (ESI) calcd for C50H52Cl2O12Na (MNa+) 937.2734, found 937.2726.
Bisbenzyl ether (M,S,S)-74 was prepared in the same manner as diastereomer (M,R,R)-74 and was obtained as a yellow resin (182 mg, 77%): −22.0 (c 0.25, CH2Cl2, >99% ee); 1H NMR (360 MHz, CDCl3 δ 0.96 (d, J = 6.1 Hz, 6H), 2.56 (dd, J = 6.3, 13.3 Hz, 2H), 2.85 (dd, J = 6.7, 13.3 Hz, 2H), 3.35 (s, 6H), 3.67 (m, 2H), 3.89 (s, 6H), 3.99 (s, 6H), 4.03 (s, 6H), 4.28 (d, J = 12.1 Hz, 2H), 4.38 (d, J = 12.1 Hz, 2H), 5.01 (d, J = 9.9 Hz, 2H), 5.05 (d, J = 9.9 Hz, 2H), 6.86 (s, 2H), 7.05 (m, 4H), 7.19 (m, 6H), 7.37 (m, 2H), 7.44 (m, 4H), 7.61 (m, 4H); 13C NMR (125 MHz, (CD3)2CO) δ 20.0, 39.1, 52.6, 61.5, 61.8, 64.5, 70.7, 75.0, 77.3, 120.8, 121.4, 124.1, 124.8, 127.8, 128.1, 128.7, 128.8, 129.2, 129.4, 134.0, 136.1, 138.7, 140.2, 147.0, 151.2, 153.1, 154.6, 167.3; IR (film) 2943, 1738, 1591, 1452 cm−1; HRMS (ESI) calcd for C64H67O14 (MH+) 1059.4531, found 1059.4524.
To a chilled (0 °C) solution of (M,R,R)-74 (50 mg, 0.047 mmol) in toluene (4 mL) under argon was added DIBALH (1 M in hexanes, 0.4 mL, 0.40 mmol). The solution was stirred for 30 min, and then was quenched with de-ionized H2O and extracted with EtOAc. The organic phases were washed with aq NH4Cl, dried (Na2SO4), and the solvent was evaporated to yield a yellow resin, which was carried on to the next step without further purification.
To a solution of the bisbenzyl alcohol in EtOAc (2.5 mL) was added 2-iodoxybenzoic acid (112 mg, 0.40 mmol). The mixture was heated at reflux under argon until the alcohol was consumed as judged by TLC. The mixture was diluted with EtOAc and filtered through Celite. The solvent was evaporated in vacuo to yield a yellow oil, which was carried on to the next step without further purification.
The bisaldehyde in diglyme (3 mL) was vigorously purged with argon for 30 min. In an inert atmosphere box, an oven-dried microwave tube with a crimp top and Teflon septa was charged with ClRh(PPh3)3 (92 mg, 0.0992 mmol). The aldehyde solution was added dropwise via cannula to the argon purged microwave tube, containing ClRh(PPh3)3. The mixture was vigorously purged with argon for 20 min and then was heated at 90 °C for 17 h. The mixture was cooled, diluted with EtOAc, and washed with saturated aq NH4Cl. The organic phases were dried (Na2SO4) and the solvent was evaporated to yield a yellow resin. Purification was accomplished by chromatography (10–25% EtOAc/hexanes) to yield (M,R,R)-75 as a yellow resin (33 mg, 75%): −10 (c 0.25, CH2Cl2, >99% ee); 1H NMR (500 MHz, CDCl3) δ 1.03 (d, J = 6.1 Hz, 6H), 2.41 (dd, J = 7.3, 13.4 Hz, 2H), 3.06 (dd, J = 5.8, 13.3 Hz, 2H), 3.64 (m, 2H), 3.67 (s, 6H), 3.89 (s, 6H), 3.99 (s, 6H), 4.33 (d, J = 12.0 Hz, 2H), 4.39 (d, J = 12.0 Hz, 2H), 5.06 (d, J = 10.1 Hz, 2H), 5.09 (d, J = 10.1 Hz, 2H), 6.75 (s, 2H), 6.80 (s, 2H), 7.17 (m, 4H), 7.21 (m, 6H), 7.37 (m, 2H), 7.45 (m, 4H), 7.62 (m, 4H); 13C NMR (125 MHz, (CD3)2CO) δ 20.0, 38.7, 56.3, 56.7, 61.3, 70.8, 75.7, 76.4, 96.5, 112.7, 117.2, 123.7, 127.8, 128.1, 128.2, 128.8, 128.9, 129.0, 134.1, 134.4, 139.8, 140.6, 148.1, 149.2, 155.8, 157.9; IR (film) 2927, 2858, 1645, 1591, 1460, 1336, 1259, 1205 cm−1; HRMS (ES) calcd for C60H63O10 (MH+) 943.4421, found 943.4413.
Compound (M,S,S)-75 was prepared in the same manner as diastereomer (M,R,R)-75 and was obtained as a yellow resin (80 mg, 90%): −6.0 (c 0.25, CH2Cl2, >99% ee); 1H NMR (500 MHz, (CD3)2CO) δ 0.96 (d, J = 6.1 Hz, 6H), 2.56 (dd, J = 6.0, 13.5 Hz, 2H), 2.76 (dd, J = 6.7, 13.5 Hz, 2H), 3.64 (m, 2H), 3.67 (s, 6H), 3.83 (s, 6H), 4.02 (s, 6H), 4.26 (d, J = 12.2 Hz, 2H), 4.35 (d, J = 12.2 Hz, 2H), 4.99 (d, J = 10.1 Hz, 2H), 5.03 (d, J = 10.1 Hz, 2H), 6.82 (s, 2H), 6.97 (s, 2H), 7.09 (m, 4H), 7.18 (m, 6H), 7.36 (m, 2H), 7.45 (m, 4H), 7.64 (m, 4H); 13C NMR (125 MHz, (CD3)2CO) δ 20.1, 38.9, 56.3, 56.8, 61.2, 70.8, 75.6, 76.4, 96.7, 112.8, 117.2, 123.7, 127.7, 128.2, 128.3, 128.8, 128.9, 129.0, 134.1, 134.4, 139.8, 140.4, 18.1, 149.2, 155.9, 157.8; IR (film) 2927, 2858, 1730, 1591, 1460, 1336, 1259, 1205 cm−1; HRMS (ES) calcd for C60H63O10 (MH+) 943.4421, found 943.4431.
To a solution of (M,R,R)-75 (10 mg, 0.011 mmol) in THF (0.7 mL) and MeOH (0.7 mL) was added 10% Pd/C (15 mg). The mixture was stirred while purging with H2 (H2 balloon). After completion as judged by TLC, the mixture was filtered through Celite, rinsing with EtOAc and CH2Cl2. Concentration yielded an unstable brown oil which was used directly in the next reaction.
To a solution of the binaphthol in anhydrous THF (1 mL) was added MnO2 (20 mg, 0.23 mmol). After completion as judged by TLC, the mixture was diluted with EtOAc, filtered through Celite, and concentrated to yield the perylenequinone. Purification was accomplished by chromatography (5% MeOH/CH2Cl2) to yield the perylenequinone as red resin (5 mg, 82%).
To a solution of the above perylenequinone product (1.5 mg, 0.0026 mmol) in THF (1 mL) under an argon atmosphere was added a solution of MgI2 in Et2O (0.07 M, 80 μL, 0.0055 mmol). The dark purple mixture was stirred 10 min (until the mixture turns from purple to black), diluted with EtOAc, washed with saturated aq NH4Cl, and dried (Na2SO4). Concentration yielded a red residue, which was chromatographed (5% MeOH/CH2Cl2) to yield product ent-2 as a red resin (1 mg, 70%): See Supporting Information for CD spectrum; 1H NMR (500 MHz, CDCl3) δ0.54 ( d, J = 6.1 Hz, 6H), 2.96 (dd, J = 6.3, 12.7 Hz, 2H), 3.42 (m, 2H), 3.61 (dd, J = 6.6, 12.7 Hz, 2H), 4.07 (s, 6H), 4.22 (s, 6H), 6.59 (s, 2H), 15.8 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 23.4, 42.4, 56.7, 61.7, 68.7, 101.7, 106.1, 117.3, 126.4, 127.3, 135.5, 152.0, 167.0, 173.7, 177.8; IR (film) 3298, 2927, 2858, 1730, 1607, 1452, 1413, 1375, 1267, 1220 cm−1; HRMS (ES) calcd for C30H29O10 (MH−) 549.1761, found 549.1776.
The perylenequinone ent-1d was prepared in the same manner as diastereomer ent-2 and was obtained as a red resin (1.3 mg, 57%): See Supporting Information for CD spectrum; 1H NMR (300 MHz, CDCl3) δ0.94 (d, J = 6.1 Hz, 6H), 2.92 (dd, J = 8.1 Hz, 13.3 Hz, 2H), 3.54 (dd, J = 3.3 Hz, 13.4 Hz, 2H), 3.74 (m, 2H), 4.05 (s, 6H), 4.22 (s, 6H), 6.54 (s, 2H), 15.9 (s, 2H); 13C NMR (125 MHz, CDCl3) δ 23.8, 42.5, 56.6, 61.5, 69.3, 101.8, 106.5, 117.9, 125.8, 127.9, 136.3, 151.4, 167.2, 172.4, 179.1; IR (film) 3375, 2927, 2858, 1607, 1522, 1452, 1413, 1282, 1244 cm−1; HRMS (ES) calcd for C30H30O10Na (MNa+) 573.1737, found 573.1757.
We are grateful to the NIH (CA-109164) for financial support. Partial instrumentation support was provided by the NIH for MS (1S10RR023444) and NMR (1S10RR022442). We thank 3D Pharmaceuticals (C.A.M.), Novartis (B.J.M.), Eli Lilly (E.O.B.), and the Division of Organic Chemistry of the American Chemical Society (C.A.M., B.J.M.) for graduate fellowships. We thank Dr. Joseph Kozlowski for helpful discussions regarding the epoxide alkylation chemistry. We acknowledge the work of Michelle Clasquin in the synthesis of starting materials. We thank Virgil Percec for assistance with CD measurements.