MSc Chemistry
Molecular design, synthesis and catalysis
Literature Thesis
Developments in the synthetic studies on englerin A
by
Ibrahim Aydin
10883568
September 2016
12 ECTS
22-09-2016
Supervisor/Examiner:
Examiner:
Prof. Dr. H. Hiemstra
Dr. J. C. Slootweg
Van ’t Hoff institute for Molecular Sciences
Division of Synthetic organic chemistry
Kidney cancer is a growing medical issue and is in the top ten most prevalent cancer. Englerin A, a guaiane sesquiterpene diester isolated from Phyllanthus engleri, shows promising biological activity and selectivity against renal cancer cells. We present the developments on the total synthesis and formal synthesis of englerin A via various routes. These routes are based on three different strategies, which are compared with each other on feasibility and efficiency. Finally, we discuss the structure-activity relationship studies on englerin A to gain insight on the intrinsic cytotoxic activity on a molecular level.
1 Introduction 1
1.1 Mechanism of action of englerin A 2
1.2 Outline of the literature thesis 3
2 Isolation and characterization 4
2.1 Isolation of (-)-englerin A and B 4
2.2 Characterization of (-)-englerin A 5
3 Total synthesis of englerin A 6
3.1 Access to the non-polar polycyclic core by a 2-step Michael addition A 6
………d.…….and samarium iodide radical cyclization
3.2 Access to the non-polar polycyclic core by various cycloaddition 7
………..reactions
3.2.1 Metal-free cycloaddition reactions 7
3.2.2 Gold catalyzed cycloaddition reactions 12
3.2.3 Rhodium catalyzed cycloaddition reactions 16
3.2.4 Platinum catalyzed cycloaddition reaction 18 3.3 Access to the non-polar polycyclic core by metathesis reactions 20
4 Formal synthesis of englerin A 29
4.1 Precursor preparation derived from various cycloaddition reactions 29 4.2 Precursor preparation derived from a ring-closing metathesis reaction 31 4.3 Precursor preparation derived from a reductive-Heck reaction 34
5 Structure-activity relationship of englerin A 36
1
1. Introduction
(-)-Englerin A (1, figure 1) and englerin B (2, figure 1) are natural products that belong to the class of guaiane sesquiterpenes. These compounds were identified by the group of Beutler in 2009 from the stem bark of Phyllanthus engleri, a flowering plant that can be found in Tanzania, Malawi, Zambia, Zimbabwe, Mozambique and South Africa (see figure 2).[1] Englerin A consists of a trans-fused oxatricyclic ether core bearing two methyl substituents, an isopropyl group and two ester substituents: a cinnamoyl ester at the C-6 position and a glycolic ester at the C-9 position. The englerin B structural analogue lacks a cinnamoyl ester moiety, replacing the latter with a hydroxyl group. It was found that (-)-englerin A selectively inhibits renal cancer cell lines, which makes englerin A a suitable candidate as an anticarcinoma drug.[1] Due to this intrinsic property and the fascinating molecular architecture (-)-englerin A possesses, it is of considerable interest for the synthetic community and has been subject to a great number of synthetic research.[2]
O H O O Ph O O OH 1 O H O O Ph OH 2 1 3 5 6 8 9 11 13 14 15 H H
Figure 1: structures of (-)-englerin A (1) and englerin B (2).
Kidney cancer is an ongoing medical issue and among the top ten most prevalent cancer with a probability of 1 in 63 to develop the disease. It is estimated that in 2016, 62.700 new cancer cases will occur in the US alone for which 14.240 patients will die and it is presumed that these numbers will rise over the years.[3]
Figure 2: The bark of Phyllanthus engleri, a shrub tree found in South/East Africa.[8]
Current ways of treatment of renal cell cancer are surgery, radiation therapy, immunotherapy and molecular-targeted therapy. However, only surgical resections are
2 effective for localized carcinoma. And merely ± 50% patients with early stage renal cancer are cured and treatment of late stage IV renal cancer being very poor.[4] Additionally, chemotherapy is very limited, with severe side effects and response rates lower than 15% eventually creating an economic burden of multi-billion dollars in the US. Therefore, it is of great importance to find effective and in particular selective (for the reduction of side effects) new chemotherapeutic drug for the treatment of kidney cancer (see figure 3 for a micrograph of healthy and tumor renal cells). This literature thesis will cover the developments in the synthetic research of one such potential drugs, englerin A and/or structural analogues. Prior to the synthetic overview is a section that discusses the mechanism of action of the biologically active englerin A and the goal of this literature thesis.
Figure 3: Micrograph of a healthy kidney illustrated on the left and renal cell carcinoma illustrated on
the right side of the image.[7]
1.1 Mechanism of action of englerin A
Not much is known about the exact mechanism of the biological action of englerin A. The first detailed study of the biological mechanism of englerin A came from the group of Ramos in 2012.[5] They found that when subjecting renal cells to englerin A, only the tumor cells were affected by the natural compound without altering healthy renal cells. Additionally, the group also discovered that pyroptotic and autophagic proteins are unaltered by 1 and, perhaps the most interesting, that concomitant with necrotic signalling the increase of calcium influx into the tumor cellular cytoplasm was observed, indicating a close correlation with each other.
This close correlation was further investigated by Waldmann et al.[6] They demonstrated that (-)englerin A activates the transient receptor potential canonical (TRPC) calcium channels TRPC4 and TRPC5 in a selective and highly efficient way. This affects the calcium permeability thus increases the intracellular calcium concentration, which leads to cell death after only a few minutes. The selectivity comes from the fact that these channels are to be found in the membranes of renal cells and are expressed in abundance in renal tumor cells. However, it is not excluded that overproduction of TRPCs are the only mechanism of action leading to tumor cell death.
3
1.2 Outline of the literature thesis
In this thesis an overview will be provided on the synthetic studies performed on englerin A. First, the reports on the isolation and characterization of 1 and 2 will be presented and discussed. Subsequently, the reports on the total synthesis of englerin A will be discussed and compared with each other on their feasibility and efficiency. Additionally, the formal synthesis reports will also be provided and discussed. These reports will be compared with the reports on the total synthesis of englerin A using the same precursor. Finally, to understand the intrinsic cytotoxic activity of 1 on a molecular level and to understand how this activity can be further improved, the structure-Activity Relationship (SAR) studies will be presented and discussed.
4
2. Isolation and characterization
Beutler et al. identified two natural products (-)-englerin A and B when searching for plant extracts that exhibited activity towards renal cancer cells.[1] The extracts of 34 natural products were analysed and tested on renal tumor cells in the NCI 60-cell panel (National Cancer Institute tumor panel screening program) and compared with eight non-renal panels in the screen (leukemia, breasts, colon, non-small cell lung cancer, prostate, CNS, colon and ovarian organ panels). In order to identify natural products that exhibit the highest selectivity towards the renal panel, two filters were applied. The first filter selected only those extracts that passed a certain bioactivity threshold, which is needed to exclude all extracts that act to any type of tumor cell without making a distinction, either by complete inhibition of all cell types or now inhibition at all. The second filter selected extracts that showed bioactivity profile higher than the group average activity, thus excluding those extracts that did not or to a lesser extent exhibit selectivity towards the renal panel. Those extracts that passed the two filters were subsequently analysed for within-renal cancer cell selectivity by comparing the bioactivities measured in GI50 (concentration needed to inhibit the proliferation of cancer cells by 50%). From these extracts, the statistical outcome showed the extract from Phyllanthus engleri to possess highest selectivity and potency against renal cancer cells. The active compound in this extract was identified as (-)-englerin A, which showed 1000-fold selectivity towards 6 out of 8 renal cancer cell lines of the NCI-60 panel with GI50 values between 1 – 87 nM.
2.1 Isolation of (-)-englerin A and B
The isolation of englerin was carried out from specimens of Phyllanthus engleri collected near Ilembula in the Njombe district of Tanzania. The natural compounds can be extracted from both the root bark and the stem bark of the shrub tree. For the isolation from the root bark, the dried and grinded root bark was extracted with a 1:1 (v/v) CH2Cl2-MeOH mixture. The two solvents were separately coated on a diol bonded phase media and separated by a solid phase extraction (SPE) method. The fractions from SPE were tested for bioactivity by bioassay guided fractionation (in which fractions are separated on the basis of cell growth inhibition activity) using sensitive renal cell lines and resistant CNS cell lines. This showed that the methylene chloride solubles contain the active component. Flash chromatography of the methylene chloride fraction further purified the sample, resulting in the collection of one fraction that possessed the highest cell growth inhibition activity. Further purification of this fraction by high pressure liquid chromatography (HPLC) led to the isolation of two natural product, identified as englerin A (1) and B (2). The same technique was applied for the isolation of englerin A and B from the extracts of the stem bark of Phyllanthus engleri. Also with this method the isolation of englerin A and B was achieved, although with lower yield.
5
2.2 Characterization of (-)-englerin A
The structures englerin A (1) and B (2) were elucidated by 1D and 2D NMR techniques by Beutler et al.[1] However, the absolute configuration of both the englerins could not be elucidated. The absolute configuration of englerin A was established a year later by the work of Christmann et al.[9] They synthesized the (+)-englerin A analogue, for which the absolute configuration was established by an X-ray structure of the employed (+)-englerin A precursor. Consequently, the optical rotation of (+)-englerin A showed an opposite value to that of the natural product (figure 4).
H Me H O O OH O O Ph O (-)-englerin A H Me H O O OH Me O O Ph O (+)-englerin A [α]20= -61 [α]20= +51
6
3. Total synthesis of englerin A
This section covers the literature regarding the total synthesis of englerin A. Various metal catalysed and metal-free key reactions were hereby employed to access the non-polar polycyclic core, a key structure forming the carbon backbone of englerin A. The following chapter will be subcategorized according to the method that implements these key reactions to access the core structure of 1.
3.1 Access to the non-polar polycyclic core by a 2-step Michael addition and samarium iodide radical cyclization.
The group of Chain discovered a short synthetic route for the preparation of (-)-englerin A.[22] The key step for the synthesis of englerin A was the formation of the tricyclic keto-alcohol structure. The carbonyl and keto-alcohol moieties in this compound are placed in a 1,4- and a 1,5- relationship, enabling the corresponding synthesis to be formed from a carbonyl-enabled cyclization sequence between an α,β-unsaturated aldehyde 17 and a 3-furanone 5 (scheme 1). O CH3 O O 17 5 + O O iPr H3C H H CH3 O H O OH iPr H3C H H CH3 O O H O OH c 1 c 2
Scheme 1: Carbonyl-enabled cyclization between an aldehyde and a ketone, forming 2 new C-C
bonds
c
1 andc
2 .The α,β-unsaturated aldehyde 17 is prepared by a ring-closing olefin metathesis reaction of
31 by a second generation Grubbs catalyst (scheme 2). The preparation of 3-furanone
started from the reaction of 3 with ethyl 2-chloropropanoate, which after intramolecular cyclization resulted in the formation of the 3-furanone 5.
O O CH3 O O Ru NMes MesN Ph PCy3 Cl Cl O 31 3 5 99% 2 steps 64% 17
Scheme 2: Preparation of the aldehyde and ketone intermediates.
The first step to form the core structure of englerin is by the Michael addition of the enolate of 5 with 2 to give the diastereoselective formation of compound 6, with a new C-C bond formed at c1 (scheme 3). The diastereoselectivity of this step is controlled by both the chelation of the aldehyde and enolate to the Li+ cation and by the C-4 methyl-bearing chiral center on 2 by pointing to a lateral direction due to its otherwise steric interaction with 5, preferring only one of the eight possible diastereomers with a d.r. of 2:1. Note that the stereochemical outcome of the corresponding reaction is significantly sensitive towards
7 additives that may disrupt the chelation of the Michael addition intermediate I-A, as was the case for LiCl and HMPA, which quite interestingly gave the other epimer of 6 at C-10. The second step for obtaining the core structure was by generating a radical centered at the aldehyde moiety of 6, which after engaging C-7 of the furanone should result in the formation of the second C-C bond c2. This reductive carbonyl-alkene cyclization was achieved with samarium(II) iodide forming the keto-alcohol 7. Also in this case, the corresponding C-C bond forming reaction is considerably sensitive towards additives, in which HMPA proved to be the only suitable additive. As an example, addition of LiCl afforded a pinacol-type carbonyl coupling product and alcohols afforded a reduced product, in which the aldehyde is transformed into a primary alcohol.
O CH3 O O O CH3 O H H CH3 H 5 then 17, 75%, dr 2:1 6 7 LDA O Me H O Me Li O O H O OH SmI2 HMPA 43% I-A d.r 2:1 1 3 5 7 10
Scheme 3: Michael addition and reductive carbonyl-alkene cyclization.
Next, 7 was esterified under Yamaguchi conditions with cinnamyc acid followed by the reduction of the C-9 ketone with NaBH4. Finally glycolic esterification afforded the target material (-)-englerin A (scheme 4). Remarkably, this convergent total synthesis of englerin A was performed in only eight steps with an overall yield of 20%, thereby being the shortest route towards the natural compound. Perhaps the most astonishing feature of this route is that the stereochemistry at three carbon centers (C-1,5 and 10) are established in one single step, making it a route towards (-)-englerin A with the least number of stereoselective steps.
7 (-)-englerin A O H O OH O H O O Ph O O OH 4 steps 63 %
Scheme 4: Synthesis of (-)-englerin A.
3.2 Access to the non-polar polycyclic core by various cycloaddition reactions. 3.2.1 Metal-free cycloaddition reactions
One of the earliest reports for the preparation of an (-)-englerin A precursor came from Nicolaou et al. [11] In 2010, they reported the total synthesis of racemic (±)-englerin A through
8 a metal-free [5 + 2] cycloaddition reaction, which generated the oxabicyclic core structure of the target material.[11]
Their asymmetric synthesis commenced from propargylic alcohol 82, which was converted into furan 83 in six steps (scheme 5). Achmatowicz rearrangement of 83 led to its corresponding dihydropyran derivative 84. Upon treatment with Mscl and base, the oxopyrilium species I-C was formed, which underwent [5 + 2] cycloaddition reaction with ethyl acrylate, furnishing oxabicyclic enone 85β with a d.r of 8:1 though with fair to poor yields. Subsequent 2-step hydrogenation reaction generated the hydroxy ketoester 86.
OBn OH 6 steps 52% OBn O OH mCPBA 84% OBn O O OH MsCl iPr2NEt CO2Et O CO2Et O CO2Et HO 1. PtO2, H2 2. Pd/C, H2 C9β:C9αca. 8:1 9 46% BnO O O H 82 83 84 85β 86 80% OBn O O I-C
Scheme 5: Hydroxy ketoester 86 synthesis.
The trans-fused 5m-ring was introduced in 5 steps. First, the ether moiety in 86 was dehydrated and the resulting terminal olefin was oxidized in Wacker conditions affording 87 (scheme 6). Intramolecular aldol condensation generated the oxatricyclic enone, which was stereoselectively reduced to give the C-6 allylic alcohol. Finally, stereoselective hydrogenation of the C4-C5 double bond afforded the trans-fused oxatryciclic compound 88.
O CO2Et HO 1. ArSeCN, nBu3P then mCPBA 2. Wacker [O] O CO2Et O H 1. KHMDS 2. NaBH4, CeCl3 . H2O 3. Crabtree cat. H2 O CO2Et HO H H 72% 67% 86 87 88 O O
Sheme 6: Preparation of trans-fused 88.
Subsequently the C-9 ester functionality of 88 was converted into a ketone through Weinreb ketone synthesis and subsequently transformed into hydroxy acetate 89 by oxidation with mCPBA (scheme 7). Subsequent esterification and deprotection reactions afforded the target material (±)-englerin A in a total amount of 21 steps with a total yield of 2% starting from 82. Even though the total yield is very poor and the target material is racemic, it should be mentioned that this total synthesis was the first being published.
9 O CO2Et HO H H 88 43% O OAc HO H H 89 1. iPrMgCl MeNH(OMe) . HCl 2. MeLi 3. mCPBA 4 steps 61% ( )-englerin A
Scheme 7: C-9 hydroxy acetylation and target material synthesis.
Another total synthesis report of 1 through a metal-free cycloaddition reaction was published in 2013 by the group of Y. Shang.[15] They designed a unified synthetic strategy for a divergent preparation of 1, 2, orientalol E and F, and oxyphyllol through an organocatalytic [4 + 3] cycloaddition reaction.
Focussing only on the preparation of compound 1, furan 113 and dienal 114 were prepared, following known protocol, in respectively three and one step(s).[16] Subsequently, [4 +3] cycloaddition catalyzed by chiral 117 generated the ether bridged bicyclic product 115a and its regioisomer 115b with moderate selectivity (115a/115b = 2.4:1; ee [%] = 67:82, see scheme 8). The stereoselective and regioselective outcome can be explained as followed. The reaction of dienal 114 with catalyst 117 results in the intermediate formation I-D. The sterically demanding TMS group of the substrate occupies the β-face (top face) of intermediate I-D and the less bulky benzyl- and tert-butyl groups of 117 cover the α-face (bottom face), thereby establishing the stereochemical outcome of the reaction. Intermediate I-D is approached by furan 113 from the less crowded α-face and the [4 + 3] cycloaddition reaction proceeds through an endo addition. However, because this substrate is not symmetrical, the furan substrate can coordinate in two ways, resulting in two transition states TS1 and TS2 with the latter being less favoured due to the steric interaction of the isopropyl group of the substrate with the tert-butyl group of the catalyst. Consequently, subsequent cycloaddition results primarily in the formation of 115a and its regioisomer 115b.
10 O OTMS CHO + NH N O tBu Ph TFA 63% O O OHC O O OHC + N N O O Si N N O O Si O I-D TS1 N N O O Si O TS2 113 α-face approach endoaddition 113 114 115a 115b 117
Scheme 8: Organocatalytic [4 + 3] cycloaddition reaction of furan 113 with dienal 114.
Proceeding the synthetic route, the ether bridged guaiane carbon framework was prepared starting from the alkylation of 115a with vinylmagnesium bromide, followed by allyl alcohol acylation to give compound 116 (scheme 9). The relative stereochemistry of the latter compound was determined by X-ray analysis, showing i.a. the C-2 carbon to be R configured (figure 5). Subsequently, Pdo catalyzed reductive deacetoxylation generated the C-1 butylene substituted analogue. A two steps intramolecular Heck reaction of this intermediate afforded the oxatricyclic core structure 117.
O OHC H 115a MgBr 1. CeCl3 O 2. Ac2O, TEAD MAP O O 116 AcO 1. Pdo HCOONH4 2. LDA, PhNTf2 3. [Pd(Ph3P)4 Et3N O H 117 88% (R)
Scheme 9: Construction of oxatricyclic 117.
11 With the carbon framework of 1 formed, the first introduction of a hydroxyl group was established by selective epoxidation of the more electron rich trisubstituted double bond followed by SN2 type reduction. This afforded the C-6 hydroxy substituted 118 (scheme 10). It is worthy of note that oxidation and reduction of 118 with Hg(OTFA)2 and NaBH4 afforded diol 32, a precursor compound used for the total synthesis of 1 by Ma et al. (scheme 14, section 3.2), thereby forming a formal synthesis.[26]
O H 117 1. mCPBA 2. DIBAL-H O H 118 HO 34% from 116 Hg(OTFA)2 NaBH4, 38% O H 32 HO OH
Scheme 10: Construction of englerin A precursor diol 32.
Proceeding with the plan, to acquire the desired stereocenters at the C-6 alcohol, the aforementioned center was epimerized by a two-step oxidation and reduction reaction yielding 119 (scheme 11). Subsequently, the introduction of a second hydroxyl moiety was effectuated by a regioselective and stereoselective hydroboration-oxidation manipulation of the sterically less demanding double bond. This afforded diol 120 with 12:1 regioselectivity (r.s.). Subsequent hydrogenation using Pfaltz catalyst afforded the trans-fused ring 121, though with low diastereoselectivity (d.r. 1:2.5). This insufficient outcome was attributed to the directing effect of the C-9 hydroxyl moiety. Finally, a three-step esterification and deprotection step generated the target compound 1. This total synthesis required 15 steps starting from 113 and 114 in 3% overall yield.
O H 118 HO 1. TPAP, NMO 2. NaBH4 70% O H 119 HO 9-BBN, H2O2 r.s. 12:1, 89% O H 120 HO OH Pfaltz cat., H2 d.r. 1:2.5 93% O H 121 HO OH 32% 3 steps 1
12
3.2.2 Gold catalyzed cycloaddition reactions
Another widely used method for the formation of the core structure of englerin A is by metal catalyzed cycloaddition reactions. One such example is the study conducted by the group of Echavarren. They developed an Au(I) complex[24] that stereoselectively catalyzed a [2+2+2] alkyne/alkene/carbonyl cycloaddition reaction from 1,6-enynes to form the key oxatricylic diol, which was used as precursor for the synthesis of (-)-englerin A as illustrated in scheme 12.[25] RO OR' O [AuL]+ O H RO H OR' AuL H RO AuL OR' O - [AuL] + O OR' H RO (-)-englerin A
Scheme 12: Mechanism of the Au(I) catalyzed [2+2+2] cycloaddition.
In order to perform this key step, the 1,6-enyne was prepared from the cheap starting material geraniol (scheme 13). This 8 step synthesis started from the Sharpless asymmetric epoxidation of 14, which is followed by an apple reaction. Subsequent treatment with nBuLi afforded the propargylic alcohol 15. After protecting the hydroxyl group the stereoselective Sharpless dihydroxylation afforded the 1,2-diol (95:5 e.r.), which is then oxidatively cleaved to form the corresponding aldehyde product 16. The Wittig reaction of the ylide 25 with 16 afforded exclusively the (E)-enal, which then underwent a stereoselective Denmark aldol reaction with 26 in the presence 24 to eventually form the 1,6-enyne 18 (> 14:1 d.r.). Surprisingly, the following Au(I) catalyzed [2+2+2] cycloaddition reaction of the enyne worked best for the unprotected aldol 18. Thus treating the corresponding compound with [IPrAuNCPh]SbF6 resulted in the formation of the key oxatricylic diol 19a as a single diastereomer. Subsequent deprotection and chemo- and regioselective protection resulted in the formation of 19c, which showed >99% ee.
13 HO OH OTES O O Ph3P OSiCl3 O OR' H RO 19a:R = TES, R' = H 19b:R = H, R' = TBS 14 15 16 25 26 N Me P Me N O N Ph H Ph H 24 [IPrAuNCPh]SbF6 3 mol%, 58% 2 steps 89% 3 steps 81 % 3 steps 97 % OTES O OH 18 69 % 1. 25, 2. 26,24
Scheme 13: Synthesis of 1,6-enyne and Au catalyzed cycloaddition.
The isomerization of 19b into 21, in which the C-11 alcohol is removed and again introduced into the C-6 position pointing to the same face of the ring as previous, was executed in two steps (scheme 14). The 3o alcohol was oxidized into an epoxy alcohol with CrO3 and subsequently deoxygenated with WCl6 affording 21. Hydrogenating the corresponding compound with Pfaltz’s Ir(I) catalyst 24 afforded 22, which after a series of esterification and deprotection afforded the natural product 1. The total synthesis of englerin A employing the Au catalyzed cycloaddition reaction required 18 steps with an overall yield of 7%. Compared to the work of Chain, this method required more than twice the amount of reaction steps and an additional stereoselective reaction step, thus lowering the efficiency in yield and complicating the selectivity. However, it does provide access to intermediate 22 in satisfying reaction steps, which can be used to prepare a variety of englerin analogues by varying the substituents on both the hydroxyl groups.
O OTBS H O OH O OTBS H OH O OTBS H OH H 20 21 22 19b Ir Py Cy3P 24 BArF CrO3 (2,5-dimethylpyrazole) 3 eq. 73% WCl6, nBuLi H2(80 bar) quant. 1:1 d.r 82% 4 steps 79 % 1
Scheme 14: Oxydation, deoxydation and hydrogenation reactions for the preparation of the natural
14 A similar strategy was utilized by Ma et al. in the same year as the work of Echavarren, of which the key step was also the gold catalyzed [2+2+2] cycloaddition from a 1,6-enyne (scheme 15).[26] However, their substrate lacked an additional hydroxyl moiety, thereby preventing the formation of a functional group at C-4 as was observed in 19a/b.
Their strategy started with transforming the aldehyde functionality of the starting material (R)-citronellal 26 into a propargyl moiety and oxidizing a methyl group into an aldehyde to give 27 in 3 steps (scheme 12). Subsequent (-)-Dip-chloride mediated enantioselective aldol reaction with the enolate isomer of 3-methyl-2-butanone gave the 1,6-enyne 29 in 2 steps (d.r ≈ 4:1). For the gold catalysed [2+2+2] cycloaddition, compound 29 was protected with several protecting groups (TBS, TES and Me), for which almost all of them gave the by-product 51 when treated with AuCl or [Au(PPh3)Cl]/AgSbF6. It was found that only the deprotected derivative 29 gave the oxatricyclic compound 30 when treated with AuCl and as a single diastereomer. Note that using [Au(PPh3)Cl]/AgSbF6 as the catalyst for the unprotected substrate gave a lower yield.
O Me Me CHO O OH 26 29 O OH H 30 O 1. P(OPh)3, Br2, Et3N 2. 1. tBuOK, 18-cr-6 (-)-Ipc2BCl, Et3N 2. MeOH, pH 7, H2O2 AuCl 10 mol% 1. 3. SeO2, TBHP salicylic acid 27 (R)-citronellal
Scheme 15: Synthesis of oxatricyclic core structure from (R)-citronellal.
A proposed explanation for this divergent behavior of the reaction of protected and unprotected derivatives of 29 with Au(I) complexes is illustrated in scheme 16 below. When treating the 1,6-enynes with Au(I), two anti-cyclopropyl gold carbene intermediates A and B may form. In intermediate A, the methyl substituent on the 5m-ring is trans to the sterically crowded cyclopropyl ring, in contrast to intermediate B, thus A being the more stable isomer. This major product reacts further to form the single diastereomeric product 30 in a highly stereoselective manner for R = H. Note that the observed stereoselectivity for generating compound 19a (scheme 10) may also be rationalized with this proposed mechanism. However, when the hydroxyl group is protected (R ≠ H), the steric hindrance induced by the protecting group prevents a nucleophilic attack of the keto moiety at the cyclopropanyl ring, thus hinders the formation of C. instead, C-C bond cleavage of A results in the formation of the by-product 51.
15 Me O OR Me [AuL] Me [AuL] O OR H Me [AuL] O OR H + Me [AuL] O OH H Me H [LAu] O OH 30 for R = H Me [AuL] O OH H 51 29 for R = H A B C OR O Me
Scheme 16: Proposed reaction mechanism for the formation of x and 30.
After unsuccessful attempts with [Cp2TiCl2]/Zn for the preparation of the trans-fused ring from the prilezhaev epoxide product of 30, a different route using 4 steps was explored (scheme 17). Treatment of the epoxide analogue of 30 with camphorsulfonic acid afforded
32 (scheme 14). Inversing the stereochemistry of both the C-6 and C-9 hydroxyl carbons by
oxidizing the alcohols with TPAP and subsequently reducing the ketones with NaBH4 afforded diol 33. Finally, subjecting 33 to H2 with Raney Ni afforded the trans-fused ring 34. To go towards (-)-englerin A, the less hindered C-9 hydroxyl group is selectively oxidized to a ketone with DMP affording 35. This step was necessary to regioselectively esterificate the 6 hydroxyl group with cinnamic acid without employing protecting groups. Thereafter, the C-9 ketone was reduced and subsequently esterified with cesium glycolate to give the end product 1 in 3 steps. Remarkably, this 15 step synthesis with an overall yield of 8.1 % did not require any protecting groups. This, in contrast to the similar strategy employed by Echavarren, thereby reducing the total number of steps.
O OH H 32 HO O H 33 OH HO O H OH HO H 34 1. mCPBA 2. CSA 1. TPAP NMO 2. NaBH4 Raney Ni H2(90 atm) 3 steps 1 1. Dess-Martin periodinane, NaHCO3 O H HO H 35 O O OH H 30
16
3.2.3 Rhodium catalyzed cycloaddition reaction
Another total synthesis report came from the group of Hashimoto et al.[10] They developed a 1,3-dipolar cycloaddition reaction between an α-diazo-β-ketoester and a vinyl ether dipolarophile catalyzed by a chiral dirhodium(II) complex [Rh2(S-TCPTTL)4] 71 that constructed the oxabicyclo[3.2.1]octane carbon backbone of 1.
Their synthetic strategy commenced from succinic anhydride, which was converted into α-diazo-β-ketoester 69 in 5 steps (scheme 18). When reacted with dirhodium complex 71, the 6 membered carbonyl ylide derived from 69 undergoes 1,3-dipolar cycloaddition with electron rich ethyl vinyl ether 70 through intermediate I-B. The LUMO of the carbonyl ylide interacts with the HOMO of the dipolarophile, favouring the exo-cycloadduct 72 with high enantioselectivity (95% ee) and diastereoselectivity (exo/endo = 87:13).
O O O 5 steps 53 % O O CO2tBu N2 69 OEt O OEt tBuO2C 72 O 70 Rh O O Rh N H Cl Cl Cl Cl O O O O CHO2tBu Rh OEt LUMO HOMO 71 I-B 76%
Scheme 18: Rh(II) catalyzed 1,3-cycloaddition reaction.
The introduction of a 5m-ring to the oxabicylic core of 72 would be established by an intramolecular aldol condensation reaction. In order to perform this, 72 had to be converted into a diketone. This started with introducing an acetal-protected ketone to C-1 in 3 steps forming alcohol 73 (scheme 19). Subsequently the ester moiety was reduced to an ether functionality in 2 steps affording 74, for which its necessity will be explained in further steps. Oxidation of the allylic alcohol using PCC generated the α,β-enone 75. Subsequent hydrogenation of the C1-C5 double bond generated ketone 76 with the desired epimer at C-1 as a singler diasteromer, similar to that of the target material C-1.
17 O OEt tBuO2C 72 O 3 steps 59 % O OEt CO2tBu 73 HO O O 2 steps 83 % O OEt 74 HO O O OTs PCC NaOAc 86% O OEt 75 O O OTs O H2, Pd/C 99% O OEt 76 O O OTs O O O EtO 1 5 10 15 TsO H 1 O O
Scheme 19: Diketone 76 synthesis.
If the ester moiety would not be converted into an ether functionality, hydrogenation of the derivative of 72 (77) would generate the ketone 78 with the wrong epimer at C-1 (scheme 20). This divergent stereochemical outcome was rationalized by molecular mechanics calculations, which showed that the exo face of the double bond was sterically blocked by the bulky tBu-ester moiety. Thus hydrogenation would be executed from the endo face (d.r = 2:1), resulting in the observed undesired epimer at C-1. However, this was not the case for
75, in which the exo face was indeed accessible for hydrogenation.
O O O O EtO O O1 5 10 15 77 H2, Pd/C O OEt CO2tBu 78 O O O H 1 H -H
-Scheme 20: Hydrogenation of tert-butyl ester substituted ketone 77.
Following the synthetic route, the acetal protecting group of 76 was removed followed by intramolecular aldol condensation of the resulting diketone, thereby constructing the oxatryclic core of 1. Subsequent stereoselective reduction generated 79 as a single isomer (scheme 21). This was followed by hydroxy group-directed hydrogenation, affording 80 also as a single isomer. With this conversion, the fused ring adopted the correct epimers at the junction similar to that of 1. The corresponding compound was protected and followed by chemoselective oxidative conversion of the ethyl ether functionality to its acetate derivative by ruthenium tetraoxide-catalyzed oxidation. This acetyl moiety together with the tosylate group were subsequently removed with LiBET3H to give its reduced analogue 81. Sequential 4-step esterification and deprotection reactions furnished the end product 1. Following this
18 synthetic plan, natural product 1 was prepared in 25 steps with an overall yield 5.2%, thereby consisting of a strategy with the most amount of steps. Additionally, compared to the gold catalyzed cycloaddition reactions discussed above, another drawback of this strategy is the inability to generate the tricyclic structure in one step from a linear molecule, thereby contributing to the large amount of steps required for the synthesis of 1.
O OEt 76 O O OTs O H 3 steps 82% O OEt 79 OTs H HO H2, Pd/C 67 % O OEt 80 OTs H OH H 1. TESOTf, 2,6-lutidine 2. RuCl3, NaIO4, pH = 7 3. LiBEt3H O OH 81 H TESO H 74% 4 steps 88% 1
Scheme 21: Intramolecular aldol condensation, reduction reactions and preparation of 1. 3.2.4 Platinum catalyzed cycloaddition reaction
It was illustrated that the 1,3-dipolar cycloaddition mentioned above was able to introduce a functionality to the C-9 carbon, which was later on converted into a C-9 alcohol. However, it required a large amount of steps. One of the reasons being the ethyl ether functionality at C-9 and the ester moiety at C-10, both of which had to be converted by a number of manipulations in subsequent steps. To reduce the amount of steps, Iwasawa et al. developed a platinum catalyzed [3 + 2] cycloaddition reaction of a Pt-carbene ylides with electron rich alkenes[28] that directly introduced a methyl moiety at C-10 and a protected alcohol at C-9.[29]
Their strategy started by subjecting the starting material γ,δ-ynone 49 to the cationic platinum(II) complex PtCl2(P(m-tolyl3)2 in the presence of AgSbF6, by which the Pt-carbene ylide intermediate I-A is formed (scheme 22). This affects the LUMO of the ylide moiety due to the intrinsic electron-withdrawing property of the Pt-carbene moiety, thereby decreasing the LUMO energy level. This initiates an exo selective [3 +2] cycloaddition with an electron rich benzyl vinyl ether. A subsequent 1,2-H shift of the formed ocabicyclic carbene intermediate generates the core englerin exo-cycloadduct 50 with a protected hydroxyl functionality at C-9 as a single isomer.
19 O OBn O [Pt] O [Pt] OBn + [Pt] OBn O OBn O [3 + 2] cycloaddition 49 50 I-A - [Pt] [Pt] PtCl2(P(m-tolyl3)2) 5 mol% AgSbF6 5 mol%, 90% 1,2-H shift
Scheme 22: Pt(II) catalyzed [3 + 2] cycloaddition.
In order to generate the oxatricyclic core of 1 by introducing a 5m-ring to 50, the corresponding compound was transformed into an enone. This was performed in 3 steps by stereoselective epoxidation (d.r 15:1), epoxide ring opening and oxidation to eventually form enone 52 (scheme 23). Subsequently, a butenyl group was introduced to the carbonyl carbon of the enone moiety by a 1,2-addition reaction. This gave predominantly the α isomer of 53 with a diastereomeric ratio of 3:2.
OBn O OBn O O OBn O OH 50 52 53 1. mCPBA 2. nBuLi 4-iodobut-1-ene sBuLi α:β = 60:39 3. PDC, quant 93% (3 steps)
Scheme 23: Enone formation followed by the introduction of a butenyl moeity.
Proceeding the synthetic strategy, oxidative 1,3-allylic rearrangement of 53α followed by regioselective copper mediated reduction of the alkene in the enone moiety and Wacker oxidation of the terminal alkene of the butenyl moiety afforded the diketone 56 (scheme 24). Subsequent intramolecular aldol condensation generated the oxatricyclic compound 57.
OBn O OH OBn O O O H OBn O O H 53α 56 57 NaOEt quant. 1. PCC
2. CuI 30 mol%, MeLi, DIBAL 3. PdCl220 mol%, CuCl, O2
57%
Scheme 24: Preparation of a diketone and aldol condensation reaction.
The following steps comprised of two stereoselective reduction reactions: the first step was the reduction of the ketone to an alcohol using NaBH4 and CeCl3 and the second step was the hydrogenation of the tetra-substituted olefin using Raney Ni giving 59, though under
20 harsh reaction conditions (90 atm, 100 oC; scheme 25). Worthy of note is that the C-9 alcohol was deprotected alongside the olefin hydrogenation, thus saving one additional deprotection step. Finally, a three step regioselective esterification and deprotection furnished the natural product 1. Starting from 49, this strategy needed 14 steps with an overall yield of 16% OBn O O H OH O H HO H 57 59 1. NaBH4, CeCl3.7H2O
2. Raney Ni, H2 3 steps
82%
69%
1
Scheme 25: Dual reduction and subsequent synthesis of target molecule 1. 3.3 Access to the non-polar polycyclic core by metathesis reactions.
Grubbs II catalyzed intramolecular ring-closing metathesis
Previous reports showed 2-step Michael addition and samarium iodide radical cyclization and various metal(-free) initiated cycloaddition reactions for the preparation of the polycyclic core of 1. However, another strategy for the preparation of the (-)-englerin A backbone is by ring-closing metathesis reaction. One such example was illustrated by Christmann et al. for the synthesis of (+)-englerin A for the determination of the absolute configuration of its epimer 1.[9] Their strategy commenced from the starting material cis,trans-nepatalactone 3, for which the carbon framework showed close similarities with the target material. In particular the stereocenters at C-1 and C-4. The key steps of their strategy consisted of an epoxylactone contraction followed by a diastereoselective Barbier reaction and a ruthenium catalysed intramolecular ring-closing metathesis (RCM) for the construction of the core structure of englerin A.
Their initial plan started with the use of trans,cis-nepatalactone starting material with not 2 but 3 similar stereocenters (C-1, C4 and C-5) compared with the target material. However, epoxidation with mCPBA afforded the epoxylactone product with the desired configuration at C-10 in unsatisfactory yields (d.r. 1:7) (scheme 1 A). Thus, they employed cis,trans-nepatalactone 3 to form the epoxylatctone 4 upon treatment with mCPBA with a d.r of 3:2. Epoxylacton rearrangement of 4 using NaOMe gave formyl lactone 5, which after a diastereoselective Barbier reaction with allyl bromide 6 provided 7 in 93% yield with a d.r of 5:1 (scheme 26 B). An X-ray structure of 7 confirmed both the absolute and relative configuration of this compound (figure 6).
21 O Me H H O cis,trans -nepetalactone (3) mCPBA O Me H H O O d.r. 3:2 4 MeO O H H Me O CHO 90% 55% Br 6 5 Zn, sonication aq. NH4Cl 93%, d.r 5:1 O H H Me O 7 HO 1 3 5 10 O Me H O trans,cis -nepetalactone 10 H mCPBA 10% O Me H O O d.r. 1:7 H A B
Scheme 26: two different attempts for the preparation of homoallylic alcohol 7 starting from
trans,cis- and cis,trans- nepetalactone (A and B resp.).
Figure 6: X-ray structure of two molecules of 7.
The synthesis proceeded with reducing the γ-lacton moiety to give a triol, for which then the vicinal hydroxyl groups were protected by acetylation in order to transform the primary alcohol to an aldehyde by IBX oxidation. Subsequently, the C-5 proton was epimerized to a trans-fused bicyclic structure with DBU (d.r 3:1) in order to establish the same configuration as the target molecule. Finaly, the aldehyde was transformed to diene 8 by a Wittig olefination (scheme 27). Subjecting the corresponding molecule to a Grubbs II catalyst afforded the ring-closing product guaiane 9 with an E-configured double bond.
O H H Me O 7 HO 5 steps H Me 8 H Me O O Ru N N Mes Mes Cl Cl PCy3Ph 20 mol% H Me 9 H Me O O 99% 59%
Scheme 27: Diene formation and ring-closing metathesis.
After a 3 step deprotection, esterification and epoxidation reaction, the epoxide product (d.r. 2.3:1) transformed spontaneously into the oxatricyclic ether 10 by transannular epoxide
22 opening (scheme 28). Finally, subsequent esterification and deprotection afforded (+)-englerin A. Optical rotation measurements showed a value of [α]20 = + 51 (c = 0.58, MeOH), which is the opposite to the value measured for the natural product ([α]20 = - 61; c = 0.13, MeOH), thereby proving the absolute configuration of the natural product to be (-)-englerin A. H Me 9 H Me O O H Me 10 H O O OTBS Me O HO H Me H O O OH Me O O Ph O (+)-englerin A 4 steps 2 steps 71% 55%
Scheme 28: Synthesis of (+)-englerin A.
In the report described above the total synthesis of (+)-englerin A by Christmann et al. was discussed for the determination of the absolute configuration of 1.[9] This report was also the first study on the total synthesis of englerin A. However, two years later M. Christmann also reported the total synthesis of the biologically active (-)-englerin A 1 and its corresponding analogues for a Structure-Acvitivity Relationship (SAR) study, which proceeded through a RCM step.[20] In this section, the discussion will solely be focused on the total synthesis of 1. The SAR study of this report will be discussed in chapter 5.
As previously shown, the synthesis of (+)-englerin A commenced from (+)-nepatalactone. Self-evidently, 1 can be prepared from (-)-nepatalactone. However, in contrast to (+)-nepatalactone, which is commercially available, (-)-nepatalactone had to be prepared. Following Schreiber’s protocol[21], the latter compound was prepared from (+)-citronellal in four steps. Subsequently, following ref. 9 (see the previous report discussed above), aldehyde 147 was synthesized (scheme 29). Barbier-type reaction of 147 with three different allyl bromides afforded 148a-c. This step was necessary for the SAR study (see chapter 5), in which the influence of the isopropyl group would be probed and to investigate whether modifications at C-7 would enhance the biological activity. The dialkenes 149a-c needed for the RCM step were prepared from 148a-c through a reduction and again following ref. 9. Subsequently, treatment with Grubbs II afforded 150a-c.
23 CHO 4 steps 22% ref 21 O (-)-nepatalactone H H O H H O O CHO ref 9 Br R Zn, sonication aq. NH4Cl H H O O R OH (+)-citronellal 147 148 a: R = iPr (93%, d.r. 5:1) b: R = Et (66%, d.r. 6:1) c: R = Me (89%, d.r. 7:1) 1. LiAlH4 2. ref 9 H 149a-c H R O O Grubbs II 150 a: 99%, b: 88% c: 99% O O H H R
Scheme 29: preparation of dialkenes 149a-c and subsequent RCM reaction.
Focussing only on 150a, the introduction of the oxo-bridge was established in four steps. The acetonide moiety was removed and the secondary alcohol protected prior to epoxidation, affording 151 (scheme 30). These manipulation were necessary for two reasons: acetonide removal of the epoxidized product led to degradation and 2o alcohol protection was needed to facilitate the desired facial selectivity of the epoxidation step. Proceeding the synthetic strategy, exposure of 151 to elevated temperature afforded the trans-annular oxatricyclic product 152. This compound is a key intermediate for the preparation of various englerin analogues for the SAR and antiproliterative optimization study, discussed in a later chapter. Finally, a four-step esterification and deprotection sequence of 152 afforded 1. Again, the crystal structure of 1 affirmed its absolute configuration.
O O H H 150a 2 steps H H 151 OTBS OH 1. mCPBA d.r 5.4 : 1 H H 152 OTBS O 2. ∆ HO 91% 4 steps 82% 1
Scheme 30: Synthesis of 1 through among others epoxidized and trans-annulated product 152. The reports above demonstrated a Barbier type reaction and the key RCM reaction for the construction of the bicyclic core structure needed for the synthesis of englerin A. Likewise, this method was used by the group of P. Metz for the total synthesis of 1.[30]
The group employed a partial chiral pool setup, starting from (-)-isopulegol. This starting material was converted into (R)-photocitral by the oxidizing agent lead tetraacetate followed by treatment with Pdo(PPh
3)4 to form (R)-photocitral as a racemic mixture (scheme 31). This material has three stereogenic centers possessing the same absolute configuration as that on the target material, thereby reducing the introduction of chiral centra’s from seven to four. The Reformatsky reaction of the corresponding compound with bromo ester 70 afforded the dialkene 61 as a diastereomeric mixture. Note that separation of the
24 diastereomers is not needed due to transformation into sp2-hybridized centers in the later
steps OH CHO Br CO2Et HO (R)-photocitral A 61 (-)-isopulegol 70 Zn, 86% 1. Pb(OAc)4, CaCO3 2. Pd(PPh3)4, pyrrolidine Et3N, d.r = 47:43:10 31% EtO O
Scheme 31: Preparation of dialkene 61.
With the dialkene 61 formed, the core oxabicyclic structure 62 is formed by intramolecular ring-closing metathesis of the former compound using a second generation Grubbs catalyst (scheme 32). Note that this metathesis reaction is far more efficient than that employed for the synthesis of 9, which required 20 times the amount of catalyst material (scheme 27). Subsequently the ester moiety in 62 was transformed in an acetyl group (for which the latter is needed to form the isopropyl moiety of the target material in later steps) by a 2 step weinreb ketone synthesis and followed by enone transformation through an additional 2 step reaction process eventually affording 64.
H H OH EtO O H H O 62 64 1 mol% 99% 1. MeONHMe.HCl iPrMgCl 2. MeLi 3. MsCl, Et3N 4. DBU 62% Ru N N Mes Mes Cl Cl PCy3Ph OH 61 H H EtO O
Scheme 32: Intramolecular ring-closing metathesis and transformation into an enone.
The ether bridge is introduced by a 4 step reaction process. The first step comprised of a chemo-and diastereoselective epoxidized of the enone alkene moiety with H2O2 (scheme 33). The chemoselectivity comes from the fact that the enone double bond is more electron deficient due to the neighboring electron withdrawing carbonyl moiety, thus being more prone to epoxidation under weitz-Scheffer conditions. Subsequent sharpless dihydroxylation of the second alkene moiety using OsO4/NMO resulted in the formation of cis-diol 66 (although with moderate diastereoselectivity). After esterification of the C-9 alcohol, treatment with Ph3P=CH2 and then with HCl resulted in concomitant Wittig reaction and acid mediated epoxide ring opening, in which the C-10 hydroxyl group attacks the C-7 C(CH3)2 carbon from the same face of the molecule, resulting in the diastereoselective formation of ether bridged compound 68. With the oxatricyclic core structure formed, alkene hydrogenation followed by a 2 step esterification process generated the end product
1. With only 14 steps, this work together with the work of Iwasawa et al.[d] is the shortest route reported after the route developed by Chain et al.[v]
25 H H O H H O O OH O 68 64 H H O OH OH O 66 1. NaOH, H2O2 2. Ph3P=CH2, then 2N HCL TBSO Cl O 1. pyridine 3 steps 1 TBSO 2. K2OsO4, NMO d.r 2: 1 88% 65% 85%
Scheme 33: Introduction of an ether bridge by a 4 step oxidation, dihydroxylation, esterification and
Wittig reaction process and target material synthesis.
A closely related study was reported by S. Hatakeyama. They also constructed a cyclopentane structure with a dialkene moiety, from which after ring closing metathesis afforded the bicyclic carbon framework.[12]
However, for the construction of the heavily substituted 5m-ring, the group started from ketone 93, in which only one of the seven chiral centers were embedded. This starting material was converted into 94 through a three-step Baeyer-Villiger oxidation, DIBAL reduction and Wittig reaction sequence (scheme 34). Subsequent tosylation, ester reduction and cyanation afforded compound 95. Stereoselective Sharpless epoxidation of the double bond (.95% de) followed by alcohol protection afforded the 5m-ring precursor epoxynitrile
96. O OH CO2Et 3 steps 1. TsCl2. DIBAL-H 3. NaCN CN OH 1. (-)-Det, Ti(OiPr)4 TBHP 2. TBSCl CN OTBS O 82 % 81% 93% 93 94 96 95
Scheme 34: An eight-step epoxinytrile synthesis.
With the epoxynitrile formed, treatment with LiHMDS resulted in epoxynitrile cyclization, forming 97 and its C-1 epimer with d.r = 8:1 (scheme 35). The stereochemical outcome was rationalized by proposing the formation of two transition states T-1 and T-2, for which the former is preferred due to the absence of steric interactions. This T-1 accounts for the observed epimer in 97. To introduce the first alkene moiety, 97 was transformed in the corresponding alkene by a three-step manipulation involving 3o hydroxyl protection, nitrile reduction into an aldehyde and Wittig reaction. Subsequent deprotection and Swern oxidation furnished keto-aldehyde 99.
26 CN OTBS O 96 LiHMDS CN OTBS 97 H OH 1 87% H Me Me OTBS H C N O H Me Me OTBS O H C N T-1 favoured T-2 disfavoured proposed transition state
1. MOMCl 2. DIBAL-H 3. Ph3P=CH2 98 H OMOM 2 steps 91% CHO 99 H OMOM OTBS +
Scheme 35: Keto-aldehyde 99 synthesis through an epoxynitrile cyclization.
Barbier-type allylation of 99 with bromo-alkene 10 was established with Indium powder. This furnished the heavily substituted 5m-ring 101 together with its C-9 epimer 102 with d.r = 8:1 (scheme 36). The high diastereoselectivity was rationalized by the attack of the allylindium intermediate to the carbonyl moiety following a Felkin-Ahn model, as illustrated in scheme 36. For the separation of these two diastereomers, they were deprotected with HCl, yielding chromatographically separable diol 103 and its corresponding C-9 epimer.
CHO 99 H OMOM Br 100 In (s) 95% O H Me MOMO InBr d.r 8:1 + H OH OH 103 + epimer (11%) 84% 102 H OMOM OH 9 101 H OMOM OH 9 HCl
Scheme 36: Indium mediated Barbier-type allylation.
Proceeding the synthetic plan, diol 103 was converted its cyclic carbonate analogue, after which Grubbs II catalyzed ring closing metathesis generated bicyclic 104 (scheme 37). Subsequently diastereoselective epoxidation by the reaction with magnesium monoperoxyphthalate (MMP) resulted in (separable) 105. Removal of the cyclic carbonate moiety (which went without any difficuly in contrast to its removal by Christmann et al.[9]) followed by a 4-step sequence involving esterifications and thermally initiated transannular cyclization furnished the target compound 1. In contrast to the similar report of Metz et al.[30] discussed above, this strategy needed 10 extra steps, though with a higher overall yield of 14%.
27 H OH OH H H O O O 1. triphosgene 2. Grubbs 2nd 0.89% MMPP 81% H H O O O O 5 steps 60% 1 103 104 105 d.r 10:1 Scheme 37: Ring-closing metathesis and target material synthesis
A similar report to that of Metz et al.[30] was published by Z. Shen[14], who also started from a 6m-ring, after which manipulations were performed from its corresponding ring-contracted 5m-ring. This, in contrast to the synthetic strategy of Hatakeyama et al. described above, who performed the manipulations on the linear, ring opening product of the 5m-ring starting material, thereby requiring an additional step to obtain the cyclopentane ring needed for the key RCM step.[12]
The group of Z. Shen started form R-(-)-carvone, which was converted into aldehyde 106 following a modified protocol.[13] Note that the protocol itself follows a ten-step route towards compound 106, but the amount of steps used by Z. Shen cannot be deduced from the paper due to their modified approach, which was not elucidated. To obtain dialkene 108 from aldehyde 106, pinacol coupling reaction with ketone 109 was tested, which gave an undesirable result (scheme 38). Attributing this to the bulky isopropyl group attached to the carbonyl moiety, resulting in steric hindrance with 106, another approach was performed in which the bulky isopropyl group was directly introduced to the aldehyde moiety of 106. This three-step approach started from the treatment with 1,3-dithiane 110 followed by dithiane deprotection, resulting in 107 with low diastereoselectivity (scheme 38). However, the diastereomers were chromatographically separable, affording pure 107, after which Grignard reaction with allymagnesium bromide furnished dialkene 108.
H O H S S 1. nBuLi, 2. HgCl2, CdCO3 H H O d.r 3:1 BrMg H H HO HO 106 88% 107 108 92% OH O 109 110 O ref 13 R-(-)-carvone
Scheme 38: The steps involved in the synthesis of dialkene 108.
Proceeding with the plan, dialkene 108 was subjected to Grubbs II, affording 111 in high yields (scheme 39). Subsequently, treatment with NaIO4 resulted both in a C-9 iodide substituent and in the formation of the ether bridge, thereby furnishing the oxatricyclic core structure of 1. X-ray analysis of this iodo-compound 112 confirmed its absolute configuration
28 (figure 7). Finally, following a five-step sequence of esterification, acetylation and deprotection reactions, compound 1 was generated with a total yield of 11% starting from aldehyde 106. One advantage of such a strategy is the preparation of intermediate 112. The iodine moiety in 112 can easily be substituted with a wide variety of different (non-ester based) substituents. This in turn can be used for the preparation of a large variety of C9-substituted analogues of 1. H H HO 108 NaIO4 Grubbs II 94% OH H 111 H HO 94% O H 112 H HO I 5 steps 63% 1
Scheme 39: Synthesis of 1 through ring-closing metathesis and halogenation.
29
4. Formal synthesis of englerin A
The following chapter discusses reports on the formal synthesis of (-)-englerin A through various metal catalysed key reactions for the construction of the cyclic framework of the corresponding natural product. The chapter will be subcategorized according to the type of strategy employed to access the key cyclic framework of the englerin precursor and will be discussed from here on.
4.1 Precursor preparation derived from various cycloaddition reactions.
It was already mentioned in section 3.2.1 that the group of Nicolaou developed a metal-free [5 + 2] cycloaddition reaction for the total synthesis of racemic (±)-englerin A.[11] However, in the same paper they also demonstrated a formal asymmetric total synthesis of 1 through the preparation of the optically active form of the oxabicyclic structure. From here on, the details of that work will be presented and discussed.
Starting again from 84, treatment with Mscl and base generated again the oxopyrilium species I-C, which now underwent [5 + 2] cycloaddition reaction with chiral sulfonamide acrylate 92, by which the norbornane moiety of the acrylate served as a chiral auxiliary. This furnished the oxabicyclic enones 90 and 91 as a racemic mixture with a d.r of 1:2 (respectively) though with poor yields (unoptimized, see scheme 40). However these two exo products are separable by chromatography, thus yielding 91 in enantiomerically pure form. OBn O O OH 84 MsCl iPr2NEt, 30% O O A* O CO2A* 9 O OBn O CO2A* 9 O OBn +
90 exo products90:91ca. 1:2 91 separable
A* =
SO2 N 92
Scheme 40: Preparation of oxabicyclic enones 90 and 91.
The ester moiety in 91 was converted into an aldehyde by a 2-step reduction and oxidation sequence, affording 92 (scheme 41). Note that this detour was taken due to possible
epimerization at C-9 in basic environment. Subsequent oxidation and esterification resulted in optically active (-)-85. O CO2A* 1. Dibal-H 2. DMP 9 O OBn 91 O CHO 9 O OBn 92 1. NaClO2 2. EtOH, H+ O CO2Et 9 BnO O (-)-85 88% 75%
30 Soon after the report of Nicolaou et al. another formal synthesis report was published from the group of Theodorakis.[27] Their strategy comprised a Rh-catalyzed [4 + 3] cycloaddition reaction between a disubstituted furan and a diazo ester bearing a chiral auxiliary (scheme 42). The furan 38 was prepared from 2-methylfuran and the diazo ester 37 was derived from (R)-pantolactone, both of which were prepared in 3 steps. When reacting furan 38 with the chiral diazo ester 11 catalyzed by Davies-Rh complex, the oxatricyclic compound 39 was formed with moderate diastereoselectivity (d.r = 3:1).
O O N2 TBDMSO O O + O O O O O TBSO O 37 38 39 d.r 3:1 2 mol% Rh2(Ooct)4 90% O 38 O 3 steps O O HO 3 steps O O N2 TBDMSO O O 37
Scheme 42: Davies-Rh catalysed cycloaddition of furan with diazoester.
Subsequent treatment with the electrophilic reducing agent DIBAL-H removed the chiral auxiliary, obtaining the corresponding β-hydroxyl enol ether. When exposed to Lewis acidic BF3, rearrangement of the enol ether affords the enone compound 40 (scheme 43). Rubottom oxidation transformed 40 into the α-hydroxyl enone 41. The strereochemical outcome of this latter reaction is controlled by the steric environment of the two faces of the molecule: the hydroxyl group adds to the less crowded face of the structure, establishing the stereochemistry at the hydroxyl carbon. It is worthy of note that m-CPBA epoxidizes the TMS enol moiety selectively without affecting the two double bonds, due to the former being more electron rich than the latter moiety.
O O O O HO 40 41 DIBAL-H then LDA, TMSCl then mCPBA, NaHCO3 BF3.OEt2, 59% then (COOH)2 39
87%
Scheme 43: Synthesis of the α-hydroxyl enone 41.
The next step of their strategy comprised of introducing and transforming the 5m-ring to form the oxatricyclic core structure. Starting from the 5m-ring formation, a diketone was prepared from 41 by a Stetter reaction with propanal (scheme 44). This afforded 42 as a single diastereomer. Subsequent intramolecular aldol condensation of the diketone afforded the tricyclic compound 43 in 2 steps by treatment with base.
31 O O HO O O TBSO H O O TBSO H O 41 42 43 1. TBSOTf, Et3N 2. propanal, thiazolium salt, Et3N, 75% 2 steps 43%
Scheme 44: Preparation of a diketone and aldol condensation reaction.
Subsequently, a hydroxyl group was introduced to the cyclic enone at the C-9 position in a stereo- and regioselective fashion in 5 steps by a series of reduction and oxidation chemistry affording triol 45 (scheme 45). The next step of the synthesis consisted of the removal of the C-3 hydroxyl group. However, performing this step under Barton-McCombie conditions afforded low yields. An alternative route was dehydration with Burgess reagent, generating an alkene. This less substituted alkene was subsequently hydrogenated with H2, Pd/C. The final step was the microwave-accelerated removal of the di-TBS-ethers with TBAF at elevated temperature, yielding the final (-)-englerin A precursor compound 32 in excellent yield. This compound was used by Ma et al. for the total synthesis of 1 (scheme 14).[26]
O TBSO H O O TBSO H HO OTBS O HO H OH 43 45 32 5 steps 41 % 1. Burgess reagent, PhMe, 90% 2. H2, Pd/C quant. 3. TBA, microwave, 93%
Scheme 45: Generation of (-)-englerin A precursor 47.
Altogether, this formal synthesis consists of 15 steps with 5% overall yield starting from the readily available starting materials 37 and 38. Note that Ma et al. needed only 8 steps to prepare this precursor compound. One drawback of the key rhodium catalyzed [4 + 3] cycloaddition reaction step is the moderate diastereoselectivity. In contrast, the rhodium catalyzed 1,3-cycloaddition reaction of Hashimoto et al. discussed in section 3.2 showed high diastereoselectivity (exo/endo = 87:13). Additionally, in contrast to the 1,3-dipolar cycloaddition reaction the [4 + 3] cycloaddition reaction did not introduce a functionality at C-9 carbon of the bicyclic structure, which thus had to be incorporated separately after subsequent steps.
4.2 Precursor preparation derived from a ring-closing metathesis reaction.
In section 3.3 it was illustrated that the total synthesis of 1 was implemented through ring closing metathesis reactions. However, all of the key reactions utilized a Grubbs II complex that only constructed the fused seven-membered ring.[30][12][14] The five-membered ring had to be introduced separately in previous steps, either by synthesis or by choosing an appropriate starting material. For the concurrent formation of the bicyclic guaiane ring system of 1, the use of a polyolefinic substrate is unavoidable, in particular one that is sterically hindered as is the case for 1. One strategy to circumvent these problems and use
32 these recalcitrant substrates to construct the bicyclic carbon framework in a single step is a Relay Ring Closing Metathesis reaction (RRCM) employed by K. A. Parker for the formal synthesis of 1.[17]
The polyolefinic substrate needed for the key RRCM step requires an enyne moiety at C-1 position in order to form a fused bicyclic structure in a single step. The preparation of this substrate started from commercially available geraniol 122. The introduction of a terminal olefin, followed by methyl oxidation and subsequent Sharpless epoxidation (93% ee) afforded 123 with the right stereochemistry at C-1 (scheme 46). The introduction of the C-1 enyne moiety was carried out by epoxide ring opening with lithitated acetylene and subsequently Parikh-Doering oxidation of the terminal alcohol afforded compound 124. The second terminal olefin was introduced by a Barbier reaction with alkyl halide 125 generating diol 126a together with its C-9 epimer. The two diastereomers were converted to its cyclic carbonate analogues with CDI (in order to establish the relative stereochemistries in further steps), furnishing the ene-yne-ene substrate 126b and its C-9 epimer 126b’.
HO O OH O H 3 steps 38% Li HMPA 1. 2. SO3, pyridine DMSO O CHO H OH 123 93% ee Br Zn dust O OR OR' 122 124 126 125 H 58 % a/a': R, R' = H (80%)
b/b': R, R' = C=O NaHCDI
9
+ 126'
Scheme 46: A seven step synthesis of the Relay ene-enyne-ene substrate 126.
With the ene-enyne-ene substrate 126b and its C-9 epimer 126b’ formed, the Stewart-Grubbs catalyst can bind to the C-4 ene moiety through the terminal olefin, generating a C-4 ruthenium carbene intermediate that undergoes a ring-closing metathesis with the enyne moiety, forming the first five membered ring system. Subsequent ene-ene metathesis generates the second seven membered ring system, constructing the guaiadiene framework. This affords 127 and its C-9 epimer 127’, which are chromatographically isolable (scheme 47). Compound 127 was further manipulated into its decarbonated and protected analogue
33 O 126b H 9 + 126b' Ru O N N Cl Cl O _ H 127 OR OR' 9 + 127' a: R, R' = C=O (45%) b: R, R' = H (97%) c: R = H, R' = TBS (93%) TBSOTf NaOH 4 O O separable
Scheme 47: Relay ring-closing metathesis and subsequent manipulations.
To construct the ether bridge the two double bonds should be differentiated. This could be established with the addition of a soft electrophile followed by transannular etherification. Thus 127c was treated with Hg(O2CCF3)2 and subsequently with a base (scheme 48). Since the trisubstituted double bond on the five membered ring is more electron rich it is less prone for soft electrophile mediated oxymercuration, in contrast to the less electron rich double bond on the seven membered ring. Additionally, the region- and stereoselectivity to form a bridge with C-7 comes also from the fact that in both the unstrained intermediates
I-Ea and I-Eb, where the C-10 bride adopts two different positions, the alkoxide cannot
perform a backside attack at C-6 (depicted with red arrows). However, the alkoxide in I-Eb can perform an attack to C-7 (depicted with a green arrow), thereby forming the desired oxatricyclic complex 128. Interestingly, the next step involving a radical mediated oxidative demercuration afforded an oxygen functionality at C-4 instead of at the desired and expected C-6, forming compound 129 together with its C-4 epimer 129’ (separable). The former compound was an intermediate for the synthesis of 1 performed by Echavarren et al., as discussed in section 3.2.2 (see 19b, scheme 13).[25] Therefore, this route constitutes a formal synthesis of 1. Note that Parker et al. needed the same amount of steps as Echavarren et al. to prepare 129. However, Echavarren afforded 129 diastereoselectively, in contrast to Parker et al, who obtained a diastereomeric mixture.
H 127c OH OTBS 1. Hg(O2CCF3)2 2. aq. NaHCO3 NaCl H 128 OTBS O2 ,NaBH 4 ClHg O H 129 OTBS O HO 55% from 127c + 129' H E O Me H OR H H H Me 6 7 4 5 H E H OR H H Me 6 7 4 5 O Me 6 7 + I-Ea I-Eb 37% from 127c 4
